Natural Gas Energy Measurement

NATURAL GAS ENERGY MEASUREMENT Edited by

Amir Attari Donald L.Klass Institute of Gas Technology, Chicago, Illinois, U.S.A.

ELSEVIER APPLIED SCIENCE PUBLISHERS LONDON

INSTITUTE OF GAS TECHNOLOGY CHICAGO

ELSEVIER APPLIED SCIENCE PUBLISHERS LONDON This edition published in the Taylor & Francis e-Library, 2005. “To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk.” INSTITUTE OF GAS TECHNOLOGY CHICAGO Sole Distributor Outside the USA and Canada ELSEVIER APPLIED SCIENCE PUBLISHERS LTD Crown House, Linton Road, Barking Essex IG11 8JU, U.K. Sole Distributor in the USA and Canada INSTITUTE OF GAS TECHNOLOGY 3424 South State Street, Chicago, Illinois 60616, USA The selection and presentation of material and the opinions expressed in this publication are the sole responsibility of the authors concerned. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of IGT. © Institute of Gas Technology 1987. ISBN 0-203-21619-9 Master e-book ISBN

ISBN 0-203-27243-9 (Adobe eReader Format) ISBN 1-85166-1522 (Print Edition) (Elsevier edition) ISBN 0-910091-61-7 (Print Edition) (IGT edition)

PREFACE

This volume contains papers presented at the first and second IGT symposia on natural gas measurement. The first program was held on August 26–28,1985 in Chicago to provide a forum for dissemination of information on all aspects of calorific value measurement of natural gas. Bowing to popular demand, however, selected papers on volumetric measurement of natural gas were also included in the second program held April 30-May 2,1986 in Chicago to broaden its scope. Speakers from the U.S. gas industry, U.S. Federal and academic thermochemical research centers, and U.S. instrument manufacturers, as well as speakers from West Germany, The Netherlands, Japan, and Indonesia presented papers. The conference papers in this book are organized under several topics that reflect the original program sessions of oral presentations. The topics of the first symposium (Natural Gas Energy Measurement, I) are: Thermal Energy Measurement (papers 1–5), Calibration Standards (papers 6–10), Automated Energy Measurement (papers 11–15), and Energy Measurement Accuracy (paper 16). The general topics of the second symposium (Natural Gas Energy Measurement, II) are: Measurement Fundamentals (papers 1–3), Volumetric Measurement (papers 4–5), Thermal Energy Measurement (papers 6–14), and Field Applications of Energy Measurement (papers 15–20). Full texts of the papers are produced with minimal editorial changes. The most significant event that came to light at the first conference was the compilation and acceptance of a single set of enthalpies of combustion for light hydrocarbons by the National Bureau of Standards and the Thermodynamic Research Center of Texas A&M University. This news is a welcome development, in that if these newly published thermochemical constants are generally adopted by the gas industry for use in the calculation of calorific value (or other properties) of fuel gases from their compositions, it will help alleviate a big source of confusion and ambiguity at the custody transfer points. We hope that these IGT programs will spark an awareness of the need for standardized measurement practices and the eventual use of a single system of units of measurement and reference conditions throughout the energy industry.

iv

The success of the two conferences in large measure was due to the efforts of the speakers and the authors to whom we are especially thankful for giving their time, energy, and experience in providing the high caliber of papers that were presented. We plan to continue Natural Gas Energy Measurement as a series. Amir Attari Donald L.Klass Institute of Gas Technology

TABLE OF CONTENTS

NATURAL GAS ENERGY MEASUREMENT I Paper No.

Page —PREFACE A.Attari and Donald L.Klass

iii

THERMAL ENERGY MEASUREMENT 1

GAS MEASUREMENT—THE NEED AND THE REALITY P.A.Hoglund

1

2

NATURAL GAS PROPERTIES CALCULATIONS FROM COMPOSITION K.R.Hall

6

3

HEATING VALUES OF COMPONENTS OF NATURAL GAS D.GarvinE.S, Domalski, R.C.Wilhoit, G.R.Somayajulu and K.N.Marsh

13

4

EFFECT OF C6+ HYDROCARBONS ON DEWPOINTS AND HEATING VALUES R.F.Bukacek

20

5

EFFECT OF WATER VAPOR ON HEATING VALUE R.J.Rau

26

CALIBRATION STANDARDS 6

PREPARATION OF STANDARDS FOR GAS ANALYSIS G.C.Rhoderick and E.E.Hughes

31

7

NEW STANDARDS FROM THE INSTITUTE OF GAS TECHNOLOGY B.H.Solka and A.Attari

39

8

THE ROLE OF CALIBRATION STANDARDS IN THE ANALYSIS OF NATURAL GAS LIQUIDS S, L.Brandt

46

9

COMMERCIALLY AVAILABLE CALIBRATION GASES D.P.Norris

55

COMPOSITE SAMPLING OF NATURAL GAS

70

10

vi

T.F.Welker AUTOMATED ENERGY MEASUREMENT 11

ENERGY MEASUREMENT WITH AN ON-LINE GC W.Dean

77

12

FIELD EXPERIENCE IN MEASURING ENERGY DELIVERED WITH GASEOUS FUELS W.H.lingman, L.Kennedy, K.R. Hall and J.Holste

84

13

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT C.J.Louttit

92

14

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER J.M.Minich

120

15

REAL TIME ENERGY MEASUREMENT A.P.Foundos and A.F.Kersey

134

ENERGY MEASUREMENT ACCURACY 16

ENERGY MEASUREMENT ACCURACY R.N.Cury

139

NATURAL GAS ENERGY MEASUREMENT II MEASUREMENT FUNDAMENTALS 1

CALCULATION OF THE HEATING VALUE OF NATURAL GAS FROM THE PHYSICO-CHEMICAL PROPERTIES OF THE PURE COMPONENTS K.N.Marsh

142

2

GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS G.E.Mattingly

148

3

CROSS REFERENCE SERVICE OF NATURAL GAS STANDARDS IN THE UNITED STATES J.C, Shapiro, G.R.Burkett and W.A.Crowley

167

VOLUMETRIC MEASUREMENT 4

NEW CONCEPTS IN GAS CALIBRATION E.J.Dahn

176

5

MICROSTRUCTURE SENSORS FOR FLOW, DIFFERENTIAL PRESSURE AND ENERGY MEASUREMENT R.Higashi, R.G.Johnson, A.K.Mathur, A.N.Pearman and U.Boone

182

vii

THERMAL ENERGY MEASUREMENT 6

THE SMART-CAL/CUTLER HAMMER CALORIMETER COMBINATION FOR BETTER PERFORMANCE AND DATA PROCESSING A.F.Kersey

193

7

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY J.J.Singh, D.R.Sprinkle and R.L.Puster

196

8

AVAILABILITY OF NBS—TRACEABLE CALIBRATION GAS STANDARDS FROM IGT—PROGRAM UPDATE B.H.Solka and A.Attari

212

9

THE EFFECT OF MOISTURE CONTENT ON NATURAL GAS HEATING VALUE T.J.Glazebrook

222

10

ON-LINE CHROMATOGRAPHY J.L.Shafer

226

11

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE R.M.Batubara and K.Tatang

232

12

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT T.A.Clark

256

13

AUTOMATED GAS ANALYSIS SYSTEM H.L.Humbke

269

14

LINEARITY AND RELIABILITY DETERMINATION OF BTU ANALYSIS BY PROCESS GAS CHROMATOGRAPHY L.Fields

276

FIELD APPLICATIONS OF ENERGY MEASUREMENT 15

ON-SITE ENERGY MEASUREMENT R.A, Price

284

16

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS M.Haruta, K.Uekuri and Y.Kiuchi

287

17

DIRECT MEASUREMENT OF ENERGY FLOW—RECENT FIELD EXPERIENCE C.H.Griffis, W.H.Clingman, L.R.Kennedy, K.R.Hall and J.C.Holste

305

18

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS K.Homann and H.J.Krabbe

314

19

HEAT QUANTITY DETERMINATION IN LONG DISTANCE GAS TRANSMISSION

330

viii

G.Strulik and T.Fischer 20

STATUS OF FEDERAL REGULATION OF NATURAL GAS AND ITS IMPLICATIONS FOR FIELD APPLICATIONS OF ENERGY MEASUREMENT P.D.Hubbard

341

Symposium Papers Natural Gas Energy Measurement I Presented August 26–28,1985 Chicago, Illinois

GAS MEASUREMENT—THE NEED AND THE REALITY Paul A.Hoglund Senior Vice President, Operations Washington Natural Gas Company Seattle, Washington

ABSTRACT

There is a need today to improve the level of precision of our gas measurement practices. Our current technology is some 50 years old. Over the past several years, our industry has committed millions of dollars in research in the basic technology area. We are currently beginning to see the results of this research effort, however, much more remains to be done. In the final analysis, research, standards and procedures will not be enough. As an industry, we’re going to have to make a major technical commitment in this area. The measurement engineer of tomorrow will need a sound base in fluid dynamics and mathematics with a little computer science thrown in. Without that high degree of interest, we’ll be unable to achieve an improvement in precision. Our topic: Gas Measurement—The Need and the Reality, presents quite a number of challenges. The need is relatively simple—there is a need today to improve the level of precision of our gas measurement practices. Reality is an entirely different matter. Psychologists tell us that it’s a matter of personal perception. In that context, this personal perception approach, there are 3 concerns in the field of basic engineering that impact us in our gas measurement practice: 1. The substitution of standards and procedures for judgement and responsibility—we’ve all used that crutch. Back in our college days we’d try to memorize how a problem was solved in the book, substitute the numbers in the problem and turn the crank for an answer. This is really the substitution of memory for brains (there is a difference). The approach has been compounded in recent times by the legal profession—you’ve got to have standards to protect yourself from a lawsuit. “On a broader scale, government is continually promulgating standards in order to fit all of us into little nitches. Their own bureaucracy is so effective, they have to impose it on the rest of us. The net effect of this is that the

2

GAS MEASUREMENT—THE NEED AND THE REALITY

successful engineer is the one who knows where it is in the book, not the one who can apply the common sense and logic. We’ve substituted standards and procedures for judgement and responsibility. 2. Engineers have become so involved with numbers that they’ve lost touch with reality. Back in the “olden days” the standard tool of the engineer was a slide rule. This would give him 3 significant figures of accuracy which was generally within the required precision of the calculation. The slide rule disappeared with advent of the hand held calculator; which in turn has been subplanted by the personal computer. A 32 bit computer will give, say, 12 significant figures. This is great for number crunching, but we have to step back from time to time and ask whether or not those 12 significant figures are within the precision of our basic data. My personal observation is that we don’t step back often enough. We become so enraptured with the numbers and our ability to manipulate them that we lose sight of what their real significance is. 3. Life has become so complicated and knowledge so broad that as individuals we tend to concentrate in specific areas—this is the “you can’t see the forest for the trees” syndrome, or its correlary “you can’t see the trees for the forest”. Sometimes in working a problem (the forest) we fail to look at its elements (the trees) to determine their impact on the solution. We can perhaps place these points into perspective by looking at the needs and the realities of gas measurement. The need here, to improve the level of precision in gas measurement, relates to changes that have taken place in the market over the last ten years. What was applicable in the 60’s is questioned in the 80’s. Yet from a measurement standpoint our technology is more in keeping with the 30’s. Today we deal with a measurement technology that is fifty years old. A technology that produces results that are +/− 0.5%. Or to place this in engineering terms—a technology that will produce results to three significant figures. Common practice, however, marries computer techniques to our calculation and we regularly produce billings to nine or more significant figures. We fail to recognize that a $1,000 bill is really +/− $5 and a $1,000,000 bill is +/ − $5,000, This technology lag is nothing new. On the other hand, it was only with the major escalation in the price of natural gas and other related hydrocarbons that occurred in the late 70’s that our industy really displayed an interest in the precision, or accuracy of our volume measurement. Actions were initiated to address these concerns. Any advance in technology requires two things—time and money. We don’t really have enough of either one. On the other hand, what we are commiting is begining to produce results. Over the past several years, our industry has committed millions of dollars in research in this basic technology area. Specific programs have been adopted and funded by the American Gas Association, the American Petroleum Institute, the Gas Research Institute, vendors, the National Engineering Laboratory of Great Britain, Gasunie in Holland, Ruhr Gas in Germany, Gaz de Franze and probably several others that don’t immediately come to mind. This work has been done in cooperation with the National Bureau of Standards at both the Boulder and Gaithersburg facilities, the University of Oklahoma, Texas A&M, the Colorado Engineering Experiment Station, Inc., Southwest Research and Gulf Research. We’ve seen a whole new emphasis on Standards as a direct result of these concerns: 1. ANSI/API 2530 (A.G.A. Report No. 3) covering orifice metering of natural gas and other related hydrocarbon fluids, was revised 1½ years ago and in May of 1985, was approved by ANSI. 2. A.G.A.Transmission Measurement Committee Report No. 7 covering turbine meters was published earlier this year and provides a sound basis in that area. 3. ISO 5167, the International Standard on orifice metering, has been with us now for about 5 years.

3

4. ASME recently approved a Standard covering Measurement of Fluid Flow in Closed Conduits with Orifice, Nozzle and Venturi. This has been proposed to ANSI. From our industry standpoint, the first two of these help. They don’t really improve the level of precision available, but they do provide a sound foundation for future change. The ISO and ASME Standards tend to add confusion in this area, but on the bright side, underscore the importance of a consistent technical approach to large volume measurement. These particular methods provide two alternate means of evaluating orifice meters. The answers are different but in most cases within the precision mentioned earlier. The key point, is they do not provide any improvement in that precision. From a political standpoint, we’ve seen a far broader interest in gas measurement. Up through the 60’s this was almost the exclusive province of the American gas industry. Today we see a high level of involvement from producer groups (GPA and API), plant operators (ASME), affiliated organizations (ISA), district heating organizations, consumer advocates (through state utility commissions), government (BPL and DOE) and international organizations (ISO, OIML, GERG and others). Unfortunately some very pointed questions are being asked—questions for which we, as an industry, have no immediate answers. An International Measurement Symposium, under the joint sponsorship of many of the organizations previously mentioned, both national and international, is scheduled to be held in the fall of 1986. The purpose of this is both to discuss the current state of the art, but on a more critical level to bring together these varied interests into a cohesive unit that can address shared concerns on a technical basis. Our industry has something to share. The research efforts that have been underway for the past several years are at long last beginning to bear fruit: 1. The supercompressibility project being conducted by University of Oklahoma, under the sponsorship of GRI, is releasing their initial report covering a far broader spectrum of gases that have been applicable to our old NX-19 evaluation. Contracts have been entered into for publishing an application guide with current thinking that this will be released this fall as A.G.A. Transmission Measurement Committee Report No. 8—Compressibility and Supercompressibility of Natural Gas and Other Hydrocarbon Gases. This is really a tremendous step as our industry regularly works with gases well outside of the applicable range of NX-19. The value of this work goes well beyond measurement. The compressibility data will provide a better basis for compressor design and testing. Reservoir engineers should find the expanded range of particular value. Line pack calculations will be more precise. This work is really a key step in basic knowledge. 2. The NBS Boulder flow disturbance work, again funded by GRI, is still in the early stages, but is already providing some answers. Papers presented at the 1985 A.G.A. Operating Section conference provide some real insight into current concerns and future efforts in this area. This work is directed not only to identifying the impact of these disturbances, but to means of mitigating those impacts. 3. The API orifice data project is well underway and we hope to have some of the early data reports available during 1985. This work includes orifice data on water, viscous fluids and natural gas. 4. The NBS Boulder data (funded by GRI), which was released in 1983, has been expanded with work at CEESI and a contract has been negotiated with Gulf Research for further data acquisition. All of this experimental data should give us a sound basis for a consensus orifice flow equation of higher precision that we enjoy today.

4

GAS MEASUREMENT—THE NEED AND THE REALITY

These varied programs were set out to address specific problems. They are advancing technology in limited areas. In the broad over-view we find that similar programs are underway in the European Economic Community and by other industry organizations on a national level. In that context, our approach is disjointed. Even today, there are a number of defined concerns that demand solution but are not being adequately addressed: we are not taking advantage of the technology that exists in other industries and has application; we do not have a real means of determining what is “accurate” and perhaps most importantly, we need a comprehensive program to both direct our actions in this vital area and to communicate our concerns to others in order to broaden the base of activity. It was in this context that an industry solicitation was made among individuals having an active interest in the measurement field. Those contacted were asked to respond with items that sould be considered as elements of this comprehensive measurement plan. The many thoughts and ideas returned were combined into twenty specific elements under the generic headings of: 1. Basic Research, 2. Laboratory Needs, 3. Hardware Needs, 4. Instrumentation, 5. Standards and 6. Associated Relationships. Respondents were then asked to prioritize these elements in the context of their perception of industry needs. Seven of the twenty were set out for immediate action with four others being categorized as high priority. Finally, respondents were asked to provide specific actions in each of the seven areas that would address the concerns and move our industry toward solutions. The response here provided the basis for a plan that was approved this spring by the A.G.A. Operating Section. The plan combines the thoughts and ideas of many people and appears to address the concerns as we identify them today. It is recognized as an initial approach. All of the listed items are important. The priorities only attempt to direct our limited resources to areas that appear to have the greatest immediate return for our industry. Hopefully by articulating the broad list, we can gather support from others and work toward solution in these many areas. The seven primary elements of this plan include: 1. We need to expand our knowledge of the physical properties of natural gas, synthetic gas and the hydrocarbon constituents of these gases. This knowledge must include the influence of contaminants in various quantities, together with the impact of pressure, temperature, etc. 2. We must continue the work to develop an expanded base of research quality data on orifice flow coefficients. This work is currently underway but its completion is essential to several other aspects of the broad measurement plan. Turbine and rotary meters have immediate application in many areas with wider rangability and to some degree less bother that the orifice meter. We can’t loose sight of the fact that there are hundreds of thousands of orifice meters in operation. The replacement of most of these is not practical from either an economic or a physical standpoint. We have no choice but to improve the precision of this basic method. 3. We need to expand our knowledge of how measurement, by whatever meter is employed, is influenced by external factors and what corrective measures can be taken to mitigate those influences. Key items here include approach conditioning, expansion factor, instrument connections, instrument locations, pulsation, type of instrumentation, transducer response, etc. 4. We need to develop an independent qualified flow test facility with basic measurements traceable to the National Bureau of Standards. This facility should be operated under the sponsorship of the gas industry. It must be capable of evaluating metering devices, both new and existing and provide documented data on performance characteristics at various temperatures, pressures, flow rates, Reynolds Number, etc.

5

Several independent facilities exist that have partial capability in this area. The individual and collective limitations dictate the need for a single comprehensive facility. This sounds simple but really isn’t. It will take a great deal of effort to make this a reality. 5. We must place additional emphasis in the area of real time measurement. We must know today the quantity of gas being used today. This involves marrying computer technology to our gas measurement concepts and the development of on site techniques to correct for pressure, temperature and gas constituents. 6. We need an acceptable secondary standard that will provide results traceable to the National Bureau of Standards. The device must be capable of being used in situ for field calibration or as a shop test facility that will evaluate a meter under its operating conditions. This is really an extension of the transfer proover concept to a broad range of flow. With the new speed of sound data the critical flow nozzle may have application. With recovery sections they have low pressure loss and multiple nozzles can provide for a wide range of flow and pressure. 7. We should develop a better understanding of the basic fluid dynamics of gaseous flow through various meters through the development of computer models that can be verified by non-intrusive devices. The computer models could result in an entirely new generation of measurement devices. The plan itself, with all twenty elements and details on action plans, has been published as an Engineering Technical Note by the A.G.A. Operating Section. Copies are available direct from A.G.A. for those with an interest. Fortunately, several of the items in the plan are covered, at least in part, by the current research efforts. This research must be completed and expanded programs implemented if we are to improve our technology in this area. Improve it, we must. If we as an industry, don’t accomplish this, it will be done for us by others. Such action could be to our serious detriment. The need then—to improve the level of precision of our gas measurement practices—becomes far more complex as we break it into it’s specific parts. The statement encompasses a broad goal. As we break it into it’s elements we can examine it’s complexity. The reality is that the needs have been identified and work is under way to accomplish the goal. What is being attempted here is to improve the level of precision by almost a full order. This cannot be accomplished by standards and procedures alone. As an industry, we’re going to have to continue a major technical commitment in this area. We’re going to have to understand what our base measurements really are and the impact of gravitational force, barometric pressure and temperature. We’re going to have to apply a better understanding of the nature of our instruments. Do they really measure absolute values of pressure and temperature as we say they do? Where does reality fit in, in terms of the precision of the instrument itself and the precision of the calibration? Do we really know the nature of the fluid we’re measuring, or it’s changes with time? Do it’s changes due to outside influences have a significance in our measurement process? The measurement engineer of tomorrow will not be able to substitute standards and procedures for judgement and responsibility. He will have to have a broad understanding of the values and the significance of the data he’s working with. And finally, he’ll have to understand both the forest and the trees. We won’t need a botonist, but we will need an individual with a sound base in fluid dynamics and mathematics, with a little computer science thrown in. Without that base, your company will suffer the economic consequences.

NATURAL GAS PROPERTIES CALCULATIONS FROM COMPOSITION Kenneth R.Hall, Ph.D. Professor of Chemical Engineering Director, Thermodynamics Research Center Texas A&M University College Station, Texas 77843

ABSTRACT

Let us assume for purposes of this paper that we can know the composition of a gas from some suitable measurement. It is then possible, in principle, to calculate pertinent properties such as: heating value, relative density, compressibility factor. Unfortunately, as often happens in practice, these supposedly unambiguous calculations become clouded by accepted procedures, misconceptions, misguided regulations. The discussion in this paper attempts to dispel the misconceptions and to clarify the accepted procedures. As for the regulations, it is only possible to wish that they would not always choose the path of maximum irrationality and to try to perform the calculations in the least offensive (technically) manner possible. The calculations described in this paper reflect those suggested in GPA Standard 2172–85. However, this paper contains considerable amplification and discussion of the techniques. INTRODUCTION While we usually consider the composition of a natural gas as a known, it is really at best approximately known because of measurement errors. In this paper, let us assume that we do in fact know the composition and, ignoring composition errors, discuss how to approach properties calculations from this knowledge. Various standards in the natural gas industry address this issue, among them are GPA-2172, ANSI-2530 and ASTM D-3588. This paper closely follows GPA-2172 but contains considerable amplification. The primary properties of interest are, compressibility factor, relative density (or more specifically relative molar mass) and heating value. In the following discussion, we examine definitions, misconceptions, recommended calculation procedures, shortcut approximations and sources of uncertainty.

7

DEFINITIONS The definitions of the pertinent properties are not necessarily totally unambiguous. The reasons are some unfortunate federal regulations and some misguided standard practices. The adopted definitions are consistent with GPA-2172 which is the most recent standard published covering this topic. Compressibility Factor The compressibility factor is the ratio of the ideal gas density to the real gas density when both are at the same temperature and pressure. The primary function of compressibility factor is to indicate the deviation of real gas behavior from that of the ideal gas. Relative Density The relative density is the ratio of the density of the gas at its temperature and pressure to that of dry air at its temperature and pressure. The relative density is primarily a means to establish the molar mass of the gas. Heating Value Heating value refers to the total energy transferred as heat in an ideal combustion reaction at base temperature and pressure. For net heating value, the water formed in the combustion appears as vapor in the products; for gross heating value, the water formed in the combustion appears as liquid in the products. COMPRESSIBILITY FACTOR The compressibility factor is usually the most convenient form in which to express the equation of state P=P (T,p). From the definition, it is (1) where Z is compressibility factor, p is density per mass, M is molar mass, P is pressure, R is the gas constant, T is temperature and super * denotes ideal gas. At conditions of temperature and pressure near ambient, the truncated virial equation of state adequately represents the volumetric behavior of natural gas (2) In Equation 2, B is the second virial coefficient for the natural gas mixture and is a function only of temperature and composition

(3)

where x is mole fraction and n is the total number of components. In Equation 3, the Bii are pure virial coefficients and Bij are interaction virial coefficients. For natural gas components (except He and H2) at near ambient conditions, the virial coefficients are negative. The Bii and Bij are functions only of temperature.

8

NATURAL GAS PROPERTIES CALCULATIONS FROM COMPOSITION

Several equations exist for estimating values for Bii and Bij. GPA-2172 contains one that is particularly convenient over the temperature range which includes all base condition temperatures used for natural gas. Invariably, these expressions require a computer for efficient use. However, on a computer, they are rather simple to calculate. For hand calculations, a simplification provides very reasonable values. The basic assumption is that the interaction virial coefficient is the geometric mean of the pure virial coefficients. (4) Under this assumption, the mixture virial coefficient becomes (5) Unfortunately, the geometric mean assumption is not correct, but for natural gas components in the range it is a reasonable approximation. Even this simplification is amenable to hand calculations only at a single temperature. Otherwise, it requires exactly the same equation as the rigorous procedure. RELATIVE DENSITY Having the compressibility factor expression allows us to calculate relative density, G: (6) where subscript a refers to air. We should note immediately that the sole function of the relative density is to provide the molar mass of the gas. Reference to air is an artifact associated with devices used for field measurements; it is not necessary but it is practical. If the gas and air have identical temperatures and pressures, Equation 6 simplifies to (7) The value of Z , often assumed to be unity, is actually about 0.9996. While 0.04% error may not be of much concern, assuming the temperatures and pressures to be identical can introduce serious errors unless the measurement instrument specifically addresses this assumption. Should the air and gas be in ideal gas states, the values of Za and Z are exactly unity and Equation 7 reduces to (8) Therefore, the molar mass of the gas is (9) Knowledege of the composition eliminates the necessity for this measurement because

(10)

and the pure molar masses are available. Division of Equation 10 by Ma produces an expression for ideal relative density.

9

(11)

HEATING VALUE The primary thing to remember about heating value is that it is an ideal gas property. It reflects energy transfer in an ideal combustion reaction (12)

(13) where * denotes ideal gas and denotes liquid. Clearly, the difference between net and gross is the ideal enthalpy of vaporization of the water, , which is slightly larger than the enthalpy of vaporization, . Data for such reactions is invariably at 25 °C and in SI units. The enthalpy of combustion data usually appears on an amount (per mole) basis. For a mixture, the value is (14)

To convert the enthalpy of combustion per mole to a per mass basis requires division by the molar mass (15) Multiplication of the enthalpy of combustion per mass by the ideal gas density provides the enthalpy of combustion per volume (16) For a dry gas, the enthalpy of combustion is the negative of the heating value. Wet or saturated gas presents a question of interpretation. Correction for Wet or Saturated Gas Because analyses of natural gas are invariably on a dry basis, if the gas is wet or saturated, a correction is necessary to account for the water. On a basis of one mole of dry gas, the mole fraction of water is (17) or rearranging

10

NATURAL GAS PROPERTIES CALCULATIONS FROM COMPOSITION

(18) where nw denotes the moles of water. Therefore, the correct analysis of the gas requires adjusting the various mole fractions. The total moles of dry gas is 1 while the total moles of wet or saturated gas is (19) The mole fractions of dry gas are (20) therefore the mole fractions of wet or saturated gas are (21) Thus, to obtain corrected mole fractions, multiply by one minus the mole fraction of water and then use the xi(cor) in the equations for enthalpy of combustion. If it is not possible to determine the mole fraction of water empirically, it is common practice to assume Raoults’ law for the saturated gas (22) where is the vapor pressure of water at the temperature selected (for practical purposes the base temperature). This expression assumes either a) ideal gas vapor, ideal solution liquid, liquid volume independent of pressure or b) all non-idealities cancel identically. It is probably a reasonable assumption for saturated natural gas between 0 and 25°C and near 1 atmosphere. For net heating value at a given base temperature and pressure, the correction term 1-xw is simply a constant multiplier . However, a problem arises when using gross heating value because technically water has a gross heating value from the “reaction” (23) equal to the ideal enthalpy of vaporization. Thus, to calculate the gross heating value per volume for a wet or saturated gas involves and additive term. (24) (25) COSTING NATURAL GAS The cost of natural gas results from multiplying the price per energy delivered by the rate of energy delivery by the time of delivery (26) where p is the price, is the ideal rate of energy released as heat upon combustion and At is a time period. can be either on a net or gross basis requiring only an adjustment of p:

11

Of course, custody transfer is on a gross basics by law or agreement. The heating value is input data to calculate . Various procedures are possible. The preferable, but never used, procedure uses (27) where is the molar flowrate. It is simple to convert the calculation to a mass basis (28) or to volumetric bases (29) where is the mass flowrate, is the ideal gas flowrate and Equations 26, 27, 28 and 29 become

(30) is the real gas flowrate. Summarizing,

(31) Equation 31 illustrates some misconceptions. Use of molar flowrate, mass flowrate or ideal gas flowrate is actually simpler than using real gas flowrate to calculate . An objection might be that flowmeters, notably orifices, produce , This is true, but with exactly the same information it is possible to calculate , or with slightly less uncertainty. In the case of and , it is not even necessary to establish a base pressure because and are independent of pressure while is not. A more serious misconception is that division of the ideal heating value, , by Z produces the real heating value. This is not true. Division of by Z produces which is the required value in calculation of . It is possible but impractical to calculate the real heating value and from it the real energy released, However, Q is not the actual energy released. is really a weighting factor in Equation 25, is equally as valid as weighting factor as would be is much simpler and less ambiguous to calculate than . CONCLUSIONS We have examined calculation procedures to obtain compressibility factor, relative density and heating value from composition analysis. In each case, we have tried to illustrate common misconceptions and to demonstrate rigorous procedures where possible. For compressibility factor, we have seen that the commonly used summation factor approach relys upon a not unreasonable but unfounded assumption. We have noted that the whole purpose of relative density is to provide the molar mass of the gas. While most field instruments measure relative density, it is possible to devise instruments which measure M directly. The heating value is the object of the most mystique. A logical and rational basis for heating value is to use the net value per mole with molar flowrate. The next most logical route is to use net value per mass with mass flowrate; then net value per volume with volumetric flowrate; then gross value per mole with molar flowrate; gross value per mass with mass flowrate; gross value per volume with volumetric flowrate. All the above procedures assume a dry gas calculation. The order of preference repeats on a lower level for saturated gas calculations. Finally, we must conclude that the least rational and most cumbersome calculation is the saturated, gross, volumetric basis. However, this is precisely what we must do because of historic reliance upon the Cuttler-Hammer calorimeter, existing contacts, and misguided federal regulations. Finally, we have noted that division of the ideal heating value per volume (which is the commonly tabulated value) by compressibility factor does not produce the real heating value per volume. It only permits use of the real gas flowrate in calculating .

12

NATURAL GAS PROPERTIES CALCULATIONS FROM COMPOSITION

GPA 2172–85 utilizes all the procedures discussed in this paper. It also permits the calculations at any temperature between 0 and 25°C (60°F=15.56°C) and for any pressure up to about 2 atmospheres.

HEATING VALUES OF COMPONENTS OF NATURAL GAS D.Garvin and E.S.Domalski Chemical Thermodynamics Data Center National Bureau of Standards Gaithersburg, MD 20899 R.C.Wilhoit, G.R.Somayajulu and K.N.Marsh Thermodynamics Research Center Texas A&M University College Station, TX 77843

ABSTRACT

New recommendations for the heating values of components of natural gas are reported. These are based on a reassessment of the available experimental data. The new recommendations are valid for the temperature range from 0° to 25°C and for pressures up to about one atmosphere. The sources of the data are indicated and factors considered in the reassessment are presented. Heating values of components of natural gas are known to about 0.02 to 0.04 percent. INTRODUCTION For the gas industry, the most important physical property of natural gas is its heating value. This value can be determined by calorimetry or by calculating the value from a knowledge of the composition of the gas and the heating values of the pure components. Except for the most careful of calorimetric measurements, the value obtained by calculation is more accurate. A typical example is an instrument that measures the composition using a gas chromatographic technique and then computes the heating value. For direct calorimetric measurements, it is also necessary to know the heating values of the major components of natural gas, although the dependence is indirect. Calorimeters used in the industry are calibrated by burning samples of natural gas of known heating value supplied by IGT. These heating values are determined calorimetrically at IGT in instruments calibrated against a reference sample of high purity methane, the heating value of which has been certified by NBS Armstrong [1], This certified heating value is calculated from the composition of the sample and the heating values of the components. Thus the field measurement is both traceable to NBS and to the heating values of the components.

14

HEATING VALUES OF COMPONENTS OF NATURAL GAS

We have updated these heating values. This has been a joint project of two groups of thermodynamicists who specialize in the evaluation of thermodynamic data. The new values take into account all of the currently available data and are applicable to the entire range of standard temperatures and pressures used in the gas industries of the world: 273.15 to 298.15 K (0 to 25 °C), 0.1, 0.101325 and 0.10156 MPa (1 bar, 1 atm and 14.73 psia). They differ only slightly from values previously recommended, but probably have a smaller systematic bias. In addition, careful attention has been paid to the question of the uncertainties to be assigned to the data. They supersede the values given in Armstrong and Domalski [3], Armstrong and Jobe [2], Domalski [6], the TRC Thermodynamic Tables 1982 [37], GPA 2145–77 [11] and related publications. These recommended data have been prepared for the International Group of Importers of Natural Gas (GIIGNL) and the (U.S.) Gas Producers Association (GPA) and will appear in their manuals and standards [12,21]. The recommendations will also be submitted to the American Society for Testing and Materials and to the International Standardization Organization for their consideration. It is our hope that the same numbers can be used for heating values where ever natural gas is bought and sold. Emphasis is placed in this paper on the data for the properties of the pure compounds that are components of natural gas. How these properties are combined to produce heating values for natural gas mixtures at appropriate standard conditions is the subject of another paper at this symposium and will not be reviewed here. Instead, the required data will be specified, their sources indicated and examples given of the data selection process. (The calculation of properties of mixtures is also treated in detail in GPA 2172 [13] and Armstrong and Jobe [2]. The latter also includes substantial material on background issues related to the specification of heating values.) THERMODYNAMIC PROPERTIES Several physical properties are combined to produce the heating values at the standard conditions used in the gas industry. The most important ones are: (1) the ideal gas enthalpies of combustion at 25 °C, ΔCH°(298.15 K), (2) the ideal gas enthalpy differences between 25 °C and the various industry standard temperature, H°(T)H°(298.15 K), (3) the enthalpies of formation of auxiliary substances such as CO2, H2O(l & g), CO, SO2, and their enthalpy differences, (4) ideal gas volumes, and (5) molar masses. Of lesser importance are the properties of the pure compounds in the real gas state. These are used to calculate volumes needed to interpret flow measurements and to correct ideal- to real-gas enthalpies. Only simple equations of state are required. For pressures up to about one atmosphere, these PVT data can be represented by the second virial coefficients of the pure components and their interaction coefficients with methane. Heating Values The selected heating values for the hydrocarbons from methane through the hexanes are given in Table 1. These are for the complete reaction at constant temperature and pressure of the gaseous hydrocarbon with

15

gaseous molecular oxygen to form gaseous carbon dioxide and liquid water, that is they are “gross calorific values” or “higher (superior) heating values”. Adjustment for Temperature. The enthalpy differences between 298,15 K and the various industry standard temperatures are based on the heat capacity data in the TRC Hydrocarbon Tables [36]. These are, in turn, based on statistical calculations and critical evaluations made by other groups. For reference, also see [2], Heat capacities of gases are rarely measured these days. The values for enthalpy differences at standard temperatures are implicit in tables 1 and 2. In order to make the set of recommendations useful at all likely conditions, the heat capacity data are given in the full reports as analytical expressions valid from 0 to 25 °C [2, 13, 21]. Table 1. Ideal Gas Heating Values for the C1, to C6 Paraffin Hydrocarbons, at Four Reference Temperatures. The second decimal place is not significant, but is provided as an aid to rounding in calculations on mixtures. −ΔCH°/kJ.mol−1 Compound methane ethane propane n-butane iso-butane n-pentane iso-pentane neo-pentane n-hexane 2-methylpentane 3-methylpentane 2,2-dimenthy butane 2,3-dimethyl butane

298.15 K 25 °C 890.65 ±0.37 1560.69 ±0.25 2219.17 ±0.45 2877.40 ±1 .00 2868.20 ±1.00 3535.77 ±0.46 3528.83 ±0.58 3514.61 ±0.50 4194.95 ±0.67 4187.32 ±1 .00 4189.90 ±1.00 4177.52 ±1 .00 4185.28 ±1 .00

288.71 K 60 °F 891.53 1562.06 2221 .99 2879.63 2870.45 3535.84 3531.51 3517.27 4198.05 4190.43 4193.03 4180.63 4188.41

288.15 K 15 °C 891.58 1562.14 2221 .10 2879.76 2870.58 3538.60 3531.68 3517.43 4198.24 4190.62 4193.22 4180.83 4188.60

273.15 K 0°C 892.99 1564.34 2224.01 2883.82 2874.20 3542,89 3562.98 3521.72 4203.23 4195.61 4198.24 4185.84 4193.63

Table 2. Enthalpies of Formation of Auxiliary Substances ΔfH°/kJ.mol−1 Compound carbon monoxide carbon dioxide water (gas) water(liq) hydrogen sulfide sulfur dioxide

298.15 K −110.53 ±0.17 −393.51 ±0.13 −2H1.814 ±0.042 −285.830 ±0.042 −20.63 ±1.00 −296.81 ±0.69

288.71 K −110.59 −393.50 −241.715 −286.148 −20.46 −296.69

288.15 K −110.59 −393.50 −241.721 −286.131 −20.47 −296.70

273.15 K −110.69 −393.49 −241.561 −286.634 -----

Note: enthalpies of formation of the following elements in their standard reference states are zero at all temperatures: Ar (g), C (cr, graphite), H2 (g), He (g), N2 (g), O2 (g), S(cr, rhombic).

16

HEATING VALUES OF COMPONENTS OF NATURAL GAS

Auxiliary Data. The thermochemical data for auxiliary substances at 298.15 K have been taken from the CODATA Key Values for Thermodynamics [4], the reference base recommended for use in accurate work. Values at other temperatures are derived using enthalpy differences as indicated above, except for water, for which the enthalpy differences are consistent with the new “NBS/NRC Steam Tables” [16]. These are shown in Table 2. Molar masses are calculated from the IUPAC 1981 Atomic Weights [17]. The value for the gas constant, R=8.31448 J/(mol.K)−1, which is needed in the calculation of volumes, is from the set that is currently being recommended to CODATA by its Task Group on Fundamental Constants [35]. Thus all of these data are up-to-date and internationally recognized. Real Gas Properties. The equation of state data in the recommendations is given in terms of second virial coefficients for the pure substances, Bi(T), and their interaction coefficients with methane, Bij(T). (Because of the low pressures used in the industry, the simple equation PV/RT=1+ B(T)/V is sufficiently accurate). There are very few hydrocarbons for which there are measured virial coefficient data in the temperature range of interest, and appreciable extrapolations are needed. Also some of the virial coefficients are based on indirect methods, and may be of marginal reliability. To overcome these problems, a correlation developed by K.R.Hall for virial coefficients for all of the hydrocarbon data has been used. It is valid for the range 0 to 25 °C and is based on a reduced equation of state. A computer program in GPA 2172–1985 [13], uses this correlation to calculate real gas properties. Evaluation of Data. The selection of the enthalpy of combustion data is outlined here in order to show the factors that are involved. A more detailed report is in preparation. In some cases molecular structure correlations have been used to help select from among discordant experimental values. Methane. In 1931, F.D.Rossini determined the enthalpy of combustion of methane [28]. His work outclassed all earlier measurements for precision of the heat measurement, analysis of the sample, the determined completeness of combustion by weighing the water produced. An adequate correction was made for the main impurity, carbon monoxide. These experiments, together with Rossini’s earlier measurements of the enthalpy of combustion of hydrogen in essentially the same apparatus have been the basis for most recommended values since that time. It is difficult to fault them. The scatter of the six determinations is shown on the left side of Figure 1. Rossini rejected the highest point, on sound statistical grounds. In 1972 Pittam and Pilcher [24] remeasured the enthalpy of combustion of methane using the same method (flow calorimetry) but with a much purer sample and a better analytical method (weighing the CO2). Their six determinations are shown on the right side of Figure 1. This is a better measurement, but the scatter in the data is similar to that in the study by Rossini; methane is very difficult to burn completely in a calorimeter [33]. When the two data sets are considered together neither stands out, and it is no longer certain that the point rejected by Rossini can be dropped. In the new assessment, the recomended value is an average of all twelve points.

17

Figure 1. Comparison of the measurements of the enthalpy of combustion of methane made by Rossini [28] and by Pittam and Pileher [24] and average values using different numbers of points.

Ethane and Propane.For both of these substances the measurements of Pittam and Pilcher [24] have been selected, in preference to those of Rossini in 1934 [29]. This choice was made because the former used an improved method, had purer gases and obtained more regular differences between enthalpies of combustion in the series of compounds methane—ethane—propane butane. Butanes. Three studies on n-butane have been reported: Rossini [29], Prosen et al. [25] and Pittam and Pilcher [24]. They are of similar quality, and the selected value is the average, without weighting, A larger uncertainty is assigned to the selection compared to those reported by the authors. For iso-butane, two calorimetric studies, Prosen et. al. [25] and Pittam and Pilcher [24] and two isomerization equilibrium results, Scott [32], and Pines [23], are in moderate agreement. The four values have been averaged. Rossini’s value, reported in 1935 [30], is not in agreement and has been discarded. Pentanes. There are three pentanes: n-pentane, 2-methylbutane, and 2,2-dimethylpropane. Calorimetric measurements have been made on one or more of them by Rossini [29], Roth and Pahlke [31], Prosen and Rossini [27], Pilcher and Chadwick [22] and Good [14]. There also are gas phase isomerization data for 2methylbutane measured by Pines et. al. [23] and reinterpreted by Scott [32]. We have adopted the selections made by Good [14], which give priority to his very careful measurements, Hexanes. There are five hexanes: n-hexane, 2-methylpentane, 3-methylpentane, 2,2-dimethylbutane and 2,3-dimethylbutane. For n-hexane there are three combustion measurements, all on the liquid, Jessup [18], Prosen and Rossini [27] and Good and Smith, [15]. These are in reasonable agreement. The results of Good and Smith [15] are selected, on the basis of improved technique and sample purity. The liquid phase value is corrected to the gas phase using enthalpy of vaporization data [Majer, 20] and an adjustment to the ideal gas state. The data for the isomers of n-hexane introduce additional complications. Four types of data are used. (1) Enthalpies of combustion which have only been measured relative to n-hexane and only for the liquids

18

HEATING VALUES OF COMPONENTS OF NATURAL GAS

[Prosen and Rossini, 26]. In effect, these are measurements of the enthalpies of isomerization. (2) Equilibrium measurements on the distribution of the isomers in the gas phase, 294–365 K [Evering and D’Ouville, 10] from which more enthalpies of isomerization can be derived, using (3) statistical mechanical thermal functions for the gases, and (4) correlated trends in enthalpies of formation among the aliphatic hydrocarbons [34]. These represent a smoothing of data that takes into account much more than the hexanes. The calorimetric, equilibrium and correlation results are in reasonable agreement for 2- and 3methylpentane, but not for the 2,2- and 2,3-dimethylbutanes. For 2- and 3-methylpentane, the calorimetric results have been adopted. For the 2,2- and 2,3-dimethylbutane, an average of the calorimetric, equilbrium and correlation data has been adopted. The recommended values carry appropriately large uncertainties. We would be pleased to see these compounds remeasured. ACCURACY OF THE DATA TODAY AND FUTURE NEEDS The estimated accuracies of the heating values at 25 °C are shown in Table 1. These are larger than the errors in the other data. They are also optimistic; as small as is reasonable. They are meant to be analogous to the 95 percent confidence level used for measures of precision. A generalization from the numbers is that the heating values are known only to 0.02 to 0.04 percent, with methane being the least well known. Any need for more accurate values will have to be met by a new measurement program, with new or improved techniques for burning the hydrocarbons. A new study will be needed if (1) calorimetric accuracy is a limiting factor, as opposed to measurement of sample mass or volume, and (2) there are improvements in the field measurement techniques for heating values. We start with the assumption that reference values should be better by a factor of 5 to 10 than those achievable in an application. Today the commonly used flow calorimeters can achieve 0.2 percent reproducibility under optimum conditions [8] and are expected to operate at 0.5 percent on a routine basis [5]. Gas chromatographic compositional analysis in the field may achieve 0.1 percent [9]. Thus, improvement in the reference heating value for methane by a factor of 4 to 10 may be a desirable investment for the future. Improvements for the trace gases are not likely to be important in the present context. REFERENCES CITED 1. 2.

3. 4. 5. 6. 7.

Armstrong, G.T.; “Calculation of the Heating Value of a Sample of High Purity Methane for Use as a Reference Material”; Nat. Bur. Standards Tech. Note 299 (1966). Armstrong, G.T.; Jobe, T.J., Jr.; “Heating Values of Natural Gas and its Components”; Nat. Bur. Standards Report NBSIR 82–2401 (1982), Armstrong, G.T.; Jobe, T.L., Jr.; “Heating Values of Natural Gas and its Components: Conversion of Values to Measurement Bases and Calculaton of Mixtures”; in Stationary Gas Turbine Alternative Fuels, ASTM STP 809, J.S.Clark and S.M.DeCorso, Eds.; American Society for Testing and Materials, Philadelphia, (1983) pp. 314–334. Armstrong, G.T.; Domalski, E.S.; Minor, J.I.; “Standard Combustion Data for the Fuel Gas Industry”; American Gas Association 1972 Operating Section Proceedings, Arlington, Va., pp D-74 to D-88 (1972). CODATA Task Group on Key Values; “CODATA Recommended Key Values for Thermodynamics 1977”; CODATA Bull. No. 28 (1978); J. Chem. Thermodyn.; 10, 903–906 (1978); Idem, draft final report (1985). Cutler-Hammer Co., Milwaukee, Wise., “Gas Measurement and Mixing Equipment”; pg R1-1 (4/1/1978). Domalski, E.S.; J. Phys. Chem. Reference Data; 1, 221–277 (1972). Dymond, J.H. Smith, E.G.; “The Virial Coefficients of Pure Gases and Mixtures: A Critical Compilation”; Oxford University Press, (1980).

19

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

Eiseman, J.H.; Potter, E.A.; J.Res. Nat. Bur. Standards; 58, 213–226 (1957). Electronic Associates, Inc., West Long Branch, N.J.; “ENCAL Energy Information Systems”; (1983). Evering, B.L.; d’Ouville, E.L.; J. Am. Chem. Soc.; 71, 440–445 (1949). Gas Processors Suppliers Association, “Physical Constants of Paraffin Hydrocarbons and Other Components of Natural Gas” in Engineering Data Book GPA 2145–77; (1977). Gas Processors Suppliers Association, “Physical Constants of Paraffin Hydrocarbons and Other Components of Natural Gas” in Engineering Data Book GPA 2145–85; (1985). Gas Processors Association; “Calculation of Gross Heating Value, Relative Density and Compressibility Factor for Natural Gas Mixtures from Compositional Analysis”; draft of GPA 2172 (1985). Good, W.D.; J.Chem. Thermodynamics; 2, 237–244 (1970). Good, W.D.; Smith, N.K.; J.Chem. Eng. Data; 14, 237–244 (1969). Haar, L.; Gallagher, J.S.; Kell, G.S.; “NBS/NRC Steam Tables”; Hemisphere Publ. Corp., Washington (1984). Holden, N.E.; Martin, R.L.; Pure Appl. Chem.; 55, 1119–1136 (1983). Jessup, R.S.; J. Res. Nat. Bur. Standards (U.S.); 18, 115–128 (1937). Knowlton, J.W.; Rossini, F.D.; J.Res. Nat. Bur. Standards (U.S.); 43, 113–115 (1949). Majer, V.; Svoboda, V.; Hala, S.; Pick, J.; Coll. Czech. Chem. Communications; 44, 637–651 (1979). Mann, D., editor; “LNG Measurement. A User Manual for Custody Transfer”, Nat. Bur. Standards, in preparation (1985). Pilcher, G.; Chadwick, J.D.M.; Trans. Faraday Soc.; 63, 2357–2361 (1967). Pines, H.; Kvetinskas, B.; Kassel, L.S.; Ipatieff, V.N.; J. Am. Chem. Soc.; 67, 631–637 (1945). Pittam, D.A.; Pilcher, G.; J. Chem Soc.; Faraday Trans. I, 68, 2225–2229 (1972). Prosen, E.J.; Maron, F.W.; Rossini, F.D.; J. Res. Nat. Bur. Standards (U.S.); 46, 106–112 (1951). Prosen, E.J.; Rossini, F.D.; J. Res. Nat. Bur. Standards (U.S.); 27, 289–310 (1941). Prosen, E.J.; Rossini, F.D.; J. Res. Nat. Bur. Standards (U.S.); 36, 269–275 (1944). Rossini, F.D.; J. Res. Nat. Bur. Standards (U.S.); 6,37–49 (1931), 7, 329–330 (1931). Rossini, F.D.; J. Res. Nat. Bur. Standards (U.S.); 12, 735–750 (1934). Rossini, F.D.; J. Res. Nat. Bur. Standards (U.S.); 15, 357–361 (1935). Roth, W.A.; Pahlke, H.; Z. angew. Chem.; 49, 618–619 (1936). Scott, D.W.; J. Chem. Phys.; 60, 3144–3169 (1974). Skinner, H.A.; Private communication; (1982). Somayajulu, G.R.; Zwolinski, B.J.; J. Chem. Soc. Faraday II; 72, 2214–2224 (1976) Taylor, B.; Cohen, E.R.; Private communications; (1983–84). “TRC Thermodynamic Tables—Hydrocarbons”, Thermodynamics Research Center, College Sta., Tx., 23–2–(i. 200); 23–2–(1,20j)–v, loose leaf, extant 1985. “TRC Thermodynamic Tables—Hydrocarbons”, Thermodynamics Research Center, College Sta., Tx., 23–2–(1, 200)–p; 23–2–(1.20j)–p, Oct. 31 (1982).

EFFECT OF C6+ HYDROCARBONS ON DEWPOINTS AND HEATING VALUES Richard F.Bukacek Institute of Gas Technology

ABSTRACT

It is shown that irreversible condensation of calibration gases due to trace amounts of C6+ hydrocarbons can be important to the calibration of calorimeters but is not likely to be significant in the calibration of devices for composition measurement. Background Gas samples in general and calibration gases in particular are subject to the possibility of condensation during storage and transit. Gaseous hydrocarbon mixtures, whether natural gases or man-made mixtures, will typically contain traces of the heavier hydrocarbons. These trace materials can raise the hydrocarbon dew point of the mixture and thereby alter the heating value of the sample if the condensation process is not fully reversible. With the high market value for natural gases, consideration of third order effects on composition can become important and several questions arise with regard to traces of the C6+ hydrocarbons. 1. What constitutes a trace amount that need not be accounted in determining a heating value? If irreversible condensation takes place in a calibration sample and trace components are present: 2. At what temperature would a significant change in composition of the vapor phase be observed? 3. Is it important to characterize the C6+ hydrocarbon traces before estimating their effect on a calibration gas?

21

Method of Study In this work, three hydrocarbon mixtures were studied. Table 1. COMPOSITION AND HEATING VALUE OF THE GASES STUDIED COMPONENT

GAS 1

GAS 2

GAS 3

Methane Ethane Propane nButane iButane nPentane iPentane Nitrogen Carbon Dioxide Heating Value

90.7% 4.0 1.0 0.3 0.3 0.1 0.1 2.5 1.0 1043.99

84.1 % 6.0 4.0 2.0

76.5% 10.0 7.0 3.0

0.5

1.0

2.4 1.0 1147.1 6

1.5 1.0 1271.04 BTU/SCF

In all tables, the higher heating values reported are calculated by the method of IGT Bulletin 32 and NGPA publication 2145–71. The Soave/Redlich/Kwong equation of state was used to calculate the dewpoints of these mixtures and the equilibrium composition of their vapors at temperatures below the dew points. The dew point of these mixtures was calculated at several pressures. On the basis of 1 mole of mixture, 0. 0001 mole of trace component was added and the mole fractions of the resulting mixture were then calculated. To see the effect of different trace materials, the trace 0.0001 mole was nC6, nC8 or nC10. That is, the trace components were characterized by assuming their properties those of a single component, either nC6, nC8 or nC10. An iterative process was used to establish the temperature at which a mixture with the trace component produced, at equilibrium, a vapor phase with a heating value 0.05% less than the heating value of the base mixture. That is, for Gas 1, the temperature was found at which vapor phase heating value was 1043.47 BTU/SCF with trace components present at the level of 0.01%. By this procedure the effect of irreversible condensation can be shown and the range of temperatures in which it might occur is illustrated. The basis for this procedure is as follows: If a calibration gas contains trace components that cause condensation that is irreversible, the liquid formed will extract components from the vapor phase that are intended for calibration. If the the most important purpose of calibration is to establish heating values, then the most appropriate criterion of significant change is the effect on heating value measured or calculated. Results Results of computation are summarized in Tables 2 through 11. In these tables T* is the temperature at which the vapor phase heating value is 0.05% smaller than that of the base mixture. Hb is the heating value of the base gas, Ht is the heating value of the mixture that includes the trace material, and H* is 0.9995 (Hb). DP (base) is the dew point of the base mixture, and DP (trace) is the dew point of the mixture which includes the trace. All temperatures are degrees Celsius, and pressures are atmospheres.

22

EFFECT OF C6+ HYDROCARBONS ON DEWPOINTS AND HEATING VALUES

Tables 2, 3 and 4 summarize results for Gas 1 . TABLE 2. GAS 1, 0.01% nC6 AS TRACE, Ht=1044.44 BTU/SCF Pressure, Atmospheres

20 30 50

DP (base)

DP (trace)

T*

C

C

C

−34.8 −30.2 −30.2

−32.0 −28.7 −27.9

−33.8 −30.4 −29.5

TABLE 3. GAS 1, 0.01% nC8 AS TRACE, Ht=1044.59 BTU/SCF Pressure, Atmospheres

20 30 50

DP (base)

DP (trace)

T*

C

C

C

−34.8 −31.2 −30.2

−12.8 −11.9 −12.9

−29.7 −26.1 −25.2

DP (base)

DP (trace)

T*

C

C

C

−34.8 −31.2 −30.2

17.0 17.4 15.8

−28.5 −24.5 −22.9

TABLE 4. GAS 1, 0.01% nC10 AS TRACE, Ht=1044.74 BTU/SCF Pressure, Atmospheres

20 30 50

Examination of Tables 2, 3 and 4 shows that for Gas 1, cooling to until the heating value of the vapor is 0. 9995 that of the base gas brings the temperature quite near the dew point of the base gas. When heating value is calculated from composition it seems unlikely that gases with very low dew points like Gas 1 would suffer a heating value loss as large as 0.05% because the base dew point is below the temperatures a sample is likely to experience, and totally irreversible condensation seems unlikely. Although trace components characterized by nC10 or higher hydrocarbons can raise the dew points can be raised to summertime air temperatures, the temperature at which calibration components would significantly move to the liquid phase is much lower. However, cooling to −25 C with nC10 as trace yields a vapor phase with a heating value 1.27 BTU/SCF smaller than that of the starting material, a loss of 0.12% in the heating value of the mixture that includes the trace. Thus a distinction must be made between a gas sample used to calibrate a calorimeter and one use to calibrate a device measuring composition. Tables 5, 6 and 7 summarize results for Gas 2.

23

TABLE 5. GAS 2, 0.01% nC6 AS TRACE, Ht=1147.61 BTU/SCF Pressure Atmospheres

DP (base)

DP (trace)

T*

C

C

C

3 6 9 12

−39.7 −28.3 −21.4 −16.4

−39.3 −27.9 −20.5 −15.5

−38.7 −27.4 −21.0 −16.0

T*

TABLE 6. GAS 2, 0.01% nC8 AS TRACE, Ht=1147.76BTU/SCF Pressure, Atmospheres

DP (base)

DP (trace)

C

C

C

3 6 9 12

−39.7 −28.3 −21 .4 −16.6

−25.5 −16.4 −10.8 −6.9

−35.8 −24.4 −17.5 −12.6

T*

TABLE 7. GAS 2, 0.0156 nC10 AS TRACE, Ht=1147.91 BTU/SCF Pressure, Atmospheres

DP (base)

DP (trace)

C

C

C

3 6 9 12

−39.7 −28.3 −21 .4 −16.4

1.7 9.5 13.9 16.8

−34.6 −22.6 −15.3 −10.1

As expected, the effect of increasing molecular weight of the trace components is to raise the dew point of the mixture. Also in accord with the results for Gas 1, although dew points can be raised to summertime air temperatures by the presence of traces characterized by nC10 and heavier, the effect of irreversible condensation on composition is apt to be very small. For example, suppose a sample of Gas 2 is stored at 12 atmospheres and trace materials characterized by nC10 are present in amount 0.01% . From Table 7 we see even if this sample was cooled to −10 C and irreversible condensation occurred, this gas used to calibrate a chromatograph would be accurate to within 0.7 BTU/SCF, 0.05%, for calculation of heating value. If this same gas were used to calibrate a calorimeter, the maximum possible reduction of heating value due to irreversible condensation would be from 1147.91 to 1146.59 BTU/SCF, a loss of 1.3 BTU/SCF or 0.12%. Tables 8, 9 and 10 summarize results for Gas 3. TABLE 8. GAS 3, 0.01% nC6 AS TRACE, Ht=1271.43 BTU/SCF Pressure, Atmospheres

DP (base)

DP (trace)

T*

F*

C

C

C

Moles/Mole

BTU/SCF

2 4 6

−36.2 −24.2 −16.6

−35.67 −24.0 −16.1

−36.0 −23.7 −16.4

0.00044 0.00046 0.00048

H1

3580 3500 3420

24

EFFECT OF C6+ HYDROCARBONS ON DEWPOINTS AND HEATING VALUES

Pressure, Atmospheres

DP (base)

DP (trace)

T*

F*

H1

C

C

C

Moles/Mole

BTU/SCF

8

−11.0

−10.6

−10.9

0.00049

3310

H1

TABLE 9. GAS 3, 0.0155 nC8 AS TRACE, Ht=1271.58 BTU/SCF Pressure, Atmospheres

DP (base)

DP (trace)

T*

F*

C

C

C

Moles/Mole

BTU/SCF

2 4 6 8

−36.2 −24.2 −16.6 −11.0

−26.6 −16.4 −9.8 −5.0

−32.9 −21.0 −13.6 −8.2

0.00043 0.00046 0.00048 0.00050

4020 3860 3740 3640

H1

TABLE 10. GAS 3, 0.01% nC10 AS TRACE, Ht=1271.73 BTU/SCF Pressure, Atmospheres

DP (base)

DP (trace)

T*

F*

C

C

C

Moles/Mole

BTU/SCF

2 4 6 8

−36.2 −24.2 −16.6 −11.0

−1.8 6.5 11.5 15.0

−31.3 −18.7 −10.9 −5.2

0.00040 0.00042 0.00044 0.00046

4570 4415 4290 4190

The pattern of condensation for Gas 3 is very like that of Gases 1 and 2. Shown in tables 8, 9 and 10 are the fraction of sample that has condensed at T* and the heating value of the equilibrium liquid, H1. It is of interest to note that at T* the moles condensed are 4 to 5 times greater than the amount of trace material, even though T* is always greater than the dew point temperature of the base gas. This illustrates the point made earlier that the non-trace components dissolve in the liquid formed. Thus, the heating value of the liquid is always much less than that of the trace materials because the lighter hydrocarbons dissolve in the liquid formed and dominate its composition. It will be observed that in all cases, T* approaches the dew point temperature of the base gas. That is, even when the dew point temperature is raised greatly by trace components and the liquid formed is predominantly materials from the base gas, the most important fact is that the amount of liquid formed at temperatures above the dew point is quite small so long as the amount of trace materials is small. This is illustrated in Table 11 where the amount of trace material is doubled to 0.02%. TABLE 11. GAS 1, 0.02% nC10 AS TRACE, Ht=1045.42 BTU/SCF Pressure, Atmospheres

15 30 40

DP (base)

DP (trace)

T*

F*

C

C

C

Moles/Mole

−34.8 −31.3 −30.2

27.4 28.3 27.1

−23.2 −18.8 −17.3

0.00056 0.00063 0.00071

25

It can be seen from Table 11 that the now 10 or 12°C greater than the dew point The amount of liquid formed at T* is still seems unlikely that calibration of devices measurement would be affected by irreversi under the usual conditions of storage and Also, the heating value with the trace nC adds 1.43 BTU/scf to the heating value of the base gas, 0.13% of its heating value. The effect of characterizing hydrocarbon on the contribution of trace components to calorimetric measurements is shown in Table 12. TABLE 12. CONTRIBUTION OF TRACE COMPONENTS TO HEATING VALUE Characterizing Hydrocarbon

Molar % that contributes 0.1% to Heating Value

nC8 nC10 nC12 nCl4

0.018 0.014 0.012 0.010

It is clear that if a gas is to be used for calibration of a calorimeter, then it will be necessary to determine the amount of trace components present and characterize their heating value if accuracy in heating value is to be better than 0.1%. Conclusions The results of this work suggest the following: 1. It is important to distinguish between the uses made of calibration gases: Trace components summing to 0.0156 are not likely to affect calibration of devices for measuring composition so long as the gas free of trace components would not have reached its dew point temperature before use. For gases used to calibrate calorimeters it will be important to determine the amount of trace components and characterize their heating value if accuracy of the order of 0.1% in heating value is to be obtained. 2. The dew point of the base gas is a fair measure of the temperature at which irreversible condensation would significantly affect a gas used to calibrate a device for measuring gas composition. 3. The possibility of irreversible condensation in gases used for calibration of devices for measuring composition should not constitute a significant problem so long as the dew point of the base gas is below the temperatures the gas is likely to experience during storage and transit, and the total of trace components are present in amount of the order 0.01%.

EFFECT OF WATER VAPOR ON HEATING VALUE Robert J.Rau Division Measurement Manager Transcontinental Gas Pipe Line Corporation Baton Rouge, Louisiana 70806

ABSTRACT

This paper discusses the basic principles, laws and methods used to calculate heating value and the effects of water vapor on these calculations. Methods shown include volumetric corrections as well as heating value corrections using the IGT method. Introduction Today, this particular subject that has shifted somewhat to a back burner due to Order 93 not being put into effect. Some specific contracts and companies are still doing this. Water, in a natural gas pipeline, is and has a detrimental effect on various items. Some of these items are: a. Hydrate formulation b. Corrosion and internal corrosion c. Accumulation of liquid water d. Medium for solids movement. The above items caused varied problems such as inaccurate measurement, flow rate interruption, equipment repair and failure, loss of capacity, loss of control of pressure and instituting a corrosion protection plan of high cost. As we all know, this causes loss of revenue and added operating costs.

27

Water Vapor Determination This particular phase of testing can be done by various methods, but the most widely used methods are visual and electrolytic water analyzers. Both systems are presently being used in the industry. Gas Sampling Gas sampling is probably the most important thing that can affect the water determination results. Errors in sampling could be a major contributing factor to either high or low results. A probe is necessary to get a good representative result. The four main considerations in a sampling system design are: 1. 2. 3. 4.

Location of sample probe Moisture build-up within the system and its retention Atmospheric as well as flowing conditions Materials used in sample probes and related equipment (length of sample line is also important). Effect of Water Content on Btu Determination

First, it should be noted that by several basic laws water determination is developed. These are Dalton’s Law of Partial Pressures, Boyle’s and Charles’ Law. These are basics and do not need further discussion. Dew point is defined as the temperature at which water vapor begins to condense out of gas. Gas can be characterized in three conditions. These are dry, partially saturated and saturated. Theoretically, the maximum volume fraction and percent of water that a gas can contain is directly proportional to the partial pressure of water at the temperature of the gas water mixture. In short, the mol fraction of water vapor in a volume would be:

where— W=mole percent of water vapor Vp=vapor pressure of water at dew point temperature of the mix. P1=total pressure on the volume psia. This condition holds true at low pressure but has some deviation at higher pressures. This can be shown by the following example. Basic Data 1. Gas Volume=1,000,000 cu ft=1 MMSCF 2. Dew Point=55°F 3. Pressure 1000 psig W=.21397=.0002109 1000+14.40 Vol H2O=.002109×1 MMCF=211 cu ft Dry Gas Vol=1,000,000–211=999,789 SCF A water vapor graph shows a content of 18.0#/MMCF. An established value is that 1 lb/MMCF=21.0181 cu ft.

28

EFFECT OF WATER VAPOR ON HEATING VALUE

Therefore— There is a 44.2% difference. As we all know, temperature changes occur all along in a pipe line for various reasons. If a gas is not allowed to drop below the dew point temperature, no fall out will occur, but if it drops below the dew point, water will condense. For this reason pipelines should have a minimum temperature that they can operate at. The main equations used at present are as follows: 1. Water Volume Fraction

where— Pw=pressure of water vapor (psia) Pt=total pressure psia Vw=volume of water vapor SCF Vt=total gas volume SCF Sample Problems Dew point temperature=55° at 200 psig From vapor pressure table—Pw=.2130 2. Also, a common factor used for calculation showed 1 Ib of water at 14.73 psia and 60° equals 21.0181 cu ft. Therefore utilizing the standard manner of stating water content used by the gas industry then

3. Several different concepts must now come into play to utilize the water content. These concepts are: a. Water vapor in saturated gas b. Water vapor above contract limit c. Water vapor as measured d. Example of water vapor in saturated gas Gas volume=1,000,000 MCF Dew point=flow temp.=70° at 1000 psig #/MMCF=24/MMCF Fw=24×.000021=.0005 vol fraction Wv=1,000,000×.0005=500 MCF e. Utilizing the above same problem but including the correction for contract limit, Fw=(24−7)×.000021=17×.000021=.00036 Wv=1,000,000×.00036=360 MCF This means .036% water vapor.

29

When gas is not dehydrated, it is assumed to be saturated. Under these conditions, the volume calculation can be factored to give a monthly average pressure and temperature, thus making correction easy. So far, the methods discussed have utilized volume corrections that can be subtracted from measured volumes but IGT developed an empirical method that corrects dry Btu basis to wet basis. This formula is as follows:

where—

An example of this method is as follows: 1. Gas Temp=78°F Gas Press=500 psia Btu=1020 dry @ 14.73 psia A. A=22,500 from Table B=NGPA ref. std. manual. B=9.11 B. Pressure Base Correction C. Correction of A & B for press, base, 1. 2. D. Water Content E. F. Corr. Btu=1020×.99886=1019 Btu corrected for water vapor af sat. gas at flowing conditions. G. MMBtu=BTUcorr×MCF/1000 with same press. base. The major items that must always be considered so that errors in payment for Btu1 s do not occur can be summarized as follows: 1. Always have the Btu on the same basis as volume. If Btu dry volume must be dry and vice versa. 2. Always have Btu and volume on the same pressure basis. The following may be used as a guideline in summarizing Btu conversion from wet to dry. Most calorimeters record Btu/cu ft at 14.7346 psia and 60°F saturated with water vapor. The following corrections need to be made. 1. Press, base factor= 2. Contract Sat. Btu=Calorimeter Btu saturated×press. base factor

30

EFFECT OF WATER VAPOR ON HEATING VALUE

3. If all water were taken out of saturated gas, the dry Btu would increase by factor of 1.0177. 4. In some cases, 7#/MMCF is considered dry by contract and an adjustment must be made for more than 7#. Therefore dry Btu: A. B. C. An example of this calculation is Calorimeter meas. Btu = 1005 Btu/cu ft (Sat) Press. Base=15.025 psia Contract water content 7#/MMCF Actual water content 54#/MMCF Delivered Btu=

Table 1 shows that the major item to watch in use of Btu’s for calculation is the different bases involved. In short use oranges with oranges and apples with apples, and not oranges with apples. Table 1. DIFFERENT BASES FOR CALCULATION USING BTU’S Contract

Calorimeter Measured

Chromatograph Measured

Condition

Converted From 14.735 Wet

Converted From 14.73 Dry

14.65 dry 14.65 wet 14.73 dry 14.73 wet 14.735 dry 14.735 wet 15.025 dry 15.025 wet

1.0118 0.9942 1.0174 0.9997 1.0177 1.000 1.0377 1.0197

0.9946 0.9773 1.000 0.9826 1.0003 0.9829 1.0200 1.0023

Pressure Base Factor (Pbf) = References 1. 2. 3.

R.N.Curry, “Btu/Water Vapor Content Adjustment,” 1980 American Gas Association Fall Meeting. W.F.Barker, “Determination of Water Vapor Content and Correcting for Its Effect on Volume and/or Btu Determination.” AGA Transmission Conference, 1981. Gas Processors Association Technical Reference Bulletin 181–81, Technical Standards Book, 1981.

PREPARATION OF STANDARDS FOR GAS ANALYSIS G.C.Rhoderick and E.E.Hughes Center for Analytical Chemistry National Bureau of Standards Gaithersburg, Maryland 20899

ABSTRACT

A primary standard is prepared by an absolute method with the gravimetric method being the preferred technique. The basic technique involves weighing minor and major components into a cylinder. This basic method can be modified to prepare samples at very low concentrations (parts per trillion) and containing very specific and complex mixtures such as hydrocarbons. The uncertainty in the preparation procedure may be determined by preparing a “family” of primary standards and intercomparing those standards by analysis. As will be illustrated, it is very important to have standards which are very close and bracket the concentration of the unknown in question. This is especially important when analyzing natural gas. In this case, two methods may be used to determine the concentration of methane in natural gas. However, one method will lead to a much lower uncertainty and thus, a lower uncertainty in the BTU value of the natural gas. INTRODUCTION A standard is a material which allows a measurement to be made in such a way that there exists a pathway between the measurement and one of the fundamental units of measurement. A Standard Reference Material (SRM), issued by the National Bureau of Standards, is such a material. Through the use of standards any two independent measurements may be compared providing each has a traceability pathway to the same fundamental unit of measurement. The pathway may be through the use of an absolute standard and a precise method of measurement or an absolute method of measurement alone. In the case of gases there are few absolute methods which are applicable and the usual approach to standardization is to use a standard which is compared to the sample by use of a precise method of intercomparison. Standards for gas analysis may be prepared by mass, pressure or volume, all of which are

32

PREPARATION OF STANDARDS FOR GAS ANALYSIS

ultimately calibrated by mass. The most direct and accurate method is by mass (gravimetry) in which each component is separately weighed into a gas cylinder. Preparation of standards by pressure is fast and somewhat less accurate than gravimetry and is further complicated by uncertainties in the corrections for the non-ideality of gases and gas mixtures. Standards prepared by volume are generally less accurate. EXPERIMENTAL Gravimetric Preparation There are three major factors which contribute to the error of a gravimetric standard. These are: 1) weighing errors, 2) errors in the assessment of the purity of reagents and 3) instability of the mixture in the container. The first two sources of error can be minimized by careful attention to the details of the weighing process and by rigorous analysis of the reagents. The last can only be observed experimentally. Weighing Procedure and Errors Gas containers are necessarily large and heavy but the constituents added to them are often quite light in weight compared to the container. Consequently, the balance used in the gravimetric preparation of standards must have a high capacity and high sensitivity at full load. We normally use a balance with a capacity of 10 kg and a sensitivity of 1–2 mg at that load. When preparing a mixture the weight of the minor component will be small compared to the weight of the container and the error resulting from buoyancy effects and/or exterior surface absorption of moisture may be considerable. Thus the weighing process must be designed to minimize these effects. A tare of almost identical volume and surface is used against which the cylinder containing the sample must be weighed. Both barometric pressure and temperature should be measured during each weighing in order to recognize changes in ambient conditions that might require adjustments to the observed weights. Finally, it is good practice to carry a blank or control cylinder through the process so that the effect of changing environment can be recognized. In practice a 1% by mole gas mixture would be made by adding several grams of the minor component to a preweighed cylinder. If for example 3.000 g were added, the uncertainty of the weighing on the balance mentioned above would be ±0.002 g (0.07%). It would be necessary to add about 300 g of the major component, (assuming that both minor and major components have similar molecular weights) which could be weighed to ±0.002 g or with an error of 0.0006%. Thus the controlling error is the uncertainty in weighing the minor component. Mixtures of lower concentration may now be prepared by dilution of portions of the 1% mixture. The weighing errors will accumulate depending on the number of dilutions required to reach the final concentration. However, we have modified the gravimetric method in such a manner that concentrations as low as 100 ppm can be prepared in one attempt. We refer to this as the two balance method. A closed-end stainless steel tube with a valve on one end is evacuated and weighed. The minor component is then introduced into the tube and then weighed on a microbalance with a readability to 0.01 mg. The cylinder in which the standard is to be made is evacuated and weighed on a balance with a 10 kg capacity and readability to 1–2 mg. The minor component is then added to the cylinder and pressurized with the major component. Think of this procedure as analogous to the preparation of a standard solution where one component is weighed into a container from a small weighing bottle and is then diluted with a measured volume (really mass) of the major component. This procedure results in a very small error due to weighing. For example, a mixture

33

requiring 0.5 g of the minor component weighable to ± 0.00001 g results in an error of only 0.002% in weighing. Reagent Purity and Errors Reagent purity is very important when preparing gravimetric standards, especially at very low concentrations. The most troublesome is the measurement of the minor component when it is present as an impurity in the major component. For instance, if a 10 ppm mixture of carbon monoxide in nitrogen is being prepared from reagent nitrogen containing 1 ± 0.1 ppm carbon monoxide then the uncertainty in the final mixture due first to the error in analysis of the reagents is 1%. On the other hand, the amount of nitrogen in the “pure” carbon monoxide is usually not as significant. A typical supply of carbon monoxide may have 0. 10 ±0.01% nitrogen present, which would result in only a 0.01% relative error for a 10 ppm mixture. Stability of Mixtures The next consideration is stability. There are two types of stability: 1) short term which considers adsorption and absorption which occur on the interior surface as soon as the gas is admitted to the cylinder and 2) long term which involves slow reactions between the components or between the container (+H2O) and a component. To test for short term stability, the sample is analyzed as soon as it is prepared. Then some of the sample is transferred to a new or treated cylinder and the contents of this cylinder is then analyzed. If the measurements for the two cylinders agree within the precision of the method of intercomparison then stability is assumed. When determining long term stability, a set of primary standards is prepared and the sample is compared to the standards. Then at some time interval, say 3 months, a fresh primary standard is prepared and the sample and all standards are compared. If the sample is within 0.5% of the original value and the primary standards agree to within 0.2%, then stability of the sample may be assumed. Total Error of the Standard The total uncertainty depends on the uncertainty of the various steps in the preparation. The uncertainty of the gravimetric standards can only be inferred in the absence of an “absolute” method and this is done by preparing a “family” of standards and intercomparing them. While an absolute value cannot be obtained, considerable information is obtained by use of a relative method whose response characteristics are known. Of course it would be necessary to randomize the analyses and the preparation as much as possible to reduce the presence of “biases”. (Different reagents, operators, methods of analyses—GC or NDIR). Table 1 illustrates the precision of mixing (±0.4%) but also reveals a possible bias at the low end which would have resulted from a trace of CO undetected in the diluent nitrogen, a reason for reanalyzing or substituting another sample of nitrogen. Total Errors for a Sample When analyzing a sample against primary gravimetric standards there are other errors as well as the uncertainty of the standards that are involved in the total error for the sample. The major contributing error is the imprecision involved in comparing a sample to the standards. Typically, for a sample of carbon dioxide in nitrogen, the following errors apply: the impurity of the carbon dioxide is 0.03 ppm, the imprecision of the preparation of the gravimetrics is 0.13 ppm (maximum), and the imprecision of

34

PREPARATION OF STANDARDS FOR GAS ANALYSIS

intercomparison is 0.09–0.17 ppm. Summed in quadrature these errors total about 0.2–0.3 ppm or about ±0. 1% relative for a 300 ppm sample. Organic Standards Preparation Technique What has been discussed is general and describes the essentials of producing a gas standard. Applications to specific materials such as the preparation of very low concentration hydrocarbon standards requires considerable modification. This is the case for preparing organic standards where a small known quantity of a liquid must be introduced into a cylinder. The process of preparing gravimetric organic standards consists essentially of the elements of the two balance method. The organic material is sealed in a small glass tube (20 mm×1.6 mm) in which it is weighed. It is transferred to a large cylinder to which is also added a weighed amount of nitrogen or other diluent. In practice, a capillary tube (100 mm length by 1.6 mm o.d.) is drawn out to a fine point using a microburner flame. About 20 mm from the drawn out end, the tube is again drawn out to a fine point and then broken off from the main tube. One end of the tube is sealed in the flame yielding a tube about 20 mm long. Control tubes sealed on both ends are prepared at the same time and are used to detect any drift in the balance during the weighing of the tubes. The tubes are weighed empty on a microbalance (sensitivity to 0.1 microgram) using a scheme where the control is weighed first and last. Corrections are made to the weights of the tubes due to any drift occuring between the first and last weights of the control tube. It is also good practice to carry along through the whole process some empty tubes so that the effect of changing environment can be recognized. The tubes are weighed several times in this manner and an average empty weight is calculated with a known precision. In order to get the organic liquid into the capillary tube, the tube is placed open end first into a vial containing the liquid. A syringe is then fitted over the vial and the plunger is pulled out a small amount to pull air out of the capillary tube. When the plunger is released, liquid displaces the air that was pulled out of the capillary. The capillary is then removed from the vial and placed in a centrifuge to force the liquid to the sealed end of the tube. Then the open end is heat sealed. A blank tube, simply a tube with no liquid but sealed on both ends, is also prepared. This tube is important in that it determines whether there is any weight loss due to sealing of the open end of the tube after the liquid is introduced. The average weight loss due to sealing is about 0.0013 mg ±0.0008 mg. Therefore, for 0.5 mg of organic material, this results in a relative error of 0.2%. After the tubes are sealed, they are weighed in the same manner as for weighing empty. An average weight is determined and the weight of the organic is calculated by difference. Corrections are made for loss of weight due to sealing. Table 2 shows the weights for empty and filled tubes for a typical mixture of benzene at approximately 10 ppb. These data show that the tubes can be weighed to within 0.001 mg standard deviation or as in this case for 0.432 mg of benzene, 0.2% relative standard deviation. The total error of weighing, which takes into account the standard deviation of weighing the benzene and the blank tube, is about 0.5% relative which is excellent at this level. Typically, small cylinders (0.1 or 30 cubic foot capacity) are used for gas standards. However, for preparing ppb level organic standards, a much larger cylinder, such as an aluminum cylinder of 150 cubic foot capacity, is employed. Thus more organic material can be introduced into the cylinder which results in smaller errors due to weighing. The cylinder is pressurized and flushed several times with dry nitrogen to remove any organic residual left from the manufacturing and cleaning process. The cylinder is then evacuated to about 1.0×10−2 cm. An electronic toploading balance of 54 kilogram capacity and a 1 gram sensitivity is used to weigh the cylinder empty. Then a nut and nipple, which is packed with stainless steel

35

filings to keep glass out of the cylinder, are attached to the cylinder valve and a piece of tubing of 1.6 mm i.d. and about 30 mm long is attached to the nipple. The capillary tube containing the Table 1: Sensitivities of Carbon Monoxide Gravimetric Standards Concentration

Sensitivity

ppm

peak area/ppm

9.33 10.30 47.09 51.33 94.54 103.54 477.9 477.9 517.7 1007.

1109 1109 1103 1106 1 100 1 101 1098 1098 1097 1101 1102 R.S.D. = 0.4%

Table 2: Weights of Benzene for Typical 10 ppb Mixture Tube wt. ,mg

Weights of Unfilled Tube

Average = S1 (Std. Dev.) = Weight of Filled Tube

Average = s2 (Std. Dev.) = Weight Differencea Imprecision aSample

Benzene

Blank

31.326 31.325 31.326 31.326 0.001 31.758 31.759 31 .758 31.758 0.001 0.432 0.5%

30.650 30.647 30.651 30.649 0.002 30.648 30.648 30.648 30.648 0.001 −0.001

weight uncorrected for blank due to weight=

bImprecision

organic sample is then fitted into the tubing and results in a tight enough fit to prevent room air from being pulled past the tube. The cylinder valve is then opened and the capillary tube is broken by manual force on the end nearest to the valve. The tube is heated slightly in order to vaporize the organic liquid at which point it is pulled into the cylinder by the vacuum. After all the material has been vaporized the other end of

36

PREPARATION OF STANDARDS FOR GAS ANALYSIS

the capillary tube is broken and room air is allowed to flush any residual organic material into the cylinder. As many as nine tubes, each containing a different organic component, have been added to a single cylinder by this technique. The cylinder is then weighed to determine the amount of air introduced, A matrix gas, nitrogen, is then added to the cylinder. Typically a standard will contain about 4200 grams of nitrogen matrix gas. The balance is readable to 1 gram, thus the error in weighing the matrix gas is about 0.02%. Thus the controlling factor for weighing error lies in the weighing of the organic material. Blank cylinders have been prepared in which no organics were introduced but room air was pulled into the cylinder to atmospheric pressure. The blank cylinder was then pressurized with nitrogen and analyzed to determine if any detectable impurities were introduced into the cylinder by this technique. No organic impurities were detected, with the limit being 0.05 ppb for benzene which results in a maximum error due to impurities in the matrix gases of ±0.5% for a standard at 10 ppb. Several gravimetric standards containing benzene in the 5–100 ppb concentration range were prepared and then intercompared. The data were then plotted, concentration versus area and correlated by linear regression. The mean percent difference of the calculated and expected concentrations was determined to be ±0.9%. That value takes into account all the errors in weighing as well as any systematic errors, such as any loss of the organic material during the transfer to the cylinder which can not be estimated. This value should be larger than the estimated uncertainty of the gravimetric standards and it is a better estimate of the uncertainty of the gravimetric standards. Taking all the other errors such as the imprecision of replicate analysis of a sample and the error of intercomparing a sample to the standards, the total uncertainty of a typical benzene sample at this level is about 2.0% relative. Natural Gas Standards In order to prepare gravimetric standards of natural gas we used a combination of the above techniques. A modification to the organic method was used in which a Pasteur pipette was employed, rather than a capillary tube, to obtain a large enough quantity of the liquid hydrocarbons for the concentrations needed. The tubing was attached to the cylinder and the pipette containing the liquid hydrocarbon was fitted into the tubing. This “free” end was then connected to a cylinder of pure methane. After the pipette was broken and the liquid hydrocarbon vaporized and pulled into the cylinder by the vacuum, the pipette was flushed with methane. The hexanes and pentanes were added first in this manner to an evacuated, weighed cylinder. Then the other hydrocarbons were added one by one starting with the lowest vapor pressure and ending with the hydrocarbon of the highest vapor pressure. Concentrations were calculated by mole using the weights of each of the compounds added. A preliminary analysis was done on an unknown sample of natural gas to determine a nominal concentration of the methane. Then several primary standards of simulated natural gas ranging from concentrations of 75 to 100% methane were prepared to evaluate the unknown more accurately. Two of the standards were prepared to “bracket” the methane concentration of the unknown to within 1%. The standards and the unknown were intercompared using a gas chromatograph equipped with a thermal conductivity detector. The data were plotted, concentration versus area, and correlated by linear regression. The concentration of the unknown was determined from that line. The uncertainty of those standards was calculated to be 0.1% which represents the mean of the differences in observed versus expected concentrations. The concentration of the methane in this unknown was calculated to be 90.64+0.36% (0.4% relative uncertainty). As a check on the analytical value, the concentration of methane can also be determined by difference. All the other compounds in the natural gas sample were determined by comparison to the primary

37

standards. The concentrations and uncertainties were calculated and then the total concentration and uncertainties were calculated for those compounds minus the methane. The methane was then calculated simply by subtracting the total concentration for the components from 100, The uncertainty for the methane was calculated by summing all the individual uncertainties. Since these uncertainties represent the maximum and we have determined the concentration and uncertainty of all components present in the unknown, we can use the total uncertainty of all the compounds as the uncertainty in the methane value. Thus the concentration for methane as determined by difference is 90.64±0.06% (0.07% relative, table 3). This method is valid only if the sample is analyzed for other impurities and those impurities are taken into account in the total value. The response for the methane in the unknown was ratioed to each primary standard with the results giving us a range of values from 90.10% to 90.96%, or about 1% relative differences. Let us consider that the natural gas industry would like to determine the heating value (in BTU) of natural gas to about an uncertainty of 0.5 BTU. Then with such a range of values between standards, 1%, the uncertainty of the methane in this case would be about ±9 BTU. If we determine the concentration by difference, the same concentration for methane is obtained but the uncertainty that can be applied is substantially different: 90.64 +0.06%. In terms of BTU this would result in an uncertainty of ±0.6 BTU. Table 3: Concentrations of a Natural Gas Sample as Determined by Comparison to Gravimetric Standards Component

Concentration, % by molea

Ethane Propane n-Butane i-Butane n-Pentane i-Pentane n-Hexane Carbon Dioxide Nitrogen Totals = Methane Methane

4.008 ± 0.011 1.008 ± 0.007 0.301 ± 0.005 0.298 ± 0.005 0.101 ± 0.004 0.100 ± 0.004 0.052 ± 0.003 0.997 ± 0.007 2.499 ± 0.009 9.364 ± 0.055 90.64 ± 0.36 by GC 90.64 ± 0.06 by difference

aUncertainties

following the concentrations represent the 95% confidence limits.

CONCLUSIONS Primary standards may be prepared with very good accuracy and very low uncertainties. By modifying the basic method for preparing gravimetric standards, very complex and specific mixtures can be prepared. It is possible to prepare mixtures at very low concentrations, with the limiting factors being the uncertainty that can be tolerated and the sensitivity of the instrument or detector to be used for analysis. It is very important that, when making primary standards for the analysis of an unknown, the characteristics of the instrument, such as linearity, be considered. Generally, standards that bracket the

38

PREPARATION OF STANDARDS FOR GAS ANALYSIS

concentration of the unknown to approximately ±1% are sufficient. This can be seen with the methane in natural gas. The individual standards, ranging from 75 to 100%, when ratioed against the unknown gave a range of concentrations of ±1%. This possibly shows the error in the preparation of the standards, a slight detector non-linearity, or detector saturation over this concentration range. Thus, the difference method for determining the methane is the more suitable method for determining the concentration. This method also results in a much lower uncertainty in the concentration, which in turn shows up in the BTU uncertainty of the natural gas: ±0.6 BTU compared to ±9 BTU by direct analysis. BIBLIOGRAPHY Schmidt, W.P. and Rook, Harry L., “Preparation of Gas Cylinder Standards for the Measurement of Trace Levels of Benzene and Tetrachloroethylene,” Analytical Chemistry, 55(2): 290–294 (1983). Rhoderick, George C., Cuthrell, William F. and Zielinski, Walter L. Jr., “A Gravimetric Technique for the Preparation of Accurate Trace Organic Gas Standards,” Quality Assurance in Air Pollution Measurements, 239–246, 1985.

KEY WORDS gravimetric, primary standard, absolute method, intercomparison, traceability pathway, Standard Reference Material

NEW STANDARDS FROM THE INSTITUTE OF GAS TECHNOLOGY Bruce H.Solka Senior Analytical Chemist and Amir Attari Associate Director, Chemical Research Services Institute of Gas Technology Chicago, Illinois 60616

ABSTRACT

Since 1961 the Institute of Gas Technology (IGT) has provided the gas industry with natural gas standards of certified heating value and specific gravity for calibration of calorimeters and gravitometers, respectively. The establishment of this program at IGT was the result of an earlier project initiated by A.G.A. in cooperation with the National Bureau of Standards (NBS), This program was recently expanded at IGT with the support and the initiative from the Gas Research Institute (GRI) to produce additional natural gas standards including a gas chromatographic calibration standard gas. INTRODUCTION For the past 20 years, the Institute of Gas Technology (IGT) has provided the gas industry cylinders of standard gas with certification of heating value and specific gravity of the contents. These cylinders have been widely used in to calibrate their measurement instruments. During this 20 year period, however, users have seen many changes in the requirements of their natural gas characterizations. One significant change has been the increased utilization (and hence, measurement) of fuel gases having heating values higher than the nominal 1000 Btu/SCF. Another change is the growing interest in obtaining detailed compositional analyses of natural gas by routine gas chromatography (GC). These developments have resulted in industry demands for reliable calibration standards for gas quality measurements by GC. We at IGT have been made aware of this demand by increasing number of inquiries from our current certified gas customers. As a result of these inquiries, in 1981 we mailed a questionnaire to 55 of the largest customers of the present program. This questionnaire addressed their anticipated use of reference standard mixtures for gas chromatographic analysis of natural gas. Of the 31 respondents, 80%

40

NEW STANDARDS FROM THE INSTITUTE OF GAS TECHNOLOGY

were using GC at that time. Satisfaction with available standards was mixed. A clear majority desired NBS traceability of standards as well as calorimetric verification of the heating value of the GC standard mixture. The Gas Research Institute (GRI) also became aware of industry interest in new reference standard mixtures certified by an independent agency. As a result, in August of 1984, GRI initiated the present program at IGT with the ultimate goal of supplying the gas industry with a series of NBS traceable calibration gases. This talk is a summary of the present status of that program. The program’s goal is to make available several types of calibration standards to the gas industry by the end of 1985. These include a GC calibration mixture with direct NBS traceable certification of composition and similarly traceable calorimetric calibration standards with heating values in range of 800 to over 1200 Btu/SCF. Selection of GC Mixture Composition Selection of Mixture Composition, GC Standard.A GC calibration gas will be most useful if its composition approximates that of the test sample, as closely as possible. The survey referred to above gave respondents the opportunity to express their wishes regarding composition of a GC calibration standard. With this information as well as our own knowledge of typical composition of pipeline-quality natural gas, numerous possible gas mixtures were considered as candidates for this program. Given recent interest in the heavier ends of natural gases, careful consideration was given to the advisability of including hexane and heavier components in the standard. Table 1. lists five mixtures typifying the composition ranges which we considered for a GC calibration gas. A primary requisite for a usable standard is that its composition remain constant during the period of its use. While the presence of heavy components may be desirable for calibration purposes, they increase the possibility of composition changes due to condensation or adsorption. In the northern climates, certified gas cylinders could well be exposed to winter temperatures on the order of −30°F during shipment or storage. We therefore, first calculated SRK equation-of-state dew points as a function of pressure for some 75 gas mixtures and based on that consideration selected gas mixture D as our primary standard gas. Dew point curves for the five gas mixtures of Table 1 are shown in Figure 1. Cost and convenience considerations in cylinder purchase and shipping of an adequate volume of gas will necessitate the gas mixture to be pressurized. It can be seen that the dew point of the mixture D will remain below −10°F at pressures up to 300 psig. A gas cylinder with an internal volume of one cubic foot at this pressure will contain approximately 21 SCF of gas, a quantity large enough to provide an extensive number of GC calibration runs. Selection of Calorimetry Standard Compositions. The composition criteria for low and high Btu calorimetric standards are less severe than for the GC standard. This is because the calorimeter measures the gross heating value of a fuel gas independently of its composition. Thus, simple binary mixtures of methane/ ethane and methane/nitrogen can be selected to attain any desired Table 1. COMPOSITION OF VARIOUS NATURAL GAS BLENDS (Dew Point Diagrams Appear in Figure 1) ------------------------Mole %-------------------------Components /Sample

1

2

3

4

D (IGT)

Methane Ethane Propane

70.55 8.98 5.95

90.58 3.50 1.00

88.73 3.50 1.00

90.65 3.50 1.00

90.65 4.00 1,00

41

------------------------Mole %-------------------------Components /Sample

1

2

3

4

D (IGT)

i-Butane n-Butane i-Pentane n-Pentane neo-Pentane n-Hexane n-Heptane n-Octane Helium Nitrogen Carbon Dioxide Propylene

3.02 2.98 1.00 1.00 ----0.46 4.95 1,09 0.02

0.50 0.50 0.15 0.15 0.05 0.05 0.01 0.01 -2.50 1.0 --

0.40 0.40 0.15 0.15 0.10 0.05 0.02 --2.50 3.00 --

0.50 0.50 0.15 0.15 -0.05 ---2.50 1.00 --

0.30 0.30 0,10 0.10 -0.05 ---2.50 1.00 --

Key to Sample Numbers: 1: NGPA calibration gas from Phillips Petroleum 2, 3, and 4: Suggested by gas industry sources. D: Gas composition selected by IGT for the forthcoming GC calibration gas.

heating value for calibration standards in the range of 800 to 1200 Btu/SCF. However, the criteria of stability still requires a dew point consideration. Calorimetric analysis also consumes several orders of magnitude more gas volume than GC analysis. Thus a calorimeter standard is normally compressed to near 2000 psig in a 1-A size cylinder in order to deliver a volume of about 225 SCF to the user. Table 2 lists the compositions and heating values of several mixtures which were considered for these standards. Figure 2 illustrates the SRK. dew point vs pressure curves for these mixtures. It can be seen that, while the 812 Btu/SCF gas (mixture B) has a dew point temperature of −144°F, the 1261 Btu/SCF gas (mixture 5) approaches the −5°F dew point region. Thus the mixture represented by curve “C”, a 75% methane, 25% ethane blend with a heating value of 1185 Btu/SCF was chosen to be as near 1200 Btu/SCF as possible and yet maintain a reasonably low dew point temperature of −26°F/. Certification Process Sequence of Events Leading to Certification. Figure 3 diagrams the overall sequence of activities leading to delivery of IGT certified calibration standards with NBS traceability. At this time, the four primary standard mixtures have been analyzed and certified by NBS. We are currently formalizing the protocol for certification of the GC calibration standard gas at IGT. The traceability of these certified standards to the NBS primary standard will be the analysis of each cylinder containing the certified gas on a GC system which will be calibrated on a daily basis with the NBS primary standard. Protocol has not yet been established but our experience with the long-standing calorimetry standard problem will serve as a guide. In that program a cylinder is tested on three separate days, using two separate calorimeters which are calibrated daily with an NBS primary standard. The results of the three runs must agree to within 1 Btu before the cylinders′ contents are considered certifiable.

42

NEW STANDARDS FROM THE INSTITUTE OF GAS TECHNOLOGY

Figure 1. HYDROCARBON DEW POINT DIAGRAMS FOR SEVERAL NATURAL GAS BLENDS, AS PPEDICTED BY SRK EQUATION-OF-STATE

We do anticipate offering two types of GC standards; a small cylinder with composition certification only and a larger cylinder with composition certified as above plus a calorimetrically certified heating value. Table 2. APPROXIMATE COMPOSITION AND HEATING VALUE OF SEVERAL BINARY METHANE MIXTURES CONSIDERED AS CALORIMETER CALIBRATION STANDARDS (Dew Point Diagrams Appear in Figure 2) ----------------- Mole % ------------------Components Sample

5

6

7

A (IGT)

B (IGT)

C (IGT)

Methane Ethane Nitrogen Heating Value, Btu/SCF at 60°F, 14.735 psia and Saturated

65,00 35.00 -1261

70.00 30.00 -1223

72.50 27.50 -1204

99.99 Trace Trace 996

81.50 -18.50 812

75.00 25.00 -1185

Key to Sample Designations

43

Figure 2. HYDROCARBON DEW POINT DIAGRAMS FOR VARIOUS METHANE BINARY MIXTURES, AS PREDICTED BY SRK EQUATION-OF-STATE ----------------- Mole % ------------------Components Sample 5 6 7 A (IGT) B (IGT) C (IGT) 5, 6, and 7: These methane-ethane blends were rejected because of their higher dew point temperatures as compared to blend C. A, B, and C: Gas compositions selected by IGT for the forthcoming new calorimeter calibration standards.

Certification of the three new calorimetry standards will follow the present protocol for the existing program. In the case of the GC standards, concentrations of 10 components (methane through hexane plus nitrogen and carbon dioxide) in the mixture will be certified to within a range of uncertainties based on thorough statistical evaluation of analytical system performance. Equipment Used. The paticular set of equipment that has been assembled and dedicated to the GC portion of this program consists of the following principal components: Gas Chromatograph: Hach/Carle AGC Series 400, Model 04192–A Data Aquisition: Perkin-Elmer Sigma 15 Chromatography Data System Sample Handling: Vacuum manifold with Validyne Model CD 223 Digital Manometer.

44

NEW STANDARDS FROM THE INSTITUTE OF GAS TECHNOLOGY

Figure 3. OPERATIONAL SEQUENCE OF IGT CERTIFICATION PROGRAM

The chromatography data station transmits results to an IBM PC computer via RS 232 communications for the purpose of data storage and evaluation. The calorimetry certification program will continue to use the Cutler Hammer recording calorimeter that has been the workhorse of the existing IGT program. This has been supplemented for the current program expansion by the addition of a third calorimeter and construction of an automated, six-port gas manifold sampling system to permit unattended measurements on batches of five cylinders. CURRENT STATUS OF PROGRAM GC Calibration Standard Cylinder Material. A major area of effort of the program has been to verify that the choice of cylinder material will not affect the stability of the calibration mixture. To test this, we filled duplicate sets of aluminum and stainless steel cylinders with the calibration gas mixture from the carbon steel source cylinder. These five cylinders have been analyzed on a monthly basis for the past six months. As expected, no significant concentration changes in any of the 10 components are evident in any of the five cylinders at this point. Precision of Certified Concentrations. The cylinder stability data, numbering some 150 separate analyses, also form a basis set for statistical evaluation of instrument precision and accuracy. Our preliminary evaluation of the collected data indicates that the componential analysis of the certified gas will produce an overall precision well within the ASTM Method D-1945–81 repeatability requirements. System Linearity. We are currently studying the overall GC system linearity for all the components over a range of concentrations expected for common natural gases. A principal reason for this study is to determine the validity of the certification procedure for gas compositions that may vary substantially from that of the NBS primary standard. Pressure Drawdown Tests. This series of experiments were designed to determine the effect of pressure reduction on the stability of certified gas composition between a maximum filling pressure of 300 psig and atmospheric pressure. Two sets of cylinders filled with a 10–component gas blend, similar in composition to our NBS primary standard, will be analyzed before and after each 50–psig pressure reduction at different gas withdrawal rates. The results of these tests will indicate the optimum rate of gas withdrawal as well as the residual cylinder pressure for the GC calibration gas.

45

Calorific Standards Three new NBS primary standards were obtained under the new program, designated as standards A, B, and C. Composition and approximate heating values for these primary standards are listed in Table 2. The primary purpose of the new standards was to measure the linearity of calorimeter response over its entire range from 800 to 1200 Btu/SCF. Preliminary tests on three calorimeters show that there may be a bias of +1.5 Btu at the midrange of the calorimeter. That means that when standard B (certified heating value of 813.3 Btu/SCF) or standard C (certified heating value of 1186.1 Btu/SCF) are used to calibrate the calorimeter and the other two standards are used as a test gas, only standard A (certified heating value of 996.6 Btu/SCF) reads high by 1.5 Btu/SCF. This study will continue for other points within the 800–1200 Btu range in order to develop an accurate calibration curve for future use.

THE ROLE OF CALIBRATION STANDARDS IN THE ANALYSIS OF NATURAL GAS LIQUIDS Stephen L.Brandt, B.S. Chromatographic Standards Blending Engineer Phillips Chemical Company Phillips, Texas 79071

ABSTRACT

The natural gas liquids industry has been striving to measure the composition of liquid streams more accurately. The chromatograph, with the use of calibration standards, has been found to be one of the most accurate means of measuring the composition of natural gas liquids streams. This paper presents the chromatograph’s need of calibration standards when analyzing natural gas liquids streams. The selection of the proper calibration standard and the production of the standard is discussed. Also covered are cylinder selection and sampling techniques which specifically include the three major cylinders used for liquid samples. The final section of the paper deals with checking liquid calibration standards to insure authenticity. INTRODUCTION Since the beginning of the natural gas liquids industry, there have been problems with the analysis of natural gas liquid (NGL) streams. These streams are composed of a wide range of hydrocarbons with a low percentage of non-hydrocarbon gases. The wide range of components make the analysis very difficult. The early methods for analyzing NGL streams were not consistent or reliable; thus, the buying and selling of NGL streams was very difficult. As the value of natural gas liquids increased due to their importance in the gas industry, it became essential that a more accurate and reliable method be devised for the analysis of natural gas liquids. The method that was devised was chromatography. Using the proper methodology, chromatograph, calibration standards, and sampling techniques, the problem of analyzing natural gas liquid streams were greatly reduced. This gave more consistent and reliable analytical results which reduced the problems of buying or selling natural gas liquids.

47

Early Analysis Methods One of the first analytical methods for determining the composition of natural gas liquids was the “Pod” method which was developed by the Gas Processors Association’s Analysis Section of the Technical Committee in 1948.1 The “Pod” method was a distillation where the natural gas liquid sample was distilled and a receiver bottle of known volume, temperature, and pressure collected the overhead product.2 The components could then be identified by the temperatures at which they boiled and the percentage of each component in the unknown sample could be calculated from the volume, pressure, and temperature of the overhead collection sample. This method was time consuming and a good analyst could run only one analysis per day barring any complications.3 Another problem with the “Pod” method was the large number of errors which could be introduced into the system, such as heat balance, overhead cooling, temperature control of boiling, and reflux. With the problems of parameters to control and time consumption, a different method was needed. In 1953, the GPA Analysis Section started work on chromatographic methods.4 By the mid– 1950’s, the chromatograph was a superior method in the analysis of natural gas liquids due to its speed, superior separation, and repeatability. The chromatograph, however, still had problems that needed to be overcome, one of which was the translation of the peak sizes into quantitative values.5 This problem was eliminated with the use of calibration standards. Chromatographs A chromatograph is a high efficiency distillation column. When a multiple component sample is injected into a chromatograph it will separate each component. In the case of natural gas liquids the sample being injected into the chromatograph will usually contain nitrogen, carbon dioxide, and hydrocarbons ranging from methane through hexanes and heavier or any combination of the components. The chromatograph can separate these components using the proper columns, but it cannot give a quantitative number. The chromatograph is totally dependent upon a calibration standard to give a meaningful analysis to an unknown. The GPA procedure 2165–75, “Standard for Analysis of Natural Gas Liquid Mixtures by Gas Chromatography” and GPA procedure 2177–77, “Tentative Method for the Analysis of Demethanized Hydrocarbon Liquid Mixtures Containing Nitrogen and Carbon Dioxide by Gas Chromatography” both stress the importance of using calibration standards to get meaningful results by gas Chromatography. For a proper analysis, an unknown would be injected into a chromatograph and run. The chromatograph would then separate each component of the sample and draw a graph of its peaks known as a chromatogram (see Figure 1).6 Each one of the peaks on the chromatogram represents a component of the unknown mixture. A known calibration standard would then be run on the same chromatograph under the same conditions and a chromatogram would be drawn of the calibration standard. The peaks of the unknown sample and the peaks of the calibration standard would be compared, thus identifying and giving quantitative values to the unknown peaks using one of two methods, Peak Area or Peak Height.7 The Peak Area method uses the areas under each peak to calculate an area percent of each unknown component in the sample. The areas under each of the peaks of the calibration standard are calculated and area percents are derived. The area percents of the calibration standard are then correlated to the known mole percents of the calibration standard and a response factor is derived. The response factor is then used to linearly correlate the unknown sample’s area percents to mole percentages. The Peak Height method works in much the same way as the Peak Area method, but uses the height of each peak to find the percentages of each component. One problem with using the Peak Area or Peak Height methods is obtaining an accurate measurement of the peak’s area or height. This problem was solved with the use of an integrator; the integrator gives a more accurate measurement and saves operator time. The results of the integrator can be fed into a computer and,

48

THE ROLE OF CALIBRATION STANDARDS IN THE ANALYSIS OF NATURAL GAS LIQUIDS

FIG.1—Chromatogram of demethanized gassoline

using proper response factors, the unknown sample can be analyzed accurately. Using one of the two procedures and the proper calibration standard, a natural gas liquid can easily be analyzed.

49

Selection of Calibration Standards When selecting a calibration standard, certain rules must be followed so that the chromatograph being calibrated will give a correct analysis of the desired unknown. “A known blend prepared by weights of high purity components (99+ percent with known impurities) to an accuracy within 0.05 percent (of the total blend) for each component is required for calibration runs. The concentration of a component in the calibration blend should be not less than one-half of the concentration of the corresponding components in the unknown; and in addition, corresponding components in the calibration blend and in the unknown should not differ by more than five liquid volume per cent (of the total mixture) if the peak height calculation method is used, nor by more than 20 liquid volume per cent (of the total mixture) if the peak area calculation method is used. In applying the rules of this section the values for the hexanes and and heavier components may be totaled and considered as a single component.”8 If these rules for selecting a chromatographic standard are not followed, the calibration of the chromatograph and the analysis of the unknown natural gas liquid will probably not be correct. In the case of an ongoing analysis of an unknown stream with a reasonably consistent composition, a calibration standard which duplicates the average of the stream composition should be used. This would give the highest degree of accuracy to the analysis by eliminating any large, linear correlation between the unknown stream and the calibration standard. Using a calibration standard which is tailored to a stream will involve more time and expense, but will improve the analytical accuracy and justify the increased expense. When selecting a calibration standard, the preceeding rules should be followed and a standard obtained that duplicates the stream to be analyzed as much as possible. Production of Calibration Standards A GPA quality NGL calibration standard requires greater than 99 percent pure blend stocks. Each of these blend stocks must have a chromatogram run on it in which the major component is guaranteed to be 99 plus percent pure and each minor component is identified. A chromatogram must be run on a depentanized natural gasoline to show composition breakdown. This product will be used as the hexane and heavier component in the calibration standard.9 Each component that is required in the calibration standard will need to have its weight calculated for the required amount of standard.10 The physical properties needed for these calculations can be obtained from the GPSA Engineering Data Book, Section 16, Physical Properties. When calcullating the amount required, it is advised to fill the cylinders only 80 percent full to allow for thermal expansion. The blend cylinder that will be used needs to be evacuated to 1mm Hg or less and weighed.11 Each blend component is added starting with the component of lowest vapor pressure and ending with the component of highest vapor pressure. A balance with a sensitivity of 0.1 gram or less must be used to produce the blend.12 A small high pressure tube is used to transfer the blend components to the cylinder. After each component has been added to the cylinder, the tubing must be disconnected to obtain a net weight of the amount added. When each net weight has been obtained, the weight percent, liquid volume percent, mole percent, vapor pressure, average molecular weight, specific gravity, and heating value can be calculated using the physical property data of each component.

50

THE ROLE OF CALIBRATION STANDARDS IN THE ANALYSIS OF NATURAL GAS LIQUIDS

Cylinder Selection and Sampling There are several types and sizes of cylinders in which a natural gas liquids calibration standard can be made. The cylinder choice and sampling technique used is one of the most crucial points in using a NGL calibration standard. When choosing the proper cylinder, it must be verified that the cylinder has a working pressure rating at least 50 psi greater than the calibration standard’s vapor pressure, that the cylinder has a double valve system, that the cylinder is large enough to obtain the needed amount of calibration standard at not more than 80 percent full, and that the cylinder is inspected and in good working condition. The three most common cylinders being used for natural gas liquids calibration standards are the doubleended cylinder, the double entry valve with eduction tube cylinder, and the piston displacement cylinder. Each of these cylinders has advantages and disadvantages in the handling of liquid standards. The double-ended cylinder’s advantages are that it is inexpensive, it has a simple design, and it has an 1800 psi working pressure rating. The disadvantages of this cylinder are that it is difficult to sample and it cannot contain a NGL standard which contains carbon dioxide. Glycol (or water) is regulated into the cylinder through the bottom valve of the cylinder to displace the NGL standard (Fig, 2).13 When the pressure is at least 50 psi greater than the blend vapor pressure, the standard can be sampled. Carbon dioxide cannot be used in this cylinder because glycol (and water) will absorb it readily; this will remove part or all of the carbon dioxide from the standard and, in turn, ruin the standard. If the cylinder is sampled without the displacement technique, a fractionation will occur in the cylinder and the lighter components (methane, nitrogen, ethane, and carbon dioxide) will separate from the heavier components and go into the vapor space of the cylinder. Also, once the glycol is added to the cylinder, it is very difficult to get the NGL standard remixed if a separation occurs. The double entry valve cylinder’s main advantage is that it is relatively inexpensive as compared to the liquid calibration standard that it contains. It has a simple design which reduces the probability of mechanical breakdown. The main disadvantages of this cylinder are that it is difficult to sample and cannot contain blends with vapor pressures higher then 400 psi or that contain carbon dioxide. When sampling the double entry valve with eduction tube cylinder, a glycol (or water) displacement method is used (Fig. 3).14 Glycol is regulated into the double entry valve cylinder at a pressure of at least 50 psi greater than the NGL calibration standard’s vapor pressure. The glycol displaces the calibration standard and when the cylinder pressure is at least 50 psi greater than the calibration standard’s vapor pressure, there should be no vapor left in the cylinder. The sample of NGL standard can then be injected into the chromatograph. The same problems that occur with fractionation and absorption of carbon dioxide in the double-ended cylinder will occur in the double entry valve cylinder. The double entry valve cylinder is also difficult to mix if the NGL standard separates. One sampling technique that has been used in the natural gas liquids industry is incorrect and should be avoided. A double-ended cylinder or a double entry valve cylinder is pressured up with the chromatograph’s carrier gas to at least 50 psi above the calibration standard’s vapor pressure. The idea is that this will keep a sufficient amount of pressure on the calibration standard and stop the fractionation effect. This method will not work and it is not recommended. The lighter and more volatile components such as methane, nitrogen, carbon dioxide, and ethane will still migrate into the vapor space to some extent and cause the standard to change when sampling. The liquid calibration standard must be in a cylinder that has no vapor space when it is injected into the chromatograph to get a representative sample. The piston displacement cylinder’s advantages are that it has an 1800 psi pressure rating, the cylinder design has eliminated the fractionation problem associated with other cylinders, and the mixing and sampling NGL standards is easily performed. The disadvantages of this type of cylinder are that it is expensive and the mechanical design is more complicated than conventional cylinders, thus creating more

51

Figure 2 Repressure System and Liquid Sampling Valve

problems. This type of cylinder has a moving piston inside which allows the NGL calibration standard to be separated from the displacing fluid; the fluid that is used on this cylinder is an inert gas such as helium or nitrogen. Since the displacement fluid is a gas, the sampling technique used is much easier; all that is needed is a bottle of inert gas and a gas regulator. The inert gas side of the cylinder is pressured to at least 50 psi greater than the product side. This in turn compresses the NGL calibration standard into a liquid. The piston displacement cylinders are equipped with mixers so the calibration standard can be mixed into a homogeneous state; this will ensure a consistent mixture going into the chromatograph. Using the piston displacement cylinder is the most accurate and reliable way to handle NGL calibration standards. Calibration Standard Check Liquid calibration standards are delicate blends and are difficult to prepare, store, and sample. Because of these difficulties, the standards can change in composition. If this occurs, the standard is no longer valid and should not be used for calibration. When a calibration standard is received and during its use, it should be checked in some manner to ensure its authenticity. A quick method to check the reliability of a calibration standard was established by the Gas Processors Association and is described below: “1. Determine mol percent response factors for normal hydrocarbons using area measurements of peaks recorded on chromatogram of reference standard run.”15 The mole percent response factor is obtained by dividing the mole percent of the component in the reference standard by the peak area or peak height.16

52

THE ROLE OF CALIBRATION STANDARDS IN THE ANALYSIS OF NATURAL GAS LIQUIDS

Figure 3 Alternate Repressure System

“2. Determine the molecular weight corresponding to each component hydrocarbon.”17 The molecular weights can be obtained from the GPSA Engineering Data Book, Section 16, Physical Properties. “3. Using log/log paper plot the response factor on the vertical scale versus molecular weight on the horizontal scale. “4. If all is in order the resultant plot will be essentially a straight line with a negative slope. For a specific instrument, the angle of the plot should remain essentially constant. A change in the angle usually indicates a change in blend composition.” (Fig. 4).18 This procedure should be run on all new NGL calibration standards to ensure reliability and to give confidence that the standard is of high quality. CONCLUSION The development and refinement of the chromatograph methods over the last 30 years has helped make improvements in the analysis of natural gas liquids. The chromatographs have improved to a point that they are run by computers and their accuracy is an industry standard. Calibration standards have improved in accuracy by using techniques and equipment which allow standards to calibrate chromatographs in the parts per million range; these standards would be worthless if cylinder technology had not progressed significantly. With the use of piston displacement cylinders, natural gas liquids samples and calibration standards can be injected into chromatographs with unsurpassed accuracy. As long as natural gas liquids are bought and sold, there will always be a need to improve measurements of these streams.

53

FIG.4 —Response factors (mol.%) vs molecular weight

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

Sutton, C., “Chromatography and the GPA”, Seventh Chromatography School, December, 1977. Ibid. Ibid. Ibid. Ibid. “GPA Tentative Method for the Analysis of Demethanized Hydrocarbon Liquid Mixtures Containing Nitrogen and Carbon Dioxide by Gas Chromatography”, GPA Publication 2177–77, 1980, p.4. “GPA Standard for Analysis of Natural Gas Liquid Mixtures by Gas Chromatography”, GPA Publication No. 2165–75, 1980. “GPA Standard for Analysis of Natural Gas Liquid Mixtures by Gas Ghromatography”, GPA Publication No. 2165–75, 1980, p.3. “GPA Standard for Analysis of Natural Gas Liquid Mixtures by Gas Chromatography”, GPA Publication No. 2165–75, 1980. Ibid. Ibid.

54

12. 13. 14. 15. 16. 17. 18.

THE ROLE OF CALIBRATION STANDARDS IN THE ANALYSIS OF NATURAL GAS LIQUIDS

Ibid. “GPA Standard for Analysis of Natural Gas Liquid Mixtures by Gas Chromatography”, GPA Publication No. 2165–75, 1980. Ibid, p.4. “GPA Tentative Method for the Analysis of Demethanized Hydrocarbon Liquid Mixtures Containing Nitrogen and Carbon Dioxide by Gas Chromatography”, GPA Publication No. 2177–77, 1980, p.8. Ibid. Ibid. Ibid.

COMMERCIALLY AVAILABLE CALIBRATION GASES Dennis P.Norris Technical Manager Big Three Industries, Inc. Mixed and Specialty Gases P.O. Box 1026 La Porte, Texas 77571

ABSTRACT

Commercial calibration gases are presently prepared by several specialty gas producers. The quality and technical support of these gases varies with the supplier. Today’s highly advanced and sophisticated instrumentation require carefully prepared calibration gases to optimize accuracy. These calibration gases must be prepared and analyzed under a stringent quality assurance program. Failure to do this can result in an integrity problem with the gas. Any error in the calibration gases can result in a high dollar loss from the products being sold using their analysis. Big Three Industries has addressed this problem by introducing the UniPhase calibration mixtures. These are single phase calibration mixtures with a technical validation package that does not rely on instrument calibration. Another approach to this variant of calibration gases from different suppliers is to reference these gases to a common source. This is now becoming available from IGT and some of the major specialty gas suppliers, such as Big Three Industries. This common reference is the National Bureau of Standards, which is world renown for their high degree of accuracy and reliability. INTRODUCTION Today’s analytical instrumentation has technically advanced into the space age. Much of the present unmanned exploration of the outer reaches of our solar system utilize analyzers not unlike those found in a commercial laboratory today. The accuracy and reliability of this instrumentation has improved many orders of magnitude since its conception. With these improvements in the capability of measurement of composition and physical properties, there is still a dependency on the reference point—the calibration gas. Not unlike the improvements in the instrumentation, the accuracy and the reliability of the calibration gases have also improved. Today there are many commercial suppliers of calibration gas competing in the

56

COMMERCIALLY AVAILABLE CALIBRATION GASES

market place. This competition has benefited the analyst by forcing the commercial supplier to continually improve the calibration gases they procure. WHAT IS A CALIBRATION GAS There still is, however, some misunderstandings about calibration gases that often cause misuse or create questions about the quality of the gases. All calibration gases must meet three minimum requirements to assure the user of their quality. Calibration gases must be homogeneous, stable and accurate. The following information will hopefully clarify some questions about calibration gases. Homogeneity The homogeneity of a calibration mixture simply means it is completely mixed. Once mixed, and if maintained within the specified temperature and pressure limitations, a calibration mixture will remain homogeneous. Only when a mixture is kept as a single phase, either all gas or liquid, will it be homogeneous. The gas/liquid phase equilibrium will effect the homogeneity of a mixture that is not maintained as a single phase. The plot in figure 1 is a typical example of a phase envelope of a multicomponent mixture. The upper curve of the envelope represents the pressure verses temperature conditions for the bubble point of the mixture. The region above this curve represents the pressure verses temperature conditions to maintain the mixture as a single phase liquid. The lower curve of the envelope represents the pressure verses temperature conditions for the hydrocarbon dew point of the mixture. The region below this curve represents the pressure verses temperature condition to maintain the mixture as a single phase gas. The center region indicates the pressure and temperature conditions which would cause the mixture to exist in both the liquid and gas phases. The conditions that would put the mixture into this center region must be avoided if the mixture is expected to remain homogeneous. Stability The stability of a mixture must be confirmed through time related testing. This is usually not necessary when dealing with hydrocarbon mixtures. Although there is some question to the reactivity between the carbon steel or aluminum in the cylinder wall and the hydrocarbons contained in a calibration mixture, the long term stability of typical natural gas product related mixtures has been proven. Only when corrosive gases such as Carbon Dioxide or Hydrogen Sulfide are present is there need for additional concern. Even these corrosive gases can be stable if the cylinder is completely dried and treated or pacified. Extra analytical support is necessary when corrosive gases are included in the mixture. To verify the mixture stability at least two analysis should bracket an incubation period of at least seven days. Any change in the concentrations of the corrosive gas compositions greater than the analysis uncertainty of the analytical measurement should be recognized as an indication of instability. The longer the incubation period between analysis the more sensitive the testing will be to slow rate chemical reactions. A mixture must have this confirmed stability to be used as a reliable calibration tool. Accuracy The results of the calibration must be accurate. The fidelity of these mixtures must be confirmed by instrumental analysis. The problem with this technique is that instrumentation must be calibrated for the

57

Figure 1.

analysis. This means a calibration mixture must be obtained to analyze the newly prepared calibration mixture. Until recently no common reference has been recognized for these calibration mixtures. This created a situation which causes a variance from supplier to supplier in certified calibration mixtures. To over come this problem, a statistical evaluation was developed by the Technical B committee of the Gas Processors Association which is designated GPA 2177–84. The method plots the molar response of each normal hydrocarbon from a gas chromatograph analysis with a thermal conductivity detector against that component’s molecular weight on log/log paper. If all is in order the results will be essentially a straight line with a negative slope. For a specific instrument, the angle of the plot should remain relatively constant. A change in the angle usually indicates an error in the calibration mixture composition. An example of this statistical plot is in figure 2. This type of evaluation allows the user to determine the fidelity of a calibration mixture independent of intercomparison to another calibration mixture.

58

COMMERCIALLY AVAILABLE CALIBRATION GASES

TYPES OF CALIBRATION GASES There are two basic types of calibration gases. These gases are usually refered to as primary and certified. The determination of which type of preparation method used is based on the accuracy requirements of the calibration gas. In an effort to supply the industry with a more reliable calibration gas, a combination of these two techniques has been developed. This combined technique is called the UniPhase calibration mixtures, which is a trademark of Big Three Industries, Inc., Mixed and Specialty Gases. The advantages and disadvantages of each is defined in the following paragraphs. Primary Calibration Gases Primary calibration gases are prepared gravimetrically, using high load, high sensitivity balances. This technique uses weights which are certified to the National Bureau of Standards and high purity analyzed raw materials. These type of calibration gases are used when a high degree of accuracy and dependability are needed. This technique however is dependent upon the quality and stability of the raw materials in the final mixture. An example of a typical certification of a primary calibration gas mixture is shown in figure 3. Certified Calibration Gases Certified calibration gases are prepared either gravimetrically or volumetrically. These gases are then certified using either a NBS Standard Reference Material (SRM) or a Gas Manufacture Prepared Standard (GMPS). These type of calibration gases are used when mixtures contain reactive compounds or when less stringent accuracy requirements are permited. These gases usually are not as accurate as a primary mixture because of the dependency upon the analytical instrumentation and the addition to the total uncertainties from the analysis comparisons. An example of a typical certification of a certified calibration gas mixture is shown in figure 4 with an accompanying chromatogram and corresponding data values in figures 5 and 6. UniPhase Calibration Gases Big Three Industries, Inc., Mixed and Specialty Gases has introduced another type of calibration mixture. We refer to this new type as the UniPhase mixtures. UniPhase means a single phase mixture, which is either all in the July 26, 1985 P.O. Number: 1234–55327 Customer: Natural Gas Calibration Mixture Lot/Item# 213 CERTIFICATION OF CYLINDER # PB-57096 Component Mole % Weight % Nitrogen 01.442 01.914 Methane 83.310 63.313 Carbon Dioxide 01.668 03.477 Ethane 04.659 06.636 Propane 02.500 05.222 Isopentane 02.074 05.711 n-Butane 02.033 05.598 Isopentane 01.012 03.459 n-Pentane 01.084 03.705 nHexane 00.105 00.429 Heptanes plus 00.113 00.536 Expiration date July 26, 1986 July 26, 1985 P.O. Number: 1234–55327 Customer: Natural Gas Calibration Mixture Lot/Item# 213 CERTIFICATION OF CYLINDER # PB-57096 Component Mole % Weight % Nitrogen 01.442 01.914 Methane 83.310 63.313 Carbon Dioxide 01.668 03.477 Ethane 04.659 06.636 Propane 02.5 Composition established usig NBS traceable “S” serieswieghts and preanalyzed high purity raw materials gas or liquid phase. This type of mixture is the most accurate type of calibration mixtures. These mixtures are prepared gravimetrically and then certified analytically to validate the blended weights. In many cases a second validation may be completed using a measurement of a physical property (such as the BTU or Relative Density) of the gas mixture. This technique is unique in that it uses the advantages of the two previously mentioned methods to resolve any error or deficiences that each presents. An example of a typical certification of a Big Three Specialty Gases UniPhase calibration gas mixture is shown in figure 7 with an accompanying chromatogram

59

Figure 2.

and corresponding data values in figures 8 and 9. Also shown is the linearity plot in figure 10 and the hydrocarbon dew point curve for the calibration mixture in figure 11.

60

COMMERCIALLY AVAILABLE CALIBRATION GASES

Figure 3.

NBS TRACEABILITY OF CALIBRATION GASES In September of 1983, a request was submitted by the Laboratory and Chemical Services Committee of the American Gas Association, which outlined the need for a calibration gas for natural gas analysis by chromatography which was certified by a recognized certifying agency. This request was undertaken by the Gas Research Institute. The result is a program that will allow calibration mixture suppliers to prepare a gas that will have a certified analysis traceable to a mixture certified by the National Bureau of Standards. This program with the NBS will create a common source to which calibration mixtures can be traced. This common source of traceability will create a reference point that can be used to resolve any dispute between parties on questions of compositional accuracy. The Gaseous Fuels committee (D-03) of the American Society of Testing Materials (ASTM) has presently initiated work to establish a standard practice outlining the requirements necessary for a calibration gas to be recognized as NBS traceable. With the availability of the NBS certified gases and

61

Figure 4.

guidelines established by ASTM for the requirements to meet NBS traceability, a calibration gas with a certification traceable to a common source can be established. CONCLUSION The most important requirements of a calibration gas are that it is homogeneous, stable and accurate. The most important job of the commercial gas supplier is to provide a calibration gas that meets these needs and the technical support package to validate that claim. At Big Three Specialty Gases, we feel that the UniPhase calibration mixtures meets these needs, and only with the technical UniPhasetm DATA VALUES FOR LINEARITY PLOT (GPA 2177–84 (Only normal components plotted) CUSTOMER Natural Gas Calibration Mixture No. Peak Mole. Wt. Amount Area Response x 10000 2 Methane 16.043 45. 101 3582650 0.12589 4 Ethane 30.070 4.659 507744 0.09175 5 Propane 44.097 2.500 325695 0.07677 7 nButane 58.124 2.033 307670 0.06609 9 n-Pentane 72.151 1.084 183160 0.05920 10 n-Hexane 86.178 0.105

62

COMMERCIALLY AVAILABLE CALIBRATION GASES

Figure 5.

19084 0.05502 The line slope is −0.4960 The % Corrilation factor is 99.9776 GAS CHROMATOGRAPHIC CONDITIONS Instrument Type: Varian 6000 TCD Carrier Gas: Helium Detector 0: TCD x 0.5 Detector: TCD x 0.5 Injection Size: CSV 0.5 ml LSV 0.2 uL Operating Conditions: Isothermal at 115 C Column Description: 30' x 1/8" DC-200/500 on Chrom PAW 80/100 mesh GAS CHROMATOGRAPHIC RETENTION ORDER Time Component Area 2.00 Nitrogen 119879 2.14 Methane 3582650 2.52 Carbon Dioxide 158657 2.80 Ethane 507744 3.78 Propane 325695 4.97 Isobutane 310334 5.83 n-Butane 307670 8. 50 Isopentane 177221 9.63 n-Pentane 183160 16.76 n-Hexane 19084 28.75 Heptanes Plus 20897 support package that it provides can a user have all the information necessary to support their analysis. The preparation of calibration gases is not unlike other present day technologies. There is a continual effort to improve their quality and accuracy. A better understanding of the effects of these gases and their limitations are constantly being improved. Methods to overcome these limitations are also being activily developed. The ever increasing presence of the computer in the laboratory is also rapidly opening the doors to a better understanding of the physical and chemical properties of gas mixtures. This understanding allows the specialty gas producers to develop and supply a more detailed and accurate technical support package with the calibration gases. This information arms the user with the tools to be more prepared to answer any question about the gas analysis they report.

63

Figure 6.

64

COMMERCIALLY AVAILABLE CALIBRATION GASES

Figure 7a.

65

Figure 7b.

66

COMMERCIALLY AVAILABLE CALIBRATION GASES

Figure 8.

67

68

COMMERCIALLY AVAILABLE CALIBRATION GASES

Figure 10.

69

Figure 11.

COMPOSITE SAMPLING OF NATURAL GAS Thomas F.Welker Executive Vice-President, Marketing Welker Engineering Company Sugar Land, Texas 77487

ABSTRACT

The object of any sampling procedure is to obtain a representative sample of hydrocarbons from the system under investigation. Any subsequent analysis of the sample, regardless of the test, is inaccurate unless a representative sample is obtained. Due to the wide variation in flowing streams and the components in these streams, the proper sampling techniques must be employed in order for the sample to be taken, transported, stored, and finally analyzed by some type of test device. The first factor that must be covered is the person chosen to physically take the spot sample, or install and maintain the sampling device. This person is the beginning of a successful sampling program. The final outcome of the sample operation will be determined by the efforts of this first link in an unbreakable chain of operations that must be performed without variances which can and will affect the outcome of the results obtained. The philosophy of the persons involved in the sample taking will need to be sound. Their techniques must conform to the requirements of the technology and accuracy required in order to effectively take a representative sample. An inaccurate sample could cost a company untold millions of dollars in lost revenue or contribute to the improper design of plants, and in addition, cause incorrect plant or pipeline balances. The therm billing that is being used throughout the industry requires that all functions of gas measurement, including sampling, be done with considerable care and utilizing specific techniques. Sampling The sampling of natural gas has been discussed and studied for many years. Serious testing on the proper sampling methods has been done in a number of locations in the recent past. From these tests, it has been determined that the sampling procedures must be carefully prepared and followed. For a person to collect a representative sample of natural gas, the procedures learned in spot sampling operations must be followed.

71

Continuous sampling is described as a method by which a representative portion of product is removed from a flowing stream and pumped into a sample container during a specific time or volume. The object of the continuous sampler is to collect the sample in the sample container without changing the chemical composition, heating value, or physical characteristics of the products being sampled. The continuous sampling system consists of a probe in the line, a sampling pump, a timing device, and a sample container. The continuous sampler is normally a mechanical device that is built to be a practical alternative to an on-line analyzing mechanism, i.e., calorimeter, chromatograph, etc. The ease of installation, simple maintenance and reduced cost make the continuous sampler an attractive alternative to spot sampling and/or continuous recording devices. The objective of any therm billing measurement program integrates accurate metering methods, including sampling, to accurately determine the heating value of the gas as delivered and sold to a customer. The heating value delivered is determined by multiplying the unit volume delivery by the heating value (BTU) of the sample extracted during the delivery period. Since natural gas is commingled from various sources prior to the actual delivery to your customer, wide variations can occur in the components in the flowing gas stream. A repeatable, representative sample of the “as delivered” gas insures the accuracy of the billing. An inaccurate method of sample heating value or the application of average figures can cost a gas company millions of dollars in lost revenue and/or contribute to the “lost and unaccounted for” volumes. Proper sampling philosophy can also lend accuracy to the sample analysis chemical composition in determining the correct supercompressibility factors in place of system averages. The fact that the price of gas is high and the profit margin in your company is low dictates that present accepted measurement methods should be updated to present day metering technology. New equipment may be costly when viewed at its first cost, however, the new equipment may overcome inaccuracies that cost companies thousands of dollars per month per location. Corporate cash flow can be enhanced. To collect a continuous or composite sample of gas, the following items must not be ignored: 1. 2. 3. 4. 5. 6. 7. 8. 9.

Sample point Sample probe Hook-up and manifold of sampler and cylinder Sampler Purging of sampler and cylinder Sample cylinder, cleaning, purging, valving Cylinder transport Leaks on sampler and cylinders and related piping Preventative maintenance of the sampler

To ensure the continuous or composite sampler will give accurate and repeatable results, the above points will be covered briefly. The Sample Point A sampler is able to produce a sample no more accurate than the sample presented to it. The main consideration in the location of the sample probe is whether it sees the center one-third of the pipeline and whether it is in an area where there is good velocity with minimum turbulence. Turbulence is an aerosol generator and therefore, liquids put into flight by the turbulence may affect the sample’s result. This turbulence makes the

72

COMPOSITE SAMPLING OF NATURAL GAS

liquids moving along the walls and bottom of the pipeline take flight and act as a gas. When aerosols are introduced into the sample container, condensation occurs. The sampler should be located in an area where the gas is moving, A sample should never be taken from a shut in or dead end line. Areas to be avoided are downstream of reduced port valves, control valves, check valves, obstructions and piping fittings. When installing the sampler downstream of an orifice plate, the probe should be as far away from the orifice as possible. Headers and blowdown stacks should be avoided as sample points. Samplers should never be connected to meter manifolds. Install the probe in a straight run of pipe as far away as possible from bends, tees, fittings or any type of obstruction in the line. A sampler should not be installed without using a sample probe. A representative sample of any product cannot be taken without the use of a sample probe. Sample Probes The use of probes in the sampling operation is imperative. Without the use of a probe in the line, an accurate sample cannot be taken. A sample probe should be in the center one-third of the pipeline and equipped with a full open ball or gate valve. The placement of the sample probe is important in all sampling applications. Probes must be kept away from piping elbows, tees, manifolds, reduced port valves and orifice plates.

Design of the Probe The probe may have a bevel or be cut flat across the end. The bevel on the probe may be faced upstream or downstream. Placement in the center one-third of the pipeline is the most important consideration. If the probe is in the meter run, the placement should be away from the inlet elbow and as far downstream of the orifice plate as possible. Headers and manifolds are poor locations for sample probes of any type. Turbulence generated by gas movement in headers and manifolds will not mix the gas uniformly. If gas comes into a header from multiple side taps, the gas moving through the header will not tend to mix readily with the gas moving in from the side. Vertical headers are turbulence generators and liquid accumulators. Horizontal headers also have turbulence problems and should be avoided. Vertical headers having runs off of the side will encourage the heavies and liquids to move through the bottom run and the lighter, dryer gas will move through the upper meter run. In the weld cap of vertical headers, there is an impingement of the liquids. Therefore, the weld cap is not a proper location for probes for any use. The actual location of a probe in the piping system is important. What is the objective? One rule is clear— the probe must be located directly in the flowing stream. Another more obscure consideration implies that the probe must be kept clear of free liquid and this includes aerosols which, in fact, are the real trouble makers. Since turbulence is the mechanism that generates aerosols, it is reasonable to make every attempt to stay away from the downstream end of turbulence producers such as reducers, elbows and measurement devices. How long a liquid remains in the aerosol state is a function of the gas velocity; however, in all likelihood, it will be a distance that exceeds 20 pipe diameters. This creates a problem when one considers that available straight and horizontal piping above ground rarely makes allowance for the ideal sampling location. For gas sampling, locate the probe in the top of a horizontal pipe.

73

Probe Construction The probe should be constructed from a material that will not react with the product. 316 stainless steel is the most practical material for probe construction. Probes are normally constructed in three different ways: 1. The stationary or permanent probe 2. The manually insertable probe 3. The automatic insertion probe The stationary probe is installed as a permanent fixture in the piping system, A full open valve should be attached to the outlet of the probe. The tube extending into the flow path should be made strong enough to resist bending. The manual insertion probe is used in locations of medium pressure where a permanent fixture cannot be left in the pipe. To insert the manual insertion probe, attach to a gate or ball valve and close the valve attached securely to the end of the probe. Open the pipeline valve and very carefully push the tube into the flow line. Tighten the fittings on the probe enough to hold the tubing in place and prevent leaks. Normally, the lower ferrule will be nylon (or PTFE) and the upper ferrule will be stainless steel. Once the stainless steel ferrule is “set,” the insertion depth of the probe is fixed. This type of probe must be handled carefully with special attention given to locking the valve on the end of the probe and securing the probe into the insertion valve. The Automatic Insertion Probe The automatic insertion probe is used in locations that require frequent insertion and retraction of the probe from the pipeline. The automatic insertion probe is built as a standard to screw into a 1-inch NPT ported valve. Other ends are available for attachment to the pipeline. The use of the automatic insertion probe style allows easy access and removal of all types of probes into the line. Probes in Wet Gas Systems The wet gas pipeline system continuously exhibits the need for probes. In wet gas systems, liquid carryover in instrument supplies, valve operations, and chemical injectors is a continual problem. Samplers, chromatographs, calorimeters and related on-line monitors should be hooked up to the line using a probe. Sampler Hook-up and Manifold From the outlet of the probe, a ball valve or large ported valve should be installed. This valve should be opened completely. Downstream of the probe valve, a short length of small diameter line should be run upgrade to the inlet port of the manifold block on the sampler. The sampler should be mounted above the sample point on a pipe stand. The line to the sampler should always be sloped back toward the valve on the sample probe. This is to allow any free liquid to drain back into the pipeline. Free liquids should be discouraged from moving into the sample container. Two phase samplers in standard sample containers are difficult, if not impossible, to handle properly in the lab.

74

COMPOSITE SAMPLING OF NATURAL GAS

The installation of the sampler should always be as close as practical to the sample point. Never sample a dead end line. On the outlet of the sampler, the sample cylinder should be connected with a short length of small diameter tubing. This is to be pumped into the cylinder, not some excessive length of tubing. Mount the cylinder in some type of holder, not on the ground or deck. The outlet tubing from the sampler to the cylinder must be carefully checked for leaks. Leaks allow the sample to dissipate nonuniformly. Gas will leak from a cylinder light ends first giving inaccurate results. Care should be taken not to put filters, drips, or regulators between the probe and the sampler. This affects the gas and it is no longer representative of the pipeline product. Sampler The sampler is a mechanism that gives the operator an opportunity to have a composite sample in a cylinder. It is an alternative to a spot sample and/or an onstream monitoring device. The sampler should take its composite sample just as an operator would put a spot sample into his container. The sampler, however, does this continuously over a specific period of time. The sampler may be a simple timed mechanism actuating the sampler periodically. It may be interfaced with measurement to cause the sample to be taken proportional to the flow electrically or pneumatically. For stations or locations where the flow varies widely or the heating content swings up and down, the sampler should be actuated proportional to the flow. For stations where the load is constant, a timer may be used without affecting the gas collected. For stations or wells that have flow, no flow operations, the sampler should be turned off with a flow switch when the flow is off. Sampling should be stopped when there is no flow in every case. After a number of years and many test locations, it is recommended that in locations where gas has a heating value of 1025 BTU or above should be considered as prime locations for the use of a continuous sampler.

The sampler should be capable of pumping the sample into the sample container, regardless of ambient conditions. The sampler should be able to purge itself prior to pumping a new “bite” into the sample container. The sampler should sample the gas at pipeline conditions. Purging the Sampler and the Cylinder When the sampling device is put into service and a sample container is attached, the sampler and all its components including the cylinder should be properly purged. The act of purging cleans the air from the sampler and associated components so they will not be present in the analysis. This purge also conditions the cylinder with the gas that is to be sampled. Note: If an evacuated cylinder is used, the associated tubing, fittings, and valves still must be purged. The Sample Cylinder The sample cylinder is the carrier of the sample; therefore, it is an integral part of the system. It should be made of a material that will not react with the gas. It should be kept clean and well maintained. Valves and reliefs

75

should be checked periodically for ease of operation and must be checked for leaks through the seat, bonnet, and threads. No leaks may be tolerated. Soft seat valves should be used on sample cylinders. The sample cylinder should be cleaned after each use with solvent and then air dried. Sample cylinders must be a proper working pressure to handle the source being sampled. Rising temperatures will increase the pressure in the cylinder so this should be considered when choosing a sample container. Cylinder Transport When the sample period is over, the sample container should be disconnected from the sampler and carefully checked for leaks. Plugs or caps should be installed on the valves. The sample information tag should be filled out fully. The cylinder should then be put into a proper case for transport. D.O.T. rules apply, even in your company trucks and autos. A cylinder should never be transported haphazardly. Leaks Leaks should not be tolerated in any portion of a sample system. Leaks will cause the sample to give incorrect results. Maintenance on the Sampler The continuous sampler is a mechanical device and should be checked each time the sample cylinder is changed. Simply remove the cover from the sampler. Check the supply, activate the sampler to check its stroke and supply regulator response. Open the vent valve on the filter (F-7) on the instrument supply to check if liquid is getting into your system. Close the outlet valve and activate the sampler to watch the pressure on the outlet gauge increase. Reopen the outlet valve. Every three months, check the sample head for chemical attack or swelling. Every year, change the o-rings and lubricate the shaft and three-way valve. Spare parts that should be kept on hand are: 1. 2. 3. 4.

Sampler head −003 o-rings O−ring kit Batteries for timer (if required) The Composite Sampler in Wet Gas Service

The composite sampler is effective in wet gas service. It is further recommended that a constant pressure sample cylinder be used in wet gas service. By using the composite sampler and the constant pressure cylinder in wet gas service, the sample can be maintained under pipeline pressure, thereby discouraging retrograde condensation in the cylinder. The sample in the constant pressure cylinder may be run in the lab under pipeline conditions, pressure and temperature to get a better result. Measurement Effect of a One BTU Error (Expressed in Dollars Per Year) Examples:

76

COMPOSITE SAMPLING OF NATURAL GAS

1. Daily Production Rate=10,000 MCFD BTU from monthly spot sample=1020 BTU BTU from composite-continuous sampling=1019 BTU—One BTU Variation Purchase Gas Cost=$3.50 per MMBTU (1000 BTU—Base) 10,000 (1.020)($3.50)=$35,700 per day 10,000 (1.019)($3.50)=$35,665 per day $ 35 per day $35 (25 days per month) (12 months)=$10,500 One BTU Variation = $10,500 per year 2. Daily Purchase Rate=200,000 MCFD BTU from spot sample=1036 BTU BTU from continuous-composite sampling=1035 BTU—One BTU Variation Purchase Gas Cost=$3.50 per MMBTU (1000 BTU—Base) 200,000 (1.036)($3.50=$725,200 per day 200,000 (1.035)($3.50=$724,500 per day $ 700 per day $700 (30 days) (12 months)=$252,000 One BTU Variation = $ 252,000 per year DOLLARS PER YEAR DUE TO MEASUREMENT VARIATION IN SPOT BTU SAMPLE VS. COMPOSITE SAMPLE (Dollars based on $3.50 per MMBTU—1000 BTU Base) Sample BTU

MMCFD—Daily Purchase or Sale Rate

Variation

5

10

20

25

30

50

1 2 3 4 5 6 7 8 9 10

$ 6,387 $ 12,775 $ 19,162 $ 25,550 $ 31,937 $ 38,325 $ 44,712 $ 51,100 $ 57,487 $ 63,875

$ 12,775 $ 25,550 $ 39,420 $ 51,100 $ 63,875 $ 76,650 $ 89,425 $102,200 $114,975 $127,750

$ 25,550 $ 51,100 $ 76,650 $102,200 $127,750 $153,300 $178,850 $204,400 $229,950 $255,500

$ 31,937 $ 63,875 $ 95,812 $127,750 $159,687 $191,625 $223,562 $255,500 $287,437 $319,375

$ 38,325 $ 76,650 $114,975 $153,300 $191,625 $229,950 $268,275 $306,600 $344,925 $383,250

$ 63,875 $127,750 $191,625 $255,500 $319,375 $383,250 $447,125 $511,000 $574,875 $638,750

Basis for above information: Daily Purchase = 5 MMCFD BTU Sample Content = 1001 Purchase Gas Cost = $3.50 per MMBTU (1000 BTU—Base) 5,000(1,001)($3.50=$ 17,517.50 per day 5,000(1,000)($3.50=$ 17,500.00 per day $ 17.50 per day $17.50 per day (365 days) = $ 6,387.50 per year

ENERGY MEASUREMENT WITH AN ON-LINE GC Warren Dean Product Marketing Manager, Analyzers Daniel Industries, Inc. Houston, Texas 77255

ABSTRACT

The energy content of natural gas is normally expressed in the United States as B.T.U. per cubic foot. This definition, therefore dictates that a method of measuring B.T.U as well as volume must be considered if total energy is to be measured. There are a number of methods and systems available to measure the B.T.U. of natural gas. This paper will describe one of those methods based on a process gas chromatograph, the Danalyzer, manufactured by Daniel Industries. SYSTEM OVERVIEW Gas chromatography is a method by which gas or vaporizable liquid mixtures are physically separated into their individual components and quantified. There are three basic elements in any industrial gas chromatograph: 1. Sample conditioning system 2. Analyzer 3. Controller The sample conditioning system provides a representative sample from the pipeline that is filtered and pressure regulated. The sample conditioning system prepares the sample for the sample inject valve. The sample conditioning system includes: Sample probe for removing a representative sample from the pipeline. Usually the probe is a length of 1/ 4” stainless steel tubing extending into the pipe approximately 1/3 of the pipe diameter. By sampling from

78

ENERGY MEASUREMENT WITH AN ON-LINE GC

Figure 1.

the center third of the pipe diameter, liquids and particulates creeping along the inner pipe wall are excluded from the analysis system thereby increasing reliability. Pressure regulator for insuring that the sample is presented to the sample inject valve at a relatively constant pressure between 2 and 30 PSIG±3%. Guard filter for removing particulates 7 microns and greater in diameter from the gas. Sample flow meter and needle valve for controlling the sample flow to the sample inject valve at approximately 50 cc/minute. Solenoid valves for stream switching of multiple analysis points and for introduction of calibration gas during automatic calibration. The analyzer is composed of five major elements. 1. Airless heat sink oven 2. Diaphragm type valve system 3. Micro packed column system 4. Micro thermal conductivity detector 5. Low usage carrier gas system A block diagram of the analyzer is shown in figure 1. The airless heat sink oven maintains the valves, columns and the detector at a constant temperature, typically 80°C. Temperature control of the above components is critical to repeatable analytical results. It is the valves and the columns that work in conjunction to effect the separation of natural gas into its standard eleven components. Temperature changes will alter the separating characteristics of the system and are therefore minimized. The heat sink oven requires no outside air eliminating the need for a high maintenance air compressor. The analyzer system is designed for harsh environments. Ambient temperature fluctuations of 0–130°F do not adversely effect the analysis. The analyzer can be mounted at the sample point with no additional enclosures, reducing cost of installation and sample lag time. To prove this, each unit is tested in an environmental chamber and must provide repeatable results to within ±1/2 B.T.U. in a thousand over a 24hour period while the ambient temperature cycles from 0–130°F. The valve system consists of three diaphragm valves, the sample inject, the backflush and the dual column valve. The valves have been proven through seven million actuations at 150°C without failure or leakage and require no 0-rings, springs, or lubrication. A key to the valve’s long life is that the pistons and diaphragm move only 0.001” and that the sample comes in contact only with the center plate and the diaphragm.

79

The sample valve performs an extremely exacting function. It must take a precise volume of the sample and inject it into the column each cycle. The sample volume and the injection rate must be identical from cycle to cycle or the results will not be repeatable. The oven temperature is also critical to sample injection repeatability. Even though the sample valve may inject the same volume each cycle, if the temperature of the sample loop or valve varies, the density of the gaseous sample will change. If the sample density changes then the number of sample molecules impinging on the detector changes. This results in an unrepeatable system. The Danalyzer is designed so that the oven, valves and detector are not effected by ambient temperatures of 0–130°F, The unit will still repeat to within ±1/2 B.T.U. in a thousand over this temperature range. The backflush and dual column valves are similar in design to the sample inject valve; however, their function is slightly different. They are column switching valves. Column switching valves allow the sample and/or carrier gas to flow forward through a column, bypass it or flow backwards. Column switching allows the analyst opportunity to increase the speed of analysis and to perform more complete separation of the components of natural gas. The column system is the heart of the separating system. The Danalyzer column system is composed of three basic columns. The first column is the backflush column and is used to separate the heavy components, hexane+ from the rest of the components. The second column separates the intermediate components, propane, iso-butane, normal butane, neo-pentane, iso-pentane, and normal pentane. The third column is used to separate the light components, nitrogen, methane, carbon dioxide and ethane. The columns are 1/16” stainless steel tubing packed with a media that selectively retards the flow of the gas components. The selectivity of the column is dependent on the boiling point or partial pressure of the individual pure components of natural gas. The lighter components will travel through the column system faster than the heavier components. Each compound travels at a slightly different rate through the column system and is eventually physically separated into a discrete band of the pure component. As the pure component bands elute from the column system they enter the thermal conductivity detector for quantification. The micro thermal conductivity detector uses the property of thermal conductivity to detect and quantify each component of natural gas. Thermal conductivity is a measure of a gaseous compound’s ability to take away heat from a hot body. In general, the lower the molecular weight the higher the thermal conductivity. The micro thermal conductivity detector is a modified wheatstone bridge with two thermistors as two opposite legs of the bridge. When the carrier gas, helium, passes over both the measuring and the referenced side, the bridge is balanced and the output from the preamplifier is zero. When a pure component band elutes across the detector, the component having less thermal conductivity than helium which is flowing through the referenced side, takes away less heat from the measuring side unbalancing the bridge. The signal resulting from the unbalanced bridge is directly proportional to concentration of the component eluting. The signal is sent to the controller for further processing and for output to a recorder, printer or a host computer. The carrier gas is the driving force in separating the components. It is the carrier gas that is introduced into the sample valve and sweeps the sample into the column. The carrier gas is also the transporting medium that pushes the components through the columns. The carrier gas also acts as a reference for the micro thermal conductivity. No plant air is required to activate the oven valves, the carrier gas pressure performs this function as well.

80

ENERGY MEASUREMENT WITH AN ON-LINE GC

Figure 2.

STEPS IN A TYPICAL B.T.U. ANALYSIS Step 1: Prior to sample injection. Prior to injection, the sample purges the sample transport tubing and sample conditioning system and vents out one side of the sample inject valve. Meanwhile, the carrier gas flows through the rest of the separating system to the detector and out to vent. With helium flowing across the detector, the system reads zero. The analysis is ready to begin. See figure 2. Step 2: Zero to 45 seconds. The sample valve switches to capture a precise volume of the sample and to allow carrier gas to sweep the sample from the valve into the columns. The columns are specially designed to hold and separate the components as follows. Column 1 – C6, C7, C8 Column 2 – C3, C4, C5, C2, CO2 Column 3 – Nitrogen, Cl See figure 3.

Step 3: Forty-five to 90 seconds. The backflush valve switches to flip column 1, the backflush column, so that the components are backed out of the column directly to the detector. C8, C7, and C6 are lumped together to form one peak designated as C6+. Backflushing C6+ rather than allowing it to travel through the entire system shortens the analysis time. By 90 seconds column 2 now contains C4, C5, and C3 and column 3 contains C2, CO2, Cl and nitrogen. See figure 4. Step 4: Ninety seconds to 350 seconds. The dual column valve switches to allow the components in column 2 to bypass column 3 and go to the detector via the empty column 1. Meanwhile, the light components in column 3 are trapped waiting their turn to elute. See figure 5. Step 5: Three-hundred-fifty to 720 seconds. The dual column valve switches again to allow the carrier gas to sweep the components in column 3 through column 1 and into the detector. The dual column valve and column 3 allows for baseline separation between nitrogen and methane, carbon dioxide and ethane. By separating each of the components completely, there is no error due to perpendicular area allocation or any other indirect measurement technique. See figure 6. CONTROLLER The controller is the “brains” of the system. It is microprocessor based for flexible powerful data reduction and report generation. The controller provides highly accurate timing for events like automatic calibration

81

Figure 3.

Figure 4.

Figure 5.

and valve switching, precision calculations, and report generation. It is also the interface for inputs from the analyzer or operator to other devices through its analog outputs or by direct digital link. A typical long analysis report is shown below. The operator can edit the report and create his own short report by selecting only those parameters that are of particular interest. Twenty-four hour averages and rolling averages of B.T.U. specific gravity, compressibility, wobbe or any other component concentration are available. Up to 15 separate 24-hour averages can be printed. A typical analysis report is shown in figure 7. Each controller is environmentally tested for 10 days. The ambient temperature within the environmental chamber is cycled 4 times a day from 40 to 140°F while the electronic components are being exercised by a quality control computer. Environmental testing of all components is essential to insure reliable field operation. ANALYSIS REPORT (AGA CALCS) DATE: 07/30/85 ANALYSIS TIME: 780 STREAM SEQUENCE: 1 TIME: 15s 17 CYCLE TIME: 900 STREAM#: 1 ANALYZER#: 5759 CYCLE START TIME: 15:03 COMP NAME COMP CODE MOLE % GAL/MCF** B.T.U.* SP. GR.* C 6 + 108 0.029 0. 0128 1.52 0.0009 PROPANE 102 1.029 0.2832 25.98 0.0157 1–BUTANE 103 0.313 0. 1023 10.20 0.0063 N–BUTANE 104 0. 313 0.0986 10.23 0.0063 NEO C5 107 0. 103 0.0393 4. 10 0.0026 I PENT ANE 105 0. 099 0.0362 3.97 0.0025 N PENT ANE 106 0. 099 0.0357 3.96 0.0025 NITROGEN 114 2.485 0.0000 0.00 0.0240 METHANE 100 89.511 0.0000 905.94 0.4958 C 0 2 117 1.011 0. 0000 0.00 0.0154 ETHANE 101 5.

82

ENERGY MEASUREMENT WITH AN ON-LINE GC

Figure 6.

009 1.3389 88.81 0.0520 TOTALS 100.000 1.9470 1054.69 0.6239 * @ 14.730 PS I A DRY & UNCORRECTED FOR COMPRESSIBILITY ** @ 14.730 & 60 DEG. F COMPRESSIBILITY FACTOR (1/2) = 1.0024 DRY B.T.U. @ 14.730 PS I A & CORRECTED FOR COMPRESSIBILITY = 1057.2 SAT B.T.U. @ 14.730 PS I A & CORRECTED FOR COMPRESSIBILITY = 1038.8 REAL SPECIFIC GRAVITY = 0.6253 UNNORMALIZED TOTAL. MOLE % = 99.97 ACTIVE ALARMS NONE SUMMARY The gas chromatograph is an excellent tool for accurately determining a number of physical properties of natural gas such as B.T.U., specific gravity and exact composition. The microprocessor and improved design of the hardware aided by environmental chamber testing have made it possible to use gas chromatographs in remote unattended locations. The Danalyzer is designed to provide the most stable, reliable performance over the widest environmental conditions, ± 1/2 B.T.U in 1000 over the entire ambient temperature range of 0–130°F. Each analyzer is thoroughly tested in an environmental chamber to assure the reliable low maintenance operation that the gas industry requires.

83

Figure 7.

FIELD EXPERIENCE IN MEASURING ENERGY DELIVERED WITH GASEOUS FUELS William H.Clingman and Lyn Kennedy Consultants, Precision Machine Products, Inc. Kenneth R.Hall and James Holste Chemical Engineering Faculty Texas A&M University

ABSTRACT

A classic method of measuring energy delivered is to use an orifice meter run and a circular chart recorder to obtain an integrated value for the flow. This flow measurement is combined with an average calorific value to obtain energy delivered. This method can be biased when changes in the calorific value correlate with changes in the flow. This classic method is next compared with two newer methods of measuring energy delivered. The first method uses a flow computer which has as inputs the pressure, temperature, and pressure differential from an orifice meter. This method also uses continuous measurements of relative density and calorific value. The output from the flow computer is the instantaneous energy flow. The second method uses a new energy flowmeter that is under development. It directly measures energy flow without measuring either calorific value or volumetric flow. INTRODUCTION The general trend in the natural gas industry is to move toward therm billing of large customers. These large customers are primarily power plants and industrial customers. Because the cost of fuel is rising and because therm billing is applied to the larger users, a premium can be paid for more precise measurement of the energy delivered. This paper covers our field experience with different methods of making this measurement. The experiments were carried out at a Lone Star Gas meter run outside of Bryan, Texas. The gas at this site is odorized before distribution and the flow measurement is used to control the automatic addition of odorant. Downstream about one mile from the site is a City of Bryan municipal power plant, which is a major user of the gas. Remaining gas is distributed to the City of Bryan. About five miles upstream from the site is a Champlin gas plant, which is the source of the gas. Gas flows at the site are of the order of 500,000

85

SCFH to 1,000,000 SCFH in an 8 inch line. The calorific value varies between 1000 and 1180 BTU/SCF depending on the amount of ethane left in the residue gas by the Champlin gas plant. Three methods for measuring energy delivered from the site are compared. The first method is the classic method using an orifice flowmeter with circular charts and an average calorific value over the time period being billed. The integrated volumetric flow over the time period corrected to standard conditions is obtained from the circular chart. This integrated flow is multiplied by the average calorific value to obtain energy delivered. The second method uses a flow computer to calculate and record both the instantaneous and integrated energy being delivered from the site. Continuous inputs to the flow computer are the pressure differential across the orifice plate, the absolute temperature and pressure at the orifice, the relative density of the gas from a gravitometer, and the calorific value from a calorimeter. The third method uses a new device being developed under a contract with the Gas Research Institute. This device is an energy flowmeter that measures the energy delivered directly without measuring either integrated flow or calorific value. The device consists of two main components, a flow separator and a Flow Titrator . The flow separator splits off a sample stream from the main flow. The flow separator operates so that the ratio of the main flow to the sample flow is independent of the flow rate and the gas composition. This ratio is called the split. The second component is a Flow-Titrator, which mixes a stoichiometric flow of air with the fuel sample stream. This stoichiometric air flow multiplied by the split is proportional to the energy being delivered in the main line. In the following section of this paper, each of the three methods is described in more detail by giving the specific parameters of the test equipment at Bryan. This section is then followed with a discussion of some of the measurement errors associated with the classical method. These errors in principle can be eliminated by using either a flow computer or the new energy flowmeter. The final section presents the experimental results comparing these three methods of energy measurement at the Bryan test site. TEST EQUIPMENT In the classical method of energy measurement at the Bryan test site, the flow measurement was made by Lone Star Gas with an orifice meter run. This run consisted of a 5 inch orifice in an 8 inch line. A standard meter run was used with a flow straightener 10 feet upstream from the orifice. The differential pressure measurement was taken from flange taps and recorded on a seven day circular chart. Also recorded on the chart was the temperature and pressure of the gas in the line. All measurements of the calorific value and relative density were made at the Champlin gas plant about five miles upstream from the orifice meter. No other fuel was added to the line, however, between the two measuring points so the gas composition was the same at both points. A Therm-Titrator was used for the calorific value measurement and a Ranerex gravitometer for relative density. The two instruments were read hourly and these readings were averaged to obtain an average calorific value and average relative density for the measurement period. The circular chart was calculated by the Lone Star measurement department. This was done by tracing the pressure and differential pressure curves with an integrater and using an average value of the gas temperature. The average relative density was used in determining the meter factor for the orifice plate. The integrated volumetric flow was multiplied by the average calorific value to obtain the delivered energy. When energy delivered was measured by the flow computer the same orifice meter run was used. A Rosemont differential pressure transducer, however, was used across the flange taps to obtain an electrical signal for input to the flow computer. A pressure transducer was also inserted in the line at the orifice meter run to provide a continuous electric input to the flow computer. The flow computer was manufactured by

86

FIELD EXPERIENCE IN MEASURING ENERGY DELIVERED WITH GASEOUS FUELS

Elliot Automation in Houston. Additional inputs to the computer were the calorific value and relative density of the gas. These were measured continuously by a GB-2000 instrument from Precision Measurement Incorporated (PMI). This instrument is a combination of the Therm-Titrator and a new PMI instrument for measuring relative density. Although the flow computer was provided with a temperature input, the temperature sensor had not yet been installed when the data presented in this paper were collected. The temperature of the gas as recorded on a circular chart at the site varied from 83° to 87°F. An average temperature of 85°F was programmed into the flow computer. From all of these continuous inputs, the calculated instantaneous energy flow in the line was displayed on a strip chart recorder. The flow computer used the relative density and NX-19 procedure to correct for supercompressibility. The equipment for the third method of measurement, the energy flowmeter, consists of two main components: a Flow Separator and a Flow-Titrator. The flow separator is illustrated in Figure 1. The flow divides into two streams, a main flow and a sample flow. In each flow stream there is a constrictor which produces a small pressure drop, In the main stream this constrictor consists of 720 parallel rectangular channels, each with a length of one inch and dimension of 1/16″×7/16″. The hydraulic diameter of each channel is 7/64″. The channels are arranged in concentric rings about the center of the flow constrictor. This constrictor is in the 8″ line at Bryan and is located just downstream of an elbow and shut-off valve. A sample line extends from the center of this flow constrictor to the outside of the spool piece which contains it. Outside of the spool piece there is a capillary in the sample line which acts as a second flow constrictor. Placed downstream from both constrictors are pressure ports which lead to a Rosemont differential pressure gauge. The sample flow to the Flow-Titrator is regulated by a control value. The minicomputer of the FlowTitrator controls the sample flow so that the pressures are the same downstream from the flow constrictors. The design insures that the upstream pressures are the same. Under these conditions, the pressure drop across the flow constrictors is on the order of ½ psi with a flow in the 8 inch line on the order of 600,000 SCFH. The ratio of the main flow to the sample flow is called the split. The split during most of the experiments at Bryan has been about 300,000. The split can be changed by altering the capillary dimensions. It is a characteristic of the flow separator design that the split is invariant with changes in the flow and gas composition. The variation of the split with flow in laboratory tests is shown in Table 1. The separator tested was of similar design to that at Bryan except that many fewer channels were used and the split was much less. The channels had the same dimensions as at Bryan. The laboratory tests were made using air. The main gas flow (air) was Table 1. SPLIT VERSUS FLOW Slotted Plug Expanding To 2” Pipe 0.080 Inch Capillary Expanding to ¼” Pipe REYNOLDS* NUMBER

MAIN FLOW (SCFH)

SAMPLE FLOW (SCFH)

SPLIT

14288 12623 11625 9081 8847 7533 7520 4406

6321 5506 5083 3934 3847 3920 3280 1942

42.64 37.67 34.69 27.10 26.40 22.48 22.44 13.15

148.2 146.2 146.5 145.2 145.7 146.3 146.2 147.7

*Reynolds number for flow in sample line channel.

87

Figure 1. ENERGY FLOWMETER REYNOLDS* NUMBER Dynamic Range for flow=3.2 Mean Split=146.5

MAIN FLOW (SCFH)

SAMPLE FLOW (SCFH)

SPLIT

88

FIELD EXPERIENCE IN MEASURING ENERGY DELIVERED WITH GASEOUS FUELS

REYNOLDS* NUMBER Standard Deviation=0.7%

MAIN FLOW (SCFH)

SAMPLE FLOW (SCFH)

SPLIT

measured with a calibrated Roots meter and the sample flow with a wet test meter. A tee was used to separate the flow into a main stream and a sample stream. Both flow streams were vented to the atmosphere after passing through their respective flowmeters which were downstream of the flow constrictors. Pressure drops across the flowmeters were negligible compared to those across the flow constrictors. Thus the arrangement assured equal upstream and downstream pressures between the main line and sample line. In Table 1 the split appears as a function of gas flow Reynolds number in the channels. It is important to have a constant split because then energy flow in the main line is always in the same ratio to energy flow in the sample line. The latter energy flow is measured by the Flow-Titrator and then the main line energy flow can be calculated from the split. The second main component of the energy flowmeter is the Flow-Titrator. This is a modified ThermTitrator and its basic layout is shown in Figure 2. The fuel sample stream is mixed with air and burned. The air flow is adjusted so that the air-fuel mixture being burned is essentially at the stoichiometric point. This stoichiometric air flow is then measured and is proportional to the rate at which energy is being delivered in the main line. A continuous output from the energy flowmeter is displayed on a strip chart recorder as the percent of calibration reading for the instrument. DISCUSSION OF PRECISION AND BIAS The classic method of measuring energy delivered lacks some precision and has some sources of bias. There is some loss in precision when following by hand the circular chart traces of pressure and differential pressure. This must be done when reading the circular charts to determine the integrated volumetric flow. This loss in precision is eliminated when using either a flow computer or the energy flowmeter, which both use microcomputers to handle the processing of the data. An instantaneous value of energy flow is integrated in the microcomputer and there is no human involvement in calculating the final result. Adding to the loss in precision is the fact that normally average values of relative density are used in reading the circular charts. The relative density is used to calculate the meter factor for the orifice plate. The gas composition can vary significantly over the measurement period. When this happens a significant error can occur if there is a strong correlation between flow variations and composition changes. The correlation can be either positive or negative. This problem is amplified when one multiplies the integrated flow by the average calorific value in the classical method. There are two approaches to eliminating these errors. The first approach is to use a flow computer supplied with a continuous measurement of calorific value and gravity. The errors are eliminated because all measurements are continuous and instantaneous. No averages are used and the energy flow itself is the only quantity integrated over the measurement period. In the second approach the energy flow is measured directly by the instrument so that the need of calculating it from other measured quantities is eliminated. EXPERIMENTAL RESULTS The first group of experimental results are shown in Table 2. These data compare the classic method with the energy flowmeter. The energy flowmeter is under development with support by the Gas Research Institute. Data obtained so far have been on a relative rather than an absolute basis. Equipment is still under

89

Figure 2. FLOW SAMPLING DEVICE

development for field calibration of the energy flowmeter and this is required for an absolute measurement. Thus, in the tables the output of the energy flowmeter is given as the percent of calibration reading. The error in any measuring instrument can be divided into three components: calibration bias, drift, and random errors. In our experiments the emphasis has been on assessing drift and random errors for the three measurement methods. Calibration error depends primarily on the quality of the standard that is being used. Random errors occur in all measuring methods and drift may or may not occur. In our experiments these are estimated by comparing two methods. Assume that M1 and M2 are simultaneous measurements of the energy delivered by methods 1 and 2. Let E be the true value of the energy delivered. In the methods that we have studied calibration involves a multiplicative calibration constant. In that case,

where C1 and C2 are calibration errors, d1 and d2 are drift, and e1 and e2 are random errors. in successive measurements with time C1 and C2 are constant, d1 and d2 would change monotonicaliy with time, and e1 and e2 would be random. If there is drift in either instrument then either d1 or d2 or both will be significant and will either continuously increase or decrease with time. Since the two instruments are independent, the drifts in each will not be the same and the measurement ratio, (M2/M1) will also drift. In our experiments no evidence has been found for a drift in the ratio (M2/M1) when comparing the energy flowmeter with either the flow computer method or the classical method. Thus, it is concluded that none of the three methods being studied have significant drift over the periods of measurement, which have been one to two weeks in duration.

90

FIELD EXPERIENCE IN MEASURING ENERGY DELIVERED WITH GASEOUS FUELS

Table 2. COMPARISON OF ENERGY FLOWMETER WITH CLASSICAL METHOD DATE LONE STAR GAS VOLUME MMCF

LONE STAR CALORIFIC VALUE BTU/SCF

MMBTU/ DAY ENERGY FLOWMETER % OF CALIBRATION

RATIO OF CLASSICAL METHOD TO ENERGY FLOWMETER

5–5 5–6 5–7 5–8 5–10 5–11 5–12 5–13 5–15 5–16

1017 1024 1016 1016 998 1012 1017 1018 1017 1016

18464 20018 19027 17029 16958 16994 16322 17653 18122 17942

203.2 191.7 210.3 206.7 193.5 212.1 206.1 201.6 195.9 198.3

18.155 19.549 18.727 16.761 16.992 15.792 16.049 17.341 17.819 17.660

90.85 104.42 90.47 82.37 87.65 80.14 79.19 87.56 92.52 90.49

Mean=202.0 STD. DEV=3.5%

The next step was estimating the magnitude of the random errors for the three methods from the measured ratios (M2/M1) and (M3/M1) . In this notation the subscripts 1, 2, and 3 refer to the energy flowmeter, the classic method, and the flow computer method respectively. Let s1, s2, and s3 equal the variances of the measured values, M1 , M2, and M3. For example if there are N measurements or M1 then an estimate of s1 is given by:

For a large number of data points the square root of the variance would be equal to the standard deviation for the method. The variance for the ratio (M2/M1) is given by (s1+ s2) when the variances are small. Thus (s1+s2) can be estimated from the standard deviation for the data in Table 2, which is 3.5%. Likewise (s1+s3) can be estimated from the standard deviation for the data in Table 3, which is 0.3%. Table 3. ENERGY FLOWMETER—FLOW COMPUTER COMPARISON RATIO OF FLOW COMPUTER TO ENERGY FLOWMETER ENERGY FLOV7METER READING %

FLOW COMP. MMBTU/ 12 HR. AVERAGE DAY

24 HR. AVERAGE

FIRST MEASUREMENT PERIOD 8/14 8/15

120.61 119.57 122.09 126.91

11525 11487 11580 12143

95.6 96.1 94.8 95.7

95.8 95.3

91

RATIO OF FLOW COMPUTER TO ENERGY FLOWMETER ENERGY FLOV7METER READING %

FLOW COMP. MMBTU/ 12 HR. AVERAGE DAY

24 HR. AVERAGE

FIRST MEASUREMENT PERIOD 8/16

118.41 11226 85.31 8205 8/17 88.29 8475 SECOND MEASUREMENT PERIOD 8/18 84.26 8216 90.52 8803 8/19 92.22 8999 90.02 8846 8/20 85.54 8246 90.80 8887 8/21 84.63 8106

94.8 96.2 96.0 97.5 97.2 97.6 98.3 96.4 97.9 95.8

95.4

97.4 97.9 97.2

From this analysis it is possible to conclude from the Table 3 data that the precision of either the energy flowmeter method or the flow computer method must be better than 0.3%. On the other hand it can be concluded from the Table 2 data that the standard deviation for the classical method would be greater than 3. 4%. CONCLUSIONS Experimental work is continuing on these methods of energy measurement. From the initial data, however, we can conclude that there are two alternatives to the classical method that can provide a significant increase in precision. One method would use an orifice meter, flow computer, and continuous measurements of calorific value and realtive density. The other is a new instrument under development which measures energy flow directly. In addition to having a high precision these methods integrate an instantaneous value of energy flow thereby avoiding errors caused by using an average calorific value. This work was partially supported by the Gas Research Institute.

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT Craig J.Louttit, B.S. Gas Engineer Pacific Gas and Electric Company San Francisco, California 94106

ABSTRACT

Several process gas chromatographs (GCs) were tested to verify whether this instrument is an accurate alternative to the calorimeter for measuring the heating value of natural gas. The test data presented in this report is divided into three groups. Group 1 compares the rangeability and repeatability of several gas chromatographs and a calorimeter to a laboratory GC, the reference standard. Group 2 compares the actual operating data for three GCs and a calorimeter, which analyzed pipeline gas over a 104–day test period. Group 3 is a separate test, which compares operating data for one GC and a calorimeter for a for 93–day test period. The results showed that the process gas chromatograph is an accurate alternative to the calorimeter. For Group 1, the GCs and the calorimeter had less than 1% error for gasses in the range of 875 to 1050 Btu/scf. For Group 2, the average heating value difference between the calorimeter and the three GCs was 2.0 Btu/scf, while the average heating value difference among the gas chromatographs was 0.5 Btu/scf. For Group 3, the average heating value difference between the calorimeter and GC was 0.1 Btu/ scf. INTRODUCTION The average wellhead price of natural gas has increased twelve-fold in the past ten years. Because of this increase, less expensive gas with heating values in the range from 700–900 Btu/scf is being purchased. The traditional heating value of distribution quality natural gas has been 1000 Btu/scf. Since gas is bought and sold on an energy basis, the accuracy of the instruments which measure the heating value of the natural gas is becoming increasingly important. Two instruments used by the natural gas industry to measure the heating value of natural gas are: 1) the combustion calorimeter and 2) the gas chromatograph (GC). These instruments measure the heating value

93

by two distinct methods. The combustion calorimeter directly measures the thermal energy generated from the combustion of a known volume of gas. In contrast, the GC determines the heating value indirectly by measuring the composition of the gas and calculating the heating value by using the heat of combustion of each gas component. The purpose of the test described in this report was to determine the ability of the GC to accurately measure the heating value of natural gas and to determine whether the GC is an acceptable alternative to the calorimeter. The data presented in this report compares three gas chromatographs with one another and also with a calorimeter. The data is intended only as a comparison between direct and indirect Btu measurement devices, and is not intended to recommend a particular manufacturer’s product. As such, all devices shall be referred to as GC-A, GC-B, GC-C, GC-D and calorimeter. BACKGROUND Historically, the Cutler Hammer Calorimeter has been the standard in the gas industry for measuring the heating value of natural gas. In fact, the Cutler Hammer Type AB Recording Calorimeter is the only calorimeter accepted in the U.S. as a referee standard (ASTM Standard and Method D-1826). The calorimeter determines the heating value by measuring the thermal energy released by the combustion of a known volume of gas. The heating value measured by the calorimeter is usually recorded on a continuous strip chart recorder. Gas chromatographs have been used in the laboratory for many years. The old laboratory GC required an operator to inject samples and to interpret the results. Recent advances in microprocessor technology have allowed gas chromatographs to become automated process instruments designed for continuous and automatic operation. The automated GC or process GC can operate on-line to a process, with no operator required and analysis times of ten to fifteen minutes. Using the principle of selective adsorption, the GC separates a natural gas sample into the various hydrocarbon constituents, measures the amount of each component, and calculates the heating value, specific gravity and compressibility by using the physical constants of each gas component (ASTM Standard and Method D-1945 and D-3588). The numerical results of an analysis are usually printed, and include the mole percentage of each component, heating value, specific gravity and compressibility. Four considerations prompted the study of gas chromatographs as an alternate method of measuring heating value and specific gravity. 1. The increasing cost of natural gas justifies an improvement in accuracy of the heating value and specific gravity measurements. 2. The high cost of maintenance and unavailability of replacement parts required the replacement of old calorimeters. 3. The cost of automated process GCs has recently become competitive with new calorimeters. 4. The use of microprocessors in process GCs allows the GCs to become an integral part of a gas company’s data acquisition system. TESTING PROGRAM A program was initiated to evaluate several gas chromatographs on the basis of accuracy, repeatability and rangeability. These terms are defined as follows:

94

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

1. Accuracy Accuracy is defined as the degree of conformity of the measured value to the standard value. 2. Repeatability Repeatability is defined as the ability of an instrument to repeat its readings over a given period of time, given a constant input. 3. Rangeability Rangeability is a measure of the practical operating limits. Chromatograph accuracy is dependent upon calibration to a known reference gas composition. The amount which a sample gas composition can deviate from the calibration gas while still maintaining accuracy is a measure of the Chromatograph’s rangeability.

Test Groups The data included in this report is from three different test groups. The data from the first group, Group 1, is from laboratory testing of several gas chromatographs. The data from the second group, Group 2, is from operating data gathered from three gas chromatographs and a calorimeter that are in field use. The data from the third group, Group 3, is operating data from a separate comparison between a gas chromatograph and a calorimeter that are in field use. All the GCs tested used a thermal conductivity detector to measure the gas components. The main difference among the three groups is the presence of a reference standard in Group 1. Group 1 Reference Standard. The control group for this phase of the testing was a laboratory GC. This device is the most accurate and precise heating value measurement device in our system, and so was taken to be the reference standard to which the process GCs and calorimeter are compared. Calibration. The laboratory GC was calibrated with a calibration gas which was blended gravimetrically. The amount of each component in the calibration gas was weighed with weights traceable to the National Bureau of Standards (NBS). The gas composition was verified on a GC, and composition values stated in the manufacturer’s certification were used in the calibration. The process GCs were calibrated against the laboratory GC using another independently-supplied calibration gas. Groups 2 and 3 Reference Standard. The GCs and calorimeters operated under field conditions and did not have a reference standard. Calibration. The GCs and calorimeters were individually calibrated by a separate independently-supplied calibration gas.

95

TEST PROCEDURE Group 1 The testing procedures were developed to determine the accuracy, repeatability and rangeability of the automated process gas chromatographs. These parameters were compared for several gas chromatographs and also to a calorimeter. Repeatability. For this test, two GCs were supplied by a cylinder of gas with a constant heating value. The test was performed over a seven-day period, to allow for the effects of any instabilities or zero shifting of the instruments to be measured. Since the cylinder contained a finite volume of gas, the normal sampling time of ten to fifteen minutes would empty the cylinder before the end of seven days. Therefore, the GCs were programmed to sample and analyze once an hour for a total of 170 analyses during the seven-day period. Rangeability. For this test, gas chromatographs A, B, and C, and a calorimeter were compared to the laboratory GC at different heating value levels. The calorimeter was calibrated using the standard calorimeter calibration gas from IGT. Group 2 Group 2 data was gathered from instruments that were in actual field use. Specifically, the data from three GCs and one calorimeter are analyzed. The results compare both chromatographs to calorimeters and chromatographs to other chromatographs. This data lacks a reference device and, as such, reflects only the relative difference between the respective instruments. For this portion of the testing, three gas chromatographs and one calorimeter were compared against one another. Each instrument was set up to sample and analyze gas from the same pipeline. Each GC was calibrated with a separate calibration gas with an approximate heating value of 1023 Btu/scf. Group 3 Group 3 is a separate comparison of operating data from a gas chromatograph and a calorimeter operated in parallel for a 93-day period. This test was performed using a gas chromatograph (GC-D) and a calorimeter that were set up in parallel. A single sample probe and sample line were used. Both instruments were calibrated at the start of the test and were subsequently calibrated on a weekly basis. The gas chromatograph was calibrated automatically while the calorimeter was calibrated manually. The heating value of the calibration gas for the calorimeter and the GC were 1015 Btu/scf and 1027 Btu/ scf, respectively. The range of gas measured varied from 1044.2 to 1086.1 Btu/scf for the 93-day test period. TESTS RESULTS The test results are presented in two parts. The first part describes results from the laboratory testing (Group 1), and the second part shows data gathered from the instruments in field use (Groups 2 and 3).

96

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Results—Guoup 1 Repeatability. The results shown in Figure 1 show the normal curve for GCs A and C. This is the distribution of the 170 data points of the constant Btu gas for each of the two instruments. As can be seen, the difference in the average heating value measured was 0.07 Btu/scf, or 0.007% of range, which is well within the expected tolerances. The standard deviation is a measure of the repeatability of a given measurement. As shown, GC-C has a smaller standard deviation and is therefore more repeatable than GC-A. It should be noted that both devices are within their manufacturer’s quoted tolerances of ±1 Btu/scf. Rangeability. As can be seen in Figure 2 , the accuracy of all the units is very good (0.2%) at heating values above 950 Btu/scf. Below 950, the accuracy of GC-A and GC-B decreases rapidly with decreasing heating value. In contrast, the accuracy of GC-C and the calorimeter remains fairly accurate even at the low heating values. This discrepancy at the low Btu levels is related to the device’s rangeability or its ability to accurately measure gas samples whose composition vary significantly from the calibration gas. The composition of the lowest heating value gas differs significantly from the composition of the calibration gas. Specifically, the low Btu gas has 23% nitrogen, while the calibration gas has 2.5%. Results—Groups 2 and 3 Group 2. Presented below are the results of the 104–day test period: Measurement Device Calorimeter GC-A GC-B GC-C

104-Day Average (Btu/scf) 1068.2 1066.2 1066.0 1066.5

The test results are divided into two categories: 1. Comparison of gas chromatographs to calorimeters 2. Comparison among several chromatographs This data is presented graphically in Figures 3 through 20 and represents the following pairs: Calorimeter versus GC-A Calorimeter versus GC-B

Calorimeter GC-A GC-A GC-B

versus GC-C versus GC-B versus GC-C versus GC-C

For each pairing of the data the following graphs are used:

97

Figure 1. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Repeatability Test)

– A comparison of the daily average heating value for each pair. – The daily average heating value difference for each pair. – The distribution of the daily average heating difference for each pair. Comparison of Gas Chromatographs to Calorimeter Group 2. Figures 3–11 represent the data comparing the calorimeter to GCs. Figures 3, 6 and 9 show the ability of the calorimeter and GCs A, B and C respectively to track the pipeline gas. Figures 4, 7 and 10 are graphs of the calorimeter’s daily average heating value minus the respective GC daily average heating value. Figures 5, 8 and 11 show the distribution of the daily average heating value difference for the respective instruments. The data in Figures 3 through 11 are summarized by Figures 5, 8, 11 and the table below. In all cases, the calorimeter has a consistently higher daily average heating value. The 104-day daily average heating value difference in this group of data is as follows: Data Cal.—GC-A Cal.—GC-B Cal.—GC-C

104-Day Average Difference 2.0 Btu/scf 2.2 Btu/scf 1.7 Btu/scf

Percentage of Data Points Within ±1 Btu 29% 27% 40%

98

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 2. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Rangeability Test)

As can be seen for this test period, the GCs all measured lower than the calorimeter. Group 3. Figures 21 through 23 graphically represent the data for this 93-day test. The daily average heating values for the calorimeter and GC-D are 1072.5 and 1072.4 Btu/scf, respectively. This amounts to a daily average difference of 0.1 Btu/scf, showing much closer agreement between the two measuring devices. Comparison Between Several Chromatographs This data is presented in a similar fashion to the above information. Figures 12, 15 and 18 show the ability of the respective GCs to track the pipeline gas. Figures 13, 16 and 19 show the daily average heating value difference between the respective GCs. Figures 14, 17 and 20 show the distribution of the daily average heating value difference. Again, the data shown in Figures 12 through 20 is summarized by Figures 14, 17, 20 and the table below. Data GC-A—GC-B GC-A—GC-C GC-C—GC-B

104-Day Average Difference 0.2 Btu/scf 0.3 Btu/scf 0.5 Btu/scf

Percentage of Data Points Within ± 1 Btu 97% 94% 99%

As can be seen from Figures 14, 17 and 20, the agreement among the chromatographs is very good.

99

Figure 3. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter Compared to Gas Chromatograph A)

It should be noted that the Btu value for the calorimeter is read off a strip chart recorder and the GC provides a printout of the actual heating value. Reading the heating value off the strip chart may induce some errors, as it is subject to the operator’s interpretation. CONCLUSIONS The results of this series of tests demonstrate that the gas chromatograph can provide accurate and reliable heating value measurement. From the field testing, Group 2, we can see that the the 104-day average between the calorimeter and the gas chromatograph varies between 1.7 and 2.2 Btu/scf. This number is within the combined accuracy tolerance as specified by the manufacturer. In contrast, in Group 3, the 93-day test showed that the average heating value difference was 0.1 Btu/scf. These results showed a much better correlation between the two devices. Since the field data were gathered under actual operating conditions, there are many items that could not be as carefully controlled as in a laboratory testing. Some of these items consist of: operating and maintenance procedures, brief periods of equipment downtime and errors in calibration that could account for the difference in heating value. It is important to note, in Group 2, the very close agreement among the three GCs. All three GCs agreed within 0.5 Btu/scf over the test period. Also, the number of data points that were within 1 Btu/scf ranged

100

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 4. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Daily Difference Calorimeter minus GC A)

from 94% to 99% for the three GCs. In contrast, comparing the calorimeters to the GCs, the number of data points that agreed within 1 Btu/scf ranged from 28% to 40%. In Group 3, comparing the calorimeter to GC-D, 71% of the data points were within 1 Btu/scf. The most important point in all of this testing is the need to develop a high-confidence, primary standard calibration gas for use with gas chromatographs, since an instrument’s accuracy is dependent upon the accuracy of its calibration. This is particularly true for gas chromatographs that require calibration gases with up to eleven components, with C6+ as low as .05% ±2%. The difference in the results from Group 2 and Group 3 indicate the need for a proper quality assurance program to ensure accurate, reliable and repeatable heating value measurement. Finally, the GC offers additional information that will help improve flow measurements by having “live” data concerning specific gravity and CO2, N2 composition which is used in the AGA 3 calculations.

101

Figure 5. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter versus Chromatograph A)

102

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 6. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter Compared to Gas Chromatograph B)

103

Figure 7. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Daily Difference Calorimeter minus GC B)

104

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 8. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter versus Chromatograph B)

105

Figure 9. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter Compared to Gas Chromatograph C)

106

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 10. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Daily Difference Calorimeter minus GC C)

107

Figure 11. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter versus Chromatograph C)

108

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 12. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Chromatograph A Compared to Gas Chromatograph B)

109

Figure 13. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Daily Difference GC A minus GC B)

110

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 14. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Chromatograph A versus Chrorratograph B)

111

Figure 15. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Chromatograph A Compared to Gas Chromatograph C)

112

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 16. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Daily Difference GC A minus GC C)

113

Figure 17. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Chromatograph A versus Chromatograph C)

114

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 18. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Chromatograph B Compared to Gas Chromatograph C)

115

Figure 19. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Daily Difference GC C minus GC B)

116

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 20. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Chromatograph B versus Chromatograph C)

117

Figure 21. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter Compared to Gas Chromatograph D)

118

COMPARISON OF DATA FROM DIRECT AND INDIRECT BTU MEASUREMENT

Figure 22. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter Compared to Gas Chromatograph D)

119

Figure 23. DIRECT VERSUS INDIRECT BTU MEASUREMENT (Calorimeter versus Chromatograph D)

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER James M.Minich, P.E. Design Engineer Specialist NATURAL GAS PIPELINE COMPANY OF AMERICA

ABSTRACT

As the price of eras has risen substantially from the 20c/Mcf levels of days gone by, companies are painstakingly measuring the flow in order to bill their customers accurately—and charge them not only on quantity, but also on the gas’ quality or heating content. In the last few years, pipeline firms have begun to look at electronic flow measurement systems to replace the time-honored chart recorders. The electronic process of collecting data inputs and performing the calculations in determining the flow of natural gas has been engineered by Natural Gas Pipeline Company of America, a subsidiary of MIDCON Corp. The Control Department has used minicomputer based control systems since the late I960’s in the area of automation and communications along its 12,750 miles of pipeline. As equipment is retired, it is replaced with new state-of-the-art electronic equipment. Much of this ground work helped to develop an electronic flow measurement system for natural gas. The flow calculation, based on the NX-19 study and AGA Report #3, is accomplished with the use of software modules, inputs received from the gas chromatograph and operator entered parameters and transducers. All of the hardware equipment was purchased from local vendors as off-the-shelf items. Only a mininum of in-house design was required to interface field transducers, input/ output contacts and power. The bulk of the project consisted of the design, development and testing of software programs. A microcomputer accepts all these inputs and executes a set of instructions that calculate the flow of natural gas. Custody transfer refers to information obtained by electronic flow measurement and used for billing purposes. There’s been a lot said about electronic flow measurement replacing chart recorders and the potential savings in converting to electronic flow measurement. This paper describes how this is done. NGPL began considering electronic flow measurement with the installation of the Trailblazer pipeline. This accurate and repetitive calculation performed by the microcomputer allows the program to execute control actions under variable input conditions. For example, a setpoint is changed in gas

121

FIGURE 1.

flow, or there is the need for an additional run, or the value of BTU has changed. The partners involved on Trailblazer had met several times to discuss the type of information to be collected at the metering sites. The entire system is developed around two areas: Hardware and software. A hardware block diagram is shown in Figure 1. All software programs were developed and tested in-house. Natural Gas Pipeline Company has developed an electronic flow measurement system for custody transfer. The flow measuring system contains the following features: 1) Makes a flow calculation in accordance with NX-19 and AGA Report #3. 2) Receives BTU, Specific Gravity, N2 and C02 analysis data from a chromatograph. Having control over the software allows the measurement department a choice in selecting a chromatograph. Currently the microcomputer can interface with three different chroraatographs. 3) Has remote, local and standby control mode capability. 4) Produces a hard copy printout every hour, daily summary at 8:00 AM and on demand. Also all setpoint changes and alarms, when set and cleared, are printed. 5) Accepts 32 analog inputs; 48 digital inputs and 48 digital outputs. 6) Contains complete hardware and software diagnostics for on-line debugging. 7) Allows operator changes of selected parameters(i.e. orifice size, run sizes, default value for BTU and Specific gravity, override setpoints, etc) through the console unit.

122

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER

8) Performs control routine actions such as flow control, runchanging, line break detection and back pressure override. 9) Uses floating point double precision arithmetic for all computations. 10) Displays all active alarms to the operator through the console. 11) Converts and stores data into a data base using standard conventions-ASCII, Integer, and Floating point formats. Attached to each value is a current status indicator. Data can be flagged as *—old data, M— manually entered data, and b (blank)—good data. 12) A calibration routine for transducers. 13) Projects flow over a 24 hour period. 14) Hourly log displays flow factors used in calculating the orifice flow factor constant for each run. 15) Hourly and Daily log record running accumulated flow totals per run. 16) Accepts a pressure override and flow control setpoints. 17) Capable of annubar or orifice plate measurement. I would like to point out some advantages and disadvantages associated with the development of this system. ADVANTAGES a) Software changes can be made quickly and in direct response to the company’s needs. b) Modular construction allowed for quick and easy configuration of software modules. c) The capability of troubleshooting software modules using an in-house debugging aid-IDR. This has been the single most valuable asset in developing a working procrram. d) Gives accurate and reliable flow measurements. e) Software written by NGPL for earlier energy systems is compatible to new versions as they are marketed. f) Enables the real time collection of data for more accurate measurement over the conventional 24 hour chart averaging. g) hardcopy printouts are tailored to the station needs or a general format can be used. h) the microcomputer can use various types of chromatographs i) the use of two programming- languages, “C” and macro assembly. C is used in all flow calculations. DISADVANTAGES a) The process of debugging software programs has been a time consuming effort. b) Considerable time was spent in training personnel on the new equipment and in documentating procedures. c) There were single transducers (static, temperature, and pressure) per run. Multiple transducers per run would offer variable range coverages. For example, 0–50", 0–100" and 0–200". NGPL is planning to experiment with dual range transducers inputs into the microprocessor at metering sites experiencing wide fluctuations in flow rates. d) Power interruptions as a result of electrical storms have been a bier problem. The electronic flow measurement for custody transfer system consists of a microcomputer, gas chromatgraph, H20 and H2S analyzer, a differential pressure, static and temperature transducers for each

123

FIGURE 2.

run, communications modem, power supplies and battery charger, printer, and console terminal. All of this hardware, except for the printer, transducers, chromatograph and analyzers, is mounted to a free-standing 19" equipment rack. Terminal blocks are provided on this rack for analog and digital inputs (see photo). Following is a description of topics considered in electronic flow measurement as approached by the Engineering Control department of NGPL. PHOTO

HARDWARE 1.0.1 THE MICROCOMPUTER The microcomputer is the heart of the system. It is a 16-bit microprocessor with added features such as memory management, FP-11 instruction set, double precision arithmetic, and four-level interrupt bus structure. It and eight other boards collect and perform all the necessary calculations for an accurate flow measurement. A typical backplane configuration depicting computer board placement is shown in Figure 2.

124

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER

1.0.2 POWER The main source of power is a battery charger that converts AC power to DC voltage levels required by the microcomputer. Between the battery charger and backplane there are 3 DC regulated power supplies which provide +5 and +12 volt for the backplane and +28 volts for tranducers. Battery backup is provided so that the unit may operate for a number of days in the event, of a power loss. On the Trailblazer pipeline an emergency power unit was installed. Its main function was to supply enough power to the chromatograph, modem, and micro to continue giving analysis data during loner term power outages. An UPS (uninterruptible power supply) should be a requirement in rural areas where power is not clean or reliable. 1.0.3 GAS CHROMATOGRAPH Natural Gas Pipeline Company is evaluating a number of energy measurement systems that reports natural gas data as each sample is analyzed, approximately every 13 minutes. The gas chromatocrraph reports Btu, specific gravity, and gas composition by name and percent of content for each analysis and as a 24-hour average at the end of each day. The chromatograph provides hardware diagnostics by number which allow a user to check the machine before operation. 1.0.4 INPUTS/OUTPUTS Data enters/leaves the computer by various methods. First, transducers are connected to analog inputs (32 maximum inputs) via a terminal block. Measurement transducers for differential pressure, static pressure and gas temperature are installed for each run and interfaced to the microcomputer using an analog input. A zero point is established and maintained as a reference point for analog signals. Second, digital input contacts provide information about ambient temperature. AC power failure, write lock inhibit, and positional information about valves. In most every case, they are nothing more than relay contacts’. Third, digital outputs are bit mask assignments that result in a control action. For example, the action may be pulsing a solenoid that opens or closes a valve, or a time duration pulse that pulses a motor controlling a regulator. All inputs and outputs have lightning protection and are coupled to the microcomputer through optical isolators. Finally, serial communication data links are connected between the chromatograph, console, and micro. The console furnishes the means by which an operator can manually alter a number of variables, such as tap position, orifice plate sizes, run sizes, and flow setpoint. It also displays individual data items by entering the correct request code and performing the proper read or write operation. Normally these functions are performed locally. See APPENDIX E. 1.0.5 PRINTERS AND HAND-HELD (CONSOLE)TERMINAL Data values can be retrieved from the computer in five different ways. First, individual data items can be requested and displayed on the console unit. Second, multiple data items can be looked at by referring to the hourly and daily reports printed automatically. Third, data can be polled by a partner through an asynchronous serial communication link in an agreed format. Fourth, remote commands and data requests can be entered through a synchronous serial communication link over a microwave system. Gas control in

125

Lombard or Houston can poll remote sites and are capable of sending a flow setpoint, changing orifice and pipe diameters, and requesting individual/all data items. For example: alarms, flow rate, H20 content, BTU, C02, etc.. Finally, service personnel can attach a CRT to view 120 memory locations. SOFTWARE The software is the brains of the system. It is responsible for collecting transducer and chromatograph data, sensing and sending contact closures, remote and local communications and the time base and event driven execution of control routines. Routines most commonly used are flow control, runchanging, backpressure override and linebreak. The memory itself is limited to 32k-16 bit words and the last 4k is designated as the I/O page. Memory is divided into two parts: UVPROM and Random Access Memory. In the event of a power loss the program is non-volatile. Using a Prolog Prom Programmer the program is burnt on 4k ×8 UV PROMS. The number of proms used is dependent on the length of the program. Approximate size of the current program is 16k. The ram portion of memory is a static CMOS read/write memory with battery backup. This 12k of ram is used as a scratch pad area for data values, calculations, flags, buffers, pointers, counters, etc. If a power loss is experienced data is retained. Ram memory can be cleared by the operator. The software is modular in construction. That is any module can be removed or inserted with no adverse effect on the overall operation of the program. Testing of software modules requires a three-step process. First, a module is individually tested until all bugs are ironed out. Second, it is then added to the overall system where it undergoes further debugging. Finally, the program is installed and field tested. It may require additional changes because no bench test can give every possible condition. However, the program is very accurate and logical in carrying out its many sequences of operation. The EXECUTIVE routine controls the sequence in which the individual programs run. It does this by executing a series of subroutine calls first on a non-timed basis followed by those on a timed basis. The EXECUTIVE also has the duty of monitoring remote communications. In the event of a communication loss, the communication program is re-initialized. On a startup there are several modules that define hardware addresses to memory locations and various bit mask assignments, assign addresses and processor status to all the interrupt vectors, reset the system hardware and initialize all the system software and hardware. A software block diagram is shown in Figure 3. 2.0.1 SYSTEM STARTUP The configuration information required by the program is entered by the local operator during a prompt and answer session. The number and type of questions can vary from one location to another. Normally, the following information is requested: stream selection, plate and run size for all runs, static tap location (upstream or downstream), default values and upper and lower limits for BTU and specific gravity, and a flow and back pressure override setpoints. In entering the last prompt value, the microcomputer begins executing the software modules. The operator still may exercise control or change data through the console terminal. For example, he can change the control mode to either local or remote, plate or orifice sizes, tap location, time and date, and force the program into a calibration. He may also view any value, or alarms using the console terminal. To view the current flow value the operator would key in a number code and the terminal would display the current value. The read operation can be performed at any time. However, changes made to the program, as in a write operation, are controlled by a key interlock switch preventing accidential or unauthorized entries. A fail-safe feature in case of a computer malfunction is the

126

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER

FIGURE 3.

incorporation of a watchdog hardware timer. If the timer is not strobbed within a given period of time, the DC power is cut off to the equipment control circuits. 2.0.2 GAS CHROMATOGRAPH The microcomputer polls the chromatograph for gas heating value, specific gravity, gas composition for each analysis and 24-hour averages, the asynchronous digital link permits the program to remotely interrogate and gather data from the chromatograph. The link is activated by modules ECOM1 and ESTAT. Basically these modules request a message, determine its status and process the data received for the current

127

sample analysis. If the 24-hour averages are available, they too are collected and stored in the data base common. Once stored in the database, other software modules convert the data into applicable data formats and store them in the database. Of the three chromatographs used one is capable of transmitting 9 messages. Of these only five are requested. Type 4 message is polled continuously since it contains current data analysis, 24-hour average analysis and alarm information. To ensure that the micro did indeed receive the messages sent by the chromatograph, a cyclic redundancy checksum (CRC) is performed(bit error detection). The chromatograph calculates a CRC and stores it as the final two bytes—low “byte transmitted first. The micro using the same polynominal [P(x)=x16+x12+x5+1] compares its result with the CRC sent in the received message buffer. Only after the message has successfully been identified does it get processed. As an example the following data messages are requested from a chromatograph: 1) 2) 3) 4) 5)

TYPE 1 TYPE 2 TYPE 3 TYPE D TYPE 4

BTU, Specific Gravity Compositional analaysis data 24-hr averages for BTU and Specific Gravity 24-hr averages for concentration data Current status message

The dicrital link between the microcomputer and eras chromatographs operate asynchronously, 1200 baud, and in full duplex mode. 2.0.3 HARD COPY PRINTOUTS The software can support three types of printers. Each printer is capable of displayina identical information if necessary. However, the format is different for each case. The first printer has 132 characters/line the others 80 and 32. (See APPENDIX C for examples.) Printouts are supplied locally and to partners wishing to tie into an asynchronous port with their own equipment. The printouts need not be identical. That is, the information printed for our billing purposes may not be identical to others, or they just may like to see additional data points. Whatever the case may be, various files are built with the desired information and outputted using various FORMAT modules. These modules, respond to basic directive commands and are directed through a serial communication link to external eguipment. Chromatograph values on the hourly log and averaae values on the daily log are flagged with their current status. Indicators are of the following type: *—old data; M—manually entered data; and b(blank) for current good data. An example would be, BTU * 1044.0 BTU/cf or BTU M 1044.0 BTU/cf. 2.0.4 FLOW CALCULATIONS The microcomputer, together with a gas chromatocrraph and transducer inputs, computes a gas flow rate based on AGA report #3 and PAR research project NX-19 equations. A flow calculation is performed every 10 seconds and memory locations are updated based on these latest calculations. At the various metering sites, each run is handled as an independent flow calculation and the sum of these calculations provides a total flow measurement for the station. Also, in addition to a gas flow rate value, a therm flow rate value is

128

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER

also calculated by multiplying the BTU by the total flow. The final value is converted to dekatherms. In the event of a power loss, battery backup ram is provided for memory retention of system components. Orifice flow rate calculation and supercompressibility factor are calculated using floating-point double precision data type. This greatlv improves the accuracy of arithmetic operations allowing direct operation on double precision 64-bit words. All flow calculations are performed using “C” programming language. Currently the program is a hybrid using “C” and macro assembly language. Orifice flow factors are calculated and printed on the hourly report per run allowing an operator a means of rechecking the total flow computation. There are three modules that perform the necessary calculations required for flow measurement: FLWRAT, AGAFPV, and ORFLC. The first is the orifice flow rate calculation, C'. (See APPENDIX A for detailed information.) For this calculation the operator can select tap location, UPSTREAM or DOWNSTREAM for flange taps. Note: pipe taps were not considered since they are not used. The second is the supercompressibility factor, Fpv. Finally, the gas flow rate is calculated in MMCFD and MMBTUD. The accuracy of the microcomputer flow rate calculation was compared to a ‘FLOW FINDER’ modular program inserted into the hack of a TI59 calculator. This together with an overlay and instruction manual provides all that is necessary to verify a flow calculation. This module is available from the SOLTAR Corporation located in Spring, Texas. I’ve listed what appears to be the differences between their method of calculation and the one we use in Appendix B. 2.0.5 CONTROL ROUTINES Not only is the microcomputer a data acquisition system, but also a control system. It collects and transmits data and performs the necessary control actions based on current flow calculations. Modules are periodically run by the EXEC to determine if control is necessary. The control programs implemented are flow control, runchanging, backpressure override, and linebreak detection. Flow control controls a regulator that responds to a flow setpoint received from a remote location or entered locally through the console. Results of the current flow are compared to the setpoint and differences between the two will increase or decrease the position of the valve. Runchanging monitors the differential pressure on run number one for 90 inches of water. If reached, the next available run is opened only after the current flow is reduced by 60%. This prevents the orifice plates from receiving a large flow of gas and possibly damaging the plates. If there are no further runs to open and the DP is at a maximum, runchanging will control the controller so that the DP will not exceed 90 inches. Likewise, if all available runs are open and DP drops below 10″ the furthest available run from run #1 is closed. When closing a run there is no need to reduce the flow. Linebreak detection will monitor rate of flow and pressure differences on a timed basis. Currently this module is under development. Back pressure override looks at down stream pressure. If it exceeds a predetermined level, the condition is flagged and the flow controller will control the flow of gas through the regulator so that downstream pressure is below the pressure setpoint.

129

2.0.6 CALIBRATION Since tranducers are calibrated on a periodic basis, a module called CALIBC was designed to allow calibration of the transducers using the console unit and a series of front panel switches. A mode switch allows the operator to view transducer inputs (analog signals) in three formats: Engineering unite, counts, and voltages. 2.0.7 DIAGNOSTICS To ensure that the hardware and software are operational there are available to the operator hardware and software diagnostics. There are three levels of maintenance. For the first level there are hardware diagnostics available for the chromatograph and microcomputer. The measurement operator or communication technician can perform these tests before he starts the system UP. The chromatograph software has 8 self checks built into it which flag contradictions in its logic. These errors are a result of some system failure. Also the chromatograph provides a diagnostic package designed for rapid on-site verification. The microcomputer provides diagnostic programs that test the processor, the memory and the user’s console. The second level of operator maintenance is through a software module called DGNSTC. The diagnostic module operates in coniunction with front panel LED’s and toggle switches. DGNSTC interprets the 16 panel switches as a memory address and outputs the contents of that memory address on the LED’s. Peripheral device addresses (considered memory addresses by the hardware) can be viewed as well. If a byte address is selected, it is displayed in the low byte of the LET) word. Switches requesting a display of a non-existent memory location cause all ones to appear on the LED lights. The third and final level of maintenance, not made available to the operator, is an IDR or interactive display routine. IDR is a diagnostic utility which enables the user to view and change the contents of memory locations in a program being debugged. This module when assembled with the program displays on a CRT, 3 full pages or 120 memorv locations showing critical buffers, pointers, counters, etc. (APPENDIX D illustrates such a page of information that might be displayed in troubleshooting a control routine called RUNCHG.) This scanning technique was developed in-house to assist in the debugging of software modules before they are introduced into the field. This has been the most valuable tool in the development of software modules. 2.0.8 COMMUNICATIONS and PARTNER HANDOFF The communication processor routine, COMM, has the duty of handling data requests and command messages from the Master station via microwave communication circuits. Natural Gas Pipeline Company has developed its own protocol to which a cyclic redundancy checksum (CRC) is added for error checking. The generating polynominal used for CRC or error bit detection is X16+X15+X2+1. Partners wishing to retrieve information from the processor database have available an asynchronous RS232 port. They need only connect a cable between this port and their equipment. NGPL has designed a ASCII format transmission for partner handoff. This communication format that NGPL uses for data exchanges is 1200 baud, asynchronous with 1 start bit and 2 stop bits, all bytes 8 bit ASCII (except CRC), no parity, transmitted as often as requested(normally between 1 and 10 minutes), and appended with a CRC.

130

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER

2.0.9 MISCELLANEOUS ROUTINES The previously mentioned routines are by far the most important, but there are other programs necessary to complete the successful operation of the flow measuring system. I would like to mention a few of these: A software library that stores a collection of routines useable by any module within the software structure; input and output modules designed to handle contact and control signals; internal averaging and conversion routines for data handling; a real-time line clock for the date and time; a millisecond timer; a module that projects flow over a 24-hour period; and a routine that calculates a total down time of a run. Each station has these and other programs necessary for total and accuracte flow measurement. 2.0.10 CURRENTLY UNDER DEVELOPMENT Currently NGPL is looking at upgrading the measurement system to take advantage in the state-of-the-art. For example NGPL is evaluating a unit that has a 20Mb hard disk drive, 2Mb floppy, self contained power supply and 4×8 backplane all within a single package. Not only is the hardware under going a change but also the software. The software is a hybrid of macro assembly and “C”. Our target is to convert the entire program into C and install a OS and “C” compiler on the hard disk eliminating the need for UVPROM and use the floppy to record statistical and necessary data to be used by gas accounting in it’s billing of customers. The floppy is seen as possibly storing up to 30 days worth of data. 3.0 CONCLUSIONS The Control Department of Natural Gas Pipeline Company has developed an electronic flow measurement system that accurately computes gas flow data and yet allows flexibility in the overall performance and operation of the system. Features can be added to the system simply by burnincr a new set of UVPROM in exchange for the old ones (hardware changes may be necessary). The advantages of an electronic flow measurement system experienced thus far are: Information is available immediately—locally and remotely —a much more accurate flow measurement compared to chart recorders, and the flexibility of addina and/or modifyincr software programs quickly. Currently NGPL has 14 flow measurement systems in service. Four of them are on the TRAILBLAZER system. A data acquisition, control and flow calculation program is operating successfully at Natural Meter in Beatrice, Nebraska. APPENDIX A The basic flow equations follow the AGA Report #3 and NX-19 equations for supercompressibility. Please refer to the publications for further explanation. APPENDIX B

131

APPENDIX C

APPENDIX D NAME MSTIME OVRDFG VMIPFG CPRTN F.DP +2 +4 RUN #1 +2 +4 RUN #2 +10 + 12 RUN #3 +16 +20 I . STAT +2 +4 ENABLE

ADDRESS 120000 140174 140216 143124 141214 141216 141220 142042 142044 142046 142050 142052 142054 142056 142060 142062 143126 143130 143132 141526

CONTENTS 76342 0 0 35373 400 41423 40625 400 1 0 400 0 0 400 0 0 10400 2200 0 1

NAME SETTLE VCLOSE I.ALM2 +2 +4 FLWMAX SPMFLG SPOFLG CTLMFG PULSFG CTLDIR RFCLOS RFCIPF RUNA F.RUNA +2 DROUT DR IN

ADDRESS 140222 140226 143134 143136 143140 140216 140210 140212 140164 140176 140162 140236 140240 140246 142037 142040 177772 177774

CONTENTS 0 0 400 0 0 0 0 0 11 11 0 0 0 0 0 0 0 200

APPENDIX E FUNCTION DEFINITIONS FUNCTION NUMBER DESCRIPTION 1 Daily Gas Flow Rate 2 Daily Thermal Flow Rate

UNITS OPERATOR CHANGEABLE MMCF/day MMBTU/day

132

ELECTRONIC FLOW MEASUREMENT FOR CUSTODY TRANSFER

FUNCTION NUMBER DESCRIPTION UNITS 3 Daily Accumulated Gas Flow MMCF 4 Daily Accumulated Thermal MMBTU Flow 5 Yesterday’s Accumulated Gas MMCF Flow 6 Yesterday’s Accumulated MMBTU Therms 7 Daily Projected Gas Flow MMCF 8 Runs open none 9 Discharge pressure PSIG 10 Run 1 Status (1 = open, 0 = closed) 11 Run 1 Differential Pressure inches w.c. 12 Run 1 Gas Temperature degrees F. 13 Run 1 Static Pressure psi 14 Run 1 BTU value of gas BTU/cu. ft. 15 Run 1 Specific gravity none 16 Run 1 Carbon Dioxide Content mole% 17 Run 1 Nitrogen Content mole% 18 Run 1 Pipe Diameter inches 19 Run 1 Orifice Diameter inches 20 Run 2 Status (1 = open, 0 = closed) 21 Run 2 Differential Pressure inches w.c. 22 Run 2 Gas Temperature degrees F. 23 Run 2 Static Pressure psi 24 Run 2 BTU value of gas BTU/cu. ft. 25 Run 2 Specific gravity none 26 Run 2 Carbon Dioxide Content mole % 27 Run 2 Nitrogen Content mole% 28 Run 2 Pipe Diameter inches 29 Run 2 Orifice Diameter inches 30 Run 3 Status (1 = open, 0 = closed) 31 Run 3 Differential Pressure inches w.c. 32 Run 3 Gas Temperature degrees F. 33 Run 3 Static Pressure psi 34 Run 3 BTU value of gas BTU/cu. ft. 35 Run 3 Specific gravity none 36 37 38 39 40 41

Run 3 Carbon Dioxide Content Run 3 Nitrogen Content Run 3 Pipe Diameter Run 3 Orifice Diameter Run 4 Status (1 = open, 0 = closed) Run 4 Differential Pressure

OPERATOR CHANGEABLE

Yes Yes

Yes Yes

mole % mole % inches inches inches w.c.

Yes Yes

133

42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 64 65 66 67 68 69 71 72 73 74 75 77 78 80 81 83 84 85 86 87 88 251

Run 4 Gas Temperature Run 4 Static Pressure Run 4 BTU value of gas Run 4 Specific gravity Run 4 Carbon Dioxide Content Run 4 Nitrogen Content Run 4 Pipe Diameter Run 4 Orifice Diameter Run 5 Status (1 = open, 0 = closed) Run 5 Differential Pressure Run 5 Gas Temperature Run 5 Static Pressure Run 5 BTU value of gas Run 5 Specific gravity Run 5 Carbon Dioxide Content Run 5 Nitrogen Content Run 5 Pipe Diameter Run 5 Orifice Diameter Water Content H2S Content override setpoint standard deviation Therm flow value for run opening Atmospheric Pressure Contract Base Pressure Contract Base Temperature Current Date Current Time Calibrate mode Chroma tograph stream number Run failure clear flag Static tap location 0 = upstream , 1=downstream Control mode Flow Setpoint Pressure Override Setpoint H1 BTU limit Low BTU limit Hi Specific Gravity limit Low Specific gravity limit Default BTU Default Specific Gravity Meter station alarm status

degrees F. psi BTU/cu. ft. none mole % mole % inches inches

Yes Yes

inches w.c. decrrees F. psi BTU/cu. ft. none mole % mole % inches inches

Yes Yes

MMBTU/day none none psia psi degrees F. none none none none none none MMCF/day psig BTU BTU

BTU none

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

REAL TIME ENERGY MEASUREMENT Albert P.Foundos, President Alfred F.Kersey, Gas Industry Manager Fluid Data, Inc. 1844 Lansdowne Avenue Merrick, New York 11566

ABSTRACT

This paper deals with real time energy measurement from the industrial gas users point of view. It illustrates how real time energy measurement by the user allows him to optimize his combustion process utilizing automatic combustion control and high speed measurement devices for feed forward and feedback systems. Fuel Gas Quality Small gas utilities with single gate stations will distribute gas as received from the pipeline. The quality of pipeline gas is not constant. Since most gas contracts are based on MMBTU rather than MCF, the pipeline has little incentive to sell gas stable in BTU. All he must do is stay within contract BTU limits. The pipeline may have little control over the gas quality. In today’s market, he may only be the transporter where the user has purchased gas from a producer on the “spot market”. The demand for gas components will also determine the degree of stripping by the processor or producer who knows what quality of gas will come from a storage field. There are many reasons that the quality of pipeline gas is not stable. A typical span of pipeline BTU over a year might be 1000 –1066 BTU on a 30" mercury column 60°F and dry base. Larger utilities will have several gate stations and gas will be supplied to them by several gas pipelines. This means that the utility is introducing gases of varied quality into his grid. Depending on supply and gate station pressure conditions, interfaces as high as 50 BTU differential may occur anywhere in his station. In addition, he may have an LNG or a propane air peak shaving plant in operation which may further complicate the interface picture. He might still be operating his SNG plant. Table 1 lists different gases which are currently used in domestic United States. Some of these are used undiluted. Some are blended to achieve a desired calorific value.

135

Table 1—Fuel Gas Heating Characteristics Gases

Calorific Value Typical Range Wobbe Index Calorific Value Specific Specific Gravity

Specific Gravity

Pipeline Refinery LNG SNG Commercial: Propane Butane Blast Furnace Coke Oven Coal By-Product

1000–1060 600–1800 1000–1240 950–1000

1197–1298 695–2085 1240–1538 1247–1313

0.63 0.49–1.0 0.65 0.58

2450–2600 3100–3300 75– 120 400–525 150–450 150–3000

1987–2109 2220–2363 75– 120 649–852 229–686 134–2683

1.52 2.0 1.00 0.38 0.43 1.00–1.5

Measuring Heating Value It is necessary to measure heating value to determine selling price and combustion efficiency for given burners. Numerous instruments are currently used for determining the heating value of fuel gases. These include calorimeters, gas chromatographs, stochiometric instruments and gravitometers. The proper instrument is determined by the particular make-up of the fuel gas, the purpose of the measurement and the required speed of response. Some BTU instruments respond to a limited span of gas components and are not suitable for all applications. Speed of response of instruments referred to above, vary from near instantaneous to 15 minutes for 100% response. Feed forward control is best served by instruments which respond accurately to the fuel gas components and which respond quickly, preferrably in a matter of seconds. Applications A. Mixing of Gases. Mixing of gases is done for several reasons. Pipelines may inject rich gas to meet minimum BTU’s as specified in the sales contract. Gas utilities inject various mixtures and amounts of butane air, or propane air, for peak shaving. These mixtures can be anywhere from 1010 BTU to 1550 BTU depending on the percentage of peak shaving. LNG is popularly used for peak shaving. Some utilities supply LNG for base load input. There may still be in operation some SNG plants used for peaking shaving or base load. In the refinery and petrochemical area, a typical blending application involves blending of by-product or waste gases with a natural gas supply to optimize use of internally produced fuel gas. For the steel industry, a very significant user of fuel gas, optimizing involves mixing of blast furnace gas so that it will burn. Steel makers blend a variety of mixtures of blast furnace gas and coke oven gas, and natural gas plus air, for maintaining total flexibility within a plant production to optimize energy delivery and/or consumption. To mix fuel gases on line, a sample feedback control loop is often used, such as that shown in Figure 1. This depicts a ratio controlled flow sytem reset by the heat value measurement to produce a gas with a constant

136

REAL TIME ENERGY MEASUREMENT

Figure 1.

Figure 2.

heat quality. There are many other variations of the calorimeter based blending operation which includes either wild flow blending or feed forward control. In all cases, the ultimate objective is to obtain a fuel gas of constant CV or Wobbe Index based on the specific requirements and application of the blended gas. B. Feed Forward Combustion Control. Largo consumers of fuel gas are discovering that the gas flow control systems they have employed for years car no longer deliver a controlled heat flow to their burners. The typical industrial combustion process consists of pressure or flow control loop and a feedback loop— see Figure 2—to reset the pressure or flow of gas to the burners based on load. These conventional control loops were based on the requirement that both the specific gravity and calorific value (BTU/SCF) of the fuel gas remain constant. We have illustrated that this is not true for most industries and localities. As a rule, the composition of these gases varies even within relatively short periods resulting in fluctuations in calorific value and relative density. Controlling the pressure and flow of a fuel gas alone cannot deliver a constant heat flow if the gas composition varies. The control problems presented by gas composition variations can be overcome by adding a hi-speed calorimeter to monitor heat value and by using a conventional flow orifice to measure flow to an existing burner control system. These two measurements, heat value and flow rate, in combination with the conventional control and feedback loop based on loan, provide the basis for effective, continuous feed forward control of the burning process. The high-speed calorimeter referred to burns and measures the heat value of a continuous sample of the gas in terms of Wobbe Index.

137

Gas flow is determined by measuring differential pressure (P) across an orifice in the gas line. P is corrected for pressure and temperature to more correctly represent gas flow.

The feed forward control loop built around the high speed calorimeter (Fig. 3) compensates for changing gas characteristics in advance of combustion. The combination of feed forward control using Wobbe Index plus a process feedback loop assures the user of correct process heat and makes possible significant reductions in excess air. Those save fuel and do not sacrifice safety or product quality. With the advent of microprocessor “on line” high speed calorimeters can be monitored to assure reliable calibration and control for reproducible control of the desired set point. SMARTCAL (tm) is such a dedicated microprocessor to interface with calorimeters for auto-calibration and various diagnostic and alarm functions to assure “on-line” reliability. Conclusion: The ability to measure heating value accurately and quickly and apply this for immediate correction in a feed forward system is current state of the art. Systems like this that maintain optimum combustion and process efficiency are certainly the key to “Real Time Energy Measurement “and energy management.

138

REAL TIME ENERGY MEASUREMENT

Figure 3.

ENERGY MEASUREMENT ACCURACY R.N.Curry, Retired Texas Gas Resources Corporation

ABSTRACT

The theme of this panel “What Needs to be Done to Achieve Energy Measurement Accuracy” opens a host of doors! Those areas that come most immediately to mind are cybernetics, gas ehromatography, computer systems hardware and software, theory and thermodynamics, mathematical technique, and undoubtedly numerous other technological and scientific facets associated with metrology! Most of these areas have been thoroughly “wrung out” by the engineering and scientific communities and there is an abundance of published data concerning same. In essence, I believe that energy measurement accuracy is generally available to us now within acceptable engineering tolerances—at least for hydrocarbon fuel gases—and what we must now do to achieve greater accuracy is establish a precise method for gathering the data and deriving the units of energy in a manner universally consistent between all concerned—buyer, seller, transporter, etc. Consequently, what I believe must be done is in the areas of standardization and education. This can probably be best accomplished by the publication of a single document such as has been done for gas measurement—ANSI 2530. Any number of standards have been published by generally acceptable but different professional organizations but there are conflicts between these standards and while such conflicts may be of minor concern to those technically familiar, they are of great consternation of others associated with the industry, more specifically, lawyers, accountants, and non-technical executives. Some examples are: 1. The use of either .2563 or .2561 as the vapor pressure of water ; 2. The definitions of Btu and heating values; (It is interesting to note that in one of the aforementioned publications, it is pointed out that ten different “hand-books” picked at random gave ten different definitions for Btu!)

140

ENERGY MEASUREMENT ACCURACY

3. The heating value of hydrocarbon components; 4. The method of gathering and processing samples; 5. The manner in which the chromatograph is operated and maintained; 6. The manner in which the calorimeter is operated, maintained, and installed; 7. The proof of calibration of all associated devices including water vapor content instruments. And there are others. This makes the case for a universal standard—at least for the North American continent. A professional task group should be established to review and evaluate all currently available standards and compile items into one document, hopefully, with the sanction of all organizations. That will not be easy—just ask Mr. Hoglund! The educational effort would be directed toward the use of such a standard and to familiarize the public as well as hydrocarbon fuel industry employees with energy measurement. Neither will that be easy. We have people currently in the industry that think the difference between gross or total and net heating value for gas is the same thing as the difference between wet and dry. And, yes, some of them are technical employees. You know the lawyers and accountants say we don’t even know what to call it— MMBtu, Dekatherm, Joule, or what?! And, heating value is non-dimensional, consequently, one cannot tell the engineer to design a pipeline or a prime mover for “X” MMBtu per day. He must know mass or volume. Recently, after a rather lengthy explanation of an energy measurement contract clause with a client and when asked if he fully understood, he replied, yes, but how many MCF is that? So, it becomes apparent, that energy measurement accuracy will be greatly enhanced through standardization and education in addition tot he conventional improvements needed.

Natural Gas Energy Measurement II Presented April 30-May 2,1986 Chicago, Illinois

CALCULATION OF THE HEATING VALUE OF NATURAL GAS FROM THE PHYSICOCHEMICAL PROPERTIES OF THE PURE COMPONENTS Kenneth N.Marsh, Ph.D. Director, Thermodynamics Research Center Texas A and M University College Station, Texas 77843

ABSTRACT

In the gas industry, the value of natural gases are based on their gross heating values (gross calorific values). Other properties that are important for calculating the amount of gas on a volumetric scale are the relative density (specific gravity) and the compressibility factor. The gross heating value is difficult to measure accurately for a flowing stream whereas the analysis of the composition of the stream by gas chromatography can be done accurately by well established techniques. Thus it is possible to replace a direct calorimetric measurement by a composition determination and a knowledge of the gross heating values of the various components of the gas. We have recently made new recommendations for determining the heating values and the specific gravity of selected components of natural gas based on a re-evaluation of the available experimental data. The method for calculating the gross heating value from these recommended values will be outlined. This calculation method is applicable to all common types of gaseous fuels including dry natural gas, reformed gas, propane—air, carbureted water gas, and retort coal gas provided suitable methods of analysis for all the components of the gas are available. INTRODUCTION The gross heating value of a gas is the maximum amount of energy that can be obtained from burning a unit of the gas when liquid water is formed as a product. When using gases as heating agents, the relative merits of gases from different sources and with different compositions can be compared on the basis of their heating values. The gross heating value of a fuel is also used to determine the price of gas in custody transfer, and it is required to calculate the efficiencies of energy conversion devices such as gas turbines.

143

The specific gravity (relative density) of the gas is required for calculating flow through pipes and orifices and to establish the relationship between the amounts of gas measured volumetrically to the corresponding mass or molar amounts. The relative density can be determined from a knowledge of the molar volumes of the component gases as a function of temperature and pressure and selected binary mixture properties. Since the gross heating value of a gas is difficult to measure accurately in a stream, the alternative method of calculating the heating value from a compositional analysis and the heating values of the individual components is preferred. For custody transfer purposes, particularly across national boundaries, it is important to establish a set of internationally accepted values for the gross heating values and the molar volumes of the major components of natural gas (1). Natural gas consists of methane with small amounts of other hydrocarbons, principally in the range of 2 to 6 carbon atoms, carbon dioxide, nitrogen, argon, water, and hydrogen sulfide. When burned in oxygen or air, the products of combustion are carbon dioxide (and if hydrogen sulfide is present, sulfur dioxide) and water. The heating values of the components must be known over the range of reference temperatures used in each of the different countries exchanging natural gas. This temperature range is from 0 to 25°C (32 to 77°F). The gross heating value of a gas, Q, is the total energy generated in an ideal combustion reaction at a standard temperature and pressure in which all the water that is formed appears as liquid (2,3). The ideal gaseous state is chosen for the reactants and products so that the numerical value of Q is independent of the actual state of the reactants and products. For example, the heat released by one pound of methane burned at 1 atmosphere pressure at 60°F will be slightly different from that released under the same conditions if the gas was assumed ideal. The gross heating value of a gas is thus an ideal gas property and it is the negative of the standard enthalpy of combustion, . Calculation of the exact heat produced by the combustion of mixtures containing many components would require a very large amount of auxiliary data on the multicoraponent mixtures and a very complex program. Further, it is more important that all parties agree to the same method for calculating the gross heating value, rather than having an exact value, which would differ slightly from the ideal gas value. CALCULATION FROM COMPOSITION The standard enthalpy of combustion of a substance is an ideal gas quantity because the value refers to the reaction (1) where * denotes ideal gas and 1 denotes liquid at a particular reference temperature T and a pressure of 1 atmosphere. Each component in a mixture of natural gas can be represented by the general chemical formula CiHjSk, and the ideal combustion reaction for each component where the reactants are in the ideal gas state and the products in the ideal gas state is: (2) This ideal reaction produces the net enthalpy of combustion ΔHc (T,net) for the burning of one mole of CiHjSk. Measurements of this property are almost always made at 25°C (298.15 K). To convert from 25ºC to another temperature requires the ideal gas specific heats (heat capacities) per mole, (3)

144 CALCULATION OF THE HEATING VALUE OF NATURAL GAS FROM THE PHYSICOCHEMICAL PROPERTIES OF THE PURE COMPONENTS

Conversion to the gross enthalpy of combustion requires subtraction of the ideal enthalpy of vaporization for water, . This value is slightly larger in magnitude than the enthalpy of vaporization: (4) Values of standard enthalpies of combustion for the individual components are normally tabulated in kilojoules per mole and are for the formation of liquid water as a product. Because the heating value of a natural gas is an ideal gas property, the gross enthalpy of combustion for a mixture is the sum of the standard enthalpy of combustion times the mole fraction of each of the components, (5) where ΔHC (T,gross,mixture) is the gross enthalpy of combustion for the mixture, xi is the mole fraction and c is the number of components. The gross heating value of the mixture on a molar basis, is (6) and this value can be converted to the gross heating value on a mass basis by division by the molar mass of the mixture M given by: (7) Hence (8) The gross heating value on a volume basis results from multiplying

by the ideal gas density:

(9) For a dry gas, these calculations are straightforward provided the enthalpy of combustion data are available. However, the calculation for a wet or saturated gas requires further interpretation. CORRECTION FOR WET OR SATURATED GAS Because analyses of natural gases are normally on a dry basis, it is necessary to correct for the presence of water vapor. On the basis of one mole of dry gas, the mole fraction of water is (10) where nw denotes number of moles of water. Rearranging this expression gives (11) It is now possible to adjust the mole fractions to reflect the presence of water. Since the total moles of dry gas is 1, the total moles of wet or saturated gas n(cor) is (12) then the corrected mole fractions of the components in the wet or saturated gas are (13) The xi(cor) become the corrected mole functions to use in the preceding equations to calculate the properties of wet or saturated gas. If it is not convenient to determine the mole fraction of water by other means, it is common practice to assume Raoult’s law for the saturated gas, (14)

145

where xw is the mole fraction of water in the gas, and is the vapor pressure of pure water at temperature T. For purposes of this calculation, it is permissible to assume the liquid phase to be pure water ( ) and then (15) Additional assumptions required are either a) the gas and liquid phases are ideal and the liquid is incompressible or b) all non-idealities for the system cancel identically. Raoult’s law is probably satisfactory when dealing with natural gas systems at conditions close to ambient (between 0 and 25°C at approximately 1 bar). In the above calculation, water contributes to the gross heating value from the ‘reaction’ (16) which equals the ideal enthalpy of vaporization. Thus, because water vapor enters with the gas, it is possible to interpret the gross heating value as including the enthalpy difference caused by condensing this ‘spectator’ water vapor to liquid water. After much discussion, GPA Standard 2172–86 has chosen to adopt this interpretation. While it may be argued that this is not rigorous, no calculation involving wet or saturated gas can be rigorous. In fact, the method of calculation which is the least ambiguous and least susceptible of errors is that on a dry, net, mass basis. The most ambiguous and most prone to errors is the calculation based on a wet (or saturated), gross, volume basis, and this is the value required for custody transfer. Under the selected interpretation, the gross heating value for wet or saturated gas per unit volume is (17) and (18)

HEATING VALUES OF THE COMPONENTS To calculate the heating value of the gas and the ideal gas volumetric flow rate, it is necessary to have a consistent set of properties of the components of natural gas. These values have been re-evaluated and updated recently (1). The details are not reported here. The new values take into account all the currently available data and are applicable at the reference conditions used in the gas industry throughout the world: 273.15 K (0°C), 288.15 K (15°C), 298.15 K (25°C all at 0.101325 MPa (1 atmosphere) and 288.71 K (60° F) and 0.10156 MPa (14.73 psia). These new values differ only slightly from the values previously recommended in earlier compilations. These recommendations were prepared for the Groupe International de Importateurs des Gas Naturel Liquéfié (GIIGNL) and the Gas Processors Association (4), and they will appear in their manuals and standards. The recommendations are being submitted to the American Society for Testing and Materials (ASTM) and International Standardization Organization (ISO) for their consideration in an attempt to ensure that the same numbers will be used for the calculation of heating values wherever natural gas is bought and sold. Details of the calculation are given in GPA standard 2172–86. This program is available from the Gas Producers Association on a disk for an IBM compatible computer or as a fortran or basic source code.

146 CALCULATION OF THE HEATING VALUE OF NATURAL GAS FROM THE PHYSICOCHEMICAL PROPERTIES OF THE PURE COMPONENTS

COST OF ENERGY When valuing natural gas, the energy delivered to the customer is the pertinent quantity. The cost is: (18) where p is the price per quantity of energy delivered. This value can be on an ideal energy basis or an actual energy basis. As noted previously, the ideal gas value is much easier to uniformly calculate and to ultimately defend. is the energy released as heat during a time period Δt. On the ideal basis, the heating value produces from (19) where is the molar flowrate. Conversion to a mass flowrate gives: (20) where is the mass flowrate. On a volumetric basis, it becomes (21) where is the ideal gas flowrate. Note that each of these expressions is valid, but usually the gas industry wants to deal with real gas flowrates: (22) where is the real gas flowrate. Summarizing these various expressions (23) The details of calculating the compressibility factor are given in reference 4. This latter expression clarifies common misconceptions. It is simpler and more accurate to calculate using molar, mass or ideal gas flowrates. An objection might be that flowmeters, such as orifices, produce not . While this is true, it is also true that with exactly the same information required to develop it is possible to develop , , or . When using or , it is not even necessary to deal with the base pressure because and are independent of pressure while is not. Another misconception is that division of by Z produces the ‘real’ heating value. This is not true. Division of by Z produces which, when multiplied by , provides . is possible to correct the heating value from an ideal property to a real property, but it is a tedious and not very accurate calculation involving equations of state and their derivatives. In addition, the difference between real and ideal heating value is usually a small value at base conditions. CONCLUSIONS The heating value is an ideal gas property which, when used with composition, can rigorously provide a cost expression for natural gas. Rigorously, this should be a calculation based upon dry gas, net value and per mass. Unfortunately, many contracts specify the calculation basis as water saturated gas, gross values and per volume. In this latter case, it is an open question as to how to handle the water carried by the gas if it is wet or saturated. In this paper, the calculation includes this ‘spectator’ water which can add about 0.1% to the heating value. Values can be on an ideal or real basis, but the ideal basis is much simpler and probably more easily defendable. In addition, molar, mass, ideal gas or real gas flowrates are all acceptable although molar or mass are preferable followed by ideal gas, and lastly real gas. Finally, division of by Z does not provide

147

the ‘real’ heating value, but only allows the use of a real gas flowrate rather than an ideal gas flowrate in the cost equation. REFERENCES CITED 1.

2. 3. 4.

Hall, K.R., Yarborough, L, Lindsay, R., Kilmer, J., Fling, W., ‘Calculation of Gross Heating Value for a Saturated Gas from Compositional Analysis,’ International Congress of Gas Quality—Specification and Measurement of Physical and Chemical Propeties of Natural Gas, Gronigen, The Netherlands (1986), April. Gas Producers Association, ‘Calculation of Gross Heating Value, Relative Density and Compressibility Factor for Natural Gas Mixtures from Compositional Analysis,’ GPA Standard 2172–86 (1986). American Society for Testing and Materials, ‘Calculating Calorific Value and Specific Gravity (Relative Density) of Gaseous Fuels,’ D 3588–81 (1981). Gas Processors Association, ‘Table of Physical Constants of Paraffin Hydrocarbons and Other Components of Natural Gas,’ GPA Standard 2145–85, (1985).

GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS G.E.Mattingly, Ph.D. Senior Scientist for Fluid Measurements Fluid Flow Group Chemical Process Metrology Division Center for Chemical Engineering National Bureau of Standards Gaithersburg, MD 20899 U.S.A.

ABSTRACT

As the value of scarce fluid resources increases in today’s domestic and international market places and process industries so does the need for improved fluid measurement and for improved traceability to primary standards. Both buyers and sellers of fluid products are increasingly concerned about accurate custody transfer. Designers and operators of industrial processes are increasingly concerned about the precision of their fluid measurement to optimize the performance of their continuous production technologies. To satisfy these expressed needs for improved fluid measurements and traceability in the wide range of fluids and conditions required, is a considerable task. The calibration facilities which flow gas and which are currently in use at the National Bureau of Standards (NBS) are described. The performance characteristics of these facilities are given together with corresponding levels of uncertainties. The concept of measurement traceability is described. The application of this concept to fluid measurement is presented together with strategies and techniques for establishing Measurement Assurance Programs—MAP’s for flow measurements. The critical ingredients for establishing assured flow measurements are also described as well as an efficient and effective way to process the data that results from such an activity. It is concluded that improving the state-of-the-art in fluid measurement is an evolutionary progression. The fundamental bases for fluid measurements are the primary standards established in the reputable laboratories in the pertinent country. These laboratories, together with the country’s national laboratory, can and should provide for realistic, quantified, and continuing traceability of the nation’s important fluid measurements. In this way, market place equity can be properly and satisfactorily established and appropriate measurement and control can be installed and maintained for optimal productivity in the chemical process industries.

149

INTRODUCTION Standards. Flowrate standards could be significantly simplified if the fundamental bases of these measurements were as simple as those for mass, length, etc. These systems of measurement are based upon discrete standards* or artifacts. For example, the platinum kilogram known as “K-20” is the ultimate artifact to provide the fundamental basis for mass measurement in the U.S. and the platinum meter bar (or its modern-day wavelength equivalent) is the ultimate artifact to provide the fundamental basis for length measurement. Identity Standards. These mass and length artifacts can be considered “identity” standards because under the appropriate conditions of use they define the basic quantity in their respective measurement systems. However, for flow rate measurements of fluids—i.e., liquids or gases, there does not exist an identity standard such as a gallon per minute, a liter per second, or a kilogram per hour. To supply the fundamental basis upon which to establish a flow measurement system, a “derived” standard is needed. Derived Standards for Flow. For gas flow measurements—as needed to form the basis of a national reference system—calibration facilities spanning a range of fluid and flow conditions are maintained by NBS for use by industry and others. These facilities consist usually of: (1) a source of flow—generally a compressor with appropriate auxiliary equipment or a regulated, pressurized tank of gas, (2) a test section into which the meter and its adjacent piping can be installed so that the flow and fluid conditions into it duplicate those expected where the meter will actually be used, (3) a flow determination system having the required level of performance and appropriate proof of this to specify and assure the desired metering performance of the devices in question. Calibration systems are generally categorized according to the type of flow determination scheme used. Several of these schemes will be described below. Flow Determination Systems. The heart of the gas flowmeter calibration facility is the flow determination system. This generally uses a timed collection of the gas which flows through the meter being calibrated. The amount of the gas collected is determined by gravi metric or volumetric techniques. This collected gas is converted to mass flowrate using the collection time; the volumetric flowrate through the meter can be determined using the pertinent thermodynamic properties measured at the meter. This system can be made to perform at a high level of performance to determine the bulk flow rate of gas. Levels of Performance. Measurement systems can be characterized via their precision and accuracy. These are briefly defined as follows:* Precision

– the degree—generally expressed as a percent—to which successive determinations of the same quantity duplicate each other. “Precision” is sometimes further subdivided into “reproducibility”—which involves “how closely will successive determinations duplicate each other” or “repeatability”—which involves “how closely can successive determinations be made to duplicate each other”.

*The term standard is used to refer to “paper” standards which are documents; it is also used to refer to reference facilities and equipment; it is also used to refer to the specific materials needed to transfer measurement quality from or between facilities. These specific materials are referred to in what follows as “artifacts”.

150 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

Accuracy – the degree—generally expressed as a percent—to which a measured result approximates the true value. These characteristics apply to measurements made by flowmeters and to measurements made using calibration facilities. Facility Performance. For gas flow calibration facilities, the precision can be evaluated from the appropriate error budget and from the precision of the component measurements that constitute the system. Difficulty is encountered when facility accuracy is to be quantified because the true value of the gas flowrate is not easily obtained. To estimate possible systematic offsets from true value, approximations— generally very conservative, are frequently used. Alternatively, and more preferably, a realistic and highly defensible traceability scheme either is available or can be generated and is appropriately used to document the systematic offset of a calibration facility. Traceability. Traceability is defined many ways by many people, see Appendix 1. Conventional calibration procedures can establish traceability of types 1 and/or 3. Conventional Calibration Procedures. With conventional calibration programs, a testing laboratory or a meter manufacturer or a meter user might own and routinely use a master meter technique to assess the flowrate measurement performance of the laboratory. To do this, the master meter might be sent yearly to NBS for a calibration in the appropriate NBS flow facility. This done, the meter would be returned to the laboratory with a report on its performance in the NBS facility. The meter would be placed into the respective facility in the laboratory and then calibrated. The relative performance of these calibrations would hopefully compare very favorably and thereby document the closeness of agreement between the laboratory’s facility and NBS. This procedure—while widely used at the present time—can leave a considerable number of factors affecting measurement completely unassessed. Traceability of type 4 (see Appendix 1) might be established for a flowmeter calibration laboratory in the following manner. If calibrated weights (for example from a state office of weights and measures) were used to check a scale system and if a timing standard were used to check the lab’s timing system, then traceability—type 4 could be asserted for the lab’s weigh-time system. However, the overall ability of the lab to calibrate a flowmeter can be quite incomplete. For such reasons, it is widely believed that type 2 traceability is preferred. This type 2 traceability can be established and maintained via flow measurement assurance programs—i.e., flow MAPs. Flow MAPs. In the case of flow MAPs, a procedure different from the conventional calibration one is used, see [1–4].* This involves NBS (or an initiating laboratory) sending a very reliable and well characterized artifact package (i.e., tandem meter arrangements consisting of two meters in series) to the laboratory in question with the request for a calibration of the device(s) according to tightly specified and prearranged conditions. The results—which would contain the effects of all the lab’s routine calibration procedures—its facilities, its operating conditions, its personnel, and its techniques for calculating final results from raw data—are then sent to NBS. These can be objectively (and informedly) compared to NBS results or, more preferably, to similar results from a number of other comparable labs which have performed the same tests in a “round-robin” set of these calibrations. In these comparisons, NBS results are also incorporated as one of the participants. The results show quantitatively, the agreement (or disagreement) among the participants’ results. Algorithms have been developed to handle these results, see [5–8]. Fig. 1 shows a comparison of conventional calibration procedures and those that can occur with MAPs. The comparison shows that the crucial advantages of the MAP program are that: (1)all aspects of the

*A number of useful definitions are given in the Glossary in Appendix 1 .

151

FIGURE 1. CONVENTIONAL CALIBRATION VS. MAP COMPARISON.

laboratory’s measurement processes are checked, and (2) there is a “feedback” and, if necessary, a “followup” activity that can make improvements, etc. These follow-up activities are directed either at the lab’s procedures or at its calibration procedures and facilities, depending upon the results of the algorithms that can be applied to the round-robin data. Basic Calibration Procedures. Calibrations are usually performed using a facility that includes a source of flow, the meter and connecting piping, and a flow determination system, all connected in series. A typical system is illustrated in Fig. 2. Control volumes, see [9], a, b, c are shown for a meter, connecting pipe and calibrator volumes, respectively. The volumes are separated by imaginary control surfaces, 1, 2, 3 and 4. For reasons to be discussed, inlet piping to the metering element should be specified and should provide a suitable and reproducible flow pattern at the inlet to the metering element, and this pipe is considered herein as a part of the meter and volume a. Depending on the type of calibrator, control surface 4 of volume c may be a moving piston, the stationary end of a tank, etc. A description of calibration strategy follows, see [10–12]. Conservation of Mass Equation. A calibration usually requires a determination of mass rate of flow (or sometimes volume rate) through the meter. An application of the conservation of mass equation illustrates some of the problems involved. The equation as applied to a flow system of a fixed control volume V enclosed within a surface A can be written in vector form as, see [9]

*Bracketed integers refer to references given below.

152 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

FIGURE 2. TYPICAL CALIBRATION FACILITY.

(1) where, in compatible units M is the substantial time derivative of the mass in the control volume, p is the mass density, ∂p/∂t represents the partial derivative of fluid density p with respect to time, is vector

153

velocity of the fluid and is the control surface element of area with a direction taken outward and normal to surface A. The first term on the right hand side of (1) represents an integration covering total volume V to find the rate of accumulation of mass, ∂m/∂t, in volume V if density p is not constant. The second term on the right represents net mass rate of flow through the control volume surfaces. Application of Eq. (1) to each of the control volumes a, b and c of Fig. 2 gives three equations, which when added together gives mass flowrate as (2) where subscripts 1 and 4 represent surfaces illustrated in Fig. 2, subscript n represents the normal component of velocity at the surface, and subscripts, a, b, and c represent the respective control volumes. Calibration Facilities at NBS. Calibration facilities are available at NBS for offering calibration service for meters that flow gas,* see [13]. These are performed using air or other gases. Capabilities are shown in the chart shown in Fig. 3. Dashed lines in 3(a) indicate planned extensions of rate. Estimated systematic errors for the calibrations as plotted in 3(a) will be discussed in more detail in other parts of this paper. Typical reference flowmeters calibrated may add from 0.05 to 0.3% uncertainty to the systematic error shown as a result of imprecise meter readout and possible meter and flow instabilities under field conditions. Gas Flow: Static Procedure. A “P, V, T, t” (pressure, volume, temperature and time) tank is used as the NBS’ primary standard for high rate gas flow, see Fig. 4. Filtered and dried air (dew point of about −50 °C at 8.5 atmospheres absolute pressure) at rates up to 85 m3/min (3000 scfm) are the original design conditions for this system. The tank contains approximately 28 m3 (1000 ft3) and is used for measurement of air flow rate via the temporal increase in gas density in the constant volume tank. The measurements are called “static” as both volume and density changes can be measured at presumably stationary conditions. Dynamics are involved only in opening and closing motions of the diversion valves which are timed to give the collection interval. Collection volume capacity is derived from weighed gaseous nitrogen fillings of the tank in conjunction with density as determined from temperature and pressure measurements. Results from separate fillings of the tank with gas indicate that a volume uncertainty of 0.01% can be achieved. Gas temperature in the tank is measured with 10 thermistors that are calibrated against a platinum resistance thermometer which was calibrated by NBS’ Temperature Section. Two calibrated thermistors are located on each of 5 horizontal concentric circles distributed vertically so that the average temperature of the gas is accurately measured. A fan is installed to stir the air in the tank to remove stratification effects. Thermistor elevations are shown in Fig. 4. These temperature sensors, in conjunction with others for measurement of tank metal temperature, and a pressure gage provide information to derive collected gas density and an estimate of its uncertainty. Although the environment of this large tank may not be well controlled, its small surface to volume ratio and use of the circulation fan permit the volume to be derived with an acceptable uncertainty. Using this facility, NBS sonic nozzles are proved and then used as transfer standards to calibrate meters flowing air. Conservation of mass principles, as applied to this measurement system for a finite measurement interval Δt, can be written, (3)

*“Special Test” facilities exist at NBS in Boulder, CO. These use cryogenic fluids. A description of these facilities can be found in [14]. In addition, low flow rate helium permeation leaks testing capabilities will be available at NBS in 1986.

154 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

FIGURE 3(a). NBS GAS FLOW CALIBRATION FACILITY CAPABILITIES.

where, in compatible units: pc represents density of the fluid in the collection tank volume Vc, with errors represented by a leak term and by an undetected tank-volume change ΔVc. Examples for these errors are, respectively, condensation of vapor in volumes a or b which could appear as a leak, and change of Vc as caused by use of the tank at other than calibration conditions. Vb can be much smaller than Vc to make any error in measurement of Δpb of insignificant consequence. Error in measurement of Vc is related to uncertainties in measured temperature, pressure and mass, perhaps combining to be between 0.02 and 0.05% as indicated by experiments. When used as a calibration device, the same uncertainties are introduced again, except that uncertainty of mass is replaced by that of published values of super-compressibility factors Z=(P/ pRT) used in the gas law for derivation of pc. Uncertainty in Z enters if either the gas state or the gas itself is different for calibration of the tank and calibration of a meter. If uncertainty of Z and that of weighed

155

FIGURE 3(b). TABULATED SUMMARY OF NBS GAS FLOW CALIBRATION FACILITY CHARACTERISTICS.

mass are comparable (or both insignificant) total uncertainty can be in the range 0.04 to 0.10% without consideration of an uncertainty in Δt. Gas Flow: Dynamic Procedure. Filtered dried air (dew point about −40 °C at 35 atmospheres absolute pressure) is measured at rates up to 1.4 m3/min (50 cfm) with piston and with bell-type provers arrangements such as those sketched in Figs. 5 and 6. Motion of the mercury-sealed piston in its vertical glass tube is timed (after an initial acceleration period) with a timer actuated by two photo sensors when their light beams are interrupted by the piston, see Fig. 5. The light beams that traverse the glass tube are placed a known, vertical distance apart, and each is constricted to about 0.01 cm. in the direction of piston travel. Vertical motion of the bell in its annular bath of sealing liquid is timed by similar switches actuated successively by the vertical motion of the bell, see Fig. 6. Interval diameter measurements are used (via calibrated calipers or via strapping techniques with an NBS calibrated tape) to derive volume per unit of distance traveled for both provers, additional measurements being required to account for displacement motion of the sealing liquid in the bell-type prover, see [12]. Careful attention to numerous details is necessary to avoid measurement difficulties. These arise from small rates of flow and small collected volumes, from considerations involving dynamics of the measurement process, and from difficulty in making meaningful gas temperature measurements in small gas

156 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

FIGURE 4. NBS’ LARGE AIR FLOW CALIBRATION FACILITY’S P, V, T TANK.

flow systems. A nearly constant laboratory temperature, both spacewise and timewise, is used in conjunction with sufficient piping upstream from the meter and calibrator to bring gas temperature to that of the calibration system, meter and laboratory. This not only reduces temperature measurement problems but prevents heat transfer in the meter and prover—a very important requirement. Thermal insulation and/or heat exchangers are also used and are recommended for difficult environments to assure equal meter and gas temperatures. Other difficulties mentioned which cause errors can be illustrated with use of Eq. (2). This can be written for the bell prover arrangement, for a finite calibration interval Δt, as (4) where, in compatible units: piΔVc represents the mass of gas collected in the prover during its stroke of volume ΔVc as based on density pi at initiation of collection, with system gas density pe existing at termination of the calibration period. Volumes Va and Vb represent not only meter and connecting pipe volumes, respectively, but also gas volume in such things as instrument lines, and any volume collected in

157

FIGURE 5. SKETCH OF PISTON PROVER ARRANGEMENT.

the prover during its acceleration period. The last term on the right represents an error caused by a system volume change, e.g., when a pressure change affects liquid level in a manometer or sealing liquid level in a bell prover. Undetected density change (pe−pi) is an error that can receive significant multiplication whenever system volume (Va+Vb) is significant compared to ΔVc (as frequently prevails in low flowrate measurements). The last error mentioned in (4) should not be admissible; but leaks as represented by , even though demonstrated as insignificant compared to piΔVc, can be disastrous if located at a pressure tap. A leak at this location can invalidate measured pressures used to describe meter performance, and preventive tests against such leaks always are made. Eq. (4) also can be used as a basis for discussion of both systematic and random errors involved in flowrate measurements using these provers. Systematic errors may be considered as characteristics of instruments ordinarily used for measurement piΔVc/Δt. Calibration and reading errors of the instruments, not their dynamic response errors, are summarized below:* Volume Pressure

0.02 to 0.05% 0.02

158 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

FIGURE 6. SKETCH OF BELL PROVER ARRANGEMENT.

Temperature Timer, clock Timer, switching

0.03 0.005 0.06

These combine to form a possible overall systematic error in the range 0.075 to 0.165%. More reliable and credible bounds to the systematic error can be established via MAP programs. Some of the observed imprecision of rates as evaluated by repeated measurements with these provers can be regarded as connected with the other three terms on the right side of equation (4) and with sensitivity and repeatability of the flowmeter under test. Two terms in (4), (pe−pi) and Δ (Va+Vb) are connected with dynamics of prover motion as affected by its design, by the procedure used, by dynamic response of the instruments, etc. Environment also affects precision through changes of ambient pressure and temperature, the latter causing heat transfer. Combining the above uncertainties, estimated standard deviation for a flowrate determination is generally between ± 0.1 to ± 0.2% when testing better meters.

159

FIGURE 7. SKETCH OF TANDEM METER CONFIGURATIONS.

Once these necessary quantifications are completed, it remains to assess the bias, or systematic errors in laboratory measurement capabilities. This can be done via estimating but it is better achieved via traceability established via flow MAP’S—i.e., round robin testing. This topic is described and discussed below. Flow Measurement Traceability. To establish traceability of the type 2 variety (see Appendix 1) a test program must be devised so that: (1) high confidence can be placed in the artifact package—the meters assembled and the specifics of the procedures, check-points, responses to anticipated anomalies, etc., (2) the data base produced is adequate to the task of clearly evaluating the significant components of the systems that participate, and (3) the algorithm for processing the data and producing the results is an unbiased and clear procedure that is adequate to this task. Artifact confidence is established via calibration testing over an extended period of time for the kind of conditions that will be used in the round robin. This testing should occur in the initiating laboratory and it should establish a credible background data base for the units being tested. Specifically, high competence can be attained by calibrating two (2) meters in series according to tightly specified conditions. This type of configuration is shown in Fig. 7. Pre-testing of these configurations gives expected values for the respective meter factors as well as for the relative performance of the meters—i.e., the ratio of their outputs.

*These values should be considered as reasonable estimates. Currently, efforts are underway at NBS to re-evaluate the performance of all the calibration facilities that are used to offer calibration services, see [10, 13].

160 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

Adequacy of the data base is established by specifying the number of repeat calibrations done for each flowrate and meter configuration. These results should produce sufficient data so that statistical significance can be generated to exhibit the quality of measurement performance—(1) how this varies for successive calibrations done for the same conditions over short periods of time—i.e., repeatability (see Appendix 1), and (2) how this varies from day to day for conditions that may vary slightly—i.e., reproducibility (see Appendix 1). It is recommended here that the data base be generated efficiently and for the expressed purpose of testing laboratory performance. To do this, a minimum number of flowrates are used and sufficient tests at each are done. An alternative approach might be to use numerous flowrates and minimal replications at each. However, this alternative approach tends to place emphasis on meter characteristics— as opposed to test laboratory characteristics. The algorithm for data processing should be well established. This attribute is achieved when it is (has been) used for a number of MAPs for other measurement systems—i.e., the procedures produced by W.J. Youden and co-workers, see [5]. By testing in both configurations shown in Fig. 7 the upstream data (and the downstream data), individually,; have the statistical independence requirement that is needed to apply “the Youden procedure, etc. The “SFC” unit shown in Fig. 7 is a “super flow conditioner” placed between the tandem meters, see [4]. It is intended to isolate the downstream meter from flow profile (or other anomalies) that might exist in the laboratory pipeline that connects to the upstream meter.* Thus, the tandem meter configuration affords one the opportunity of generating data both without and with pipeflow profile effects because downstream meter and upstream meter performances can be treated separately. Comparisons can give unique global insights into laboratory pipeflow phenomena without having to measure these distributions. The types of flowmeters for this type of laboratory testing should be selected according to the experiences of the participating laboratories. This consensus selection should produce the type of meter, the size, manufacture, associated instrumentation, etc. This selection process should be extended to include the fluid conditions, the flowrates, etc. as well as the tolerances to be used in arranging these. The data generated via the round robin testing program is analyzed for each of the flowrates selected and for each of the meter positions. For each of these conditions, plots are produced of the respective meter performance characteristics—i.e., meter factor, discharge coefficient, etc., see [5, 6], Individual results, or averages thereof can be plotted. Each point represents the combined results for both meters for each laboratory. The data processing procedures consist of determining median values for the respective sets of data for the meters. By drawing horizontal and vertical lines through these median points, the plot is divided into four Cartesian quadrants. The origin of this Cartesian system is, according to the available data, the best estimate of the true values of the meter factors for the two meters tested according to the specified conditions, see Fig. 8. In the northeast Cartesian quadrant, the data can be considered systematically inaccurate in that points are each higher than those of the origin. Similarly in the southwest quadrant, points are lower. Thus, the degree to which data is distributed in these quadrants is a measure of the systematic off-sets prevailing in the laboratory data. In the northwest and southeast quadrants the data can be considered inconsistent or random in that one value is low while the other is high. Therefore, the degree to which the data is distributed in a northwest to southeast manner about the median intersection is a measure of the random variation in the data.

*This placement is the suggestion of Dr.E.A.Spencer, O.B.E.

161

FIGURE 8. SKETCH OF YOUDEN PLOT.

The preferred result indicating good control would be to find that the measure of systematic distribution (northeast to southwest) is equal to the random distribution (northwest to southeast) and that these measures are acceptably small, see Appendix 2. The respective levels of uncertainty can be quantified. Where, as is usually the case, the two meters are identical, a procedure for quantifying the respective random and systematic levels of the data can be used as follows, see [5]. A line of slope +1 is drawn through the intersection of medians on Fig. 8. The data is then projected perpendicular to and parallel along this diagonal line. The respective projections are then used to produce standard deviations:

The ratio of these quantities produces the degree of ellipticity of the data. When this ratio is larger than unity, the interpretation is that systematic variations prevail among the labs; this is quantified by the magnitude of e. Analogous conclusions can be drawn for e<1. Depending upon the results obtained for ellipticity, a number of reactions can occur. If e is large and this is produced by one or more laboratories, then the reaction should be to examine the components of their flow measurement processes to find systematic causes, etc. If e is small and this is produced by one or more laboratories, the reaction should be to examine the components of their processes with respect to their precision. If e is near unity but the levels of uncertainty are considered too large, then the appropriate response would be for the labs responsible to search and repair the pertinent components. When such search and repair efforts are completed, the round of tests should be repeated for the same conditions so that improvements can be quantified . Even when such search and repair efforts are not needed, repeat testing is needed to produce the continuous data record required by type 2 traceability.

162 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

CONCLUSION The standards philosophies for flowrate measurements have been presented, briefly. The specified gas flow calibration facilities currently in operation at NBS at Gaithersburg, MD have been described. Levels of performance have been given. Techniques for establishing and maintaining flowrate measurement traceability have been presented. A specific scheme has been described in some detail so that realistic data, produced on a continuing basis, can be generated so that a laboratory’s entire flowrate measurement process can be assessed. It is concluded that once these types of traceability chains are produced so as to link flow measurement laboratories within and across national borders and boundaries, satisfactory fluid measurements can be achieved at specified levels. In this manner, the increasingly critical and costly measurements of gas resources and products can occur satisfactorily in custody transfers and in industrial processes. APPENDIX 1 —GLOSSARY OF TERMS ACCURACY—The closeness of the agreement—usually expressed as a percentage—between the result of a measurement and the true value of the quantity being measured. Accuracy of flow measurement is defined as the total uncertainty which consists of the sum of the systematic error, or bias, and the precision. PRECISION—The closeness of the agreement—usually expressed as a percentage—between the results of two or more measurements of the same (or similar) quantity being measured. Precision of flow measurement is defined as the random uncertainty at the 95% confidence level in ASME-MFC-2M. a) REPEATABILITY—The closeness of the agreement—usually expressed as a percentage between the results of two or more successive measurements of the same quantity subject to all of the following conditions: – – – – – –

the same method of measurement, the same observer, the same measuring instrument, the same location, the same conditions of use, repetition over a short (specified) period of time.

Example: Repeatability expresses “how close duplicate measurements can be made to be.” b) REPRODUCIBILITY—The closeness of the agreement usually expressed as a percentage between the results of two or more measurements of the same (or similar) quantity(s) under changing conditions such as: – – – – –

method of measurement observer measuring instrument location conditions of use

163

– time A valid statement of REPRODUCIBILITY requires specification of the conditions changed. Example: Reproducibility expresses “how close duplicate measurements normally agree.” LINEARITY—In flow metering, LINEARITY refers to the constancy of the meter factor over a specified flowrate (or Reynolds number) range. This constancy is quantified by an uncertainty band specified by the upper and lower limits of the meter factor over the specified range; the constant meter factor is determined by averaging the upper and lower limits of the meter factor. RANGEABILITY—Flowmeter rangeability is the ratio of the maximum to minimum flowrates specified by the meter’s linearity (see above). Rangeability is frequently termed “turn-down”. 4— 3— Type 2— Type 1— Type TRACEABILITY—Traceability to Type Traceability Traceability designated standards (national, Traceability is Measurement a the implies have means international, or well characterized the ability to s to capability to traceability to ability reference standards based upon demonstrate express relate designated fundamental concepts of nature) is conclusively quantitatively a standards if individual an attribute of some that

164 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

measurements. However, there are several interpretations or types of traceability, see [2]:

particular instrument or artifact standard has either been calibrated by a national reference laboratory (i.e., NBS), or has been calibrated against another standard in a chain or echelon of calibrations, ultimately leading to a calibration performed by the national reference laboratory (i.e., NBS).

and only if scientifically rigorous evidence is produced on a continuing basis to show that the measurement process is producing measurement results (i.e., data) for which the total measurement uncertainty relative to the designated standards is quantified.

measurement results to designated standards through an unbroken chain of comparisons.

the results of a measurement in terms of units which are realized on the basis of accepted reference standards, usually national standards.

APPENDIX 2 —TANDEM METER STATISTICS Calibrating meters in tandem enables a number of checks to be made on the flowrate measurements, see [5, 6]. Besides affording the opportunity of having a redundant result for the same flowrate, the tandem meters can produce not only “on-line” indications of relative meter performance, but also “off-line” assessments of uncertainties associated with the meters themselves and the system in which they are being calibrated. On-line Indications.Instantaneous meter responses from a tandem configuration of meters can be compared via ratios or differences in order to assure that each meter is operating satisfactorily. By satisfactorily is meant that current meter performance matches that obtained when the meter was calibrated previously. When previous calibration data for tandem meter configurations—in a single laboratory or in a number of laboratories—is processed properly, expected values can be produced for the respective meter factor ratio, etc. This expected value and the associated uncertainty can be used to evaluate the relative performance of these meters when they are used in round robin calibrations. Interpretations. When the ratio of meter response is found to differ from the expected value by more than the expected tolerance, the interpretations can vary. For example, one is not sure which meter is the cause for the difference; one is not sure of the reason for the deviant behavior. However, one can be sure that something is sufficiently wrong that the calibration should not be continued until this situation is remedied. When the relative meter performance produces the expected value, it is assumed that the prevailing meter conditions satisfactorily duplicate those that occurred during previous calibrations. Intrinsic in this

165

assumption is that whatever meter anomalies that may occur, they do not occur simultaneously so as to affect each meter in the same manner. Long term variations can be observed both in meter factor and in relative meter performance data plotted in time-series format. Statistics. For successive calibrations done for constant flowrate conditions, the meter factor results for each meter can be processed to give averages and the total standard deviations, i.e.,

For each specific flowrate, the data from each tandem meter configuration, can be processed to give the correlation coefficient, i.e.,

where subscripts 1 and 2 refer to the upstream and downstream meters, respectively. The total uncertainty for each of the meters can be categorized according to that which is correlated with the other meter and that which is uncorrelated. These components can be designated σm and σs respectively referring to the meter and the system used for the calibration. These are quantified for each meter, via and so that Analysis. The time series data traces for each meter should, via the above decomposition for σm, indicate valid expectation data to corroborate meter well-being during round robin tests as well as longterm longevity. These results are critical to show what type of and perhaps when refurbishments are needed to keep the artifacts in proper condition. The values for σs determined in each lab provide another interesting parameter for intercomparing performances. These data, together with the results of the Youden analyses described above, further establish the degree of agreements among participants. ACKNOWLEDGMENT The author acknowledges his predecessors at NBS who in the 1960’s and before have designed, built, and characterized the facilities and procedures described above. Of special significance in this group are the efforts of M.R.Shafer and K.R.Benson; they have been essential in establishing and maintaining the high reputation that the fluid flow group at NBS enjoys in the fluid metering community—both national and internationally.

166 GAS FLOW MEASUREMENT: CALIBRATION FACILITIES AND FLUID METERING TRACEABILITY AT THE NATIONAL BUREAU OF STANDARDS

The author also acknowledges the fine secretarial efforts placed toward this product by Mrs.Susan L.Johnson as well as her patience through it all. REFERENCES CITED 1. 2. 3. 4. 5. 6. 7. 8.

9. 10. 11.

12. 13. 14.

Cameron, J.M., “Measurement Assurance,” U.S. National Bureau of Standards Int. Report No. 77–1240, April 1977. Belanger, B.C., “Traceability—An Evolving Concept,” ASTM Standardization News, February 1979. Mattingly, G.E., and Spencer, E.A., “Steps Toward an Ideal Flow Transfer Standard,” FLOMEKO Symposium, Groningen, The Netherlands, 1978, published by North Holland Publishing Co., Amsterdam, NL. Mattingly, G.E., “Dynamic Traceability of Flow Measurements,” Invited Lecture, IMEKO Tokyo Flow Symposium, 1979, Society of Instrument and Control Engineers, Tokyo, Japan. Youden, W.J., “Graphical Diagrams of Interlaboratory Test Results,” Journal of Industrial Quality Control,” 15 (110, May 1959, pp. 133– 137. Strohmeier, W.A., “Notes on Turbine Meter Performance,” (unpublished treatise), TN #17, March 1971, Fischer and Porter Co, Warminster, PA. Mattingly, G.E., “An Interlaboratory Round Robin Flowmeter Test Using Turbine Meters Flowing Water,” in preparation as a U.S. National Bureau of Standards Int. Report. Mattingly, G.E. and Kinghorn, F.C., “An NBS-NEL/UK Orifice Meter Crosscheck Program,” see Mattingly, et al., “Workshop Report on Fundamental Research Issues in Orifice Metering,” Gas Research Institute Report 84/ 109, 1984. Streeter, V.L., Fluid Mechanics, 5th edition, McGraw-Hill Book Co., New York, NY, 1971. Ruegg, F.W. and Shafer, M.R., “Flow Measurement: Procedures and Facilities at NBS,” Proc. Semi-Annual Meeting at ASHRAE, San Francisco, CA, 1970. Olsen, L., and Baumgarten, G.P., “Gas Flow Measurement by Collection Time and Density in a Constant Volume,” Proceedings Symposium on Flow—Its Measurement and Control in Science and Industry,” Paper No. 3–8–138, Pittsburgh, PA, 1971. Ruegg, F.W. and Johnson, O.P., “Dynamics of the Bell Prover,” Proceedings Symposium on Flow—Its Measurement and Control in Science and Industry”, Paper No. 3–8–152, Pittsburgh, PA, 1971. Kieffer, L.J., ed., “Calibration and Related Measurement Services at NBS,” NBS Special Publication 250. Mann, D.B., et al., “Gas Orifice Meter Discharge Coefficients as Determined by Mass Flow Measurement,” NBSIR 83–1685, August 1983.

CROSS REFERENCE SERVICE OF NATURAL GAS STANDARDS IN THE UNITED STATES Julia C.Shapiro, Ph.D. Laboratory Manager Gerald R.Burkett, N.B., B.S. Plant Manager William A.Crowley, M.B., B.S. Operations Manager Scott Specialty Gases, Inc. Troy, Michigan 48083

ABSTRACT

The United States Natural Gas Industry requirements for high accuracy analysis of natural gas pointed out a need for interlaboratory comparison of compositional analysis. Scott Specialty Gases, in response to this need, has developed a Cross Reference Service to correlate analytical results between laboratories. The statistically summarized results of the interlaboratory correlations have identified strong and weak points of existing analytical techniques. The accuracy of analysis of specific components varies from a relative standard deviation of 2% to as high as 20%. Although the Natural Gas Industry would like to analyze energy to ±1 BTU, the results of the Cross Reference Service indicate an accuracy of ±2.5 BTU at a 95% confidence level. The problems causing inter!aboratory disagreement are discussed with particular emphasis on differences in analytical techniques, instrumentation, and absence of uniform standards throughout the Industry. The United States Natural Gas Industry’s goal for precision in measurement of BTU is less than one BTU. The need for such extreme accuracy equal to ±0.1% can be understood from the magnitude of United States natural gas consumption (Table 1). TABLE 1 UNITED STATES 1984 NATURAL GAS MARKET

Residential Commercial Industrial Miscellaneous

Energy, BTU, Quads

(Billions) Dollar Estimate

7.4 7.0 3.2 .5

43.73 27.51 11.62 0

168

CROSS REFERENCE SERVICE OF NATURAL GAS STANDARDS IN THE UNITED STATES

Energy, BTU, Quads

(Billions) Dollar Estimate

18.1

82.86

1 Quad—1015 BTU

Estimates of United States gas production in 1984 are approximately 18 quadrillion BTU or 83 billion dollars. In this case an error of one BTU has a potential of gains and losses to transmission and pipeline companies and public utilities of 83 million dollars a year. The extreme accuracy requirements of measuring BTU encouraged development of compositional analysis of natural gas and calculation of the theoretical BTU value which is now the standard practice. The majority of the Natural Gas Industry uses gas chromatography as a fast, efficient and reliable analytical technique. There are several analytical procedures published; these include American Society for Testing and Materials, Gas Processors Association, as well as those developed by various G.C. manufacturers and users. Many G.C. manufacturers have developed instruments specially designed for natural gas analysis. As it is known, one of the most essential factors in G.C. analysis is the existence of a reliable standard. Many specialty gas companies, including Scott, provide gravimetric compositional standards. The accuracy of standards available varies. A primary standard reference from the National Bureau of Standards has not existed until recently. Thus, all three factors mentioned above, various analytical techniques, different instrumentation and absence of a primary standard reference material can cause a significant difference in analytical results between laboratories. The objective of implementation by Scott of the Natural Gas Cross Reference Service was to improve interlaboratory correlation by allowing laboratories to compare themselves to others in the Industry. The Natural Gas Cross Reference Service is performed on a quarterly basis. The process consists of: 1. 2. 3. 4. 5. 6. 7.

Manufacture of the Natural Gas mixtures. Transfer of the mixture to multiple, small cylinders. Analysis of samples to insure their homogenity or uniformity. Supply of samples to subscribers. Compilation of the reported data. Preparation of a statistical report of results submitted. Submission of the report to participants.

The BTU Cross Reference Service cylinders are manufactured from a master cylinder containing at least 8– 9 natural gas components in a Methane balance gas. This mixture is then transferred to 25 or more smaller cylinders. In the process of manufacturing, the major problems appear to be the same as has been met in the field: transferring of natural gas mixtures without condensation, separation of heavy hydrocarbons, and pressure and temperature changes which occur inside of the sample system. Each sample cylinder was analyzed, in order to assure their homogenity, using a Varian Vista 6000 Gas Chromatograph. When developing this technique we encountered the common G.C. problems: separation of Ethane from Methane in the presence of 90% Methane, separation of Isobutane from Butane, Isopentane from Pentane and condensation of higher hydrocarbons (Hexane, Heptane) on intermediate lines before injection.

169

All hydrocarbons were analyzed using an 80/100 mesh n-octane on porasil C stainless steel column (6′×1/ 8″) and a FID detector with temperature programming. The temperature is programmed from 30°C (2 minutes) to 110° at 10°C/min. Nitrogen and Carbon Dioxide are analyzed on a Carbosieve S column (10′×1/8″) at 30°C isothermally using a TCD detector. Both analyses used Helium carrier gas at 30 ml/min., a 0.25 cc loop and a heated valco valve. An external gravimetric standard was used for calibration throughout the process of analysis. A sample of the chromatogram is shown in Figure 1. This analytical procedure, along with proper manufacturing technique, gave us a guarantee of homogenity of the gas with spread less than 0.5% relative standard deviation of each component from cylinder to cylinder. This deviation includes uncertainties in transferring mixtures into cylinders and uncertainties in the analysis (Table 2). Each participant returns his analytical results to Scott for compilation. We normalize all of these results to 100%. All BTU values are recalculated using the GPA procedure. This process of compiling results is necessary to make them comparable. The data is then statistically summarized and the resulting report issued to subscribers. The report consists of the following: All participants data normalized to 100% (Table 3), the statistical analysis of reported results (Table 4). The statistical summary includes the mean, median, estimate of standard deviation, estimate of standard error, number of observations and number of outliers. The balance of the report includes histograms of the participants data. Figures 2–3 show sample distributions for some components. Besides the regular Cross Reference Service provided by Scott on a quarterly basis, we arranged in 1985 a Round Robin testing program among nine laboratories in the Natural Gas Industry. In this program a cylinder with synthetic natural gas mixture was manufactured and analyzed by Scott, then sent to participating laboratories for analysis and returned back to Scott for final analysis. The results were published. Table 5 shows the precision of analysis by Scott and by the nine laboratories. The results of all Cross Reference services (Table 6&7) which has been provided during the last two years can be summarized in the following conclusions: 1. There is a gradual improvement in the analysis of most components in the last series of BTU Cross Reference Services compared to the first. Especially significant are changes for Ethane, Propane, Carbon Dioxide and Nitrogen. 2. The error in component analysis increases with carbon content of the molecule from 0.18% Relative Standard Deviation for Methane to 52% for Hexane. Table 2 Relative Standard Deviation from Cylinder to Cylinder for Each Component in One Series of BTU Service (Analyzed by Scott). Methane Ethane Propane Butanes Pentanes Hexane Nitrogen Carbon Dioxide

0.38% 0.31% 0.37% 0.52% 0.46% 0.53% 0.24% 0.23%

170

CROSS REFERENCE SERVICE OF NATURAL GAS STANDARDS IN THE UNITED STATES

FIGURE 1 Sample of Chromatograms

171

TABLE 3 Tabulation of Reported Analytical Results

TABLE 4 Statistical Summary of Reported Analytical Results

No. of Observ. Min. Observed Value Max. Observed Value Mean Value Median Value Scott Value Est. Std. Dev. Est. Std. Error Coeffic. of Variance (%) No. of Outlying Values

Methane (Vol %)

Ethane (Vol %)

Propane (Vol %)

Butane (Vol %)

BTU (Vol %)

14

15

15

15

14

88.5243

4.8824

1.0362

.1620

1024.5400

89.0780

5.1699

1.1406

.2601

1030.4699

88.8917 88.9406

4.9838 4.9670

1.0725 1.0670

.1931 .1840

1027.9697 1028.3449

88.7845 .16105

5.0884 .07412

1.0691 .02668

.1768 .02619

1029.0400 1.64478

.04304

.01914

.00689

.00676

.43959

.18118

1.48730

2.48729

13.56396

7.65446

1

0

0

0

2

172

CROSS REFERENCE SERVICE OF NATURAL GAS STANDARDS IN THE UNITED STATES

FIGURE 2 Frequency Distribution of Reported Analyses—Hexane (Vol %)

FIGURE 3 Frequency Distribution of Reported Analyses—N2 (Vol %) TABLE 5 Precision of Analysis by Scott and Industry (Round Robin) Components

Analysis by Scott

First

Last

Methane Ethane Propane

95.5280 1.6140 0.4973

95.5220 1.6140 0.5019

% Diff. Between First & Final

Relative Standard Deviation of Industry % (n=9)

0.006 0 0.925

0.22 3.08 1.17

173

Components

Analysis by Scott

First

Last

i-Butane n-Butane i-Pentane n-Pentane Hexane Heptane Octane N2 CO2 He BTU

0.1676 0.2156 0.0547 0.0571 0.0240 0.0179 0.0045 1.0500 0.7400 0.0212 1029.64

0.1676 0.2149 0.0551 0.0572 0.0245 0.0180 0.0040 1.0500 0.7426 0.0225 1029.69

% Diff. Between First & Final

Relative Standard Deviation of Industry % (n=9)

0 0.326 0.730 0 2.100 0.560 11.100 0 0.350 6.100 0

1.72 1.75 18.09 69.52 32.74 (C6+) 4.39 16.97 14.14 0.06

3. The largest errors were observed for components requiring accurate G.C. separation technique, as isobutane and n-butane, iso-pentane and n-pentane. The combined error for the sum of Iso-butane and n-butane and sum of iso-Pentane and n-Pentane is significantly less than for the individual components. The high relative standard deviation for Hexane is explained by the fact that some customers analyze heavy hydrocarbons as Hexane+ mixtures, some analyze them separately. 4. Despite the relatively high error in minor components approaching 20–30% relative standard deviation, the process of calculating BTU value from compositional analysis results in a relative standard deviation of 0.16%. 5. The analysis of BTU value show a significant improvement in relative standard deviation from 0.47% two years ago to 0.16% most recently. Table 6 Relative Standard Deviation % for All Components in Seven Cross Reference Services BTU

BTU

BTU

BTU

BTU

BTU

BTU

Component

#1

#2

#3

#4

#5

#6

#7

Methane Ethane Propane Iso-butane n-butane Butanes Iso-Pentane n-Pentane Pentanes Hexane

0.21 8.25 6.41 5.22

0.25 3.68 9.11

0.36 2.45 2.68 7.70

0.17 1.65 2.19 30.91 8.57 7.8

0.32 2.26 6.40 3.11 20.23 4.8

0.16 2.22 1.84 10.86 23.56 7.6

28.20

9.14

26.35

10.87

50.29

20.37

0.18 1.49 2.49 13.56 7.65 4.60 18.09 69.52 14.50 52.10

20.86

19.03

174

CROSS REFERENCE SERVICE OF NATURAL GAS STANDARDS IN THE UNITED STATES

Component Heptane Hydrogen Carbon Dioxide Nitrogen BTU

BTU

BTU

BTU

BTU

BTU

BTU

BTU

#1

#2

#3

#4

#5

#6

#7

3.33

4.91

3.19 0.34

1.95 0.16

25.20 7.45

8.87

75.88 1.92

3.26

123.3 9.13 1.99

2.45 0.47

2.32 0.51

1.13 0.34

2.35 0.24

3.01 0.19

24.64

6. All observed Industry’s errors were much higher than Scott’s estimated accuracy. In an attempt to rationalize the high errors of compositional analysis with the relatively low errors in BTU analysis the following model of contributing BTU errors from component analysis was developed (Table 7). The Table shows that at 95% Confidence level (assumed two standard deviations), Scott’s precision for mixtures of 1028 BTU is estimated at 0.40 BTU compared to the Industry’s 2.5 BTU. 7. Although Nitrogen has a relative standard deviation of 2% which is significantly less than many of the heavy hydrocarbons, it was the largest contributor in BTU error because it has no heating value. Table 7 Comparison of Standard Deviation of BTU for Scott’s and Industry Analysis of Natural Gas Industry Precision Mean %

C1

88. 93 4.98 ± 1.07 ± 0.40 ± 0.03 ± 4.03 ± 0.56 ±

Std. Dev.

Scott Precision Contr ib. in BTU

1009 .7 C2 0. 37. 074 80 ± C3 0. 16. 027 18 ± C4 0. 8.99 021 ± C6 + 0. 1.34 012 ± N2 0. −40. 079 65 ± CO2 0. −5. 027 61 ± BTU 1027 .81 ± Estimated Accuracy of BTU Analysis (95% Confidence at ±2% Standard Deviation)

Contri b. in BTU of Std. Dev.

0.56 0.40 0.47 0.49 0.79 0.27 1. 28* 2.5 BTU

Mean %

88. 77± 5.09 ± 1.07 ± 0.39 ± 0.06 ± 4.05 ± 0.57 ±

Std. Dev.

0. 0193 0. 0033 0. 0020 0. 0004 0. 0122 0. 0015

Contr ib. in BTU

1009 .7 38. 64 ± 16. 13 ± 8.77 ± 2.43 ± −40. 89 ± −5. 75 ± 1028 .33 ±

Contri b. in BTU of Std. Dev.

0. 146 0. 050 0. 045 0. 015 0. 123 0. 015 0. 20* 0.4 BTU

175

Industry Precision Mean %

Std. Dev.

Scott Precision Contr ib. in BTU

Contri b. in BTU of Std. Dev.

Mean %

Std. Dev.

Contr ib. in BTU

Contri b. in BTU of Std. Dev.

* These numbers are calculated:

8. The fact the Industry precision is much lower than Scott’s can be explained by three factors contributed to it: various analytical instrumentation, different analytical techniques and absence of a “Primary Standard”. Table 8 shows the technique and instrumentation employed by the participants. Some participants used regular G.C.’s, some dedicated natural gas chromatographs. There were analysis done at isothermal, as well as T-programming conditions. Most participants analyzed the mixtures on two columns, but some used as many as four columns. Table 8 Analytical Technique Used by Participants Number of Participants

Instruments

Amount of Columns

Detectors

3 8 1 3 1

G.C. G.C. G.C. G.C.Carle Gas Monitor

2 2 4

TCD Isotherm FID & TCD T—Progr. FID & TCD T—Progr. In Series Dedicated BTU Analyzer

Conditions

CONCLUSIONS Two years of providing the Cross Reference Service has proved the fol lowing: A) Clear evidence of improving correlation between laboratories. B) Industry precision is approximately ±2.5 BTU. C) Industry should be able to achieve ±1 BTU or better accuracy. D) The Cross Reference Services is justified as a method of obtaining improved interlaboratory correlation.

NEW CONCEPTS IN GAS CALIBRATION Edward J.Dahn Director—Sales/Marketing Flow Technology, Inc. Phoenix, AZ 85040

ABSTRACT

The paper describes a new and unique design of a primary Gas Flow Calibrator. It is an automatically controlled precision volumetric flowmeter calibrating device specifically designed for use with a variety of gases over a wide range of pressures. The Aerotrak is a closed loop positive displacement system able to minimize both the quantity of gas used and the energy required to create the desired flow rate. The entire operation of the Aerotrak is under the automatic control of a uniquely programmed Personal Computer (PC). When initiated by the operator, the pre-selected program ensures the calibration data is taken after the pressure and temperature surge created by the initial piston motion has completely decayed and the flow rate is constant. This is achieved by the computer comparing successive calibration points and accepting data only when it has repeated within specified limits. As the actual data is being generated, it is continuously displayed for the test operator and stored in memory. Upon completion of the test, all data can be viewed on the computer screen by the operator and immediately evaluated. The completed data is then stored on a diskette which can be used to generate hard copy printouts and the plots of curves. DESCRIPTION The Aerotrak Gas Calibrator design, not a modification of some existing calibration technique, is a new and truly unique approach to Gas Calibration. It is a closed loop positive displacement calibrator, combined with this new design concept is a powerful marriage with a Personal Computer and menu driven software program. The program automatically controls the total operation of the calibrator. In addition to rapid data reduction the PC Screen displays Final Test Report and Curves. This new Closed loop principle affords the user an opportunity to calibrate both gas and liquid with the same basic system. However, this paper will address only the Gas Application.

177

OPERATION The system is charged with the selected gas ranging in pressure from 5 psia through 1440 psig @ 100 degrees F. The gas flow is created by a solid sealed piston which is moved through the length of a bored and honed flow tube with 12 RMS finish. The preselected velocity of the piston is established by a servo controlled motor. A rotary encoder mounted on the motor shaft provides the feedback necessary for constant rate control of the advancing piston. (Figure 1) The flow transducer under test is installed in an external flow path with the upstream and downstream legs geometrically symmetrical. As the piston is advanced it creates the gas flow through the flow transducer and then to the upstream side of the piston in the flow tube. At the time of initial motion of the piston a pressure/temperature surge is experienced by the flow transducer. As the pressure/temperature surge decays a steady state flow condition exists at the flowmeter. The Computer samples the variables required to make the flow rate calibrations such as transducer frequency, encoder frequency, pressure, temperature, delta pressure, and delta temperature. Successive calibration points are calculated by the PC and when repeatability at the desired flowrate is achieved in a selected number of points the data is stored and the velocity of the piston is automatically changed to the next desired calibration point. This pattern is repeated until the required number of preselected consecutive points are gathered and are repeatable within the desired limit. When the piston reaches the extreme downstream portion of the flow tube, a motor driven ball valve in the return line is opened and the gas is returned to the upstream side of the sealed piston while the piston is being moved to its original starting position. It is then ready for another calibration run. SYSTEM CHARACTERISTICS The advantages of the closed loop gas calibration are as follows : A. Minimizes the amount of gas required for test. The system is charged to the required pressure and remains at that pressure level throughout the test. B. Minimizes the amount of energy required to create desired flow rate at working pressure. C. Capable of calibrating flowmeters from partial vacuum to 1440 psig. D. Capable of utilizing a variety of calibration gases. E. Capable of calibrating at required operating densities with a selected gas. F. Easily checked for seal leaks. G. Time required for calibration is reduced to a few minutes versus hours for the present conventional methods. H. Rapid change from one density to another. I. Does not require correction for changes in atmospheric pressure. J. Provides a high degree of consistency in Test Program through standardization of technique. K. Improved ease of operation by less skilled people. L. The symmetrical design of the flow paths allows flowmeters to be placed in a position which minimizes the amount of pressure/temperature correction required. M. The displacement is determined by conventional water draw technique using test measures traceable to National Bureau of Standards.

178

NEW CONCEPTS IN GAS CALIBRATION

FIGURE 1

SOFTWARE PACKAGE Upon completion of the automated calibration runs the raw and calculated data is displayed on the Computer Screen. The operator is afforded the opportunity to immediately view and approve the test results. The completed calibration data is stored on a diskette and then used to generate hard copy print-outs and plots of curves. (Figure 2) The Aerotrak Gas Calibrator is capable of determining the changes in flowmeter calibration characteristics, where the calibration charge pressure is changed from one level to another. The test data is stored, resulting from a test at one pressure level. The calibrator is pressurized at another level and an identical test is conducted. The PC has the capability of plotting both curves on the PC Screen, or generating a hard copy via a plotter. (Figures 3, 4 and 5) The operator has the option to view a variety of plots on the

179

FIGURE 2

PC Screen. The variables includes frequency, K-factor, flow rate, density, linearity limits, and time. Additionally the operator may select his units of volume and time. TYPES OF TRANSDUCER The Aerotrak has an inherent capability to calibrate a variety of flow transducers, such as turbine, variable area, nozzles, vortex shedders, orifice plates, laminar, thermal mass flow controllers. The type of inputs include both frequency and non-pulsing analog inputs such as 0 to 5 VDC, 4 to 20 MA or 10 to 50 MA. Manual inputs are required for the variable area flowmeters. FUTURE DEVELOPMENTS Development work is being carried on to drastically increase the maximum flow rate capability. This increase in flow rate is required to support the needs generated for the Natural Gas Industry.

180

NEW CONCEPTS IN GAS CALIBRATION

FIGURE 3

FIGURE 4

181

FIGURE 5

FIGURE 6

MICROSTRUCTURE SENSORS FOR FLOW, DIFFERENTIAL PRESSURE AND ENERGY MEASUREMENT R.Higashi, R.G.Johnson, A.K.Mathur, A.N.Pearman, and U.Bonne Honeywell Inc., Physical Sciences and Technology Centers Minneapolis, Minnesota

ABSTRACT

The point measurement of velocity of gaseous fluids has been made for some time with hot wire anemometer type sensors. We have developed a microflow sensor on silicon which features a wide dynamic range (1–4000 feet/min), fast response (5 msec), low operating power (~10mW), and batch fabrication for reproducibility and cost effectiveness. Its small size and low power dissipation make it well suited for a velocity sensor to indicate mass flow or differential pressure across a small orifice. Theoretical aspects of this approach and laboratory experiments to characterize the performance of this interesting microsensor are discussed. Safety test results are reported. We are also exploring the sensitivity of certain micro-sensors to variations in gas properties such as composition, gas density, specific heat, and viscosity. This sensitivity to gaseous fluid parameters might form a basis for the measurement of energy flow under advantageous conditions of automatic compensation for pressure and temperature changes. The Gas Research Institute sponsored part of this work involving correlations between various natural gas properties. INTRODUCTION The micromachining of small structures in silicon became an attractive approach to the development of new types of microtransducers following publication of papers by Bassous(1), Bean(2) and Peterson(3) in 1978. These papers outlined the principles and methods for precision anisotroplc etching of single crystal silicon and presented many examples of structures fabricated by these techniques. Through innovations on this early work, we have developed techniques for the fabrication of precisely dimensioned dielectric film structures containing temperature sensitive resistors that can be thermally isolated from the chip. This development has enabled locally high temperatures to be attained with very low input power and little heating of the chip. It has also led to novel concepts of integrated thermal sensors and transducers that can be

183

built with the low-cost, thin-film batch processing characteristic of integrated circuit fabrication. In addition, there is the option of incorporating active silicon circuit elements on the chip with the microtransducer, thus achieving a larger scale device integration with lower cost, lower device size, and easier applicability. These expectations are now close to being realized in practical devices at Honeywell with the development of an air flow and differential pressure sensor. This paper explores how these techniques may be used for natural gas flow, differential pressure and energy flow measurement. DISCUSSION Microtransducer Design for Flow Sensing of Gaseous Fluids The choice of materials for the transducer was strongly influenced by the objective of making the finished device easy to manufacture. For example, a high temperature ceramic was chosen as the dielectric because it can be deposited to tightly controlled specifications and is an excellent insulator and passlvatlng film. A metal alloy was chosen for the resistor metallization because deposition technology was well established. In addition, the alloy’s higher resistivity than pure metals allows the delineation of resistors in the 500 to 1000 ohm range on a 150×350 micron area using 5 micron line widths while maintaining a substantial temperature coefficient of resistance of about .003 per degree centigrade. Standard (100) silicon was chosen for the substrate because of its desirable anisotroplc etching properties and its suitability for later circuit integration with the transducer. The basic flow transducer consists of a pair of ceramic film bridge microstructures, as shown in Figures 1 and 2 containing a central heater resistor divided equally between the two bridges, and two identical detector resistances placed adjacent to and symmetrically with respect to the heater resistance. In operation, air flows across the chip perpendicular to the long axes of the bridge structures, cools the upstream detector and heats the downstream detector. The resulting temperature and corresponding resistance differential yields a circuit output measurement of the flow rate. The use of identical detector resistors eliminates zero point offsets. The bridge structures are less than one micron thick, and their small heat capacity provides a short thermal time constant, typically .005 second or less, and a high sensitivity to rapid flow changes. The air space beneath the bridges thermally decouples the heater and detectors from the silicon to a large extent. Consequently, large heater temperature differences in the 100 to 200°C range relative to the silicon can be sustained by small power inputs. The thermal efficiency is typically 15°C per milliwatt of input power under no-flow conditions. The thin ceramic film permits the detector resistors to be placed closely adjacent to the heater so that they can operate at about 60 percent of the heater temperature elevation, and can develop large temperature differentials under small gaseous flow conditions. The etched cavity below the bridge pair is precisely limited at the sides by the etch-resistant (111) planes of the silicon, and at the bottom of the cavity and the ends of the bridges by accurate timing of the etch duration. The symmetry and effectiveness of the etched undercut of the bridges is maximized by orienting the axes of the bridges at 45 degrees to the <110> directions in the silicon. Drive Circuit Operation Figure 3 shows heater and sensor circuits used to operate the flow sensor. The sensor circuit is a conventional Wheatstone bridge circuit. However, the heater circuit, as shown in Figure 3, is uniquely adapted to the flow sensor to provide an output proportional to mass flow, and to minimize errors due to ambient temperature changes. The circuit is designed to keep the heater temperature at a constant

184

MICROSTRUCTURE SENSORS FOR FLOW, DIFFERENTIAL PRESSURE AND ENERGY MEASUREMENT

FIGURE 1. PLAN VIEW OF AIR FLOW SENSOR MICROSTRUCTURE

differential above the ambient air temperature under conditions of ambient-temperature variation and air flow variation. The ambient temperature is sensed by a similar heat sunk resistor on the chip, and the chip temperature (which remains within about one degree of the air) is a satisfactory approximation to the ambient. This mode of heater operation also provides means to sense changes in air or gas composition (with or without flow) which alter the thermal conductance and thus change the operating temperature of the heater and detector resistances. For example, extreme changes of thermal conductance, such as that encountered between air and helium, cause large differences in power needed to hold the heater at its constant temperature differential. This is one input for the determination of natural gas heating value, which in its first approximation is proportional to its density or average molecular mass. The error integrator is the active component in the heater circuit. It integrates the voltage differences seen on the Wheatstone bridge, and changes the voltage to the heater to maintain the bridge circuit balance. The error in heater temperature can be kept less than 1.0 percent of its differential above ambient over an ambient temperature range of −40°C to +80°C. Chip Packaging and Application Considerations The chip is fabricated with passivation techniques to permit its exposure to gaseous environments during a long operating life. To prevent corrosion, gold pads and gold wire bonds are used, and all other surfaces are passlvated with silicon nitride. The chip housing protects the microstructure from mechanical damage and

185

FIGURE 2. CROSS SECTION, A–A, OF AIR FLOW SENSOR MICROSTRUCTURE

provides the smooth flow channel that is desired for precise flow measurements. Mass flow measurements require calibrations that depend on channel size and the onset of turbulence, or velocity-dependent flow profiles across the channel, which limits the accuracy of mass flow measurement. Therefore, it is important to keep Reynolds numbers below the turbulent values, and it is desirable to have fully-developed laminar flow at the chip for all channel sizes. The small channel sizes, which can be comparable to or less than chip dimensions, and the resulting high flow impedance, facilitate the sensing of dynamic differential pressures between two rooms, across a duct flow orifice, or any application in which a small flow through the sensor is permissible. The detector microtransducer provides an output response proportional to the mass flow rate. This response can be calibrated to yield a differential pressure measurement in which the range and shape of the response curve can be determined by the dimensions that determine the flow impedance of the housing. In addition, the small channel dimensions that can be obtained provide a high impedance relative to the impedance of common pneumatic connecting lines. Consequently, it is possible to remotely locate the differential pressure sensor by using convenient connecting lines, thus reducing installation costs and, in some applications, reducing environmental stresses on the sensor. The effects of dust coming through the flow line are minimized by the chip surface being parallel to the flow direction. In contemplated differential pressure sensing applications, a long life in typical industrial atmospheric environments can be assured by removing most of the dust by using a suitable filter. This approach is practical because of the low maximum throughput rate, typically 20 to 50cm−3 per minute, made possible by the small channel dimensions and the high sensitivity of the detector. Accelerated life tests with heavy dust concentrations have demonstrated an equivalent life in normal industrial air in excess of 20 years with no dust accumulation on the chip and no deterioration of response.

186

MICROSTRUCTURE SENSORS FOR FLOW, DIFFERENTIAL PRESSURE AND ENERGY MEASUREMENT

FIGURE 3. HEATER AND SENSOR CIRCUITS FOR THE AIR FLOW SENSOR FOR MASS FLOW

Performance Characteristics An outstanding property of the sensor is the large temperature differential that develops between the two detector resistances with only small air flows, thus making it practical to use a simple alloy film to sense

187

FIGURE 4. TEMPERATURE DIFFERENTIALS UNDER FLOW CONDITIONS IN A TYPICAL SMALL CHANNEL AS A FUNCTION OF PRESSURE DIFFERENCE BETWEEN CHANNEL PORTS

temperature even if its temperature coefficient is much smaller than that of diodes or thermistors. Figure 4 shows the temperature changes of the detector resistors in a typical differential pressure measurement, and the resulting temperature differential between the upstream and downstream detector resistances. Figure 5 shows pressure differential characteristics at four ambient temperatures and illustrates the temperature compensation that can be achieved over a broad operating temperature range. The sensor measures mass flow quite accurately under density changes caused by variations in ambient temperature and ambient pressure. Variations of gas composition involve changes in many properties of the gas such as specific heat, molecular weight and size, and others. Therefore, because the gas flow sensor responds to changes in the gas other than just molecular mass, changes in composition may not always lead to sensor outputs indicative of changes in molecular mass, (Fig. 6). Figure 7 shows a comparison between helium (He) and air as an example in which the sensor output appears consistent with the mass flow rather than volume flow, and is independent of gas composition. For minor gas composition differences, the mass flow errors are generally quite small. Figure 8 shows the excellent reproducibility in signal vs. differential pressure across the flow channel obtained for four sensor units from one process batch. Because of the close relationship between heating value of natural gas and its density, we believe that energy flow sensing is directly possible with this device. A correlation between natural gas heating value and density, shown in Figure 9, is less satisfactory than generally assumed, since errors of over 10% occur, even if propane-air peaking mixtures are excluded. If included, the accuracy deteriorates further. However, means to improve these accuracies are possible and are being studied. Applying corrections to density can lead to results which may be useful for natural gas energy flow measurement, as shown in Figure 10. A safety test was made at PSC to demonstrate that a microbridge overvoltage burnout does not ignite a stoichiometric gas-air mixture. The microbridge chips were placed in a premlxed (0.8 stoichiometry) gas air

188

MICROSTRUCTURE SENSORS FOR FLOW, DIFFERENTIAL PRESSURE AND ENERGY MEASUREMENT

FIGURE 5. TEMPERATURE COMPENSATION ACHIEVED IN DIFFERENTIAL PRESSURE MEASUREMENT OVER 0 TO 60°C RANGE

mixture 0.5 inch above a large bunsen burner flowing at 9 l/min. Seven burnouts were done with no ignition in any case. Even at a pulse voltage 10× the voltage required for slow burnout, no ignition occurred. The . 005" spacing of the resistor from the silicon chip and the low burnout energy of about 10−5 joules (vs. minimum ignition energy of 2×10−4 joules (4)) make the cooling of the heated volume quite rapid and flame propagation cannot develop. CONCLUSIONS We have developed a novel, highly-sensitive air or gas flow sensor which performs well as a mass flow sensor and differential pressure sensor. It is especially suited to applications in the low differential pressure range from 0 to 1.0" of water column. It provides a more direct approach for temperature and pressure compensation than other presentlyavailable mass flow sensors requiring measurement of temperature and pressure. For some gas mixtures of varying composition, mass flow is indicated accurately (e.g. CO2 and He) without calibration corrections. Because it can be fabricated by conventional thin film deposition and silicon processing techniques. It offers the possibility of lower cost and broader applications than present commercially available gas flow sensors. Mixtures of other gas constituents require corrections. It appears to be possible to generate these corrections (using auxiliary real time measurements) to provide a mass flow output under varying temperature, pressure and composition conditions. Refinements of this approach may enable a similar measurement of energy flow in a natural gas stream. We are currently studying these approaches.

189

FIGURE 6. RESPONSE TO MASS FLOW IN A TYPICAL SMALL CHANNEL FOR AIR AND CO2

Because this sensor can be fabricated by conventional thin film deposition and silicon processing techniques, it offers the possibility of lower cost and broader applications than present commercially available gas flow sensors. REFERENCES 1. 2. 3. 4.

E.Bassous, “Fabrication of Novel Three Dimensional Microstructures by The Anlsotroplc Etching of (100) and (110) Silico’’. IEEE Transactions on Electron Devices, ED-25. No.10, 1178–1185 (1978). K.E.Bean, “Anisotropic Etching of Silicon”, Ibid., p. 1185. K.E.Peterson, “Dynamic Micromechanics On Silicon: Techniques and Devices” , Ibid., p. 1241. B.Lewis and G. v.Elbe, “Combustion Flames and Explosions of Gases”, 2nd Ed., Acad. Press, NY and London, (1961), pp. 323–346.

UB/jh IGT.wp 04–16–86

190

MICROSTRUCTURE SENSORS FOR FLOW, DIFFERENTIAL PRESSURE AND ENERGY MEASUREMENT

FIGURE 7. MICROSENSOR OUTPUT VS. A: VOLUME FLOW AND B: MASS FLOW

191

FIGURE 8. RESPONSE CURVES OF FOUR DIFFERENTIAL PRESSURE SENSING DEVICES, TO DEMONSTRATE REPRODUCIBILITY OF SENSOR FABRICATION

FIGURE 9. GAS HEATING VALUE VERSUS MOLECULAR WEIGHT OF NATURAL GAS SAMPLES SUPPLIED BY GRI

192

MICROSTRUCTURE SENSORS FOR FLOW, DIFFERENTIAL PRESSURE AND ENERGY MEASUREMENT

FIGURE 10. PLOT OF GAS HEATING VALUE VERSUS CORRECTED SPECIFIC GRAVITY FOR GRI GAS SAMPLES

THE SMART-CAL CUTLER HAMMER CALORIMETER COMBINATION FOR BETTER PERFORMANCE AND DATA PROCESSING Alfred F.Kersey Gas Industry Manager Fluid Data, Inc. 1844 Lansdowne Avenue Merrick, New York 11566

Figure 1

ABSTRACT

Ever since its invention in 1921, the Cutler Hammer recording calorimeter has utilized a stripchart recorder to display its instantaneous BTU and to provide a continuous BTU record. In the early years, the BTU record was merely for information, to check manufactured gas and gas mixing processes, or to assure that BTU minimums are met. In this day of MMBTU, the BTU is a major component in the cost of gas sales. Those people using Cutler Hammer Calorimeters use the BTU information in the following different ways : — They average the BTU from the stripchart by eye. — They average the BTU from the stripchart using an electronic chart averager. — They transmit a milliamp signal proportional to BTU to a host computer for MMBTU calculation.

194 THE SMART-CAL CUTLER HAMMER CALORIMETER COMBINATION FOR BETTER PERFORMANCE AND DATA PROCESSING

To review the operation of the tank unit, Figure 1 is a pictorial view of the tank unit. This you have seen before. The important features of this tank unit include: — Precise metering of gas to be burned. — Controlled flow of primary and secondary combustion air with 40% excess secondary air for complete combustion over a wide range of calorific values. — Precision metering of the heat absorbing air media. — The metering using water sealed meters at atmospheric pressure, makes the instrument immune to atmospheric humidity changes, immune to atmospheric pressure changes and the water provides a common temperature base for the metering. — Combustion in the enclosed heat exchanger causes the instrument’s response to be to the total calorific value (gross, higher). The water vapor formed during combustion is condensed and the latent heat of vaporization is included in the BTU measurement. All products of combustion are returned to the initial temperature. — The temperature rise of the heat absorbing air is a measure of the BTU of combustion. — The temperature rise is measured by two nearly identical nickel thermometers. — The BTU is proportional to the resistance difference between the two thermometers. The inputs to SMART-CAL are the inlet and outlet thermometers (RTD’s) of the tank unit. Measuring the resistance difference between these two thermometers, a transmitter in SMART-CAL generates an analogue output proportional to the resistance difference. This analogue output is proportioned to BTU. Then an analogue to digital converter sends a digital output proportional to BTU to the microprocessor. The SMART-CAL keyboard provides operational directives to the microprocessor: — To perform in BTU, K-Cal or Mega-joules mode. — To set the time and date. — To set auto-calibration times. — To call for special calibration. — To print instantaneous calorific value on demand. — To set alarm limits for calorific value. — Manual calibration. The SMART-CAL as directed, will give the following: — Instantaneous calorific value indication. — Average calorific value for one hour, eight hours and 24 hours printed on tape. — Daily auto-calibration including control for the customers sample line transfer values. — High and low calorific value. — Milliamp output proportional to calorific value. — Alarm indication and associated contact closure. — RS232C output of calorific value. It is appropriate to point out that there are two calibration techniques used with SMART-CAL. One technique is used for start up and for periodic service such as that done on a six month basis. The second technique is the one used for auto-calibration.

195

Photo 1

The start-up calibration and periodic service calibration uses cold balance and certified gas to position the zero and span of the performance curve of the instrument. A cold balance switch introduces into the circuitry a resistor network. This is necessary because of the higher degree of curvature of the performance curve as the outlet thermometer approaches the cold condition. A zero adjustment is made after the outlet thermometer is sufficiently cool. Then the instrument is put on certified gas and allowed to come up to heat. At this point, a keyboard adjustment is made so that the SMART-CAL display indicates the certified calorific value. The keyboard is used again to put the calorimeter back onto line gas. Auto calibration is done without operator assistance. The microprocessor will control the transfer from line gas to calibrating gas. After an appropriate time for the instrument to stabilize on calibrating gas, the microprocessor will note any deviation and correct the SMART-CAL reading. Then automatically, the calorimeter will go on to line gas and after an appropriate period of time, the calorific values will again be used for averaging. The improvements in accuracy which we expect to achieve with SMART-CAL come from the more frequent calibra-tion and from the averaging. Whereas most of our users calibrate weekly and some calibrate as infrequently as monthly, automatic daily calibration should achieve a higher degree of precision or at least a higher degree of confidence level in the measured calorific value. The automatic averaging provided by SMART-CAL should achieve a higher degree of accuracy than the eyeballing of a stripchart or even the electronic chart averaging. Chart errors due to chart expansion and contraction are eliminated. We are sure that those persons who use SMART-CAL will achieve our improved performance claims and feel their money was well spent.

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY Jag J.Singh, Ph.D. Danny R.Sprinkle Richard L.Puster NASA Langley Research Center Hampton, Virginia 23665

ABSTRACT

As a spin-off of a system developed for’ monitoring and controlling the oxygen concentration in the Langley 8-Foot High-Temperature Tunnel, a highly accurate on-line technique has been developed for determining heats of combustion of natural gas samples. It is based on measuring the ratio m/n, where m is the (volumetric) flowrate of oxygen required to enrich the carrier air in which the test gas flowing at the rate n is burned, such that the mole fraction of oxygen in the combustion product gases equals that in the carrier air. The m/n ratio is directly related to the heats of combustion of the saturated hydrocarbons present in the natural gas. A measurement of the m/n ratio for the test gas can provide a direct means of determination of its heat of combustion by using the calibration graph relating the m/n values for pure saturated hydrocarbons with their heats of combustion. The accuracy of the technique is determined solely by the accuracy with which the flowrates m and n can be measured and is of the order of 2 percent in the present study. The theoretical principles and experimental results are discussed. INTRODUCTION We recently developed a system for monitoring and controlling the oxygen concentration in the Langley 8Foot High-Temperature Tunnel (ref. 5). It is based on a y2O3-stabilized ZrO2 electrochemical sensor, The system is capable of maintaining oxygen concentration at 20.9±1.0 percent in the methane-oxygen-air combustion product gases. During the development of the system, it was noted that the degree of the required oxygen enrichment of the air was very strongly dependent on the purity of the combustible gas (CH4). For example, it was noted that the presence of noncombustible components—such as N2, CO2, or H2O —noticeably affected the amount of oxygen that needed to be added to the air to make the mole fraction of oxygen, X(O2), in the combustion product gases equal to that in the standard air. It was quickly realized that

197

the reverse was also true, that is, the purity of the combustible test gas could be inferred by measuring the amount of oxygen needed for X(O2) equalization (ref. 6). It is now demonstrated that the technique can be further extended to infer the presence of heavier saturated hydrocarbons in the combustible test gas (natural gas) as well. It is assumed that saturated varieties are the only types of hydrocarbons found in the natural gas samples. According to reference 3, saturated hydrocarbons are the predominant varieties present in U.S. natural gas samples. Since the heat of combustion of natural gas depends on its effective hydrocarbon content , measurement of the amount of oxygen needed for X(O2) equalization can be used for direct determination of the thermal content of the test gas (natural gas). Details of the theory of operation of the new system and the experiments conducted to verify its accuracy are discussed in the following sections. The heats of combustion of gaseous hydrocarbons are presently determined by using a constant pressure flame calorimeter (ref. 8). However, calorimetric measurements cannot be made on-line and require information about the thermal properties of the combustion products of the test sample. The technique reported here, on the other hand, is direct, can be performed on-line, and requires no prior knowledge about the composition of the test sample. SYMBOLS A

A depends only on the composition of the test gas and the mole fraction of oxygen in the air used for combustion. CxHy Hydrocarbon (for saturated hydrocarbons, y=2x+2) f Combustible fraction of test gas I Noncombustible impurities (such as N2, CO2, and H2O) in test gas Volumetric flowrate of carrier air stream m Volumetric flowrate of oxygen n Volumetric flowrate of test gas (.mixture of saturated hydrocarbons and noncombustible impurities) sccm Standard cubic centimeters per minute; values listed are for 20°C and 101.3 kPa X(O2) Mole fraction of oxygen α m/n for X(O2)=0.2095 (standard air) ΔH Heat of combustion THEORETICAL PRINCIPLES A general expression for the combustion of a natural gas sample in oxygen-enriched air can be written as follows:

(1)

where

198

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY

CxHy+I CxHy I f

Test gas Effective hydrocarbon in test gas Noncombustible impurities in test gas Combustible fraction of test gas

The mole fraction of oxygen in the products of equation (1) is given by

(2)

If X(O2)=0.2095, we obtain (3) If f=1(i.e., the test gas has no noncombustible impurities), equation (3) reduces to (4) This equation is identical to equation (6) in reference 6. If, on the other hand, f=0 (i.e., no combustible fraction is present in the test gas), equation (3) reduces to (5) Generalizing equation (3) for nonstandard air, we obtain

(6)

The values of m/n for some of the pure saturated hydrocarbons in standard and nonstandard air are summarized in Table I. Figure 1 shows the correlation between A and m/n for selected saturated hydrocarbons. It is apparent that A and m/n are linearly related. If f≠1, m/n values for various impurity-containing hydrocarbon gases will be different from those for pure hydrocarbons. An experimental measurement of m/n will then give f, the combustible fraction in TABLE I. SUMMARY OF m/n VALUES FOR SELECTED SATURATED HYDROCARBONS [X(O2)=0.2042 and 0. 2095] Hydrocarbon

Chemical Formula

X(O2)=0.2042

X(O2)=0.2095

Methane Ethane Propane

CH4 C2H6 C3H8

m/n for

2.770 4.783 6.796

2.795 4.825 6.855

199

Hydrocarbon

Chemical Formula

X(O2)=0.2042

X(O2)=0.2095

Butane Pentane Hexane Heptane Octane Nonane Decane

C4H10 C5H12 C6H14 C7H16 C8H18 C9H20 C10H22

m/n for

8 .809 10.823 12.836 14.849 16.862 18.875 20.888

8.885 10.915 12.945 14.975 17.005 19.035 21.065

the test gas. The correlation between A and m/n for binary gas mixtures is shown in figure 2. Figures 1 and 2 clearly demonstrate that the m/n values for gases containing saturated hydrocarbons are uniquely related to their hydrogen and carbon contents. If the test gas is not a pure single hydrocarbon but is a mixture of two or more hydrocarbons, equation (6) can be used to readily calculate the m/n value for “effective” test hydrocarbon. For example, for a test gas containing equal mole fractions of CH4 and C2H6 , the effective equivalent hydrocarbon would be C1.5H5.0, giving m/n=3.81 for X(O2)=0.2095. If this mixture also contained an incombustible impurity, the corresponding m/n value would be lower, as indicated by equation (6). Figure 3 shows A as a function of m/ n for f ≠ 1 for test gases containing several saturated hydrocarbons as well as noncombustible gases. Again the uniqueness of the correlation between A and m/n is evident. Because m/n values for various test gases containing saturated hydrocarbons are related to their chemical composition, they can serve as the basis for direct determination of heats of combustion. Table II lists the gross heats of combustion of several pure saturated hydrocarbons (refs. 1, 2, 4, 9). Figure 4 shows heat of combustion (in kilocalories per mole) versus m/n value for several selected saturated hydrocarbons. Similar results for mixtures containing selected saturated hydrocarbons and noncombustible gases are illustrated in figure 5. It is apparent that the heats of combustion of various gases are directly related to their corresponding m/n values. An experimental determination of m/n for the test gas is, therefore, expected to provide an on-line determination of its heat of combustion. EXPERIMENTAL PROCEDURES AND RESULTS The experimental plan involved measuring m/n values for the following kinds of test gas samples: 1. Pure saturated gaseous hydrocarbons (methane, ethane, propane, butane, etc.) 2. Binary mixtures of selected saturated hydrocarbons (i.e., no noncombustible components) 3. Binary mixtures containing a saturated hydrocarbon and an inert gas. 4. Gaseous mixtures containing several saturated hydrocarbons and selected noncombustible impurities. Figure 6 shows a schematic diagram of the experimental system used for measuring m/n values for the test gas samples (refs. 5 and 6). TABLE II. GROSS HEATS OF COMBUSTION ΔH OF SELECTED SATURATED HYDROCARBONS Hydrocarbon

m/n [X(O2)= 0.2095]

ΔH, kcal/mol

CH4 C2H6

2. 795 4.825

212.80 372.82

kcal/mol 76.14 77.27

200

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY

Figure 1. A as a function of m/n for pure unbranched acyclic hydrocarbons. X(O2)=0.2042; A=4X(O2) (1−f) + f{4x+[1 +X(O2)]y} Hydrocarbon

m/n [X(O2)= 0.2095]

ΔH, kcal/mol

C3H8

6. 855

530.61

kcal/mol 77.40

201

Figure 2. A as a function of m/n ofr binary of saturated hydrocarbons and noncombustable gases. X(O2)=0.2042; A=4X (O2) (1−f) + f{4x+[1+X(O2)]y} Hydrocarbon

m/n [X(O2)= 0.2095]

ΔH, kcal/mol

C4H10 C5H12

8 .885 10.915

687.65 845.10

kcal/mol 77.39 77.43

202

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY

Figure 3. A as a function of m/n for complex mixtures of saturated hydrocarbons and noncombustible gases. X(O2) =0. 2042; A=4 X(O2) (1−f)+f{4x+[1+X(O2)]y}. Hydrocarbon

m/n [X(O2)= 0.2095]

ΔH, kcal/mol

C6H14

12.945

1002 .55

kcal/mol 77 .45

Determination of m/n Values for Pure Hydrocarbon Gases Since m/n values are strongly dependent on the mole fraction of oxygen, X(O2), pure bottled dry air was used to supply the carrier air stream. A gas chromatographic analysis of this air gave an X(O2) value of 0. 2042±0.0050. The ZrO2 sensor output was first recorded for thecarrier air stream. Next, the products of combustion of selected hydrocarbons in oxygen-enriched air were directed through the O2-sensing system.

203

Figure 4. Heats of combustion versus m/n values for saturated hydrocarbons. X(O2)=0.2042.

The oxygen flowrate (m) was adjusted until the ZrO2 sensor output matched that for air. These

204

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY

Figure 5. Heats of combustion versus m/n values for gas mixtures containing saturated hydrocarbons and noncombustible impurities. X(O2)=0.2042.

205

Figure 6. Experimental system for determining heats of combustion of hydrocarbon gases.

206

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY

measurements were repeated for a number of hydrocarbon flowrates (n) while the air flowrate ( ) was held constant. It should perhaps be emphasized that n in the parameter m/n represents the volume flowrate for the test gas rather than the conventional mass flowrate. It will therefore be necessary to use a volume flowmeter for metering the test gas flowrate. We used a positive displacement dry test meter for measuring the true volume flowrate (n) of the test gases. A dry test meter was preferred over the wet test meter, since the latter would allow the test gases to come in contact with water (or some other selected fluid) during passage through the meter and thereby affect the combustible fraction in the test sample. Three gases with widely different thermal characteristics were used to measure the accuracy with which the dry test meter measured their flowrates. The data summarized in Table III demonstrate that the dry test meter gives as accurate a volume flowrate value as the widely used thermal mass flowmeters. A comparison of the experimental and the calculated values for various saturated hydrocarbons is given in Table IV. Determination of m/n Values for Test Gas Mixtures Values of m/n for various test gas mixtures were determined in exactly the same way as for pure hydrocarbons. The test gas flowrates were calculated by using a flowrate conversion factor computed on the basis of known test gas composition (ref. 7). These computed flowrates were in agreement with the values obtained by direct measurements with a positive displacement dry test meter. The results for the various test gas mixtures are summarized in Table V. From the data summarized in Tables IV and V, it is apparent that the calculated and experimental values of m/n for various types of mixtures are in excellent agreement. As indicated earlier, the m/n values for the mixtures can be used to infer their heats of combustion. The reported (refs. 2,4,9) gross heat of combustion values for saturated hydrocarbons listed in Table II were used to derive the following relationship between m/n values and heat of combustion expressed in kilocalories per mole: TABLE III. COMPARISON OF DRY TEST METER AND THERMAL MASS FLOWMETER FLOWRATES IN THE RANGE OF 500–1500 seem Test Gas Dry Test

Flowrates, sccm, measured by metera

Air 792.3+10.0 651.8+10.0 502.7+10.0 Helium 1118.7+10.0 840.7+10.0 557.6+10.0 Carbon Dioxide 636.4+10.0 518.5+10.0 aErrors bErrors

Thermal mass Flowmeterb 999.3+10.0 794.7+10.0 646.0+10.0 499.9+10.0 1400.0+10.0 1119.3+14.3 842.1+14.3 558.1+14.3 729.0+10.0 629.3+ 7.4 512.8+ 7.3

994.6+10.0

1413.3+14.3

731.9+ 7.4

in the dry test meter flowrates are based on uncertainties in the time required for the flow of preset volumes. in the mass flowmeter flowrates represent 1 percent of the full scale of the meter (Full scale=1000 sccm).

207

TABLE IV. SUMMARY OF EXPERIMENTAL AND CALCULATED m/n VALUES FOR SELECTED HYDROCARBONS IN AIR [X(O2)=0.2042] Test Hydrocaron Hydrocarbon Flowrate,a n ,sccm Experimental

Calculated (from Eq. (6))

CH4 C2H6 C3H8 C4H10

71.80±0.72 45.85±0.46 29.18±0. 30 24.20±0 .24

aThe

Equalizing Oxygen flowrate ,a m, sccm

Carrier Air Flowrate ,a , seem

m/n value

200.80±1.00 214.67±1.00 199.71±1.00 213.00±1.00

500.0 500.0 500.0 500.0

2 .797± 0.042 4.682±0.070 6.843±0.103 8.801±0.132

2. 770 4.783 6.796 8.809

seem values are for 20°C and 101.3 kPa. The accuracies of the flowrates reflect ±1 percent of the full scale of the flowmeters.

TABLE V. SUMMARY OF EXPERIMENTAL AND CALCULATED m/n VALUES FOR VARIOUS TEST GAS MIXTURES (COMPLETED) [X(O2=0.2042)] Test Gas Mixture, Percent Volume

Mixture Flowrate,a n, sccm

Experimental

Calculatedb

Equalizing Oxygen Carrier Air Flowrate ,a m, sccm Flowrate,a , sccm

m/n Value

39.7+0.4

199.0+1.0

500.0

5.014+0.075

4.894+0.083

49.9+0.5

198.0+1.0

500.0

3.968+0.060

3.975+0.034

50.3+0.5

193.0+1.0

500.0

3.837+0.057

3.936±0.046

49.2+0.5

199.5+1.0

500.0

4.055+0.061

4.070+0.036

Complex Mixturesc 44.63%C2H6+25. 88%C3H8+ 10.81% C4H10+5.09%CO2+ 13.59% N2 53.04%CH4+26. 09%C2H6+ 18.42% C3H8+1.43%CO2+ 1.01% N2 50.84%CH4+17. 12%C2H6+ 24.87% C3H8+3.34%CO2+ 3.83% N2 51.51%CH4+21. 24%C2H6+ 23.82% C3H8+1.74%CO2+ 1.69% N2 aThe

seem values are for 20°C and 101.3 kPa. The accuracies of the flowrates reflect ±1 percent of the full scale of the flowmeter. bThe errors in the calculated values results from errors in the mixture composition. cThe accuracies of the concentrations are of the order of +1 percent of the values listed.

TABLE V. SUMMARY OF EXPERIMENTAL AND CALCULATED m/n VALUES FOR VARIOUS TEST GAS MIXTURES (CONT.)

208

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY

[X(O2=0.2042) Test Gas Mixture Percent Volume

Mixture , Flowrate,a Equalizing Oxygen Carrier Air Flowrate ,a m, sccm Flowrate, a , sccm n, sccm

Experimental

Calculatedb

m/n Value

Binary Mixtures 50.10% CH4+49. 90% C2H6 50.19% CH4+49. 81% C3H8 50.40% CH4+49. 60% C4H10 51.39% C2H6+48. 61% C4H10 50.13% CH4+49. 87% N2 50.08% CH4+49. 92% CO2 49.92% C2H6+50. 08% N2 49.87% C2H6+50. 13% CO2 49.96% C3H8+50. 04% N2 49.91% C3H8+50. 09% CO2

96.00+0.83 68.70+0.75 56.18+0.75 52.58 ± 0.53

362.3+1.0 326.3+1.0 322.0+1.0 353.0+1.0

500.0 500.0 500.0 500.0

3.774+0.043 4.751+0.066 5.732+0.092 6.713+0.087

3.770+0.044 4.783+0.065 5.7890.094 6.740+0.030

100.30+1.23 100.40+1.01

153.5+1.0 153.3+1.0

500.0 500.0

1.530+0.028 1.527+0.025

1.517+0.034 1.515±0.031

99.80+1.10 99.90+0.84

252.2+1.0 253.0+1.0

500.0 500.0

2.526+0.038 2.531+0.031

2.516+0.026 2.514+0.023

99.90+1.05 100.00+0.78

353.3+1.0 354.3+1.0

500.0 500.0

3.536+0.040 3.542+0.037

3.524+0.021 3.521+0.018

(7)

where a0=23.2134 a1=91.2088 a2=−2.9745 a3=0.3032 a4=0.0117 In developing

this

equation,

a

statistical weight of 5 was assigned to to reflect the fact that the heat of combustion of a noncombustible mixture is definitely zero, whereas the heats of combustion of other mixtures are not known as accurately. A comparison between the heat of combustion values determined from experimentally observed m/n values and those calculated from known chemical compositions is given in Table VI. It is apparent from the results summarized in Table VI that the agreement between the experimental and the calculated values of heat of combustion for the various gas mixtures is excellent.

209

VERIFICATION OF COMPUTATIONAL PROCEDURE The gas selected for the verification test was a sample of natural gas supplied to Langley Research Center by the Virginia Natural Gas Company (VNG). Its flowrate (n) was measured with the dry test flowmeter, and the corresponding m/n value was determined in the manner explained earlier. The test gas was later analyzed by gas chromatography (ref. 2) and its m/n value was computed on the basis of its measured chemical composition, which is given in Table VII. It should be noted that the only combustibles present in the VNG sample were saturated hydrocabrons. There were no measurable concentrations of H2, O2, and C5H10. This is consistent with a recent Bureau of Mines study (ref. 3), in which it was found that the concentrations of H2, O2, and C5H10 in almost all natural gas samples analyzed in 1982 were ≤ 0.1 mole percent. A comparison between the experimental and computed tables of m/n is summarized in Table VIII. Also included in this table are the computed TABLE VI. COMPARISON OF EXPERIMENTAL AND CALCULATED VALUES OF GROSS HEAT OF COMBUSTION ΔH FOR VARIOUS GAS MIXTURES (CONT.) Mixture Composition, Volume Percent

ΔH, kcal/mol

Experimentala

Calculatedb

Binary Mixtures 50.10% CH4+49.90% C2H6 50.19% CH4+49.81% C3H8 50.40% CH4+49.60% C4H10 50.13% CH4+49.87% N2 50.08% CH4+49.92% CO2 49.92% C2H6+50.08% N2 49.87% C2H6+50.13% CO2 49.96% C3H8+50.04% N2 49.91% C3H8+50.09% CO2

292.6±3.4 369.5±5.2 446.3±7.4 110.6±2.2 110.4±2.2 193.0±3 .1 193.4±2.6 274.3±3.8 274.7±3.0

292.2±3.9 371.1±5.2 449.2±2.2 109.0±2.8 109.4±2.6 192.2±2.2 192.0±1.8 273.3±1.6 273.0±1.4

TABLE VI. COMPARISON OF EXPERIMENTAL AND CALCULATED VALUES OF GROSS HEAT OF COMBUSTION ΔH FOR VARIOUS GAS MIXTURES (COMPLETED) Mixture Composition, Volume Percent ΔH, kcal/mol Experiment ala

Calculatedb

Complex Mixtures 44.63% C2H6+25.88% C3H8+ 10.81% C4H10+5.09% CO2+ 13.59% N2 53.04% CH4+26.09% C2H6+ 18.42% C3H8+1.43% CO2+ 1.01% N2 50.84% CH4+17.12% C2H6+ 24.87% C3H18+3.34% CO2+ 3.83% N2 51.51% CH4+21.24% C2H6+ 23.82% C3H8+1.74% CO2+ 1.69% N2 aThe

390.1±11.7

380.6±9.9

307.7± 7.8

308.3±8.0

297.4±5.6

305.4±8.0

311.5±8.0

315.9±8.3

experimental values of heat of combustion were obtained by using the experimentally measured m/n values. The errors in these heat of combustion values result from errors in in the m/n values. bThe calculated values of heat of combustion were obtained from the known mixture compositions. The errors in these values result from the errors in the mixture composition.

210

A NEW ON-LINE TECHNIQUE FOR NATURAL GAS CALORIMETRY

TABLE VII. COMPOSITION OF NATURAL GAS SAMPLE Componenta

Volume Percentb

Componenta

Volume Percenta

CH4 C2H6 C3H8 C4H10

95.91 2.14 0.37 0.20

C5H12 C6H14 CO2 N2

0.10 0.01 0.84 0.43

aNo

measurable concentrations of O2 or C5H10 were found in the natural gas sample tested in this study, though traces of these gases have been reported in natural gas samples from other sources (ref. 3). bThe accuracies of the concentrations are of the order of ± 2 percent of the values listed. TABLE VIII. COMPARISON OF COMPUTED AND EXPERIMENTALLY DETERMINED VALUES OF m/n AND HEAT OF COMBUSTION FOR THE NATURAL GAS SAMPLE Parameter

Computed

Experimental

m/n ΔH , kcal/mol

2.84±0.03 218.8±2.2

2.85±0.01 218.5±1.1

and experimentally determined values of the gross heat of combustion. The agreement between corresponding values is excellent. CONCLUDING REMARKS A new technique has been developed for determining heats of combustion of test gases containing saturated hydrocarbons. It is based on the measurement of the m/n ratio, where m is the volumetric flowrate of oxygen needed to enrich the carrier air in which the test gas flowing at the rate n is burned, so that the mole fraction of oxygen in the combustion product gases equals that in the carrier air. The m/n values are directly related to the heats of combustion of the test gases. The accuracy of the derived values of heat of combustion is determined solely by the accuracies with which the flowrates m and n can be measured. At the flowrates used in the present study, the respective accuracies are of the order of 1 percent. This leads to an error in the heats of combustion of the order of 2 percent. The heats of combustion of hydrocarbons are presently determined by using a constant volume bomb calorimeter for liquids and solids and a constant pressure flame calorimeter for gases. These measurements can be very accurate (< 1 percent), since they depend mainly on the bath temperature measurement. However, calorimetric measurements cannot be made on-line and require information about the thermal properties of the combustion products of the test sample. The technique reported here, on the other hand, is direct, can be performed on-line, and requires no prior knowledge about the exact composition of the test sample. (The only assumption made regarding the composition is that saturated hydrocarbons are the only combustibles present in the test samples). It thus appears that this new technique may be more useful for field operations where on-line measurements of the heats of combustion of the test gases are often needed. REFERENCES 1. 2.

Green, Don W., ed.: Perry’s Chemical Engineers’ Handbook, Sixth ed. McGraw-Hill, Inc.(1984). McCracken, Dudley J.: Hydrocarbon Combustion and Physical Properties. BRL Rep. No. 1496, U.S. Army, Sept. (1970). (Available from DTIC as AD 714 674).

211

3. 4. 5. 6. 7. 8. 9.

Miller, Richard D.; and Hertweck, Floyd R., Jr.: Analyses of Natural Gases, (1982). Info. Circ. 8942, Bureau of Mines, U.S. Dept. Interior. Rose, J.W.; and Cooper, J.R., eds.: Technical Data on Fuel, Seventh ed. John Wiley 8 Sons, Inc., (1977). Singh, Jag J.; Davis, William T.; and Puster, Richard L.: Proposed Fast-Response Oxygen Monitoring and Control System for the Langley 8—Foot High—Temperature Tunnel. NASA TP—2218 (1983). Singh, Jag J.; and Puster, Richard L.: New Technique for Calibrating Hydrocarbon Gas Flowmeters. NASA TM– 85792 (1984). Singh, Jag J.; and Sprinkle, Danny R.: A New Technique for Measuring Gas Conversion Factors for Hydrocarbon Mass Flowmeters. NASA TM-85676 (1983). Sunner, Stig; and Mansson, Margaret, eds.: Experimental Chemical Thermodynamics. Volume—Combustion Calorimetry. Pergamon Press (1979). Weast, Robert C., ed.: CRC Handbook of Chemistry and Physics, 65th ed. CRC Press, Inc. (1985) pp. D-275–D-281.

AVAILABILITY OF NBS-TRACEABLE CALIBRATION GAS STANDARDS FROM IGT—PROGRAM UPDATE Bruce H.Solka Amir Attari Institute ot Gas Technology Chicago, IL 660616

ABSTRACT

The Institute ot Gas Technology has provided the American gas industry with natural gas standards ot certitied calorific value and specific gravity since 1965. In August of 1984, a research program was established at IGT to study the factors that may influence the stability ot natural gas calibration standards during their preparation, storage, and use. The information obtained in the course ot this study is being used to prepare a series ot calibration standards for natural gas measurements by gas chromatography and calorimetry. This paper will review the history ot previous work and the cooperative research activities that led to the preparation and availability of the new calibration standards that are traceable to the primary standards from the National Bureau of Standards (NBS). INTRODUCTION Prior to conversion trom manufactured gas to natural gas, the standard practice of the American gas industry was to calibrate its calorimeters with nydrogen. As natural gas gradually replaced manufactured gas in the 1940’s and 1950’s, the gas industry became concerned about the use of hydrogen as a calibration gas. Although hydrogen has a calorific value reasonably close to that of the manutactured gas, it only has one third the calorific value of natural gas. This concern led to a cooperative program between the American Gas Association (A.G.A.) and the NBS to study how accurately the Thomas recording gas calorimeter (made by Cutler-Hammer) could measure gases of higher caloritic value. As a result of this study, it was concluded that the gas industry would require a continuing source of calibration gases with calorific values near that of natural gas. In recognition of that need a new program was initiated in 1957 by the A.G.A. under which the Washington Gas Light Company would supply puritied natural gas, and NBS would provide the certification service.

213

In 1961, NBS suggested that the program De transferred elsewhere; IGT was selected to provide that service. By 1965 the program was fully activated at IGT with the arrival of the first two cylinders of NBS primary standards containing high purity (99.96 mole percent) methane; the balance consisted of ethane, carbon dioxide, nitrogen and oxygen. Today there are 3300 size 1A steel cylinders used in IGT service, with 100 to 150 cylinders added annually. Since 1965, IGT has issued 8782 calorific value and 7983 specific gravity certificates. The certified gas is prepared by purifying natural gas taken from a medium-pressure service main. The purification is accomplished by passing the gas over activated charcoal to remove hydrocarbons heavier than ethane, in order to improve gas stability. The processed gas, containing methane, ethane, nitrogen and carbon dioxide, is then adjusted to the desired calorific value or specific gravity before compression to 2000 psi into evacuated cylinders. Quality control is maintained by periodic gas analysis, and by regenerating the charcoal towers when propane breaks through. During the certification procedure, each cylinder receives three test runs of 1–1/4 hour duration on three Cutler-Hammer calorimeters that are calibrated daily against the NBS primary standards. The results of the three determinations must agree to within 1 Btu/SCF betore they are certified for accuracy to 0.9 Btu/sCF and shipped. Similarly, the specitic gravity readings obtained on our gravity balances must agree to within 0. 0005 units betore they are acceptable for certitication to ±0.001 specific gravity unit. This degree of precision and accuracy is achieved through a strict adherence to sound laboratory practices and regular instrument maintenance and service. In addition, the certification laboratory has a constant temperature regulation of ±1.5°F and a 100% exhaust system that removes any traces of combustible gases or vapors that might otherwise cause errors in the calorimeter readings. NEW PROGRAM There is an ever increasing demand within the U.S. gas industry for improved accuracy in gas quality ana gas quantity measurements. An important element of these measurements practices is the determination of the caloritic value of natural gas which requires the best available instrumentation as well as reliable standard gases to calibrate them. The instruments used to measure the thermal content ot natural gas fall into two general categories: • Combustion calorimeters that directly monitor the thermal energy from the gas combustion. • Gas analyzers that first determine the gas composition, then calculate the calorific value from the composition. Both types of instruments require careful calibration with a gas of known calorific value or composition in order to generate accurate results. Standard gases have been readily available from IGT for the calibration of calorimeters and gravitometers since 1965, but as the U.S. gas industry began to use gas chromatographs (GC) in their laboratories to measure composition, they had to rely on calibration gases purchased from a variety of sources without NBS traceability. Another change in measurement practices has been the need to measure fuel gases having heating values far different than the nominal 1000 Btu/SCF. Although reliable GC calibration gases were available from some vendors as a result of the standardization activities of the Gas Processors Association (GPA), and the American Society for Testing and Materials (ASTM) these gases lacked the NBS traceability that usually becomes an important consideration in resolving difterences in the calculation of a purchased quantity ot energy at the custody

214

AVAILABILITY OF NBS-TRACEABLE CALIBRATION GAS STANDARDS FROM IGT—PROGRAM UPDATE

transfer points. Clearly, a need existed within the gas industry for a series of new calibration gases with the added confidence, consistency, and recognition that direct NBS traceability would have provided. IGT was made aware ot this demand by an increasing number of inquiries from our current certified gas customers. As a result ot these inquiries a questionnaire was mailed to 55 ot the present program’s largest customers. This questionnaire addressed their anticipated use of reference standard mixtures for gas chromatographic analysis of natural gas. Of the 31 respondents, 80% were using GC at that time. satisfaction with available standards was mixed. A clear majority desired NBS traceability of standards as well as calorimetric verification of the heating value of the GC standard mixture. The Gas Research Institute (GRI) also became aware ot industry interest in new reference standard mixtures certified by an independent agency. To fulfill that need, in August of 1984 GRI initiated an experimental program at IGT that has culminated in the availability ot a new 10-component standard gas for GC calibration ana three new calorimeter standard gases for more precise calibration ot IGT calorimeters used for gas certification in the range ot 800 to 1300 Btu/SCF. The research conducted at IGT during this program nas provided information on the following parameters that may influence the stability ot the calibration gases during storage and use. 1. Selection of Gas Composition A GC calibration gas is most useful when its composition approximates that of the test sample as closely as possible. The survey referred to above gave respondents the opportunity to express their wishes regarding the composition of a GC calibration standard. With this information as well as our own knowledge of the typical composition ot pipeline-quality natural gas, numerous possible gas mixtures were considered as candidates for this program. Given recent interest in the heavier ends of natural gases, careful consideration was given to the advisability ot including hexane and heavier components in the standard. Table 1 lists five mixtures that typify the composition ranges that were considered for a GC calibration gas. A primary requisite for a usable standard is that its composition remain constant during the period ot its use. While the presence ot heavy components may be desirable for calibration purposes, they increase the possibility of composition changes caused by condensation or adsorption. In the northern climates, certified gas cylinders may well be exposed to winter temperatures of−30°F duriny shipment or storage. Therefore, we first calculated the SRK equation-ot-state dew points as a function of pressure for some 75 gas mixtures, and based on that consideration, selected gas mixture D as our primary standard gas. Table 2 lists the nominal composition ot the GC standard currently sold along with typical uncertainties for each component. These uncertainties are based on the sum ot the uncertainty in the NBS certification plus our instrument calibration and analysis of each cylinder. Table 1. COMPOSITION OF VARIOUS NATURAL GAS BLENDS Components/Sample

1

2

------------Mol Methane Ethane Propane i-Butane n-Butane

70.55 8.98 5.95 3.02 2.98

90.58 3.50 1.00 0.50 0.50

3

4

D

%d ------------88.73 3.50 1.00 0.40 0 .40

90.65 3.50 1.00 0.50 0.50

90.65 4.00 1.00 0.30 0.30

215

Components/Sample

1

2

------------Mol i-pentane n-pentane neo-pentane n-Hexane n-Heptane n-Octane Helium Nitrogen Carbon Dioxide Propylene Estimated dew point temperature, °F at filling pressure of 300 psig

1.00 1.00 — — — — 0.46 4.95 1.09 0.02 *

0.15 0.15 0.05 0.05 0.01 0.01 — 2.50 1.0 — 24

3

4

D

%d ------------0.15 0.15 0.10 0.05 0.02 — — 2.50 3.00 — 12

0.15 0.15 — 0.05 — — — 2.50 1.00 — −3

0.10 0 .10 — 0.05 — — — 2.50 1.00 — −12

*This gas is sold at a pressure of 100 psig for which the dew point is 24°F. Table 2. TYPICAL COMPOSITION OF NEW GC CALIBRATION GAS Component

Mol %d

Nitrogen Carbon Dioxide Methane Ethane Propane i-Butane n-Butane i-pentane n-pentane n-Hexane

2.50 ± 0.02 1.00 ± 0.01 90.65 ± 0.08 4.00 ± 0.02 1.000 ± 0.008 0.300 ± 0.007 0.300 ± 0.007 0.100 ± 0.006 0.100 ± 0.006 0.050 ± 0.004 Selection of Calorimetry Standard Compositions

The composition criteria for low- and high-Btu calorimetric standards are less severe than for the GC standard. This is because the calorimeter measures the gross heating value of a fuel gas independently of its composition. Thus, simple binary mixtures of methane/ethane and methane/nitrogen can be selected to attain any desired heating value for calibration standards in the range ot 800 to 1200 Btu/SCF. However, the criteria of stability still requires a dew point consideration. Calorimetric analysis also consumes several orders of magnitude and more gas volume than GC analysis. Thus a calorimeter standard is normally compressed to near 2000 psig in a 1–A size cylinder to deliver a volume ot about 225 SCF to the user. Table 3 lists the composition and heating values ot the mixtures that were chosen for these standards. A mixture considered for a 1261 Btu/SCF gas approached the −5°F dew point region. Mixture “C” (75% methane/25% ethane blend) with a heating value of 1185 Btu/SCF was chosen to be as near 1200 Btu/SCF as possible and yet maintain a reasonably low dew point temperature of −26°F.

216

AVAILABILITY OF NBS-TRACEABLE CALIBRATION GAS STANDARDS FROM IGT—PROGRAM UPDATE

Figure 1. OPERATIONAL SEQUENCE OF IGT’S CERTIFICATION PROGRAM Table 3. PROPERTIES OF NBS PRIMARY STANDARDS FOR CALORIMETRY Component/ Standard Methane Ethane Nitrogen Other Properties Calorific Value, Btu/SCF Estimated maximum dew point temperature, °F at 900 psig

A B C d ---------Mol % ---------99.99 81.50 75.00 trace — 25.00 trace 18.50 — 996 −206

812 −136

1185 −26

2. Certification process Following the selection procedure, tour gas mixtures designated as A, B, C and D were commercially prepared to IGT specifications, then analyzed by NBS for certification as primary calibration standards in this program and returned to IGT. Figure 1 diagrams the overall sequence of activities leading to delivery of IGT certified calibration standards with NBS traceability. GC Calibration Gas

We are currently shipping this standard with certifica-tion compositions determined in the following manner. • Stock gas at the nominal composition shown in Table 2 is prepared in batches of adequate size to split on a manifold into sets of six size N150 (or 15 size N60) aluminum cylinders. These sale cylinders are prepared by a three-fold exhaustive cycle of evacuation and pressurization with nitrogen to purge any residual gases. • Each sale cylinder is then analyzed by the following procedure. First, a daily system blank run is made to verify that there are no GC system air leaks that will contribute to the oxygen/nitrogen values in the certification analyses. Then three runs are made on the NBS certified primary to check the GC calibration. Results are calculated by a “relative response factor with normalization” method. Response factors for these three runs must agree with each other ana with those of preceding days to within specified tolerances before the certification analysis can begin. The standard is again re-run midway

217

through the work day to verify that instrument response is not drifting during certification work. Certified compositions for a given cylinder are determined by averaging a set of six results from duplicate analyses conducted on three non-consecutive days. If coefficients of variance of results for each of the 10 components are within allowable tolerances, that cylinder’s composition is considered certitiabie. • Calculated calorific values and relative densities are provided with each certificate ot composition. it the customer has ordered optional certification of caiorimetric value and/or measured relative density the cylinder is transferred to the calorimetry laboratory for those determinations. New Calorimeter Calibration Gas These gases continue to be certified in the manner outlined in the Introduction. Requests deviating from the formerly nominal 1000 Btu/SCF are met by nitrogen/ethane blending with the purified stock gas. The significant improvement in the new program that permits certification over a broad range of heating values is in the use ot the new primary standards discussed above. Using the new NBS primary standards shown in Table 3, each calorimeter was first calibrated at the midrange with the standard A, then the other primary standards were used as the test gases to measure any deviation between the calorimeter response and the known calorific values of standards B and C. The results indicated a remarkable consistency among the signal outputs from the three instruments for each standard gas, including the small negative deviations shown in Table 4. The significance of this information is that it alerts the certitied gas users to the need for close matching between the heating values ot the calibration gas and that of their test gas to within 25 Btu if the linearity of their calorimeter reading is unknown. The data also provides IGT with yet another correction factor to improve the reliability ot the certitication procedure. Table 4. CALORIMETER RESPONSE FROM 800 TO 1200 Btu/SCF Standard Gas Designation

Certified Value

Measured value

Deviation

-------- Calorific value, Btu/SCF ------B E* A F C

812.0 906.4 996.5 1091.3 1185.6

811.7 ± 0.5 906.1 ± 0.2 (used as calibration gas) 1090.9 ± 0.2 1184.7 ± 0.4

−0.3 −0.3 0.0 −0.4 −0.9

*Standards E and F were certified by IGT Chromatography Analysis.

Similar efforts were made to obtain data on the linearity of response for the GC detector over a twenty-told range of concentration (e.g., propane concentration range ot 0.25 to 5 mole percent) for each one of the 10 components present in the GC primary standard, except methane for which data for a range of 65 to 95 mole percent had already been obtained. This information is essential for IGT it IMBS traceability is to be maintained for the analysis ot natural gas samples with component concentrations that vary substantially from those of the NBS primary standard. Equipment Used

The particular set of equipment that has been assembled and dedicated to the GC portion of this program consists of the following principal components:

218

AVAILABILITY OF NBS-TRACEABLE CALIBRATION GAS STANDARDS FROM IGT—PROGRAM UPDATE

Gas Chromatograph: Data Acquisition: Sample Handling:

Hach/Carle AGC Series 400, Model 04192-A Perkin-Elmer Sigma 15 Chromatography Data System vacuum manitold with validyne Model CD 223 Digital Manometer

The chromatography data station transmits results to an IBM PC computer via RS 232 communications for the purpose ot data storage and evaluation. The calorimetry certification program will continue to use the Cutler Hammer recording calorimeter that has been the workhorse of the existing IGT program. This has been supplemented for the current program expansion by the addition of a third calorimeter and construction ot an automated, six-port gas manifold sampling system to permit unattended measurements on batches of five cylinders. 3. Program validation Container Materials

For convenience in shipping and handling, IGT felt aluminum cylinders would be the most attractive choice ot the options available. To verity that storage stability would not be adversely affected, a storage stability test was conducted. A carbon steel and duplicate sets ot stainless steel and aluminum cylinders were filled with the standard gas and analyzed monthly for seven months ana then bi-monthly to the present time. During this 14 month period, no change has been observed in the concentrations ot any ot the 10 components in any of the five cylinders. Aluminum cylinders are being used in this program for convenience and ease ot transport. Pressure Draw-Down Test

A cylinder to be used in this program was tested to verify that the composition of the certified gas does not change as the cylinder pressure is drawn down. Its contents were analyzed as the pressure was drawn down from 1500 to 15 psig at a rate that would be consistent with actual use. There was no significant change observed in the concentrations for any of the 10 components. User Tests of GC Calibration Gas

Prior to ottering the GC calibration gas as a general service a test application was arranged with the cooperation of six gas industry laboratories. The objective nere was twofold. The first was to verify the suitability ot the mixture as a standard for labs using a variety of GC methodologies (especially in terms ot sampling technique and treatment of the C6 and heavier traction). The second was to verify that the mixture retained its compositional integrity during its four-month period of shipping and handling by seven different labs (including IGT) geographically distributed throughout the country. To conduct the test, two test sample cylinders of natural gas plus one cylinder of new GC calibration yas were filled and analyzed at IGT before shipment to other participants. This set of three cylinders was then sent sequentially to the six industry labs and finally returned to IGT. Each lab was asked to do two sets of triplicate determinations on separate days. Individual results of the test from each lab are shown in Tables 5 and 6. The preliminary evaluation of this set of data leads to the conclusion that the gas standard functioned quite well in this test. Data not shown in the tables but recently obtained by re-analysis at IGT after the

219

cylinders were returned resulted in values that were nearly identical in their limits of precision-variance to the third decimal place of the analysis after we obtained the results betore shipping the cylinders. Another conclusion that can be reached is that these six laboratories are doing natural gas analyses with a fairly high degree of accuracy as can be judged by the inter-laboratory comparisons and standard deviations of the set of data as a whole. within a single laboratory, the precision of day-to-day analysis is quite good. Caretul inspection of the entire set reveals certain compounds that are giving problems to one laboratory or another. These will be discussed in our final report to GRI on this program, but do serve to point out that a routine crossreference or inter-laboratory quality assurance program may be of value to the gas industry. Heating values were calculated from all of the results for Sample 1 and are included in Table 4. A range of 4.5 Btu/SCF was observed which includes four values in the 1019 to 1020 Btu/SCF region which might in fact meet the criteria for discarding as outliers when the final evaluation is done. This result as well as those for the concentrations of the individual components, indicate that gas chromatography can be a valuable analytical tool to the gas industry. 4. Future Standards From IGT At present the program is restricted to standards of the nominal composition as shown in Table 2. Within a few months, as the program becomes more establisned, we will probably consider requests for other concentrations of the listed components. If interest is sufficiently high we may expand the component list to add higher carbon number hydrocarbons or other light gases but this will require another set of program validation experiments betore they could be ottered as another NBS-traceable standard. 75/NBS/PAP

220

AVAILABILITY OF NBS-TRACEABLE CALIBRATION GAS STANDARDS FROM IGT—PROGRAM UPDATE

Table 5. INTERLABORATORY TEST (Sample 1)

221

Table 6. INTERLABORATORY TEST (Sample 2)

THE EFFECT OF MOISTURE CONTENT ON NATURAL GAS HEATING VALUE Ted J.Glazebrook, A.A.S. Superintendent of Measurement Tenngasco Corporation, Louisiana Operations Alexandria, Louisiana 71309

ABSTRACT

This paper deals with the effect that water vapor has on the heating value of natural gas. It outlines a brief history of the evolution of energy measurement, its implications to the industry and its practical application through examples of calculations of water vapor content on MMBtu calculations. This paper should familiarize the reader with various methods and instruments used to determine water vapor content, and to apply this information to practical use. A Short History The natural gas industry was founded on principles of volumetric calculation. In its infant years, the determination of volume alone was sufficient for all practical purposes. Prior to the advent of even the crudest mechanical calculators or electronic computers many of the standard gas flow calculation factors were abbreviated or ignored. As technology matured, and calculators and computers became more sophisticated, the gas flow calculations became more precise and complete. Even so, the determination of HEATING VALUE was still largely ignored. In 1978, the Natural Gas Policy Act (NGPA) stipulated that producers and pipelines set the price of gas according to the energy content of the gas. This began to get the industry’s attention. It forced many companies to at last deal with the issue. When FERC Order 93 was issued in July, 1980, it brought about a complete change in the natural gas industry. Order 93 mandated that gas measurements must reflect the true energy content of the gas. Rather than assume the Btu to be “wet” or “dry”, water saturation had to be accurately assessed and compensated for. Suddenly, “AS DELIVERED BTU” became the industry’s favorite buzzword. And now, even though Order 93 has since been rescinded, we are still left with the

223

vestiges of its effect in the form of “As Delivered” contracts and settlements of overcharges made during those years of pass-through costs. A Little Chemistry This brings us back to the question of, what is the effect of moisture content on the heating value of natural gas. When fuels are burned, chemical energy is made available as heat. This energy may be used in different ways such as power generation, heating, cooling, and cooking. The greater the heating value of a gas, the less volume will be required to perform a specific task. Each pure gas has a definite and characteristic heating value. Water contributes no heating value; therefore, when water is present in a gaseous mixture as a vapor, it decreases the heating value and thus causes an increase in the volume of gas required to perform a given task. Water is always present in the production of natural gas in both the liquid and vapor phases. If water liquid is present, molecules at the surface tend to escape into the space above. These molecules become vapor and are free to mix with the other constituent gases in the stream. Ultimately, an equilibrium, or steady state of balance is attained when the rate of escape and rate of return into solution are equal. When this condition occurs, the gas is said to be saturated with water vapor. For any gas, this condition of balance or saturation is directly proportional to the temperature (Boyles Law, T1=V1 as T2=V2) . The greater the temperature of the carrier natural gas, the greater the volume of water that can be absorbed into solution. Conversely, as the temperature of a gas begins to drop, the kinetic energy of the fluid decreases. As this happens, the water vapor begins to condense and return to the liquid phase. Water Vapor Determination The key to determining the effect of moisture on heating value is first to determine the AMOUNT of water vapor contained in the natural gas stream. There are two ways to determine the amount of water in a stream. The first is direct measurement through a variety of instruments. Some of these instruments include: Dew Point Method. The Bureau of Mines chilled mirror method determines water vapor content by reducing the temperature of a sample stream until the water condenses on a mirror. Electrolytic Water Vapor Analyzer. The MEECO method uses a Phosphorus Pentoxide (P2O5) cell: Through voltaic electrolysis water vapor pressure is established. Capacitance Hygrometer. This method measures water vapor pressure by the use of an Aluminum Oxide (Al2O3) sensor. The conductance of the sensor varies with the amount of absorption of water vapor. Water Vapor Titrator. This is a wet analysis method based on the amount of time required to neutralize a given amount of Karl Fischer Reagent. Calculations Once an accurate actual water content has been determined, the practical effect that the moisture will have on the heating value can be calculated by the following equation : Btu Reduction Factor = 1−(lbs H2O/MMSCF * 0.00002102) Example 1: During the monthly meter calibration at Acme Well No. 2, the technician determines by test the water content was 37 Ibs H2O/MMSCF. The monthly spot sample determined the Btu factor at 14.73

224

THE EFFECT OF MOISTURE CONTENT ON NATURAL GAS HEATING VALUE

dry to be 1.030. The volume for that month was 16,000 MCF @ 14.73. Acme’s contract dictates that a water content of 7 lbs or less is considered dry. 37−7=30 lbs H2O/MMSCF

If the actual water content of the gas cannot be determined by measurement, it can be readily calculated using the methods outlined in the IGT Research Bulletin 8, “Equilibrium Moisture Content of Natural Gas” by R.F. Bukacek where :

For temperatures between −10°F and 140°F, the values of W are within 1 percent of the IGT tabular values. Note that the values for A and B are given at 14.7 psia and 60°F. If any pressure base other than this is used the values for W must be corrected. Example: During the monthly meter calibration at Acme Well No. 3 the technician determined that the gas was too wet to be tested by instrument. However, he did record that the temperature of the gas was 98° F and the well had 358 psia pressure. The monthly sample showed a Btu factor of 1.010 @ 15.025 dry. The monthly volume was 9,050 MCF @ 15.025. Using the above calculations renders:

In practical application, the calculation of the Btu reduction factor can be facilitated by modern computers and data processing systems by providing for the input of either the measured water vapor content or the inputting of the temperature and pressure values with the subsequent Ibs H2O/ MMSCF calculation. These values can then be picked up by the Gas Accounting Departments to calculate the MMBtu corrected for water vapor at “As Delivered” conditions. The effect of moisture content on heating value content is a serious consideration for all parties interested in utilizing natural gas for whatever purpose. However, being aware of the diminuitive effects of water and

225

more importantly knowing how to quantify the effect of moisture content by using these calculations will leave the user prepared to respond to any contingencies or to employ any methods needed to correct deficiencies. BIBLIOGRAPHY Miller, R.W., Flow Measurement Engineering Handbook, 2–32 to 2–36. New York: McGraw-Hill Book Company, 1983. Northwest Pipeline Corporation, Internal Company memo correlating IGT Bulletin 8 with Least Squares Curve Fit equation. Parent, J.D., Natural Gas—Production and Transmission, 10–23. Chicago, I.G.T. Home Study Course, 1952.

ON-LINE CHRGMATOGRAPHY Jimmy L.Shafer Product Development Engineer UGC Industries Shreveport, Louisiana

ABSTRACT

Rising natural gas prices have lead to an increasing awareness of the BTU content of natural gas. Custody transfer is being based upon the heating value of the gas as well as the volume. Consequently, the on-line gas chromatograpn or BTU analyzer is becoming an increasingly important instrument in the industry. INTRODUCTION Measuring the energy content of natural gas in an online fashion requires a self-contained instrument that is programmable and rugged. Programmabi1ity allows the unit to run stand-alone collecting and analyzing data unattended. Ruggedness allows the unit to be operated on location whether it be outdoors or in an air conditioned control room. The gas chromatograph is not only capable of meeting the needs for on-line applications but also provides several additional features. Technological advancements in both micro-packed columns and microprocessors have allowed the development of very powertul and flexible on-line gas chroma tographs. BASIC CHROMATOGRAPHY The basic chromatograph consists of several valves and micropacked columns configured in such a way that when properly controlled will completely separate the constituents contained in a natural gas sample (See Figure 1). Pure helium flows continuously through the columns. A sample of natural gas is injected into the helium stream and is carried through the columns. The constituents in the natural gas sample have an affinity for the material used in the micropacked columns. The degree of affinity is proportional to the molecular weight of the constituent.

227

Figure 1

The constituents of the natural gas are carried through the columns at different rates, based upon their affinity for the packing material, and emerge or elute from the end of the column at different times. As they elute, the constituents are detected by means of a Thermal Conductivity Detector (TCD). A TCD is basically a thermistor or wire which carries sufficent electrical current to cause self-heating . As helium flows over the detector elements, heat is carried away by conduction. As the concentration of eluted gases changes, more of less heat is carried away resulting in a detectable change in the temperature of the element. If the columns are sufficiently long, the constituents will be completely separated as the gas elutes from the columns. To speed up the process, several different columns are used. The packing material in each column is chosen for a particular capability and proper control of the configuration will maximize the separation while minimizing the overall time. The basic chromatograph has two modes of operation, inject and backflush. Inject is the condition when a sample of gas of known volume is presented to the system. Backflush is the condition when a rearrangement of columns has occurred for maximizing the process. Reterring to Figure 1, solid lines indicate the column arrangement during an inject. A known volume of natural gas sample is presented from the sample loop to the first column. This column is chosen for its ability to hold the heavy constituents (C6, C7, C8), while passing the lighter constituents on to the second column. The second column will initiate the separation process of the lighter constituents during the first minute after inject. At the end of a minute, the controller will rotate the valves to the backflush position as indicated by the dashed lines in Figure 1. This will arrange the columns so that the first column is now ahead of the second and is reversed. This will allow the heavy constituents to be “backflushed” onto the final column. The heavy constituents, also called C6+, will remain unseparated and will be eluted onto the detector as one component. Additional separation will occur in the other constituents throughout the process.

228

ON-LINE GAS CHROMATOGRAPHY

Figure 2

Figure 3

Finally, each constituent is eluted onto the detector for analysis. The time at which elution occurs allows the microprocessor controller to indentify the constituent while the amplitude is used in calculating its MOL %. Figure 2 shows a chromatogram taken from such an arrangement. Calibration of the instrument occurs automatically. The frequency of calibration can be programmed by the user. A certified calibration standard gas is used during the calibration cycle (See Figure 3). Response factors are calculated based upon the area under the peaks taken during the chromatogram. Calibration should be performed at least once every twenty-four hours for best performance. The on-line gas chromatograph can be arranged in any of a number of systems (See Figure 4). As a mocroprocessor-based instrument, communications and programmablity allow configuration into networks.

229

Figure 4

Remote operation can be achieved using modems and telephone lines. Data collection is handled by a hardcopy printout, analog outputs, or a host computer communications link. ON-LINE GAS CHROMATOGRAPHY VS. CALORIMETRY The standard for measuring the BTU content of natural gas has been the calorimeter. A comparison of the calorimeter with an on-line gas chromatograph indicates that the calorimeter is capable of equivalent accuracy to that of the chromatograph, but is much slower to respond. As indicated by Figure 5, the calorimeter exhibits a lag in its response to fast changing BTU values. Often referred to as a “dead-band”, this lag was responsible for an average deviation between instruments of +−1 BTU/SCF. The instruments ability to track the changes in BTU content was an important data point targeted by the comparison. As shown by Figure 5, a relatively slow change in BTU content will be tracked equally well by either the on-line chromatograph or the calorimeter. Faster changing gas was not adequately tracked by the calorimeter. ADVANTAGES AND DISADVANTAGES OF ON-LINE GAS CHROMATOGRAPHY The on-line gas chromatograph has several advantages over the calorimeter. The chromatograph is microprocessor-based allowing very flexible operation and configuration into networks or other measurement systems. When configured with a flow computer, the MOL% of N2 , CO2, and the Specific Gravity can be easily provided for calculation of Fpv in AGA-3. Also, the chromatograph has very low mechanization allowing improved reliability and lower maintenance. An added bonus of the lowered mechanization is the reduced size of the instrument. Since the column and valve assembly are maintained near 100 degrees C, a controlled environment is not necessary thus allowing the instrument to be placed onsite and subjected to the elements. The explosion-proof construction makes hazardous environment use feasible. Finally, some chromatographs are available with columns for analyzing water. This allows the added flexibility of analyzing water content and BTU content with a single instrument (See Figure 6). While

230

ON-LINE GAS CHROMATOGRAPHY

Figure 5

Figure 6

approximately fifteen minutes are required for a gas analysis cycle, only five minutes are needed for water analysis. A major disadvantage that the on-line gas chromatograph has is its inability to do continuous analysis. Analysis occurs every fifteen minutes whereas the calorimeter is a continuous instrument. Second, the chromatograph bases its analysis on theoretical calculations made from the measurement while the calorimeter actually measures the BTU content by burning a sample of the gas. Room for error exists and great care should be taken when specifying calibration gas for use with a gas chromatograph, expecially with respect to the C6+ composition.

231

SUMMARY The on-line gas chromatograph has become a viable tool in the industry for measuring BTU content of natural gas. It compares favorably to the industry standard calorimeter in tracking BTU changes. A very rugged device, the on-line gas chromatograph boasts a repeatability of +−1 BTU/SCF out of 1000.

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE R.M.Batubara and K.Tatang Laboratory Division P.T.Badak NGL Co. Bontang—Kalimantan Timur Indonesia

ABSTRACT

The determination of natural gas composition by gas chromatography using reduced pressure techniques enables the gas analyst, by close monitoring of sample loop pressures, to ensure that very accurately reproducible sample volumes are injected into the analysis system. This in turn, leads to greater precision in the derivation of response factors from pure gases, and the subsequent analyses of reference standard gas and sample gas compositions which is essential if reliability is to be maintained in calculating natural gas heating values. Using these techniques, response factor repeatabilities well below 1% have been achieved. The selection of multi-column analytical system in conjunction with reduced pressure sampling technique has afforded analysts greater resolution and reproducibility well within GPA tolerance limits. This degree of reliability is demonstrated by the results of biennial inter laboratory test programs in which error of heating value from compositional calculation has been reduced to 0.1% . Accuracy in determination of LNG heating values is a major economic incentive, being the most important factor in LNG custody transfer calculation, and in presenting this paper the manner in which highly reliable data are obtained for this purpose is set out in detail. INTRODUCTION Precise gas analysis technique is essential in order to obtain compositional analysis data for the reliable determination of natural gas heating values. Since 1977, laboratories in Indonesia and Japan serving the LNG export and import trade between these countries have co-operated to develope analytical procedures, based on GPA Standard 2261, that provide the required level of reliability. This paper sets out to show that stringent adherence to technique together with carefull selection of equipment and reference materials can achieve the level of reliability required to accurately calculate LNG heating value.

NATURAL GAS CORRELATION TEST RESULT BETWEEN

233

EQUIPMENT FOR GAS CHROMATOGRAPHY The following equipments and operating parameters have been used in Bontang LNG Laboratory to obtain the desired high reliability. Gas Chromatography Any gas chromatography by a reputable manufacturer, fitted with a thermal conductivity detector and suited to multi column operation, may be used. Detector The detector shall be a thermal conductivity type with high sensitivity and stability. The selection of detector temperature depends on sample composition, but temperatures of 10–50 °C above analysis column temperature have been found to provide adequate sensitivity. Regular checks on detector linearity are essential in order to detect signs of filament ageing. The detector should be brought up to temperature and the current switched on several hours before analysis runs to ensure stability. Periodic cleaning, once every six months with suitable non corrosive solvent is recommended to further ensure detector stability. Columns The chosen column for the separation of hydrocarbon components C3 and heavier is 1/8"×5 ft packed with 30% Silicone DC 200 on 80/100 mesh acid washed Chromosorb P. CO2 and C2 are separated on 1/8"×6 ft column packed with 80/100 mesh Porapak Q. An adsorption column, 1/8"×10 ft, packed with 45/60 mesh 13 X Molecular Sieve. These columns are set in a series by-pass flow path with switching valves fitted with flow restrictors adjusted as required to maintain a stable base line. Column Temperature Control. Calibration and analysis runs are iso thermal at 70 °C ± 0.3 °C. Column Conditioning. The columns are conditioned for 24 hours prior to being installed in the chromatograph. This conditioning is done in seperate ovens set aside for this purpose, the DC 200 and Porapak Q columns are conditioned at 225 °C with a carrier flow of 30 ml per minute. The Molecular Sieve column is conditioned at 350 °C with a carrier flow of 50 ml per minute. Recording Instrument A strip chart recorder, or an electronic integrator with print-out facility may be used to display the peaks for the seperated components. Manometer An absolute pressure manometer for sample pressure measurement with vernier that can accurately read to 0.1 mm Hg, measurement of the injected sample. U-tube style manometers are not recommended for this purpose.

234

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Figure 1. SAMPLE INLET MANIFOLD ARRANGEMENT

Vacuum Pump A vacuum pump with a capability of less than 1 mm Hg absolute. Drier A drier consisting of a 4 inch length of 1/2 inch stainless steel tubing, packed with 40–60 mesh Calcium Chloride, with both ends plugged with glass wool is fitted to the sample inlet system. Carrier Gas 99.99% Helium, filtered through a Molecular Sieve moisture trap is metered through two flow controllers to provide a flow rate of ml per minute through analytical and reference columns. Flow restrictors fitted to the switching valves are adjusted to provide a stable baseline with columns in series flow or by-passed. Sample Inlet System The sample inlet system, constructed entirely of stainless steel, is fitted with a 0.5 ml sample loop. This sample size is adequate to ensure linear detector response for all components in the sample. The sample loop and injection ports are maintained at a constant temperature to ensure reproducible sample volumes. To further ensure the accurate measurement of sample volume, the sample loop is connected to a manometer and vacuum system. For manifold arrangement, see figure 1. ANALYSIS DESCRIPTION The basic method for natural gas analysis used is GPA Standard 2261. Components to be determined in a representative natural gas sample are physically separated by gas chromatography and compared to the known composition of a standardized reference gas which is determined under identical operating conditions. To obtain optimum accuracy of compositional analysis, the following steps should be executed :

NATURAL GAS CORRELATION TEST RESULT BETWEEN

235

1. pure gas purity check. 2. determination of component response factors by pure gas and linearity response curve check. 3. determination of reference gas composition. 4. reference gas calibration run. 5. determination of sample gas composition. It is not recommended to use the manufacturer’s composition analysis supplied with the reference gas mixture. Therefore, the reference gas composition has to be determined using the same gas chromatograph as used to analyze the natural gas sample. If significant changes of response factors are subsequently observed, it will be necessary to re-check the linearity of pure gas response factors. If several response factor changes are detected, re-determined of reference—gas composition shall be done. The quantification of calibration and sample analysis can be derived as follows:

Mi RF i A P

= = = = =

concentration of component unnormalized. response factor. component. peak area. sample pressure, mm Hg.

The purity of pure gas requires to be checked prior to use for response factor determination or prior to linearity check of pure component response factors. Purity of pure gas is checked by consecutive analysis runs of each pure gas. The component area divided by the sum of peak areas recorded with each component, will be used as a measure of the purity of component. Using response factors derived from pure gases, determine the composition of the reference gas by making several analysis runs at diffrent partial pressures. The peak areas obtained in these runs are converted to unnormalized mol percent (100 ± 1%) using the above equation. With the same equation, the natural gas sample composition can be determined from peak area of each component recorded following two or more consecutive analysis runs of the sample gas directly after the reference gas calibration runs for obtaining response factors. If the reference gas does not contain C6+, the concentration of hexanes plus in a natural gas sample can be calculated as follows:

It is recommended to monitor the stability of the response factors derived in each calibration run of reference standard gas prior to analyze of sample gas. If response factors vary by more than 1%, recalibration and determination of reference gas composition should be considered . The other dominant factors of analysis, analysis run time sequence and valve switching configuration will be discussed in the next paragraphs.

236

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Analysis run time sequence As previously described analytical columns, consist of Silicone DC 200, Porapak Q and Molecular Sieve 13 X, which are placed in by-pass series flow path of gas chromatograph system. Once carrier gas flow rate is adjusted and oven temperature set up, analysis run time table can be established. Table 1 is typical of run time table which shows that in 1.50 minutes, C3H8 and heavier components will elute from the Silicone column, at 15.85 minutes valve 3 ON position, will trap O2, N2 and CH4 in the Molecular Sieve column, followed by elution of CO2 and C2H6 from Porapak Q column. At 19.85 minutes valve 3 off position, O2, N2 and CH4 will be separated by Molecular Sieve column. Figure 2 is typical of a chromatogram for natural gas separation. Table 1. TYPICAL RUN TIME FOR NATURAL GAS ANALYSIS Run time, minute

Valve position

Description

0.01 0. 10 1.50 7.00 14.53

valve 3 ON valve 1 ON valve 2 ON valve 1 OFF valve 2 OFF

by-pass column 3 sample injection by-pass column 2 back flush C6+ transfer light gas

14.53 15 .85 19 .85

valve 3 OFF valve 3 ON valve 3 OFF

trap O2, N2, CH4 in column 3 . elute O2, N2 , CH4

30.00

stop

Valve switching configuration In order to have a representative and reproducible amount of sample injected into the sample loop, the sample injection system should be leak-tight . Since any part of the sampling system may leak, a through knowledge of the valve switching configuration is essential for trouble shoot ing leaks in the valves and columns system. Fig 3 to Fig 8 shows the sequence of valve switching in steps from injection to elution of light hydrocarbons C3H8 up to C5H12, backflush C6+ , transfer of light gasses, trapping O2, N2, CH4 with elution of CO2, C2H6 and completion of run with elution of O2,N2,CH4. CALIBRATION PROCEDURE AND SAMPLE ANALYSIS A comprehensive calibration procedure is essential in liquefied natural gas manufacturing in order to achieve high accuracy in the determination of heating value. Fig.9 shows a schematic diagram of a calibration procedure that has been implemented by LNG Terminal Laboratories in Indonesia and Japan since 1977. Using this procedure, a high degree of reliability of compositional analysis was demonstrated by performing several correlation test programs between Indonesia and Japan Laboratories.

NATURAL GAS CORRELATION TEST RESULT BETWEEN

237

Figure 2. CHROMATOGRAM OBTAINED FROM INDONESIA BADAK LNG PLANT NATURAL GAS FEED. COLUMN 1: SILICONE DC 200 COLUM 2: PORAPAK Q, COLMN 3: MOLECULAR SIEVE 13 X.

Purity of pure gas This comprehensive calibration procedure includes the purity check of pure gases. This should be checked by actual analysis on the gas chromatograph used. It is not recommended that the manufacturer’s label of pure gas concentration attached on each cylinder be used.

238

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Figure 3, 4 and 5. VALVES SWITCHING CONFIGURATION

NATURAL GAS CORRELATION TEST RESULT BETWEEN

Figure 6, 7, and 8. VALVES SWITCHING CONFIGURATION

239

240

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

The method of determination of pure gas concentration is to calculate the area of each peak by the following equation:

M i Ai

= = =

purity of component. component. peak area of component.

Pure gas purity shall be 99.0 % or higher except for oxygen and nitrogen. Dry air can be used directly for oxygen/nitrogen with the composition of 21.87% for oxygen and 78.08% for nitrogen. Repeatability Analysis results in mol percent shall differ by less than 1% purity as obtained from several consecutive analysis runs. Linearity of response factor of pure components It is strongly recommended that pure components be used to check the linearity of response curves. This work must be very carefully carried out. To establish response factors for the components at interest and in order to observe the linearity of each component, it is necessary to construct response factor curves by injecting varying amounts of pure gases sample into the gas chromatograph. The partial pressure for pure gas should match the expected concentration of the concerned component in the natural gas sample. All components except methane will be injected from 50 to 150 mm Hg with each run using four (4) pressures within that specific range to establish the pure component response curve. A response curve for each component shall be drawn as in Fig. 10. Each response factor of each partial pressure of pure gas run is plotted on the basis that the average response factor is 100%. Linearity check procedure. A pure gas cylinder of known purity is connected into the sample loop inlet. Evacuate the sample outlet manifold including sample loop to less than 1 mm Hg absolute pressure. Close the valve to the vacuum source and carefully open the sample metering valve to allow the gas sample to fill the sample loop up to the desired pressure as indicated by the manometer. Run the analysis by injecting the sample into the gas chromatograph. Then, repeat the same procedure for other partial pressures in recommended range. The response factor of each component can be derived from the measured pressure and peak area obtained from the separation as in the following calculation:

(RF)i M Ai Pi

= = = =

response factor of component. purity of component in Mol.% . peak area of component. measured pressure of component.

A reasonable period for linearity checks using pure components is once every 2–3 months or if any significant changes in the response factor are observed.

NATURAL GAS CORRELATION TEST RESULT BETWEEN

241

Figure 9. SCHEMATIC DIAGRAM OF CALIBRATION PROCEDURE FOR NATURAL GAS COMPONENTIAL ANALYSIS BY GAS CHROMATOGRAPH

Repeatability. In order to achieve a highly reliable compositional analysis, the error of each component shall be less than 1%, derived from the maximum and minimum response factor of the consecutive analysis runs. Table 2 shows an example of the response factor calculation. Determination of Reference Standard Gas Composition Moisture free gas mixtures for reference standard gas of required composition can be purchased or home made. The composition of the reference standard gas shall be close to the expected sample composition. All components in the mixture must be homogenous in the vapour state at the time of use. For this purpose, it is

242

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Figure 10. RESPONSE CURVE OF ETHANE PURE COMPONENT GAS

recommended to have the reference gas cylinder wall warmed up with a heating mantle several hours before the gas is to be analyzed. A resonable period for re-de termination of reference gas composition is once every six months or more frequently if significant changes in response factor are observed. Procedure. The response factor of pure component gas is derived from the peak area obtained on the gas chromatograph with recommended varying injection pressures using vacuum method as previously desribed in pure component linearity response factor. Further, after obtaining pure gas response factor, connect reference standard gas cylinder to gas chromatograph sample inlet line. Evacuate outlet line of cylinder regulator including sample loop to less than 1 mm Hg absolute pressure. Then close the valve to the vacuum source and carefully open the metering valve to allovr sample to fill the sample loop to 600 mm Hg pressure. Repeat in the same way for pressures of 650, 700 mm Hg. Composition of the reference standard gas calculated as in the following equation (see table 3).

M (RF)i P Ai

= = = =

mol percent of component in reference gas before normalized (unnormalized). response factor of pure gas component. sample pressure, mm Hg. peak area of component in reference gas.

Repeatability and Error.The error calculated for each unnormalized component mol % shall be less than 1% or the difference between peak heights for the smallest component within 1 mm. The total composition before normalization must be 100 ± 1%. The technique is to make two or more runs which have a repeatability of 100 ± 1%, then average the un-normalized totals, then normalize the average for the final result. If un-normalized total is not within 100 ± 1%, new response factors need to be re-checked. Determination of sample gas composition The composition of the sample gas is determined by applying response factors derived from a reference standard gas of known composition to the sample gas component peak areas, thus:

NATURAL GAS CORRELATION TEST RESULT BETWEEN

Table 2. CALCULATION OF LINEARITY CHECK OF PURE COMPONENT

243

244

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Table 3. REFERENCE STANDARD GAS CALCULATION

NATURAL GAS CORRELATION TEST RESULT BETWEEN

M (RF)i Ai P

= = = =

245

component mol percent (un-normalized value) component response factor derived from reference standard gas. peak area of component. sample press, mm Hg.

Two or more analysis runs are made for both the reference standard gas and the sample gas. These runs are made consecutively to ensure that the component response factors are derived from reference standard gas analysis runs made under identical conditions to the sample gas runs. There are two steps in determination of sample composition. The first being the derivation of response factors from a reference standard gas of known composition, followed by the analysis of the sample gas. Procedure. Inject a sample of reference standard gas of known composition at absolute pressure of 700 mm Hg. Repeat at least once, and calculate response factors for each component in the reference standard gas. Error shall be less than 1% for each component response factor. In the same operating condition inject the sample gas at absolute pressure of 700 mm Hg, repeat at least once and obtain peak areas for each component. In order to obtain the unnormalized composition of the sample gas, multiply the peak areas of each component by the appropriate response factor derived from the reference standard gas by using the above equation. Error calculated for each unnormalized component mol % shall be less than 1% or for the smallest component the difference between peak heights shall be less than 1 mm. The sum of all unnormalized component mol % shall be 100 ± 1%. The unnormalized component mol % are then averaged and the final sample composition is obtained by normalizing to total of 100%. Attached table 4 is an example of calculation method for natural gas composition analysis. CALCULATION OF HEATING VALUE From the composition obtained in natural gas analysis using high accuracy gas chromatography discribed above, heating value of natural gas sample can be calculated from pure component gross heating value obtained from the physical constant of natural gas published by GPA Standard 2145. For LNG quality control purposes heating value is calculated as BTU/SCF, and for commercial purposes the LNG heating value calculated as BTU/KG. The following is an example of the calculation of heating value of natural gas from the composition. Gross Heating Value Calculation For quality purpose. Express as BTU/SCF as following equation (table 5)

H Xi Hi

= = =

heating value of the gas sample in BTU/SCF. molar fraction. heating value of pure component at 60 °F, BTU/SCF.

246

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Table 4. NATURAL GAS SAMPLE COMPOSITION CALCULATION

Table 5. GROSS HEATING VALUE CALCULATION Component CH4

Molar Fraction (Xi) .

8

8

8

Value @ 60ºF BTU/cu.ft (Hi) 4

1010.0

(Xi×Hi) in BTU/SCF 8

9

7

.

3

NATURAL GAS CORRELATION TEST RESULT BETWEEN

Component C2H6 C3H8 i−C4H10 n−C4H10 i–C5H12 n–C5H12 n–C6H14 N2 O2 CO2 Total

Molar Fraction (Xi)

1

. . . . . . . . . . .

Value @ 60ºF BTU/cu.ft (Hi)

0 0 0 0 0

5 3 0 0 0

9 6 7 8 0

5 1 0 1 5

0

0

0

4

0

0

0

0

247

(Xi×Hi) in BTU/SCF

1769.7 2516.1 3251.9 3262.3 4000.9 4008.9 4755.9 0 0 0 1

1

0 9 2 2

5 0 2 6 2

1

4

4

. . . . . . . . . . .

3 8 8 4 0

6

For commercial purpose. Express as BTU/KG as follows:

H Hi Mi Xi

= = = =

heating value of the gas sample in BTU/KG. heating value of each component at 60 °F, BTU/KG. molecular weight. molar fraction of component.

Table 6. GROSS HEATING VALUE CALCULATION Component Molecular Weight (Mi) CH4 C2H6 C3H8 i−C4H10 n−C4H10 i−C5H12 n−C5H12 N2 O2 CO2 Total

16.043 30.070 44.097 58.123 58.123 72.150 72.150 28.013 31.999 44.040

Molecular Mass Xi×Mi 1

4 1 1 0 0 0 0

1

8

. . . . . . . . .

2 7 7 4 4 0

5 8 8 0 7 3

3 9 9 7 1 6

0

1

1

.

5

9

1

BTU/KG (Hi) (Ri×Hi) BTU/KG

.

. . . . . . . . .

7 0 0 0 0 0

6 9 9 2 2 0

7 6 6 1 5 1

9 4 4 9 3 9

8 0 0 3 8 4

0

0

0

5

9

0

0

0

0

0

52.671 49.236 47.737

4 0 4 4

. . .

4 7 0

5 4 9

0 6 5

46.959 46.392 46.485 0 0

1 0

. . . . .

1 0

9 9

2 0

5 1

.

6

0

0

248

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Table 7. RESPONSE FACTOR OF REFERENCE STANDARD GAS OF LNG ANALYSIS Cargo BZ. 64 #

BX. 58

BZ. 65

BX. 59

BZ. 66

BX. 60

BZ. 01

BZ. 02

BX. 01

BX. BZ. 02.02 03

BX. 03.

BZ. 04

Date

10/ 12’8 5

12/ 12’8 5

13/ 12’8 5

18/ 12’8 5

20/ 12’8 5

23/ 12’8 5

29/ 12’8 5

31/ 12’8 5

3/ 1’86

4/ l’86

5/ 1’86

8/ 1’86

12/ 1’86

. 033 35 . 038 14

. 033 37 . 038 29

. 033 23 . 037 96

. 032 99 . 038 11

. 033 00 . 038 31

. 033 23 . 038 31

. 033 19 . 037 80

. 033 69 . 038 17

. 033 25 . 038 16

. 033 34 . 038 06

. 033 90 . 037 64

. 033 03 . 038 05

. 033 27 . 037 89

. 033 25 . 038 07

.00025

. 024 22 . 018 93 . 016 10 . 014 71 . 012 37 . 012 44

. 024 35 . 018 99 . 016 16 . 014 76 . 012 47 . 012 58

. 024 16 . 018 91 . 016 07 . 014 66 . 012 34 . 012 41

. 024 33 . 018 96 . 016 13 . 014 75 . 012 38 . 012 51

. 024 37 . 019 02 . 016 18 . 014 78 . 012 42 . 012 51

. 024 45 . 019 14 . 016 26 . 014 86 . 012 46 . 012 56

. 023 98 . 018 96 . 016 19 . 014 70 . 012 35 . 012 46

. 024 17 . 018 93 . 016 10 0. 147 0 . 012 37 . 012 48

. 024 15 . 019 04 . 016 25 . 014 73 . 012 32 . 012 39

. 024 08 . 018 84 . 016 02 . 014 61 . 012 23 . 012 39

. 023 88 . 018 67 . 015 90 . 014 53 . 012 28 . 012 39

. 024 15 . 018 98 . 016 20 . 014 69 . 012 29 . 012 38

. 024 00 . 018 74 . 015 92 . 014 52 . 012 15 . 012 23

. 024 18 . 018 92 . 016 11 . 014 69 . 012 35 . 012 44

.00017

O2 N2

CH4

CO2 C2H6

C3H8

iC4H 10

nC4 H10 iC5H 12

nC5 H12

MEA STAN N DARD VAL DEV. UE

.00020

.00014

.00011

.00010

.00008

.00009

NATURAL GAS CORRELATION TEST RESULT BETWEEN

Table 8. GAS CHROMATOGRAPH OPERATING CONDITION IN P.T.BADAK LABORATORY

NATURAL GAS CORRELATION TEST RESULT BETWEEN INDONESIA AND

249

250

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Table 9. TYPE OF ANALYSIS EQUIPMENT AND OPERATEING CONDITION FOR JAPAN—INDONESIA NATURAL GAS CORRELATION TEST PROGRAM

NATURAL GAS CORRELATION TEST RESULT BETWEEN

251

JAPAN LABORATORIES Table 10. 1979 CORRELATION mol % COMP ONEN T

CHITA CHUB U ELEC TRIC

max −min2

2/1 × 100%

O2 N2 CH4

− 0.047 88. 464 6.937

HIMEJ I KANS AI ELEC TRIC

TOBA TA KITA KYUS HU LNG ELEC TRIC

0.00 0.05 88.23

0.00 0.05 88.29

− 0.05 88.34

3.347 0.535 0.510

7.01 − 3.39 0.57 0.56

6.96 − 3.43 0.56 0.53

iC5H12

0.079

0.09

nC5H1

0.081 100. 00 1140. 9

C2H6 CO2 C3H8 iC4H10 nC4H1

CHIT A TOHO GAS

SENB OKU OSAK A GAS

HIMEJ I OSAK A GAS

P.T.A RUN

P.T.B AVER Error ADAK AGE 1

TOLE RANC E ALLO WED BY GPA STAN DARD

0.034 88. 184 7.041

0.049 88. 258 6.993

0.028 57.1 0.427 0.484

0.03 0.30

6.91 − 3.39 0.57 0.56

0.062 88. 037 7.098 0.004 3.462 0.572 0.554

0.188

2.688

0.10

3.449 0.549 0.545

3.411 0.559 0.543

0.115 0.037 0.50

3.371 6.619 9.208

0.05 0.03 0.03

0.09

0.09

0.109

0.099

0.095

0.028

0.03

0.10

0.09

0.09

0.109

0.099

0.095

0.028

29. 474 29. 474

100. 00 1143. 9

100. 00 1143. 8

100. 00 1147. 1

100. 00 1145. 8

100. 00 1144. 3

100. 00 6.2

6.2

0

2

Total HHV

0.03

0.5

Table 11. 1981 CORRELATION TEST mol % COMP ONEN T

CHITA CHUB U ELECT RIC

max −min2

2/1 × 100%

O2 N2

0.02

CHIT A TOHO GAS

SENB OKU OSAK A GAS

0.02

HIMEJ I OSAK A GAS

HIMEJ I KANS AI ELECT RIC

TOBA P.T.AR P.T.BA AVER Error TA UN DAK AGE 1 KITAK YUSH U LNG ELECT RIC

TOLER ANCE ALLO WED BY GPA STAN DARD

0.02

0.02

0.03

0.02

0.02

0.02

0

252

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

mol % COMP ONEN T

CHITA CHUB U ELECT RIC

max −min2

2/1 × 100%

CH4 C2H6 CO2 C3H8 iC4H10 nC4H10 iC5H12 nC5H12 Total HHV

86.57 8.41 3.69 0.62 0.63 0.05 0.01 100.00 1159.2

CHIT A TOHO GAS

SENB OKU OSAK A GAS

HIMEJ I OSAK A GAS

HIMEJ I KANS AI ELECT RIC

TOBA P.T.AR P.T.BA AVER Error TA UN DAK AGE 1 KITAK YUSH U LNG ELECT RIC

TOLER ANCE ALLO WED BY GPA STAN DARD

86.53 8.32

86.43 8.29

86.41 8.39

86.55 8.29

86.34 8.40

86.47 8.35

0.23 0.27 0.12 1.44

0.30 0.10

3.80 0.63 0.64 0.05 0.01 100.00 1160.6

3.86 0.69 0.65 0.05 0.01 100.00 1162.7

3.86 0.63 0.64 0.04 0.01 100.00 1161.7

3.82 0.63 0.63 0.05 0.01 100.00 1162.9

3.88 0.65 0.65 0.05 0.01 100.00 1161.9

3.82 0.64 0.64 0.05 0.01 100.00 1161.3

0.19 4.97 0.07 10.94 0.02 3/13 0.01 20.00 0

0.05 0.03 0.03 0.03 0.03

3.7

0.3

Table 12. 1983 CORRELATION TEST INDONESIA AND JAPAN LABORATORIES mol % CHTT CHITA SENB COMP TA TOHO OKU ONENT CHUB GAS OSAK U A GAS ELECT RIC

max −min 2 O2 N2 CH4 C2H6 CO2 C3H8 iC4H10 nC4H10 iC5H12

HIMEJ I OSAK A GAS

HIMEJ I KANS AI ELECT RIC

TOBA P.T.AR P.T.BA AVER TA UN DAK AGE 1 KITAK YUSH U LNG

Error

TOLER ANCE ALLO WED BY CPA STAN DARD

2/ 1×100 % 0.00 88.03 6.85

0.00 87.88 6.93

0.01 87.96 6.90

0.00 87.95 6.89

0.00 87.80 6.96

0.01 87.97 6.88

0.00 87.94 6.90

0.01 0.23 0.11

− 0.26 1.59

0.03 0.30 0.10

3.60 0.68 0.79 0.05

3.64 0.69 0.81 0.05

3.58 0.70 0.80 0.05

3.63 0.68 0.80 0.05

3.67 0.70 0.82 0.05

3.61 0.69 0.79 0.05

3.62 0.69 0.80 0.05

0.09 0.03 0.03 0.00

2.49 4.35 3.75 0.00

0.05 0.03 0.03 0.03

NATURAL GAS CORRELATION TEST RESULT BETWEEN

mol % CHTT CHITA SENB COMP TA TOHO OKU ONENT CHUB GAS OSAK U A GAS ELECT RIC

max −min 2

2/ 1×100 %

nC5H12 Total HHV

0.00 100.0 1150.8

HIMEJ I OSAK A GAS

HIMEJ I KANS AI ELECT RIC

TOBA P.T.AR P.T.BA AVER TA UN DAK AGE 1 KITAK YUSH U LNG

0.00 100.0 1151.5

0.00 100.0 1151.7

HIMEJ I OSAK A GAS

HIMEJ I KANS AI ELECT RIC

TOBA TA KITA KYUS HU LNG

P.T.AR P.T.BA AVER Error UN DAK AGE 1

0.00 100.0 1152.6

0.00 100.0 1153.9

0.00 100.0 1151.2

Error

253

TOLER ANCE ALLO WED BY CPA STAN DARD

0.03 100.0 1151.9

3.1

0.3

Table 13. 1985 CORRELATION TEST mol % COMP ONEN T

CHTT CHITA SENB TA TOHO OKU CHUB GAS OSAK U A GAS ELEC TRIC

max −min 2

2/ 1×100 %

O2 N2 CH4 C2H6 CO2 C3H8 iC4H10 nC4H10 iC5H12 n C5H12 Total HHV

TOLE RANC E ALLO WED BY CPA STAN DARD

0.02 85.68 8.87

0.01 85.70 8.93

0.01 85.65 8.90

0.01 85.68 8.89

0.01 85.74 8.82

0.01 85.75 8.84

0.02 85.66 8.89

0.02 85.72 8.84

0.01 85.71 8.87

0.01 0.10 0.11

100.0 0.03 0.12 0.30 1.24 0.10

4.24 0.60 0.57 0.02

4.21 0.60 0.54 0.01

4.24 0.61 0.57 0.02

4.23 0.60 0.57 0.02

4.25 0.60 0.56 0.02

4.21 0.60 0.57 0.02

4.26 0.59 0.56 0.02

4.23 0.61 0.56 0.02

4.23 0.60 0.56 0.02

0.05 0.02 0.03 0. 0I

1,18 3.33 5.36 50

100.0 168.0

100.0 1167.0

100.0 1168.5

100.0 1168.0

100.0 1167.6

100.0 1167.3

100.0 1168.0

100.0 1167.5

100.0 1167.7

1.5

0.1

0.05 0.03 0.03 0.03

254

RELIABILITY OF COMPONENT ANALYSIS BY GAS CHROMATOGRAPH FOR CALCULATION OF HEATING VALUE

Table 14. INTERLABORATORY NATURAL GAS CORRELATION TEST REPRODUCIBILITY

RESULTS AND DISCUSSION The stability of reference standard gas response factor can be maintained at high level of repeatability since the operating parameter of the gas chromatograph, sample injection technique, and room condition are held constant. Table 7 shows the response factor stability of reference standard gas used in natural gas sample analysis for aperiod of month. Table 8 and 9 show the gas chromatograph operating conditions used for analysis natural gas sample in PT. Badak and several laboratories. The complete elution time for all peaks can be shortened by 5 minutes, if the backflush step is omitted as in the analysis of commercial LNG without C6+. The participating laboratories, which are Indonesia LNG manufacturers and Japan LNG buyers have set up abiennial correlation test program. High reliable results have been achieved well within GPA Standard reproducibility tolerance. The correlation test is carried out using the same method and procedure as described. All laboratories using the same sample and each laboratory using its own reference standard gas. The analysis conducted during a similar time period. Table 10, 11, 12 and 13 shows the correlation test result between several laboratories during the year of 1979, 1981, 1983 and 1985. The tabulated results in table 14 shows the components error of four correlation test program being reduced, and heating value overall error being reduced from 0.5% to 0,1% or 6,2 BTU/SCF to 1.5 BTU/SCF. ACKNOWLEGMENT The authors appreciate the support and encouragement of Sukardji Wibowo, Hadi Daryono and Join Management Group of Japan—Indonesia LNG Project for providing LNG Buyer & Seller inter laboratories correlation test data. REFERENCES 1.

Annual book of ASTM standards, D 1945–81, “Analysis of natural gas by Gas Chromatography”. American Society for Testing and Materials Philladelphia, USA.

NATURAL GAS CORRELATION TEST RESULT BETWEEN

2. 3. 4. 5. 6. 7. 8.

255

GPA publication 2145–85, standard table of physical constants of Paraffin Hydrocarbon and other components of Natural Gas. GPA standard 2261–72, method of analysis for Natural Gas and Similar gaseous mixtures by Gas Chromatography. Gas Processors Association Tulsa Oklahoma. GPA standard 2172–76, method for calculation of gross heating value, specific gravity, and compressibility of Natural Gas mixtures from compositional analysis. Gas Process Association Tulsa Oklahoma. Hewlett Packard Gas Chromatograph Instrument Manual Vol.1–5,HP 5880 A, April 1984. Hewlett Packard 5880 A Gas Chromatograph Data Sheet “Natural Gas Configuration”, 1979, August. Indonesia—Japan LNG Project analysis sampling system at each terminal, data sheet , Japan—Indonesia LNG Co. LTD. 1985, August. Ludwig Huber, “Rapid and Simple Analysis of Natural Gas by G.C.” J. Chrom. Sc.-,Vol. 21, 1983, November.

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT Thomas A.Clark Production Manager ENERCO, INC. Honolulu, Hawaii 96842

ABSTRACT

GASCO, INC., formerly the HONOLULU GAS COMPANY, is unique in that its entire gas distribution system is supplied by manufactured gas. GASCO bills its customers by the THERM, which normally requires two measurements: a volumetric measurement and a heating value measurement. Due to variations in the gas composition and heating values resulting from the manufacturing process, determination of “Therms” becomes more complex. In order to achieve accuracy, the SNG measurement system has grown to include the following components: 1. Recording Gas Calorimeter 2. Turbine Meters 3. Orifice Meters 4. Mass Flow Meters 5. Flow Computers 6. Gas Chromatographs 7. Computers Information from the first six components are fed daily into the computer, which uses the information to calculate the quantity of gas produced. The calculated production is then compared to the “THEORETICAL PRODUCTION,” computed by using a mass balance equation for the plant as a check. To complete the measurement data, the degree of saturation with water vapor must be determined as GASCO is permitted to sell its gas saturated at the temperature at the customer’s meter. Thus, in order for GASCO to measure its gas production, it requires a complex system of measurements, analyses, and calculations.

257

BACKGROUND The Hawaiian Islands are geologically quite young, thus there are no fossil fuels available. This unique fact of nature has put Hawaii ahead of the rest of the world by subjecting it to numerous ENERGY CRISES throughout time. When Captain Cook first landed in Hawaii in 1778, Hawaiians were using RENEWABLE SOURCES OF ENERGY, firewood and kukui nut oil. As the population grew, the supply of these items became insufficient and the first ENERGY CRISIS was faced. As the whaling fleets came to Hawaii, people began using whale oil. Again, as the world demand for this energy form increased, the price rose from 43¢ per gallon in 1823 to $2.55 per gallon in 1866; thus, Hawaii faced another ENERGY CRISIS. With the discovery of petroleum around 1860, the world as well as Hawaii began their dependence on fossil fuels. In 1904, the Honolulu Gas Company began building its first oil-fired gas plant, which further emphasized Hawaii’s dependence on imported fossil fuel. Due to Hawaii’s unique situation, all forms of energy have been expensive. The major escalation in the price of natural gas and other related forms of energy in the 1970’s which generated the massive interest in the precision and accuracy of measurement in the gas industry has been a fact of life for us for decades, and the demands for accurate metering have always been foremost in the policies of The Gas Company. How “Accurate” is “Accurate”? The Honolulu Gas Company gradually began increasing the BTU of its manufactured gas in the early 1950’s, as mainland gas utilities were switching from manufactured to natural gas. The BTU of the gas increased from 500 BTU/cf in 1950 to 900 in 1965. During this progression, fluctuations in BTU and specific gravity were normal. However, since GASCO billed its customers by the therm and controlled gas quality by using the Wobbe Index, its customers were generally unaware of the changes. How accurate were our measurements? By today’s standards—subject to question. To determine the heating value of the gas, a Cutler-Hammer Calorimeter was used. While the accuracy of this instrument is well accepted, the calibration gas was a natural gas that was diluted with inerts before certification to more closely approach the actual BTU and specific gravity of the manufactured gas stream. The accuracy of the technique used in preparing this gas is unknown; however, it was all that was available. The gas chromatographic techniques and hardware available at the time, plus the twenty plus identifiable components in the ever changing gas stream, made chromatographic determination unreli able. As for the volumetric measure, an atmospheric pressure positive displacement meter was used (Figure 1). After the measurement the gas was compressed, cooled (to remove water and light ends) and scrubbed to remove aromatic hydrocarbons. Calibration of this meter was accomplished by using a “calibrated” watersealed gas holder (Figure 2). The temperature was assumed to be constant and the pressure was recorded from a manometer. The gas holder was scientifically calibrated into the number of lifts, sheets, and rivets of gas that were available. How “accurate” is “accurate”? Well, it was accurate enough to satisfy our Public Utilities Commission and our customers, as well as provide an acceptable rate of return to the Company for most of its 70 years. INTRODUCTION In 1974 GASCO, Inc., formerly the Honolulu Gas Company, built and put into operation a Lurgi Process Synthetic Natural Gas Plant. In a single step, GASCO made the transition from the Stone Age to the Twentieth Century without having the benefit of the growth stages in between. When the dust had settled

258

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT

Figure 1.

and our production reports were compiled for the first full month of operation, the corporate profits were disastrous while the “unaccounted for gas” soared. Questions from upper management began to pour in on the accuracy of our meters. In addition, the SNG Plant was built in a Foreign Trade Zone adjacent to our sister refinery, so we had to not only satisfy Company personnel on the accuracy of our meters but also the U.S. Customs and Foreign Trade Zone officials. The quest for both “accurate” and “precise” measurement was about to begin. What follows is how we developed what we felt is “within reason” an accurate and precise metering system. There is no attempt to make a judgment as to the accuracy of the standards, or the deficiencies in the technology used, only how using off-the-shelf items a system can be built to satisfy your metering needs. SYSTEM DEVELOPMENT The interest in our metering accuracy can better be understood if you look at our gas rates. In 1974, they were $7 per million BTU, or 70¢ per therm, at the residential level and $5.30 per million BTU, or 53¢ per therm, at the commercial level. In addition, the SNG Plant was incorporated as a separate entity, selling all of its output to GASCO for distribution to its customers; thus, “unaccounted for gas” became a GASCO problem but was very much dependent on the accuracy of our meter. It seemed easier at the time to say “The SNG plant’s meter was inaccurate” than to admit there were leaks in the distribution system. In the beginning, our system was fairly simple (or so we thought). SNG (see Figure 3 for analysis) was produced at a constant rate with a consistent quality and enriched with LPG (utilizing Liquified Petroleum Gas C-2—C-5). The mixture was measured with a gas turbine meter and sent to the transmission line for GASCO (Figure 4). THE PROBLEMS As the system was examined for possible errors, many actual and potential problems were uncovered.

259

Figure 2.

1. The turbine meter had a mechanical correction system for temperature and pressure compensation. This system was subject to excessive wear and resultant inaccuracies. 2. The turbine blades in the meter cartridge had a short life expectancy. 3. The turbine bearings had a relatively short life. 4. The pressure at the meter was not constant. 5. Supercompressibi1ity was not fully understood. It was determined that SNG is significantly different from natural gas due to the hydrogen and carbon dioxide content , as well as the higher hydrocarbons, and that using data developed for natural gas was inaccurate. The availability of Enrichment LPG was dependent on the type of crude oil available to the refinery and, thus, the enrichment quality would vary, resulting in the need to vary the CO-2 content to maintain a constant Wobbe Index (Note: SNG, as produced, contains approximately 20% carbon dioxide, which is removed in a hot potassium carbonate system. Part of the CO-2 is left in the gas to control the Wobbe Index.) Thus, the change in composition changed the Supercompressibi1ity. 6. The composition of the LPG varied also with the type of crude oil. Thus, when the LPG was high in C-4’s and C-5’s, there was a question as to whether the LPG was fully vaporized when flashed into the SNG line. If the LPG was not fully vaporized, what effect did the droplets have on the turbine meter accuracy, blades, and bearings? 7. Factory calibration of the turbine meter was not reliable and no local means of calibration was available. 8. Using the orifice meters for a check was unreliable because of changing gas composition due to variable enrichment which changed the gas density. 9. If the LPG was not being fully vaporized, was the the gas stream going to the Calorimeter representative of the gas being sold? 10. Was the Dew Point being recorded a true Dew Point, or a hydrocarbon Dew Point?

260

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT

Major Components Hydrogen….. Methane….. Carbon Dioxide….. Enrich ment Gas..... Average BTU ..... Average Specific Gravity.....

10.0% 77.0% 5.0% 8.0% 1 ,050 0.675

± 2.0% ± 3.0% ± 2.0% ± 1.5% ± 10 ± .015

Figure 3. SNG Analysis

Figure 4. Original SNG Metering Schematic

11. Was our 1960’s vintage gas chromatograph giving us the degree of accuracy we needed in our gas analyses? 12. Was the meter really a problem? The primary reasons for suspicion were: the calculated plant efficiency (with no operating history, this figure was questionable) and the calculated UNACCOUNTED FOR GAS, the accuracy of which was also questionable. THE SOLUTIONS Step 1. As previously mentioned, GASCO sells its gas by the therm which requires a minimum of two measurements, a volumetric measurement and a heating value measurement. The first and simplest thing to check was the accuracy of the Calorimeter, a relatively simple matter, and then to determine if the Calorimeter was reading a representative gas sample regardless of the injection rate of enriching LPG. A test was run using the SNG Plant Calorimeter and the Calorimeter at the old manufactured gas plant approximately 22 miles away. The results showed that the BTU measurements were identical when the time delay was accounted for. We thus assumed the SNG Plant Calorimeter was measuring a representative sample. The next step was to resolve the problems, or potential problems, identified earlier. The first logical step was to put a more sophisticated meter in series with the first meter. The new meter had its own flow computer and electronic compensation for temperature and pressure. Automatic compensation for changing supercompressibility was not incorporated because it was known that SNG could not be made to fit the correction curves designed for natural gas. Results: With the plant operating at constant flow rate, temperature, and pressure, using two new factory calibrated cartridges, the meters read 3% apart! I will not go into attempts to resolve these discrepancies with the manufacturers .

261

The next step was to develop a new standard of comparison for judging meter accuracy. The SNG Plant has a theoretical mass balance equation that reduces to the following: eq. (1) Since the plant feed analysis is usually given in weight fractions (i.e., weight percent Carbon and weight percent Hydrogen) all we needed to know was the weight of the feed plus its C/H ratio and an analysis of the gas produced to calculate the plant output. The complete plant mass balance was set up on the computer. A test was run under controlled conditions, inputting all of the required data and the meter checked to within 0.4% of the mass balance. We now had a method of determining the accuracy of our meter under controlled conditions; however, the C/H ratio of the feed had to be determined in a mainland laboratory and the gas analysis using our old GC’s was very time consuming. We still had to determine the cause of the mechanical problems with the meter. A thorough analysis of the problem led to a simple solution that also eventually solved some of the other problems. A vaporizer was added to the enrichment LPG and the injection point was moved down stream of the meter. This addition solved not only the mechanical problems with the meters but also reduced the variation in gas composition seen by the meter. The Dew Point was now read before the addition of enrichment, eliminating the question of hydrocarbon Dew Points. Using this method, however, we became dependent on a liquid turbine meter to determine the quantity of the liquid LPG and a chromatographic analysis to determine the composition from which other physical properties could be calculated. We now had the beginnings of a system we could feel comfortable with; however, the system was far from acceptable. Our goal was to develop a system that was both accurate and precise, with the accuracy and precision verifiable on a daily basis. To accomplish this, we would need the following: 1. A new computer 2. Two on-line computer- or micro-processor-controlled gas chromatographs 3. An automatic sampling system 4. Data handling system for GC information 5. Software 6. Additional meters on the liquid streams 7. Automatic system checks 8. Additional information; i.e., specific gravity determi nati ons A team consisting of design engineers, chemists, computer programmers, and instrument engineers was put together. They designed and built the following system: The plant mass balance was included in a daily report. This compares the theoretical plant output to the metered plant output (Figure 5). For this to be meaningful, however, all of the inputs had to be accurate and timely. To accomplish this, two new automated Perkin-El mer Gas Chromatographs were purchased and installed, one for gas streams and one for liquid streams. The manpower and time required to generate all of the required information, however, was prohibitive for a small lab staff and while the GC’s had their own microprocessor, it was insufficient to handle all of the data. We, therefore, purchased an additional IBM PC to operate the system which also included an automatic sampling and sample conditioning system. With this system in place, we could now not only automatically analyze all of the plant input and output streams (Figure 6) but also use a daily calculated supercompressibi1ity factor to adjust the average factor used in the

262

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT

SNG meter flow computer. The overall plant efficiency is also calculated which is another check on both the process and the metering (Figure 7). To eliminate questions of accuracy of liquid meter data and specific gravity data (used to generate the inputs for the mass balance) liquid mass meters were installed as a check. The problem is, however, when you have two meters in service and they generate conflicting data, which is correct? To solve this problem, integrators and totalizers were added to the orifice meters as a check on the other meters. When at least two out of the three read the same, we assume they are correct. This limits the amount of time we have to spend calibrating meters. And, finally, the specific gravities of the gas streams were determined by Ranarex Gravitometers and the liquid streams by a hydrometer. This data is then compared with the analytical data generated by the laboratory. By comparing the calculated BTU and specific gravity values to the measured values, we can immediately spot errors in either measurements. Since all of the above information is available in the computer, we print it out on a daily basis on a Quality Control Report. This report is then reviewed by the Instrument Engineer (in charge of metering), the Plant Chemist (in charge of analyses) and the Technical Coordinator (in charge of software) to resolve whether discrepancies are a result of measurement errors, analytical error, or data input error (Figure 8). A schematic of the present system (Figure 9) represents an investment of over $500,000 and two- to three-man years of development. All of the components of the system are calibrated on a regular basis to ensure accuracy. The plant metering system computes the total number of “therms” of gas saturated at 60°F and 30 inches of mercury sold to GASCO, INC., our distribution subsidiary. GASCO, however, is permitted to sell gas saturated with water at the temperature at the customer’s meter. Since the actual SUBSATURATION CORRECTION FACTOR FOR JANUARY 1986 The water content of the street gas leaving the plant during the month of December as determined by 22 tests averaged 0.0252 volume percent by the Dew Point Method. The average gas temperature at the customers’ meters as recorded for the same period was 77.17°F. If the gas were saturated at this metered temperature and atmospheric pressure, it would contain 3.10 percent water vapor. The Correction Factor is: gas at the customer’s meter is dry, a “Subsaturation Correction Factor” is generated by using the actual water content of the gas and the actual temperature at the customer’s meter (Figure 10). This factor is then used by the Billing Department to adjust the customer’s bill. The generation of the subsaturation correction factor is the final step in the metering process. By many standards, we may be considered a small utility; however, when both plant inputs and outputs are considered, we meter approximately $40 million worth of products annually. Thus, the accuracy of our metering has the attention of the highest level of management. CONCLUSION What I have presented is the solution to a unique metering problem, where in order to measure an ENRICHED SYNTHETIC NATURAL GAS stream, requires (in addition to a volumetric and heating value measurement) two gas chromatographic analyses to calculate the supercompressibi1ity of the raw SNG and the heating value of the enriching LPG, one liquid volumetric analysis, and a specific gravity measurement

263

Figure 5. Weight Balance Report

along with a computer analysis to provide the final measurement. To maintain precision and accuracy in the above measurement system, requires five additional gas chroma tographic analyses, up to nine additional

264

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT

Figure 6. Stream G/C Analysis Report

liquid and one gas volumetric measurement, two additional heating value measurements, two manual and two automatic specific gravity measurements, and a computer to analyze the data. Whether one considers metering an art, or a science, there appears to be in recent years a new commitment on the part of IGT, GRI, and others as well as numerous vendors to provide the gas industry

265

Figure 7. Thermal Efficiency Report

with the tools necessary to solve the most complex metering problems. We will continue to look to these people for improvements in both the precision and accuracy of metering in the future.

266

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT

Figure 8. Quality Control Report

267

Figure 9. SNG Metering Schematic

268

ENRICHED SYNTHETIC NATURAL GAS MEASUREMENT

Figure 10. Moisture Content of Street Gas

AUTOMATED GAS ANALYSIS SYSTEM Harvey L.Humbke Laboratory Technologist NOVA, AN ALBERTA CORPORATION Edmonton, Alberta, Canada

ABSTRACT

This paper discusses the development and implementation of an automated gas analysis system. The major advantage of the system is its capability to process twice as many gas samples as before. Also, improvements were noted in accuracy and repeatability as compared to the results derived from 3 devices (calorimeter, gravitometer, and gas chromatograph). Substantial savings in the order of $170,000.00 per year were realized in the areas of manpower and operating cost reductions as a result of implementation of the new system. INTRODUCTION NOVA, AN ALBERTA CORPORATION owns and operates a natural gas transmission system in the Province of Alberta and handles more than 75 percent of the Canadian gas sold in North America. As of December 31, 1985, NOVA’s Alberta Gas Transmission Division recorded receipts of 2.25 trillion cubic feet with an average of 6.2 billion cubic feet being moved per day. NOVA’s system consists of 756 receipt and major delivery metering points connected by 8,559 miles of pipeline. Prior to December 1983, NOVA operated two quality control laboratories, with each laboratory analyzing approximately 600 gas samples per month. All 600 samples were analyzed on the calorimeter and gravitometer with only a portion being analyzed on the gas chromatographs to determine hydrocarbon and inert components. This was a very time consuming function as the samples had to be run on the calorimeter for 45 minutes in order to get reliable results. Samples analyzed on the gas chromatographs had to be manually injected thus limiting the number analyzed per 8 hour working day. The results obtained from the calorimeter, gravitometer, and gas chromatographs were used for custody transfer billing calculations.

270

AUTOMATED GAS ANALYSIS SYSTEM

A concept using calculated gravity and heating value from the gas analysis for custody transfer billing was presented to Consumer and Corporate Affairs (Canada’s regulatory agency) for approval in 1979. NOVA CORPORATION began development and in late 1983, implemented an automated gas analysis system capable of analyzing all samples on the gas chromatographs. This innovative concept has resulted in many tangible benefits and savings. The two laboratories were amalgamated into one centrally located facility which now analyzes 1450 samples per month. CONFIGURATION The present configuration used by NOVA consists of two stream selector valves to which thirty-one gas samples are connected. These samples are regulated to approximately 210 kPa (30 psig) and are systematically purged through either of two computer controlled 16 port stream selector valves. The stream selector valves are air actuated and are controlled by two three-way solenoids. Each sample is further regulated to 105 kPa (15 psig) before it passes through a three-way solenoid controlled valve. This solenoid valve has the capability of directing the gas to the differential pressure switch or to an outside vent. When the three-way solenoid valve is turned to vent, two things occur: 1. The pressure in the differential pressure switch begins to drop and, 2. the gas sampling valve advances to the next sample. The differential pressure switch is set so that when atmospheric pressure is sensed, it completes the circuit connecting it to the remote start on the Analog to Digital Converter (A/D) and injects the gas sample. The next component is a two-way solenoid controlled valve that is used in controlling the purging time of low pressure samples and also for shutting down the system on completion of the sample rotation. (See automated sampling configuration diagram). CAPABILITY This system is capable of analyzing 57 samples per standard work day compared to a maximum of 25 using the manual injection technique. At the beginning of each day a technician runs a calibration standard to verify the accuracy of the gas chromatograph. At the same time, the automated system is loaded with 26 gas samples which is the maximum number that can be analyzed by the end of the regular working day. Before the end of the regular working day, the automated system is again loaded with a new batch of 31 gas samples. During the day the technician processes the previous batch and that day’s analysis information while the gas chromatographs are automatically analyzing a new batch of samples. This system could be further expanded by the addition of another 16 port stream selector valve which would further increase its capability to 84 samples per 24 hour period. BENEFITS Repeatability Discussion of Results. The repeatability of heating values derived using a gas chromatograph has resulted in a lowering of the standard deviation experienced by over 60% when compared to the calorimeter’s heating value.

271

Figure 1. AUTOMATED SAMPLING CONFIGURATION

Tests conducted over an extended period of time (four months) have demonstrated that the standard deviation in the calculations for heating value derived from compositional analysis of a primary standard

272

AUTOMATED GAS ANALYSIS SYSTEM

was 0.0109 MJ/m3 (0.3 BTU/cf). Certified Standard Gas heating value: Repeatability of the four month test was:

For this same time period the heating value derived from the laboratory calorimeter had a standard deviation of 0.0285 MJ/m3 (0.8 BTU/cf). Standard Gas heating value: Repeatability for this test was:

This calorimeter was recalibrated whenever it was off by more than ±0.06 MJ/m3 (±1.6 BTU/cf) from the certified value of the Standard Gas. Accuracy Calculations of total energy for orifice metering facilities using the results from two devices (gravitometer and calorimeter) can result in a worst case error of approximately ±0.4%. However, the principle of the gas chromatograph operation places a direct correlation between relative density and heating value, thus reducing this error to ±0.01%. It is prudent to use a well maintained gas chromatograph system as a single source for these values since the total energy measured is multipled by some monetary value per GJ (MMBTU). (See Appendix A). Carbon Dioxide (CO2) Absorption in Water Carbon dioxide is absorbed by the water in the calorimeter, depending on the amount of CO2 which results in a heating value error. The lost CO2 volume is replaced by more of the natural gas sample in order to maintain the fixed volume for which the calorimeter is calibrated. The CO2 component has no adverse effects on the gas chromatograph analysis. Effects of Helium on the Calculation of Heating Value and Relative Density Discussion of Results Accurate and reliable results from one device rather than three devices for custody transfer can be achieved only by analyzing for helium. There are a few locations which must be identified where the concentrations of helium will affect the heating value of the gas analyzed. It may be prudent to do a “one time” analysis for helium in either the volumes entering or leaving the pipeline. Any helium present in a natural gas sample cannot be analyzed when helium is used as a carrier gas in the chromatograph. NOVA CORPORATION has overcome this problem of erroneous heating values through

273

the use of another gas chromatograph that analyzes for helium using argon as a carrier gas. (See appendix B). Timeliness of Results Gas Chromatograph analysis results are available in 17 minutes after injection compared to 45 minutes for the calorimeter. QUALITY CONTROL CHECK PROCEDURES Quality control is very significant since we are relying on a single instrument for very important results. The following procedures were implemented to ensure accurate reliable results: 1. Daily analysis of a secondary standard and agreement of results within the limits of repeatability stated in ASTM D1945. 2. Daily plottings of values for relative density, heating value, and some of the mole fraction components. 3. Verification using a Primary Standard whenever the secondary standard varies by more than 0.04 MJ/m (1.1 BTU/cf). Recalibration occurs when the heating value is greater than 0.03 MJ/m3. (0.8 BTU/cf). 4. Monitoring of gas samples from new stations on a gravitometer and calorimeter to ensure there is no component present (in any appreciable amount) for which the gas chromatograph is not calibrated. 5. Monitor stations with hexanes plus concentrations greater than 0.0014 mole fraction. Reason: If this amount of hexanes were actually octanes, the calculated heating value would be lower than the actual value by 0.08 MJ/m3 (2.1 BTU/cf) ECONOMICS NOVA CORPORATION has realized substantial savings in operating costs as a result of the automated gas chromatograph. The cost of materials for a 16 sample automated system was approximately $5,000.00 and $10,000.00 for a 31 sample automated system. The need to expand or upgrade an air conditioned calorimeter room to meet the increasing number of samples analyzed was eliminated. Also, the alternative of purchasing another device capable of measuring heating value was no longer required. A reduction of approximately $8, 000.00 per year in calorimeter maintenance costs resulted with no appreciable increase in the gas chromatograph maintenance costs. Substantial savings were realized as NOVA was able to combine the operation of two laboratories. A net decrease in manpower costs of approximately $170,000.00 per year was realized. As well, we were able to reduce the operating expenses of maintaining two laboratories with duplicate equipment. CONCLUSION The main advantages of an automated gas chromatograph system are: 1. 2. 3. 4.

Increased capacity. Increase in repeatability over previous conventional methods used. Increase in accuracy. Identification of some potential problem areas that can affect total heat energy.

274

AUTOMATED GAS ANALYSIS SYSTEM

5. Reduction in operating costs. APPENDIX A Example Actual values for the Gas Standard were Relative Density (R.D.)—0.560 Heating Value—37.80 MJ/m3 (997.7 BTU/cf) Results from the various instrumentation are acceptable with the following tolerances: Gas Chromatograph Relative Density—0.560 ± 0.001 Heating value 37.80 MJ/m3 ±0.02 (997.7 BTU/cf ± 0.8) Calorimeter Heating value—37.80 MJ/m3 ±0.08 (997.7 BTU/cf ± 2.1) Gravitometer Relative Density 0.560 ± 0.002 Energy Calculation = Heating Value × Corrected Volume Imperial System BTU/cf × Corrected Volume (MCF) = MMBTU Metric System

TABLE 1 Largest positive error occurs when Gas Chromatograph

Calorimeter & Gravitometer

0.559 37.77 50.517 +0.01

Relative Density (R.D.) Heating Value (MJ/m3) Calculated Ki % Error

0.558 37.88 50.710 +0.39

0.561 37.83

Relative Density (R.D.) Heating Value (MJ/m3)

0.562 37.72

275

Largest positive error occurs when Gas Chromatograph 50.507 −0.01

Calorimeter & Gravitometer Calculated Ki % Error

50.136 −0.39

The Same calculation could aslo be used for calculating k and ki in the Imperial System of Measurement by substituting BTU/cf for MJ/m3. APPENDIX B Examples—Calculation with and without helium With helium 0.0005 0.566 37.33 (985.3)

helium mole fraction Relative Density Heating Value MJ/m3 (BTU/cf)

Discrepancy between results— −0.02 MJ/m3 (−0.5 BTU/cf). 0.0011 helium mole fraction 0.704 Relative Density 39.48 (1042.0) Heating Value MJ/m3 (BTU/cf) Discrepancy between results— −0.04 MJ/m3 (−1.1 BTU/cf). 0.0091 helium mole fraction 0.651 Relative Density 35.22 (929.6) Heating Value MJ/m3 (BTU/cf) Discrepancy between results— −0.33 MJ/m3 (−8.7 BTU/cf).

Without helium 0.0000 0.566 37.35 (985.8)

0.0000 0.704 39.52 (1043.1)

0.0000 0.655 35.55 (938.3)

LINEARITY AND RELIABILITY DETERMINATION OF BTU ANALYSIS BY PROCESS GAS CHROMATOGRAPHY Leighton Fields Application Engineering Manager Daniel Industries, Inc. Houston, Texas 77055

ABSTRACT

Apparatus described uses exponential dilution method to determine range of linear measurement for BTU analysis. Every component measured by the gas chromatograph can be easily checked for both linearity and minimum detectable limits. The dilution flask is set up with an amount of each impurity in natural gas which far exceeds that found in pipeline gas. This mixture is then gradually diluted with pure methane and readings taken every six minutes until the reading is 100% methane. Although it is theoretically possible to do the absolute calibration using this technique, the apparatus available is not nearly as precise as the primary certified standards available from today’s gas blending laboratories. For this reason, the instrument is first calibrated using a standard gas of a typical natural gas blend and then the exponential dilution is run. The linearity of the instrument is independent of the calibration and could be checked without an absolute calibration, but if the instrument is first calibrated and does prove to be linear, then all the minimum detectable limits of each component are also determined. After system linearity is proven, each instrument can be tested for linearity using two calibration standards. The repeatability of each instrument should be tested over its entire ambient temperature specification prior to installation in the field. INTRODUCTION Gas chromatography is a method by which gas mixtures are physically separated into their individual components and quantified. The on-line or process gas chromatograph performs this function on a continuous basis. On-line BTU measurement of natural gas is accomplished by analyzing eleven components of the gas and determining the BTU content based on the component mixture using appropriate factors. This paper will not attempt to deal with the factors or formulas in converting mole percent to BTU,

277

but will focus on the ability of the process gas chromatograph to accurately and reliably analyze the components in natural gas. EQUIPMENT The analyzer tested was the Daniel Model 500 equipped with the Model 2251 controller and the Epson RX-80 printer to record the data. The dilution equipment is from Fisher Scientific. The primary certified calibration gas used was supplied by Liquid Carbonic. The pure Methane gas was supplied by Matheson. The gas chromatograph employs a three column system which allows a complete analysis through C6+ in six minutes. (See Figure 1A & B). LINEARITY DETERMINATION One of the best methods of determing linearity of response of any gas analysis system is exponential dilution. It is a simple method to set up and yields a large amount of experimental data over a relatively short period of time. This is especially true in the case of a multi-component mixture. This method employs a mixing flask of fixed volume with two inputs, one for starting gas, one for diluent gas and one exit for the gas mixture. (See Figure 2). The flask is initially filled with high concentrations of the components in natural gas. (See Table 1). This mixture is then diluted with pure methane at a constant flow rate. With constant flow rate and proper mixing, the concentration of each component can be related to time. TABLE 1—STARTING GAS FOR EXPONENTIAL DILUTION Nitrogen Carbon Dioxide Ethane Propane Isobutane n-Butane Neopentane Isopentane n-Pentane C6 + Methane

17% 8% 10.0% 8.0% 3.5% 3.5% 0.03% 1.0% 2.0% 0.7% Balance

The general exponential decay curve can be represented by the expression: (1) where Co is the concentration at time (t) 0 and is the decay constant equal to r/V ,r being the rate of diluent flow and V being the volume of the dilution flask. A l000cc flask was used and a diluent flow rate of approximately 9cc/min. We can predict the time for a given concentration by the expression:

278

LINEARITY AND RELIABILITY DETERMINATION OF BTU ANALYSIS BY PROCESS GAS CHROMATOGRAPHY

FIGURE 1A BTU CHROMATOGRAM

279

FIGURE 1B ANALYSIS REPORT

To determine how long it will take to dilute the starting 8% CO2 down to 10 PPM then:

or 16.7 hours. In the same manner, all the component’s values can be predicted and compared with the printout for accuracy. Starting with the mixture in Table 1 and diluting with pure methane, the linearity of every component can be determined in one dilution. By plotting the natural log versus time, a straight line will be plotted if the response of the system to the particular component is indeed linear. (See Figure 3). All components plotted were found to be linear with all points within the repeatability listed in Table 3. By plotting each component concentration as printed out from the integrator, the entire analytical system is checked out. That is the detector, separating system, and the integration software and normalization techniques are checked in the system as it will be used for BTU measurement. As a double check on the computer, the natural log of the BTU contribution of each component can be plotted against time and will yield the same graph as the concentration. Since all the components are normalized to 100%, this method also proves the linearity of the methane measurement. Once the analytical system design has been proven linear, extensive tests need not be continued on each duplicate analytical system. Every time an instrument manufacturer changes design in either hardware or software in the system, these tests must be run to insure the accuracy of the analyzer. For duplicate systems of the same design, a simple two gas check can be run on each unit. A calibration on a known standard gas of high BTU content and an unknown run using a low BTU standard gas can suffice as a linearity check for individual systems.

280

LINEARITY AND RELIABILITY DETERMINATION OF BTU ANALYSIS BY PROCESS GAS CHROMATOGRAPHY

FIGURE 2 EXPONENTIAL DILUTION FLASK

MINIMUM DETECTABLE LIMITS OF EACH COMPONENT Another useful piece of information derived from the exponential dilution is the minimum detectable limit of each component. As each component approaches zero concentration, the analyzer will cease to reliably detect it. Since the dilution is exponential, the data on the low end is very precise and the minimum detectable limit of each component can be very accurately determined. For the run mentioned above, the lower detectable limits were determined by taking the first run which read zero for a given component and backing up five runs and using the actual value at that level. By backing up five runs (30 minutes) there is a 39% safety margin in determining the value. The values obtained are listed in Table 2.

281

FIGURE 3 TABLE 2 BTU CONTRIBUTION

COMPONENT Nitrogen Carbon Dioxide Ethane Propane Isobutane n-Butane neo-Pentane Isopentane n-Pentane C6 +

30PPM* 20PPM 15PPM 20PPM 10PPM 8PPM 11PPM 20PPM 16PPM 8PPM.05

— — .03 .05 .04 .03 .06 .08 .07 .05

* Diluent methane contained 56PPM N2 so nitrogen level determined by extrapolation of curve and comparison of relative response of CO2.

As can be seen from Table 2, the maximum error due to not detecting every low level component is −0.41 BTU. This assumes that the gas analyzed has every component just below its minimum detectable limit. It is obvious that this type of natural gas would be very hard to come by, and that for the average pipeline gas the degree of error from sensitivity of the system is indeed negligible. REPEATABILITY DETERMINATION Any analytical system can only be as accurate as it is repeatable. That is, analyzing the same sample must give the same result every time. To test the reproducibility of the chromatograph a standard gas was run for 24 hours and every run (240) analyzed for accuracy. The results are listed in Table 3.

282

LINEARITY AND RELIABILITY DETERMINATION OF BTU ANALYSIS BY PROCESS GAS CHROMATOGRAPHY

TABLE 3 Hi Reading Component

Concen.

Nitrogen Methane CO2 Etfiane Propane isobutane n-butane neopentane isopentane n-pentane C6 + BTU Spc Grav.

0.455 95.780 0.942 2.021 0.437 0.109 0.103 14PPM 0.042 0.031 0.091 1030.0 0.5872

Low Reading BTU 969.18 35.86 11.09 3.55 3.38 0.06 1.68 1.24 4.80 1029.8 .5870

Concen. 0.457 95.769 0.936 2.017 0.440 0.108 0.102 7PPM 0.040 0.029 0.089

BTU 969.29 35.78 11.03 3.52 3.32 0.03 1.62 1.16 4.71

Concentration

BTU

Error

Error

± 10 PPM ±60 PPM ± 30 PPM ±20 PPM ± 15 PPM ± 5 PPM ±5 PPM ±5 PPM ±10 PPM ±10 PPM ±10 PPM

±.01 ±.06 ±.03 ±.04 ±.03 ±.015 ±.03 ±.03 ±.03 ±.04 ±.045 ±0.1 ±.0001

As can be seen from Table 3, the maximum deviation on the BTU value was 0.1 BTU. If all the individual maximum errors in non-reproducibility occur in the right direction to produce the greatest error, it is possible to obtain an error of 0.36 BTU. Although this is theoretically possible, it would be practically impossible for all eleven components to be at their outside limits at the same time. RELIABILITY DETERMINATION There are several factors which affect reliability of the process gas chromatograph. The two major sources of hardware failure are infantile failure of electronic components and failure due to wear in the moving parts in the gas chromatograph valves. Preburning of all electronic components themselves and then burning in all electronic boards in an environmental chamber will eliminate most infantile electronic failures. Failure due to valve wear has been drastically reduced by the use of the diaphragm type valve as opposed to the sliding plate valves previously used. A diaphragm valve which has no 0-rings or any need for lubrication has been tested to as many as 5 million cycles without a failure. This is primarily due to the small distance in the movement of the piston which is less than 0.001″. With the two primary hardware failures reduced to a minimum there remains one major cause of analytical failure for a process gas chromatograph: fluctuating ambient temperature. This causes no failure in the hardware of the system, but if not properly handled, can have a drastic effect on the analysis accuracy. The process gas chromatograph should be mounted near the sample point for greatest accuracy and thus will not usually be mounted in a temperature controlled environment. The fluctuating temperature can cause many problems to a gas chromatograph that has not been designed to withstand a wide variation. This is perhaps the major reason why the process gas chromatograph has been associated with high maintenance. When temperatures changed the unit had to be adjusted. The analyzer here was tested in an environmental chamber and cycled from 0 to 130°F while running analysis on known standard gas.

283

The analyzer was calibrated at 70°F, then run for 8 hours at 0°F. The temperature was then raised to 130° F which takes approximately 6 hours, and held at 130°F for 8 hours, then lowered back to 70°F. The results are listed in Table 4. TABLE 4 REPEATABILITY FROM 0 TO 130°F COMPONENT

HI VALUE

LOW VALUE

REPEATABILITY

Nitrogen Carbon Dioxide Methane Ethane Propane isobutane n-butane neopentane isopentane n-pentane C6+ BTU Spec. Gravity

0.440 0.548 97.082 1.100 0.514 0.100 0.099 0.050 0.047 0.047 0.020 1030.1 .5786

0.413 0.537 97.050 1.093 0.510 0.99 0.098 0.049 0.045 0.046 0.018 1029.9 .5784

±.014% ±.006% ±0.016% ±0.004% ±0.002% ±0.001% ±0.001% ±0.001% ±0.001% ±0.001% ±0.001% ±0.15 BTU ±0.0001

RESULTS AND DISCUSSION Once the analyzer has been determined to be linear and repeatable, it can now be used to accurately analyze unknown samples, only after it has been calibrated with a known standard. Thus the analyzer can only be as accurate as the calibration gas. It is not recommended to pull a sample off the line, send it to a lab for an analysis and then use it as the primary standard for another analyzer. If an analyzer is to be used for custody transfer, it should be calibrated on a standard gas that has been prepared gravimetrically using NBS traceable weights. This gives a cross check on the instrument which is totally independent of the analytical method. As can be seen from the above tables, the process gas chromatograph can be a very accurate and precise technique for analyzing BTU content of pipeline quality natural gas. The analyzer is linear over a broad range of composition and very repeatable over a wide ambient temperature range. References Cited 1.

Colson, Ervan R., “Flame lonization Detectors and High End Linearity,” Analytical Chemistry 58, 337–344 1986 February

ON-SITE ENERGY MEASUREMENT R.A.Price Measurement Technical Services Manager Southern California Gas Company Los Angeles, California 90051

ABSTRACT

Southern California Gas Company’s entrance into the age of electronic measurement is discussed. Prototype installations at their four out-of-state purchase points have been installed and are operating. The measurement stations are located in the California Mojave desert where temperatures may vary by as much as 50°F to 60°F or more on a daily basis and seasonably by more than 100°F. On-site energy measurement equipment has been installed at these locations and it is assumed that if the equipment will operate satisfactorily at these locations, it will operate anywhere. Planned future installations, and their inherent problems, will also be discussed at length. “On-Site Energy Measurement” as described here is as follows: It’s real time measurement, on site, incorporating instrumentation to measure flow, specific gravity, % Nitrogen (N2) % carbon dioxide (CO2), and heating value (BTU). These variables are all necessary for the measurement of energy (therms), which is computed in real time, on site. PILOT INSTALLATIONS It is intended to install on-site energy measurement at all of Southern California Gas Company utility electric generating plants as well as its wholesale customers (approximately 30 locations). Currently, however, only four pilot installations are in and operating and these are at our own facilities. They are located at Blythe, Needles, and Newberry Springs, the points where we measure the gas we purchase from El Paso Natural Gas Company and Transwestern Pipeline Company, our major out-of-state suppliers. These locations are all in the Mojave Desert and it is believed that if the equipment will operate under these severe conditions, it will operate anyv/here. Temperatures at these locations may vary on a daily basis by as much

285

as 50°F to 60°F and on an annual basis by more than 100°F. Before describing their installations in detail, the current method of energy measurement will be described. CURRENT METHOD Several Company departments are involved in the currently used method of energy measurement. Division personnel change the flowmeter charts, temperature charts and gravitometer charts on a regular basis. These charts are sent to the Measurement office for processing. The Measurement Gas Analysis Center changes sample tanks on a regular basis. These tanks are returned to the Analysis Center where they are analyzed for specific gravity, % N2, % CO2 and BTU. These values are sent to the Measurement office for inclusion in volume and energy calculations. The Measurement office is where it all comes together. The flow meter charts are logged in, matched with the appropriate temperature, specific gravity and/or pressure charts as applicable and forwarded to the cleanup section which monitors the charts, cleans them as necessary, and determines whether they will be read with the optical scanner or manually integrated. Extremely dirty charts cannot be scanned and must be read manually. At this point the chart readings obtained from scanner or integrator must be matched with the values obtained from the Gas Analysis Center. This is performed manually on a special form. This form is then sent to the EDP section where the energy units are computed and the bills prepared. Along the way, there are many checks to hopefully preclude any errors. ON-SITE METHOD The on-site method currently being tested at the four pilot locations must have the capability of handling multiple orifice meter runs (as many as twelve) as well as turbine or rotary meters. It must have the capability of retaining several days’ information in 15-minute increments. This is because tariff rates can change on short notice. It must incorporate real time specific gravity, % N2, % CO2 and BTU as well as flow. The four pilot installations are almost identical and consist of the following components: 1. A remote terminal unit (RTU) capable of retaining volume and energy units in 15-minute increments for up to 72 hours. Also it must have the capability of two-way communications with the host computer. 2. A process chromatograph which completely analyzes the flowing gas at least every 15 minutes. This is transmitted immediately to the RTU for inclusion in the volume and energy computations. 3. Differential pressure, pressure, and temperature transmitters for each meter run. 4. An uninterruptable power supply (UPS) capable of operating the system for several hours in the event of a power failure. 5. Telephone modem to communicate all data to headquarters for billing and operating purposes. MODE OF OPERATION Under normal operating conditions the system will monitor volume and energy measurement continuously, storing this data in 15-minute increments. At some regular time, the host computers (H.P. 1000) located at headquarters, will poll the RTU, and extract the last 24 hours of data for billing purposes. In the event of a

286

ON-SITE ENERGY MEASUREMENT

tariff rate change at a time other than the billing period the RTU will have to be manually interrogated through the host computer. MAINTENANCE It is expected that maintenance will be minimal but in the event of problems most diagnostic work can be performed via the communications link. Calibration data on the process chromatograph is also readily available via this link. CONCLUSION On-site energy measurement as described in the paper is being used by an increasing number of companies in the sale of large volumes of gas. This has been made possible by the availability of more powerful and less expensive microprocessors. As the upward trend in power, and the downward trend in cost continues, on site energy measurement will be applied to smaller and smaller meters. Someday, in the not too distant future, we’ll probably see your home gas meter reading in energy units rather than cubic feet.

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS M.Haruta, K.Uekuri and Y.Kiuchi Asst. General Managers Osaka Gas Company, Ltd. Osaka 541, Japan

ABSTRACT

Osaka Gas Company has applied the correlation between the density and calorific value of natural gas to indirect continuous measurement of calorific value. By developing measuring instruments and employing existing instruments, calorific control of send-out gas and calorific monitoring of process gas have been improved in our plant. (1) Rauter meter: This conventional instrument with its high reproducibility and reliability has been used for calorific control of send-out gas at our Senboku Works. Five years of actual operation have produced satisfactory results. (2) Sonic type: Developed by Osaka Gas, this equipment has higher accuracy and responsity than a gravimeter. It utilizes the general correlation between the sound velocity u and density of gas p. This system is applied to continuous monitoring during LNG unloading. (3) Optical interferometer type: This calorimeter presently under development by Osaka Gas has demonstrated high accuracy, reproducibility and responsity in the field trial. Measurement is made by transforming the shades of interference fringes induced by the difference of refractive indices between standard gas and sample gas into electrical signals. Simple and compact, the apparatus has great potentials for control and monitor operations. INTRODUCTION Every year, about 12 billion cubic meters (equivalent to 10,000 kcal/Nm3) of town gas is distributed in Japan, and about 65% of it is natural gas. Japan imports most of its natural gas in the form of liquefied natural gas (LNG). Osaka Gas Co., Ltd. imports LNG from Indonesia and Brunei to provide town gas for its customers. Town gas in Japan is legally subject to quality control to ensure that the calorific value is kept at a constant level.

288

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Because, however, the ingredients of LNG vary somewhat according to its source and the conditions of its liquefaction, the calorific value of the resulting town gas is subject to variation. Osaka Gas maintains the calorific value of its town gas at 11,000 kcal/Nm3 by adding liquefied petroleum gas (LPG) to the gas produced by vaporization of LNG. Calorific value control and monitoring of send-out town gas is dependent on the existence of continuous measuring equipment with high reproducibility of measurement and speed of response. The fact that LNG import transactions are carried out on the basis of the BTU quantity of unloaded LNG also lends importance to LNG sampling and analysis as well as to measuring the volume. The commercial calorific value is determined from the value obtained by gas chromatographic analysis of gas stored in a holder by continuous sampling of LNG during unloading. Highly sensitive continuous measuring devices that respond rapidly to fluctuations are needed to enable stable monitoring of gas in the holder during continuous sampling. In the past, the combustion method was used for the continuous measurement of calorific value: gas was burned at a constant rate, the resulting heat was transferred to water or air, and the calorific value was calculated by measuring the resulting rise in the temperature of the water or air. Although the reproducibility of measurements by this method is high, its speed of response is very low. The difficulty of maintenance and its high price are further disadvantages. The use of process gas chromatography was also once envisaged, but it proved totally unfeasible for control applications, because roughly 15 minutes were needed to complete a measurement, even with highspeed models. Thus a highly reproducible, highly sensitive, and rapidly responsive method of continuous calorific value measurement was sought. At Osaka Gas, we recognized on the close correlation between LNG gas density and its calorific value. Our method of indirect calorific value measurement by gas density for control and monitor operations has now reached the application stage. This paper presents data on the performance and operation at Osaka Gas’s Senboku Works II of the established Rauter calorimeter, a sonic-type calorimeter developed by Osaka Gas, and an optical interferometer-type calorimeter currently under development. CALORIMETRIC METHODS Relationship between Calory and Density LNG is brought to Japan in a liquefied state at −160 degrees centigrade. All impurities in the natural gas are removed at the production site. As shown in Table 1, the gas is composed of paraffinie hydrocarbons, mainly methane. As shown in Figure 1, there is a very high correlation between gas density and the calorific values of these paraffinie hydrocarbons. The calorific value can be measured by continuously measuring gas density because of the perfectly proportional relationship between the density and the calorific value of LNG, as shown in Figure 2, with the correlation coefficient of 0.997 (sample number n = 126). Indirect calorimetry by measurement of gas density is possible by the following methods: 1) Rauter meter 2) Sonic type [1]~[4] 3) Optical interferometry [5]~[7] 4) Vibrational method [8]

289

Rauter Meter Principle. Measurement by this method is based on the proportionality of gas density to the pressure created by an impeller rotating at high speed. (1) Composition and Functions of the Apparatus. Figure 3 is a schematic diagram of the Rauter meter. An impeller rotating at high speed gives the air pressure torque to the second vane wheel in the chamber. This torque is measured and compared with that produced by an identical impeller arrangement in a standard gas (air). The Rauter meter has proven its worth through years of densitometric service. However, because it is a purely mechanical device, its speed of response and sensitivity are limited. Its speed of response is seconds, and its maximum sensitivity to Table 1. TYPICAL COMPOSITION OF LNG (VOL.%) Component

A

B

C

N2 CH4 C2H6 C3H8 i-C4H10 n-C4H10 i-C5H12 n-C5H12 Gross heating value [kcal/Nm3] Gas density (Air = 1)

0.04 90.89 4.40 3.09 0.68 0.87 0.02 0.01 10,659 0.627

0.02 88.98 5.90 3.56 0.68 0.80 0.06 0.00 10,833 0.639

0.01 85.80 8.80 4.19 0.61 0.57 0.02 0.00 1 1 ,059 0.654

density is 0.001 of air (=1). According to Figure 2, this sensitivity corresponds to 15.2 kcal/Nm3. Thus there are limits to the Rauter meter’s precision, but it is adequate for calory control and monitoring purposes. To achieve a higher speed of response and a target density sensitivity of 0.0001, we began development of sonic -type[4] and optical interferometer-type calorimeters.[7] Sonic-Type Calorimeter Principle. The speed of propagation of sound waves in a gas is given by Laplace’s equation, given below. (2) The speed of sound varies with temperature as follows: (3) Using these equations, we can measure the density, and therefore the calorific value of a gas, by measuring the speed of sound in the gas under conditions of constant temperature. To measure the speed of sound, we create a stationary wave in an acoustic cell of constant length and measure the frequency (f) at a constant wavelength (λ) from Equation (4) .

290

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Figure 1. CORRELATION BETWEEN THE DENSITY AND CALORIFIC VALUE OF HYDROCARBONS

(4) The relationship between frequency and density, expressed in Equation (5), is obtained from Equations (2) and (4). (5) Composition and Functions of the Apparatus. The technique of measuring the speed of sound in the gas is a well established one, and has been the subject of many studies, but very few industrial apparatus based on this principle have been commercialized. In 1977, we began development of a sonic-type calorimeter that would be highly reliable and more sensitive than existing types. We began commercial operation of our calorimeter in 1980. A schematic drawing of the sonic-type calorimeter is given in Figure 4. The gas whose calorific value is to be measured (sample gas) flows through the acoustic cell and resonates at a certain frequency in the oscillation circuit, which is composed of a speaker at one end of the cell and a microphone connected to an amplifier at the other end. Calorimetry by this method relies on

291

Figure 2. CORRELATION BETWEEN THE DENSITY AND CALORIFIC VALUE OF LNG

comparison of this frequency with the oscillating frequency of a reference gas (CH4) of similar composition sealed in an identical second acoustic cell. The influence of the sample gas’s rate of flow is nullified by allowing the sample gas to enter the cell in the middle and escape from both ends. The acoustic cell is maintained at a constant temperature by immersion in a constant temperature bath. The gas is maintained at a constant pressure by a pressure regulator. The frequency data are recorded after correction by a linearizer to conform with Equation (5). Optical Interferometer-Type Calorimeter Calorific Value and Refractive Index of Natural Gas. In addition to a characteristic density, each substance has its own characteristic refractive index. There is a very strong correlation between the gas density and the refractive index of homologous hydrocarbons, as shown in Figure 5. It is possible to measure the density of a mixture of different gases because the refractive index of a gaseous mixture is equal to the constituent sum of the respective single refractive indices. Thus the calorific value of natural gas can be determined from its

292

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Figure 3. SCHEMATIC DIAGRAM OF RAUTER METER

refractive index. Table 2 and Figure 6 give data on the calorific values and refractive indices of five different types of natural gas, and demonstrate the linear relationship between total calorific value and refractive index. Composition and Functions of the Apparatus. The simplest, most sensitive, and most precise method of measuring the difference in refractive indices of gases is to use an optical interferometer. Portable densitometers for inflammable gases by interferometry are compact, light, and very precise, and have been in service for decades. The calorimeter we built on a trial basis was developed to provide a higher sensitivity and speed of response by combining the latest electronics technology with the optics used in portable densitometers. A schematic drawing is given in Figure 7. The gas to be tested is fed into the sample gas chamber (CHG) and the reference gas (methane, etc.) is fed into the reference gas chamber (CHR). CHr is divided into two parts that are connected by a pipe; the same gas is introduced into both parts. White light emitted by a lamp “L” goes through a condenser which turns it into parallel beams of light, which pass through a slit and incide onto a plane mirror “M” with a plated back, which reflects the light, dividing it into two beams. One beam enters CHR and another penetrates the glass before being reflected by the plated back of the mirror and sent to CHG. After going through the chambers, these light beams are reflected by the main prism “PM” and are sent back through their respective chambers and converge on the surface of M, where they create an interference fringe. The brightest part of the interference fringe is magnified, collected by a telescopic lens, and directed toward a detector. The chamber length and optics are designed to enable adjustment, so as to vary the amount of incident light in accordance with the difference between the refractive indices of the gases in the two chambers.

293

Figure 4. SCHEMATIC OF THE SONIC-TYPE CALORIMETER Table 2. REFRACTIVE INDICES OF TYPICAL LNG (VOL.%) LNG

A

B

C

D

E

Component CH4 2 C2H6 C3H8 2 1 i-C4H10 0 n-C4H10 0 i-C5H10 n-C5H10 0 Grass heating value [kcal/Nm3] Refractive index n

95 3 2 1 0.5 0.5 0 10,196 439.0

93 5 3 1.5 0.5 0 0 10,522 452.4

90 8 5 1 0.5 0.5 0 10,786 463.1

85 9 5.5 1 1 0.5 0 11,328 485.8

83

11,587 496.3

294

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Figure 5. CORRELATION BETWEEN GAS DENSITY OF HYDROCARBONS AND REFRACTIVE INDEX

From this quantitative change, it is possible to determine the difference in refractive index. Since the relationship between the difference in refractive index and the amount of incident light is not linear, the data are linearized by a microprocessor. Since the refractive index of a gas is influenced by its temperature and pressure, steps must be taken to correct for these factors. The whole measuring chamber apparatus is immersed in a constant temperature bath to control temperature within the range 40 ±1°C and obviate the need for data correction. Pressure euqalization is provided by interconnection of the outlets from both chambers and the absolute pressure of these parts is measured, followed by data correction by the microprocessor. PERFORMANCE Influence of Other Components Mixture of the sample gas with O2 ,N2 or a noninflammable gas will result in an apparent error on the positive side in the calorific value corresponding to the density and refractive index of the components. Table 3 gives a sample of errors produced using the data in Figures 1 and 5. LNG contains 0.01–0.05% N2 which within this range causes an erroneously positive result. The extent of the error, however, is not sufficient to affect the control of send-out town gas. Further, when the quantities of other constituent gases are nearly constant, a corresponding correction of the data can be made by subtracting the expected errors.

295

Figure 6. CORRELATION BETWEEN CALORIFIC VALUE OF LNG AND REFRACTIVE INDEX

Characteristics of the Different Apparatus Table 4 compares the performance of the different methods of LNG calorimetry. Due to the structure of the Rauter meter, the correlation between output and gas density is nonlinear. Data characteristics are given in Figure 8. The speed of response of the Rauter meter is relatively high, because specimen gas flow is high and diffusion of gas in the detector does not impede measurement. Compared with the Rauter meter, in principle, the sonic calorimeter, having no moving parts, has higher sensitivity and reproducibility as well as outstanding speed of response. For several months, we cross-checked the data obtained by sonic calorimetry with data by gas chromatography. Figure 9 shows the differences in 72 samples; the average calorific value is 0.038% (4 kcal/Nm3) higher by sonic calorimetry than by gas chromatography. The maximum deviation in the calorific value was 0.126% on the positive side and 0.197% on the negative side, a generally satisfactory level.

296

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Figure 7. SCHEMATIC OF OPTICAL INTERFEROMETER–TYPE CALORIMETER Table 3. ERRORS RESULTING FROM PRESENCE OF 0.1% GASEOUS O2 OR N2 N2 = 0.1% Rauter meter

+

0.14%(15.8kcal/Nm3)

O2 = 0.1% + 0.16%(17.9kcal/Nm3)

297

Figure 8. CORRELATION BETWEEN RAUTER METER OUTPUT AND GAS DENSITY N2 = 0.1% Sonic calorimeter Optical interferometer Gas density(air=1) Refractive index

O2 = 0.1%

0.14%(15.8kcal/Nm3)

+ 0.16%(17.9kcal/Nm3) + 0.04% (5.0kcal/Nm3) O2=1.1053 O2=252.2

+ + 0.05% (5.6kcal/Nm3) N2= 0.9673 N2=276.7

Table 4. RESULTS OF ACTUAL APPLICATION TO NATURAL GAS

[kcal/Nm3]

Range =density Reproducibility Sensitivity/Min. range [kcal/Nm3] Response time [T90] Volume of gas specimen Reference gas ( ): Anticipated values

Router meter

Sonic calorimeter

Optical interferometer

9,000~14,000 =0.5~0.9 ±1.5% F.S ±25 / 2500 30sec 3 /min Air

9,500~11,500 =0.55~0.68 ± 0.5% F.S ±2 / 400 1 5sec 1 /min CH4

9,500~11,500 =0.55~0.68 ( ±0.5%) ( ±2 7200) (10~15sec) 0.2 /min CH4

298

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Figure 9. DIFFERENCE BETWEEN CALORIFIC VALUES MEASURED BY GAS CHROMATOGRAPHIC ANALYSIS AND SONIC–TYPE CALORIMETER

Field tests with the optical interferometer were begun in March 1986; the figures in parentheses in Table 4 are anticipated values.

299

APPLICATION Calorific Value Monitoring during LNG Unloading At Senboku Works II, we are now using sonic calorimetry to monitor calorific values in LNG at unloading, and use Rauter meters for calorific value control of feed gas; our optical interferometer is now undergoing field tests and further development to exploit the method’s theoretical advantages in terms of high-speed response, high sensitivity, and simplicity and compactness of the apparatus. Figure 10 shows the layout of the Senboku Works II, and Figure 11 is a schematic diagram of the systems used to monitor the calorific value of sampling gas at unloading and to control the calorific value of the town gas we distribute. Senboku Works II yearly receives 3 million tons of LNG which has been liquefied at the production site and transported to the works in LNG carriers. After unloading, the LNG is pumped to LNG tanks where it is kept at −160°C. As we said, LNG supply contracts are based on the BTU quantities unloaded. Sampling at unloading involves changing the LNG to gas in a sampling vaporizer, then feeding it continuously to a sampling holder using sampling compressors. After completion of unloading, the gas in the holders is tranferred to cylinders, from which samples are taken for determination of calorific value by gas chromatography. Figure 12 shows the changes in calorific value during unloading. The calorific value of the sampling gas fluctuates according to the number of cargo pumps in operation. Therefore, gas is transferred to the sampling holder when all the ship’s pumps are in operation, LNG flow has stabilized, and sampling gas calorific values cease to fluctuate. The calorific value begins to fluctuate immediately if the number of cargo pumps in operation is altered in the course of unloading. For this reason, a sonic-type calorimeter, because of its high sensitivity, reproducibility, and speed of response, is used for deciding the moment to begin sampling and for continuously monitoring sampling conditions. The monitoring data are both recorded on a chart and inputted to the supervisory computer system, which monitors sampling conditions and generates alarms and reports when an abnormality occurs. The sonic-type calorimeter is very reliable: it has been in continuous use without maintenance for five years thus far. Calorific Value Control of Distributed Town Gas Before distribution of town gas to users, LNG must be vaporized by raising its temperature using seawater and LPG must be added to maintain its calorific value at a determined level. Calorific value control is carried out using a computerized cascade. When a discrepancy occurs between the determined level and the value obtained by continuous calorimetry, a corrective operation is performed on the ratio between LNG and LPG. Since Senboku Works II came on stream in 1977, we have used a Cutler-Hammer combustion-type calorimeter for calorific value monitoring and quality control, and a Rauter meter for gas density monitoring. As we mentioned, however, we now know the correlation between calorific value and density, and we have compared the performance of the two methods operating in parallel, confirming that the Rauter meter’s rapid response and high reliability make it fully adequate for calorific value control.

300

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Figure 10. LAYOUT OF SENBOKU WORKS II

301

Figure 11. CALORIFIC VALUE MONITORING AND CONTROL SYSTEM

Figure 13 shows sample data for the Rauter meter in control measurement applications. It controls calorific value to within ±20 kcal/Nm3. Because LNG is clean, there is little error caused by contamination of the apparatus and the apparatus remains accurate if given a regular yearly inspection.

302

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

Figure 12. TIME SEQUENCE OF GAS SAMPLING OPERATION AND CALORIFIC VALUE DEVIATION DURING LNG UNLOADING

CONCLUSION Gas densitometers can measure, continuously and indirectly, the calorific value of natural gas, can accurately control the calorific value of send-out town gas and can accurately monitor process gas. Comparison with conventional combustion-type calorimeters, based on actual experience at the works, reveals the tendencies shown in Table 5. The optimal calorimetric method must therefore be chosen according to the intended application at the works in question.

303

Figure 13. CALORIFIC CONTROL RESULTS BY RAUTER METER

ACKNOWLEDGMENT We thank Riken Keiki Corporation for the valuable cooperation in the development of the optical interferometric calorimeter. Table 5. COMPARISON OF DIFFERENT METHODS Accuracy Monitoring

Control

Combustion type Gas chromatography Rauter meter Sonic calorimeter Optical interferometer

A A B A A

Purpose

B B B A A

B C A A A

Maintenance

Cost

B B A A A

B B A B A

A : Optimum (Inexpensive) B : Good (Moderate) C : Unsuitable (Expensive)

304

CONTINUOUS MEASUREMENT OF CALORIFIC VALUE OF NATURAL GAS

NOTATION f n P ΔP t V Vo Vt γ λ Po pt

= = = = = = = = = = = =

Frequency Refractive index Pressure Wind pressure Temperature (°C) Revolution rate Velocity of sound (N.T.P.) Velocity of sound (at t°C) Ratio of the two heat capacities (Cp/Cv) Wave length Gas density (N.T.P.) Gas density (at t°C) LITERATURE CITED

[1] [2] [3] [4] [5] [6] [7] [8]

Crouthamal, C.E. and Diehl, H., “Gas Analysis Apparatus Employing the Velocity of Sound”, Anal. Chem. 20, 515–520 (1948). Haswell, R., “Development of the Sonic Gas Analyzer”, Ind. Chemist. 38, 164–166 (1962). Chimoto, S., Hirayama, T., and Yamazaki, H., “Sonic Type Gas Analyzer ER—62”, Yokogawa Technical Report 5 (Yokogawa Electric Corp.), 84–88 (1958). Haruta, M., “Combustion Property of Gas Measuring Apparatus”, U.S. Patent 4246773 (1981) Jan. 27, Japan Patent 1249799 (1985) Jan. 25. Namba, S., “Photoelectric Recording Interferometer for Gas Analysis”, Rev. Sci. Instr. 30, 642– (1959). Sugano, A., “Automatic Alarm of Methane”, Anzen Kogaku 10, 113–119 (1971) Feb. Komachi, M., “Method and Equipment of Gas Heating Values Measurement”, Japan Patent 1294576 (1985) Dec. 26. Ditcham, S., “Gas Quality Metering”, Pipeline & Gas Journal 212, 36–42 (1985) Oct.

DIRECT MEASUREMENT OF ENERGY FLOW—RECENT FIELD EXPERIENCE Carl H.Griffis Project Manager, Gas Operations Technology Gas Research Institute Chicago, Illinois 60631 Dr. William H.Clingman, Vice President Lyn R.Kennedy, Technical Staff Precision Measurement Incorporated Duncanville, Texas 75138 Dr.Kenneth R.Hall, Associate Director Dr.James C.Holste, Associate Professor Texas A&M University College Station, Texas 77843 ABSTRACT

Two methods of measuring total energy flow have been compared with one another. The first method uses a flow computer which has as inputs the pressure, temperature, and pressure differential from an orifice meter. This method also uses continuous measurements of relative density and calorific value. The output from the flow computer is the instantaneous energy flow. The second method uses a new energy flowmeter that is under development. It directly measures energy flow without measuring either calorific value or volumetric flow. This energy flowmeter has been operating in a test mode for about one year at a field metering site. In this paper some of the most recent experimental results are presented. INTRODUCTION The Gas Research Institute has been sponsoring the development of a new approach for measuring the energy delivered from a pipeline or distribution system. The heart of this method is an energy flowmeter that measures directly the Btu/Hr of gas flowing down a pipeline. Traditionally this has been determined by measuring volumetric flow, (scf/Hr), and calorific value, (Btu/scf), independently. These two quantities have been multiplied together to obtain the energy flow, (Btu/Hr). The instrument that we have been developing measures energy flow directly without measuring either volumetric flow or calorific value. This avoids the problems normally associated with high precision measurements of volumetric flow and increases the precision of the energy measurement. This method was discussed in some detail at the August 1985 conference sponsored by IGT on “Natural Gas Energy Measurement”. At that time, however, only preliminary results were available. This paper provides an update of the experimental work. The energy flowmeter has been operating for several months

306

DIRECT MEASUREMENT OF ENERGY FLOW—RECENT FIELD EXPERIENCE

on an eight inch pipeline in parallel with a flow computer that is also measuring total energy flow. The installation is at a Lone Star Gas orifice metering site near Bryan, Texas. In the following section of this paper the equipment at this installation is described. The data from both methods are compared in order to assess the precision and linearity of the energy flowmeter. In the final section of the paper conclusions are presented as well as a discussion of the continuing program. MEASUREMENT METHODS Two measurement methods are being compared at a test site near Bryan, Texas. One method is the Precision Measurement Incorporated energy flowmeter, which has been developed under sponsorship by the Gas Research Institute. The other method uses an orifice meter and calorimeter to provide inputs to a flow computer that calculates the energy flow. The Bryan test site is an odorant injection station operated by Lone Star Gas, which uses an orifice meter to control the odorant injection. The orifice run is in an eight inch line, with gas flows varying between 500,000 and 1,000,000 scfh. The source of the gas is a Champlin gas plant about five miles upstream from this site. Operations at this plant are variable, and the calorific value can change rapidly over a 1000 to 1200 Btu/scf range. Gas leaving this measuring site goes both to a municipal power plant and to the Bryan, Texas distribution system. The energy flowmeter has two main components. These have been covered in detail in the previous presentation(1) and GRI reports (2, 3). The first component is a flow separator which divides the flow into two streams, a main stream and a sample stream. A specially designed flow constrictor is placed in each stream. The constrictor in the main stream contains over 700 rectangular channels all of the same hydraulic diameter. The sample stream contains a manifold of capillaries. A valving arrangement can direct the flow through either a single capillary or a group of capillaries in parallel. A differential pressure gauge continuously monitors the differential pressure between the sample and main line downstream from the flow ccnstr ictors. The sample flow which is split off from the main stream is burned in the second component of the energy flowmeter. This second component is called a Flow-Titrator™. The Flow-Titrator performs two functions. It controls the sample stream flow to maintain zero differential pressure at the flow separator. It also mixes the sample flow with a stoichiometric quantity of air and burns the mixture. The Flow-Titrator design is patterned after the Therm-TitratorR, which is a commercial instrument manufactured by Precision Measurement Incorporated for measuring the calorific value of natural gas. It has been established in previous experiments that the ratio of the main flow to the sample flow is always the same and is independent of the flow magnitude or the gas composition. Thus the air flow in the Flow-Titrator is in direct proportion to the energy flow in the main line. At the Bryan site the main line flow constrictor was installed flush against an isolation valve and a 90° elbow. This was done to establish worst case test conditions. It was expected on theoretical grounds that the operation of the flow separator would be insensitive to upstream flow conditioning of the gas. Test data taken to date has proven this to be the case. The second method of measuring energy flow using independent measurements of volumetric flow was provided by a Lone Star Gas orifice meter. Input to the Elliot flow computer included the pressure differential across the orifice plate and the pressure and temperature of the gas at the orifice. A commercial GB-2000 instrument from Precision Measurement Incorporated provided a continuous input of specific gravity to the flow computer. Calorific value also was provided by the GB-2000. The flow computer used these inputs to calculate the volumetric flow at standard conditions. It then corrected for

307

supercompressibility using the NX-19 procedure. The corrected volumetric flow was multiplied by calorific value to obtain the energy flow. EXPERIMENTAL RESULTS Data was recorded over a period of several weeks using chart recorders to continuously record the output of the energy flowmeter and the flow computer. The Flow-Titrator calibrated itself once a day on a fixed flow of methane. The reading of the instrument was taken as a percentage of this fixed energy flow. The flow computer output was in millions of Btu per hour, (MMBtu/Hr). This output was then used to calculate an overall meter factor for the energy flowmeter. This was done every 30 minutes during a run. These meter factors were averaged over the length of the run, which was always 24 hours or less. If there were no random errors in either measurement method then this meter factor would be constant with time. The random measurement errors, however, introduce a random error in the meter factor. From a statistical treatment of the meter factor variations it was possible to place an upper limit on the errors in the measurement methods. The energy flowmeter was designed so that it could shift flow constrictors in the sample line. The ratio of the main flow to the sample flow was called the split of the flow separator. The capillary and valve manifold on the sample line allowed the split to be changed between four different values. This was done automatically in response to changes in the main line flow. The micro-computer of the Flow-Titrator selected the split as to keep the sample flow in the 2–4 scfh range which is the operating range of the combustion system of the Flow-Titrator. During the field test the split actually changed between two values. These corresponded to a high and low range for the main line energy flow. In Table 1 the average meter factors are shown for each operating range along with the date and times that the instrument operated continuously in that range. This defined an operating period. Each time the range changed a new operating period was begun and a new average meter factor was calculated. For a given range there is some fluctuation in the meter factor from period to period. This is due to the random errors in the two measurement methods. The weighted standard deviation was calculated for the meter factors for each range and these are shown in Table 2. Since operating periods in a given range were 24 hours or less, the weighted standard deviation was an upper limit on the precision of measuring the daily energy delivered by either method. It could be concluded from this data that the precision of either method would be better than 0.3%. It could also be concluded that the energy flowmeter was linear over the range of conditions at the Bryan test site. The meter factor was not affected by TABLE 1 BRYAN TEST SITE DATA Energy Flowmeter Measurement Period Average Meter Factor Date

Begin Time End Time High Flow Range Low Flow Range Standard Deviation of Average Percent

10/13/85 10/14/85 10/15/85 10/16/85

1430 0 0 0 1230 1300

2400 2400 2400 1200 1230 2130

8.38 8.42 8.46 8.49 4.51 8.44

0.69 0.87 0.68 0.39 * 0.26

308

DIRECT MEASUREMENT OF ENERGY FLOW—RECENT FIELD EXPERIENCE

Date 10/19/85 10/20/85 10/21/85 10/23/85

10/24/85 10/25/85

10/26/85 10/27/85 10/28/85 10/30/85 10/31/85

11/01/85 11/02/85 11/03/85

Begin Time End Time High Flow Range Low Flow Range Standard Deviation of Average Percent 2230 1300 0 0 1800 2000 2030 0 1230 0 900 1130 2230 2400 0 0 630 0 900 1230 0 1100 1900 0 700 1600 2200 1600 1830

2400 2400 2400 830 1900 2000 2400 1200 2400 830 1100 2200 2330 2400 2400 530 2400 1600 1200 1600 1030 1800 2400 630 1900 1800 2400 1800 2400

8.33 8.61 8.59 8.52 4.71 8.63 4.63 4.68 8.65 8.66 4.66 8.66 4.62 8.67 8.67 8.68 8.72 8.57 4.68 8.74 4.68 8.77 4.67 4.67 8.99 8.94 4.74 8.85 4.76

1.20 0.60 0.53 0.45 0.50 * 1.18 1.56 1.31 0.33 1.63 0.93 1.10 * 1.83 0.31 0.47 3.51 0.31 0.25 0.52 0.38 0.67 0.30 0.98 0.43 0.44 0.41 0.86

* One Reading Only

TABLE 1 (Continued) BRYAN TEST SITE DATA Energy Flowmeter Measurement Period Average Meter Factor Date

Begin Time End Time High Flow Range Low Flow Range Standard Deviation of Average Percent

11/04/85

0 730 830

600 730 830

4.80 4.80 4.87

0.32 * *

309

Date

11/06/85

11/07/85

11/08/85

11/09/85 11/10/85 11/10/85 11/11/85

11/12/85

Begin Time End Time High Flow Range Low Flow Range Standard Deviation of Average Percent 900 2030 900 1400 1900 2100 0 230 730 930 1930 2330 2400 0 230 530 730 830 1200 0 0 1800 0 830 2200 0 330 830 930 1530 1630 2200

1600 2400 1330 1800 2000 2400 130 630 830 1730 1930 2330 2400 130 430 630 800 1030 2400 2400 1600 2400 630 1730 2400 300 800 900 1430 1600 2130 2330

9.07 4.81 4.76 8.86 4.78 8.89 8.95 4.78 4.84 8.90 4.70 8.88 4.74 4.77 8.94 4.78 4.81 8.93 8.82 8.77 8.75 8.78 8.74 8.93 4.76 4.76 8.86 4.81 8.92 4.71 8.77

* One Reading Only

TABLE 1 (Continued) BRYAN TEST SITE DATA Energy Flowmeter Measurement Period—Average Meter Factor

4.75

0.61 0.55 0.78 1.12 0.68 0.65 0.14 0.49 4.31 1.21 * * * 0.18 0.29 1.64 1.37 0.56 0.93 0.69 1.33 1.21 0.61 0.76 0.64 0.43 0.98 0.77 0.55 0.36 1.39 0.42

310

DIRECT MEASUREMENT OF ENERGY FLOW—RECENT FIELD EXPERIENCE

Date

Begin Time End Time High Flow Range Low Flow Range Standard Deviation of Average Percent

11/13/85

0 800 1000 1030 1400 1830 2100 0 700 2300 30 100 600 830 1400 1900 2100 2200 0 230 400 0 1630 0 400 0 930 1400 0.89 1830 0 400 530 2100 2200 0 130 600

11/14/85

11/15/85 11/16/85

11/17/85

11/18/85 11/19/85 11/20/85

11/21/85

1530 11/22/85

* One Reading Only

730 930 1000 1330 1800 2000 2400 630 2230 2400 30 730 730 1230 1500 2000 2130 2400 100 330 2400 1530 2400 300 2400 900 1300 1730 2400 330 430 1500 9.07 2400 100 530 1030

8.87 4.72 8.80 4.74 8.78 4.87 4.69 4.70 8.77 4.70 4.65 8.79 4.81 4.85 4.84 4.85 9.05 4.85 4.74 4.90 8.95 8.88 4.67 4.66 8.79 8.91 4.82

0.50 0.50 * 1.99 0.99 3.41 0.42 0.80 1.31 0.28 * 0.66 0.35 0.75 0.41 0.58 0.54 0.35 1.05 0.55 1.32 0.50 0.51 0.18 0.83 0.38 0.83

8.85 4.74 4.78 9.00 4.81 0.70 4.83 4.81 9.10 4.85

0.74 0.28 0.05 1.24 0.65 0.19 0.35 0.57

311

TABLE 2 WEIGHTED STANDARD DEVIATION OF METER FACTORS Month

Range

Average of Meter Factors

Weighted Standard Deviation

October October November November

High Low High Low

8.56 4.66 8.86 4.75

0.9%* 0.3% 0.2% 0.3%

* Drift due to Capillary Leaks.

variations in flow and calorific value. There was some drift in time of the meter factor for the high range, and this was attributed to leaks that were found in the capillaries of the flow separator. These leaks were corrected by redesigning the capillary manifold, and then a second set of experiments was initiated at the Bryan site which are underway at the present time. Data collection was changed by replacing the chart recorders with remote data collection. The output of the flow computer was input to the microcomputer of the Flow-Titrator, which was programmed to calculate the meter factor continuously. One minute averages of the data were then stored in memory. The micro-computer at Bryan also had communications capability through an RS-232C port and external modem. Stored data was downloaded to diskettes at the Precision Measurement Laboratory in Duncanville, Texas. The operating parameters of the instrument as well as the program being run in the micro-computer of the Flow-Titrator could also be modified at will from Duncanville using the same communication link. This system is currently being used to further study the performance of the energy flowmeter, and a typical example of the type of data being collected is shown in Figure 1. Referring to Figure 1, the energy flow calculated by the flow computer is indicated by the symbol E, with the Flow-Titrator measurement indicated by the symbol F. The vertical centerline represents the average of all the measurements taken during the test period. Time increments on the vertical scale are two minutes, with energy measurements on the horizontal scale in increments of 0.1%. As the chart indicates, the energy flow meter and flow computer values track each other very well over a energy flow range of ±3%. CONCLUSIONS The results have demonstrated that the energy flowmeter will measure energy delivered to a precision better than 0.3%. This precision is based on a direct measurement of energy flow in a pipeline rather than by combining measurements from several instruments. No correction is required for supercompressibility at pipeline conditions and no corrections are required for pressure and temperature of the gas. A principal advantage of the energy flowmeter is the relative ease of installation compared with an orifice meter. In the experiments described in this paper the flowmeter was installed where there were worst case flow conditions. No flow straighteners or straight run in advance of the meter were required. Even so, the meter factors calculated during the experiments were independent of the total gas flow or composition. The accuracy of the energy flowmeter will be limited by the accuracy of the calibration. This is the focus of the ongoing work. A field test is being established with several participants to evaluate different approaches to calibrating the energy flowmeter components. These field test sites will correspond to different applications of the meter.

312

DIRECT MEASUREMENT OF ENERGY FLOW—RECENT FIELD EXPERIENCE

FIGURE 1 COMPARISON OF ENERGY FLOWMETER AND FLOW COMPUTER MEASUREMENTS

NOTES: E—Energy flow calculated by flow computer F—Energy flowmeter reading Centerline is average of the readings taken Increments on vertical scale are two minutes Increments on horizontal scale are 0.1% of average energy flow

REFERENCES 1.

2.

Clingman, W.H.; Hall, K.R.; Holste, J; and Kennedy, L.H., “Field Experience in Measuring Energy Delivered with Gaseous Fuels”, presented at the IGT Symposium on Natural Gas Energy Measurement, Chicago, August 1985. Clingman, W.H.; and Hall, K.R., “Energy Flowmeter Development: Phase I—Gas Sampling Methods; Phase II— Prototype Energy Flowmeter”, GRI Report 81/0200, December 1983.

313

3.

Clingman, W.H.; and Hall, K.R., “Energy Flowmeter Development”, Final Report for GRI Contract No. 5083– 271–0938, January 1986, to be published May 1986

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS Klaus Homann, Dr.-Ing. Head of Gastechnical Department Vereinigte Elektrizitatswerke Westf. AG Bezirksdirektion Dortmund Ostwall 51, 4600 Dortmund, FRG and Hans J.Krabbe, Dr. rer. nat. Head of Department for Organic Chemistry Vereinigte Elektrizitatswerke Westf. AG Hauptverwaltung Rheinlanddamm 24, 4600 Dortmund, FRG ABSTRACT

The following paper deals with methods of gas quality determination at VEW, a regional electric power and gas utility in FRG. As far as relevance for the understanding of gas quality problems discussed here, boundary conditions of the FRG’s gas market are described in further detail in chapter 1. In chapter 2 we mention some details of its marking general and gas specific regulations concerned with utilisation and account problems arising from varying gas qualities throughout a complex network. Chapter 3 gives a survey over recent developments in meeting continuously increasing requirements of customers and official authorities in this field. Chapter 4 is concerned to a VEW-development of a microprocessor (μP)–controlled gas sampling unit for solving special problems in the field of gas quality attachment, while chapter 5 finally reports some aspects of GC-apply in FRG and experiences of VEW in that field. 1. FRG’S MARKET FOR NATURAL GAS Despite a short-term decline the total primary njiergy consumption of the FRG has reached a value of 11×1018 Joule per year. Among this only a part of about 30 % is from domestic origins so that 70 % mainly in form of oil and gas have to be imported from other countries, western as well as eastern ones, with all inherent technical and political risks. The contribution of natural gas to total energy consumption is about 20 %, in the room heating sector even 27 % (Figure 1). The important role of natural gas for the energy supply is clearly pointed out by this few facts. On the other hand it is obvious, that securement and diversification are essential demands. One third of gas consumption is produced from our own resources, the other two are imported from the Netherlands, the USSR and the Norwegian North-Sea. So far we can state, that the dominating part of gas

315

Figure 1: Partition of end energy consumption and room heating

Figure 2: Structure of german gas industry

consumption is from west-european countries, procurement and its diversification are out of discussion presently. After this view on FRG’S energy consumption scene we should also put some interest on the structure of german gas transporting and distributing organisations. Figure 2 shows an illustrating diagram. Although it is pointed out, that 6 major organisations operate with their own import contracts it should be stated, that nearly all import contracts are dealt under response of one big enterprise, the Ruhrgas AG. While transmission companies are in general privately structured, the municipal distributors are often companies under official management. VEW itself is mainly engaged in production and delivery of electrical energy. But meanwhile one quarter of total turnover is caused by transportation and delivery of natural gas. Being active in two sectors of figure 2, VEW belongs to both categories of companies: transmission and distribution. 170 000 customers are supplied directly and another 630 000 via smaller municipal companies. Figure 3 shows the high pressure network (70 bar and 16 bar) under operation and the location of a big underground storage of the aquifer type, which was exploited for seasonal equalisation of gas purchase. More details of high pressure transportation are given in /3/. Let us at first have a look on the present situation and special problems caused by different qualities.

316

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS

Figure 3: High pressure network of VEW

2. DIFFERENT GAS QUALITIES—APPLY AND ACCOUNT PROBLEMS Apply problems During the past, varying gas compositions mainly caused apply problems at industrial and household burners. Being aware of the fact, that all types of combustible gases would be used in public gas markets, the private german technical advisory committee for the gas industry—the DVGW—dismissed in 1939 “G 260: Technical rule for gas quality”. This technical rules, presently in revision again, should be interpreted as a frame work for gas delivery, running of gas-burners and their development and standardisation. Beyond

317

this indications are given, up to which degree gases of different qualities can be intermingled without apply problems. G 260 devides the whole scope of combustible gases into 4 major “families”: 1.: 2.: 3.: 4.:

Hydrogen Gases;

group A: town gas group B: pipeline gas Methane Gases; group L: low heating value (~10 kwh/m3 ) group H: high heating value (~/12 kwh/m3) Liquid Gases; Propane, Butane and mixtures Mixtures of Liquid or Natural Gases with air, mainly for peak shaving purposes

Especially all inland productions and imports of natural gas can be attached to one of the two groups of the 2nd gas family. Consequently the range of quality defining values—heating value and density—is rather wide. The ratio of heating value Ho and square root of density, the so called “Wobbe-Index” may vary within the whole family about ± 17 % and within each group about ± 7,5 %, a tribute to diversification of gas procurement and integration of pipeline networks. During the whole time of application of G 260 for useability of different gases for public supply, burner technology has reached a high standard. Only inferior problems occur with radiation heaters and gas engines, sometimes not being suitable for the whole quality range permitted. On the other hand steadily increasing prices for gas in households and increasing importance of gas as a production factor in industrial processes caused an increasing interest in measurement and account accuracy. Account problems During a long time delivered gas quantities were merely accounted by measuring the gas volume. Quality details (heating value) and values of thermal state (pressure and temperature) were hidden in the gas price, fixed by the supply service. With increasing energy costs more and more suppliers went over to another way of account, more traceable for the customers: heating value, pressure and temperature were actually mentioned within the statement of consumption, so that, following official demands, the DVGW dismissed a special rule—G 685; Enforcement on thermal billing—in 1982. The primary intention of the official authorities was to define areas of flowrate and measuring pressure with prescribed types of pressure regulators and state correction procedures. Beyond this, boundaries for estimating time means or flowrate weighted means of heating value ( ± 1% !) were fixed. An additional paper to G 685, presently in work, will give further details of what to do, if different gas qualities are intermingled in integrated networks. Accuracy of heating value attachment will be a central theme. Up to now, only the minimum accuracy of actually installed measuring equipment is prescribed. We will now give a rough sketch on present FRGstandard. For further detail the reader is referred to additional literature . Accuracy of measuring equipment Household gas meters should, at first installation, have an accuracy of ± 2 %, turbine meters for higher flowrates of ± 1 %. The accuracy of state correctors i s ± 1 % and that of calorimeters, the only officially permitted way of heating value measurement, ± 0,8 %. Let us insert some remarks on state correction, as it is done presently in FRG. Mechanical state correctors are still widely in use, so that real gas behaviour is often neglected or put into account by additional

318

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS

numerical procedures. AGA NX 19—formula is usually taken for this. That means, that heating value, density and CO2-partition have to be measured. Because of microprocessors becomming more and more cheaper, electronic state correctors are in advance, on basis of a state equation as well as on basis of density measurement on working conditions. By the few numbers on measuring accuracy mentioned above it is pointed out, that official authorities in FRG require a degree of accuracy, which manufactures are just able to guarantee. On the other hand it is obvious for everyone, that handling problems of varying gas quality especially with respect to smaller household consumers has not reached that standard. The next chapter shall deal with efforts of various german companies in this field. 3. FRG-TRENDS IN HANDLING OF VARYING GAS QUALITIES Conventional Methods During a long area in german gas utilisation history coke oven gas and natural gas—inland production and imports from Netherland—were of same importance for supply. Consequently the present transportation network, which was erected during the past 40 years, arose from smaller local high pressure networks in the regions of coal and steel production. The overall length of the german pipeline network today is 165 000 km. Although the steadily growing gas market forced more and more new gas qualities to be made available for the customers, high heating value gas from the North Sea (~12 kwh/m3), russian imports with a little bit lower value (~11 kwh/m3)—it could be managed to keep them separated from each other to a certain degree. If necessary, mixing stations were built, for mixing natural gas with air or coke oven gas or, vice versa, coke oven gas with liquid gas for lower pressure ranges. Already when only coke oven gas was distributed, calorimeters became widely in use for quality control. Up to now calorimetry is the only officially permitted method to determine heating values. This is due to the fact, that relevant german law requires officially licensed and supervised equipment for measuring values, which are used directly or indirectly (!) for accounting purposes. Cutler & Hammer Calorimeters e.g. have such a license for FRG. Calorimeter technique still is a too costly method, to be used at every billing point in a complex network with local or time variations of gas quality. The installation of gas samplers at certain characteristic points within the network is a cheaper alternative. The american system “Arco Anubis” is a well known instrument also in germany and widely in use. Being independent of any electric supply its inset is advantageous in desert areas, but it has some disadvantages, as e.g. being unable to take samples in flow rate proportional intervals, that caused us to develop a new technique, which we will describe later. A new, rather sophisticated method of facing quality attachment problems shall be mentioned first. Unsteady net flow calculations Unsteady net flow calculation was developed by Weimann et.al. /5/, originally as a simulation technique in process-computers for network control. Using a finite difference method the unsteady one dimensional balance equations for mass and momentum were solved, to predict pressure and flow development in high pressure networks. An application of this method at VEW is described in /3/. Reacting on demands of the official authorities and the gas industry, Herr et.al. /2/ extended his model for handling quality attachment. Now, if flow rates of input and offtake as well as quality data of input gasflow

319

are known, the corresponding data at any points especially those of billing can be calculated. Obviously this method is only suitable in bigger transmission networks. An official working allowance was given under the condition that it is used together with a master calorimeter. Another field of apply will be in planning representative locations for conventional measuring equipments as described above. 4. μP-CONTROLLED GAS SAMPLING The high pressure network of VEW as shown in Figure 3 presently is divided into three major quality districts each separately run: the north western part is operated with natural gas, group H, mainly from the Norwegian North Sea, a smaller south western part delivers H-gas as well, but with reduced russian quality. The whole eastern part is run with natural gas, group L, of different netherland and german origins. Especially within the L-gas region local and time variations of heating value in a range of 10 % have to be handled. Further network separations were rejected with respect to supply security. Unsteady network flow calculation could not be used, because a nonneglectable part of the high pressure offtake is not permanently registered. Therefore, being aware of some disadvantages of the “Arco—Anubis Gas Sampler” VEW developed a new one, consequently adjusted to VEW demands. Some of these demands are listed underneath : Sample intervals are to be triggered externally, especially proportional to flow rate. Sample container has to be either evacuated by pipeline pressure (exhauster) or sweeped out with pipeline or inert gas. Fixing the sample size has to consider for real gas behaviour (AGA NX 19) and varying pipeline temperature and pressure. Present state of gas sampling has to be supervised via remote facilities. All operations for changing the sample container should be μP-supported. The connecting pipe between sample container and pipeline should be sweeped with “fresh” gas before each sampling. Self controlling with respect to possible undensity. After all an apparatus was constructed consisting of two parts: a control unit outside the explosion area and the sampling facilities as e.g. valves, pressure and temperature transducers and the sample container (figure 4) inside the gas station. The working principle is as follows: Sampling Method. The physical properties of natural gas in a sampling cylinder of constant volume are exactly described by the thermal equation of state. The mass in the cylinder can be calculated, if the pressure, the temperature and the composition of gas are known. Since the pressure and the temperature can be measured easily, the only unknown variable is the composition of gas for the calculation of the gas constant R and the compressibility factor Z. The composition, however, varies only in a small range, so that an averaged value can be used for calculation. This favors the evaluation of the thermal equation of state as a very effective and precise tool in gas sampler technique. At the beginning of the sampling period the microcomputer obtains the increase in mass for one sample from

320

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS

Figure 4: Installation of sample cylinder

= preselected cylinder pressure at the end of the sampling period = volume of the sampling cylinder = compressibility factor (calculated from AGA NX 19, ref. 3) = gas constant, calculated from a given composition of the gas components methane, carbon dioxide and nitrogen = temperature of the sampling cylinder at the beginning of the sampling period = number of samples Now that the mass per sample is fixed, the corresponding pressure increase p in the sampling cylinder is derived from The indices i and i+1 denote for two consecutive samples. p is generally different for every sample since T is now the actual temperature and Z depends on p and T. In order to recognize possible leakage pressure and temperature are checked before the next sample will be taken. The pressure has to correspond to the temperature at the constant volume of the sampling cylinder according to the state equation. A deviation indicates a leakage which is reported by the computer. Sampling unit. The sampling unit is connected to the gas pipeline by a line. Figure 4 shows the mechanical components of the gas sampler schematically. The essential parts are:

321

– – – – – – –

solenoid valves V1 to V5 metering valve MV Pressure regulator PR vacuum injector VI transducer for differential pressure PDI transducer for temperature TI sampling cylinder SC

The sampling line contains the micron filter MF and the oortion-volume PV between the solenoid valves V1, V2 and V3. The blow-off line, connected to the outlet of valve V3, is used to purge the sampling line and to evacuate the sampling cylinder by means of the vacuum injector VI at the beginning of the sampling period. The pressure regulator PR guarantees, that the pressure in the vacuum injector is always in the operating range. To prevent reflow from the sampling cylinder into the gas pipeline a check valve CD is mounted behind valve V2. The following metering valve MV reduces the flow rate, which results in a smooth pressure increase in the cylinder. On top of the sampling cylinder two needle valves NV1 and NV2 are mounted. To obtain a complete gas exchange by purging the cylinder, valve NV1 is equipped with a tube which goes down to the bottom of the sampling cylinder. Valve NV2 is connected to the differential pressure transducer PDI and to the pressure gauge PR, which gives a rough idea of the cylinder pressure. The operating range of the differential transducer is 500 mbar. At the beginning of each sample valve V4 is actuated, to zero the pressure difference over the transducer. The use of a differential pressure transducer in connection with a 12 Bit A/D-converter yields a high resolution of the pressure signal. Valve V5 shuts off the blow off line. As an additional security against loss of samples, caused by a possible leakage of valve V5, a hand-actuated ball valve BV2 is mounted in the line. The gas temperature, used in the determination of the pressure increase for each sample, is assumed to be equal to the cylinder surface temperature which is measured. This is a resonable assumption since the heat capacity of the cylinder is large compared with the capacity of one sample. The sampling cylinder can either be exchanged at the end of a sampling period or the samples can be transferred to a portable cylinder via the blow-off line. Control device. The control device is mounted in a standard 19" rack. It contains the power supply for the solenoid valves, the power supplies for the transducer, a 12 Bit A/D-converter and the microcomputer. The microcomputer is equipped with a battery-buffered RAM and a restart logic to restart automatically after a power fail. The faultless operation of the system is controlled by a watchdog. The device is operated by menu, displayed on a 4×40 alphanumeric LCD-display, and a keyboard with 10 numeric kevs, 5 function keys and a key operated switch to disable certain commands. At the backside of the control unit a Centronics style printer port is installed. A serial data port for communication with a host computer is optionally available. Three electrical lines reflect the states of the gas sampler and can be evaluated by a remote control station. These are: – sampling cylinder full – sampling cylinder nearly full – error

322

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS

Operation. Before the different phases of operation are described some fundamental features, of the device should be mentioned. The device has three different states of operation: the STAND-BY, the ACTIVE and the WAIT state. In the STAND-BY state the microcomputer is ready to accept the START command in order to begin the sampling period. After starting, the sampler is in the ACTIVE state. When all samples have been taken the sampler returns to the STAND-BYE state. The WAIT state is entered upon the occurence of a fatal error or by manual interaction. No samples will be taken in the WAIT state. There are two informational features which are most important. The first is the error list in which every error that occurs is entered with its error description and a time stamp. The error list provides space for 20 errors, beyond that the latest error supersedes the oldest one. Among the different error messages there are leakage, low intake pressure, malfunction of pressure and temperature probes and power fail. The second feature is the logbook in which all important events and parameters (also errors) are noted. Informations like begin and end of sampling including pressure and temperature, deviation from the calculated pressure increase, time of entering the WAIT and all operator interactions which might endanger the flow rate weighted sampling are logged. The logbook can be printed at any time to check the proceeding of the sampling period. Start of sampling period and first sample. At the beginning of a sampling period the operator can choose the number of samples to be taken and the end pressure of the sampling cylinder, which can be as much as 95 % of the lowest possible pipeline pressure. The components of H0, CO2 and density ratio can be entered according to the latest gas analysis to evaluate the composition dependent parameters R and Z. The microcomputer then calculates the mass per sample. If this mass is too small with respect to a preselected error limit the microcomputer issues a warning. The number of pulses from the state corrector or the internal clock can be entered to define the flow rate or the time between the samples. Thereafter, the device can be started to begin the sampling period. The microcomputer first checks if the cylinder is still pressurized to avoid the possible loss of unanalyzed gas. If this is not the case the device starts purging the cylinder with pipeline gas. The purging gas is blown off into the open air and the sampling cylinder takes atmospheric pressure. After the purging has been completed successfully the further proceeding depends on the level of the pressure for the first sample. If the pressure which is calculated according to equation (2) is below 1 bar (which is normally the case) the pressure in the sampling cylinder is lowered by the vacuum injector to the calculated value. The lowest possible pressure the vacuum injector can achieve is 100 mbar. If the pressure for the first sample is above 1 bar pipeline gas will be added up to this pressure. Sampling. Every time, the given number of pulses from the state corrector is counted, a sampling phase is initiated. At the beginning of this phase the microcomputer measures the pressure of the cylinder and checks if the pressure has changed significantly since the end of the last sampling phase. Pressure changes due to temperature are taken into account and not rated as leakage. If a leakage is stated the microcomputer enters the appropriate error message into the error list and the logbook and continues operation. Only in case the pressure transducer has been found faulty the microcomputer interrupts the normal operation and enters the WAIT state. At the beginning of a sample the microcomputer checks if the pressure in the pipeline is high enough to drive gas into the cylinder. This is done by passing the contents of the portion-volume (see figure 4) into the cylinder. If this does not cause a significant pressure increase in the cylinder the pipeline pressure is too low. The reason for this could be, that the end pressure of the cylinder was chosen too high or the pipeline pressure has dropped for some reason. The system then tries to recover the sample every 15 minutes. The corresponding error message will be written into the error list and the logbook.

323

Sampling continues with the continuous filling phass. Valve V1 and V2 are opened to allow the gas to flow into the cylinder until 95 % of the calculated pressure increase is reached. The system then waits until the pressure has settled within a certain limit. The remainder of the desired pressure increase is supplemented in discrete steps using the portion-volume. After each step the system waits again for a stable pressure before the next cycle is started. This process is repeated until the deviation from the calculated pressure increase is a minimum. At the end of the sampling phase valve V4 is opened to reset the differential pressure transducer to zero. As already mentioned all important data of the sampling phase are noted in the logbook in order to make the process transparent to the operator. Man/machine interface. The man/machine interface was designed to allow for an easy and fail-safe operation with a minimum training effort for the personal. The dialog with the microcomputer is performed by menu. Each menu shows clearly all functions and commands that can be activated in the actual system state. Just three keys are necessary to control the command dialog. A CURSOR is used to jump from command to command. An EXECUTE key is used to activate a command, a STOP key to terminate the command execution. All commands and messages are written in clear text. No error or function codes are used. Two different access rights can be chosen with a key operated switch. For example in the STAND-BY state with the key operated switch set to the FREE position the operator can start the sampling period, read and write all system parameters, display the actual pressure and temperature, display the error list, print out the logbook, perform test functions. A LOCKED switch allows the operator only to read all system parameters, display the actual pressure and temperature, display the error list and print out the logbook. The WAIT key can be used to force the device into the WAIT state. When a sampling phase is in progress it will be interrupted and can be resumed later by means of the CONTINUE key. This is useful for the case that for example some repair work is going on in the gas station or even at the gas sampler itself and the normal sampling procedure can not continue. The real-time/multi-tasking operating system, that runs on the microcomputer, enables parallel processing of different tasks. When for example sampling is in progress the parameter dialog can be performed as well. The ability of parallel processing is most important if the remote interface option is installed to allow the remote computer to access the gas sampler at any time. 5. SOME REMARKS ON GC Determination of gas components Figure 5 shows a block diagram of GC-analysis. The gas, which has to be analysed, is transported by means of a carrier gas (in general Helium) to the pre-cut column, containing the prechosen stationary liquid phase. An equilibrium state is achieved between stationary and mobile phase, so that according to the different partitions of the pure substances in the mixture characteristic speeds of migration can be detected. The control of carrier gas composition is done by the detector unit at the end of the column. The course of time dependent concentration can be monitored via an amplifier and penwriter by the operator. The integrator calculates the area beneath the different peaks. The detection of concentration within the carrier gas has to be without any distortion or delay. Additionally it should be mentioned, that the detector volume should not exceed that carrier gas volume, referring to the characteristic width of the smallest peak.

324

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS

Figure 5: Block diagramm of GC-Analysis

Presently at VEW still three probe injektions are necessary for one complete gas analysation. The determination of hydrocarbons higher than Propane is done on a 5,5 m, 1/8" -gaschrom- Q with 10 % silicone oil (SE 30). The separation of hydrocarbons is shown in figure 6. The analytic pre-cut column is held on a temperature of 30 C during a period of 4 min and then raised by a control program with a time gradient of 8°C/min to a value of 200 C, while the column flow is 30 ml He per minute. The detection is by flame ionisation (FID). The qualitative evaluation is done by a numerical integrator and Propane (C3) as a reference substance. The quantitative conversion of all other hydrocarbons is done by response factors. The quantitative separation and calculation of Argon/ Oxygeno (Ar/O2), Nitrogen (N2) and Methane (C4) is done by a 1,8 m—1/8" molecular sieve column under a temperature of 100ºC and a carrier qas flow of 30 ml He per minute. If a sieve column with 5 A spacing is used, i-Butane (C4) will appear before Ar, and beneath N2 also spurs of hydrocarbons will be seen. The molecular sieve 13X does’nt show this disadvantages. Carbon dioxide (CO2) and Ethane (C2) are detected behind a 3,6 m—1/8” Porapak-QIT (1:1)—column at 100°C and 30 ml He/min by a thermal-conduction-sensor TCD (figure 7). With this three injections a full scale analysation is obtained, so that the sum of all components will give 100 ±0,2 % (table 1). Although a computer evaluation in on-line mode is possible, the method described is of course not optimal. Therefore efforts are spent presently to achieve completely automatic analysation. A gas fractometer is tested with two separately controlled heating facilities. The columns are installed as figure 8 shows. Via two inlet valves two colums are fed with the following operating conditions: 1. 2 m—1/8” molecular sieve 13×, separating hydrogen .(H), Ar, N2, C4 and CO. Detection by TCD. 4 min with t=45°C then heated to 70°C with 5°C/min (figure 9). 2. 5 m—1/8” gaschrom—Q—column with 10 % silicon oil SE 30. The components CO2, C2 and C3 are separated afterwards on a 1,5 m—1/8” -Porapak -R column and detected by TCD. Beyond C3 the valve in figure 8 switches the SE 30 column to the FID so that C4 to C8 can be detected. The oven is operated at first 14 min with 35°C and then heated up with 9°C/min to 130°C (figure 9 and 10).

Table 2: Accuracy estimation Calorific value kwh/m3

January

Calorimeter

GC

9,906

9,955

+

325

Figure 6: Separation of hydrocarbons Calorific value kwh/m3

February March April May June July August September October November December

Calorimeter

GC

9 ,942 9,909 9,927 9,924 9,838 9 ,851 9,909 9,879 9 ,890 9,926 9,916

9 ,906 9,917 − 9,882 − − 9 ,984 9 ,867 9 ,867 9 ,948 9,960

− + −

+ − − + +

Calculation of calorific value and accuracy estimation. The calculation is done from the volume partitions of the pure components and their specific calorific values under standard conditions. The comparison of values calculated in this way and those obtained by a gas calorimeter is demonstrated in table 2. Probes for chromatography were taken with a sampling facility as described above. The deviation is smaller than 0,5 % and a sign test shows that systematic errors do not

326

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS

Figure 7: TCD of CO2, C2, AR, N2, C4 Table 1: Gas analysis

occur. The good accuracy shows, that GC is a suitable method for energy measurement and that all effort of manufacturers and users should be done, to achieve an official licensation for this method in FRG. REFERENCES CITED /1/

Herning, F., Wolowski, E., “Kompressibilitatszahlen und Realgasfaktoren von Erdgasen nach neuen amerikanischen Berechnungsmethoden”, ges. Ber., Ruhrgas 15, (1966)

327

Figure 8: Switching over pre–cut colums /2/ Herr, E., Scheibe, D., Schröder, P., Voss, K.F., Weimann, A., “Rechnergestutzte Zuordnung von an den Einspeisepunkten eines Ferngasnetzes vorgenommenen Brennwertmesssungen zu den an den Übergabestationen entnommenen Gasmengen”, gwf-gas/erdgas, 124–3 (1983) /3/ Homann, K., “Rechnerführung eines ausgedehnten Gashochdrucknetzes”, gwf-gas/erdgas, 122–7 (1981) /4/ Krabbe, H.J., “Einsatz der Gaschromatographie im Kraftwerkslabor”, VGB Kraftwerkstechnik, G1–1, (1981) /5/ Weimann, A., “Modellierung und Simulation der Dynamik von Gasverteilnetzen im Hinblick auf Gasnetzführung und -überwachung”, Thesis, Technical University of Munich, 1978, July 7

328

LONG TERM GAS SAMPLING AND HEATING VALUE CALCULATIONS FROM GC-ANALYSIS

Figure 9: TCD N2, CH4, CO2, C2, C3

329

Figure 10: TCD, C4…

HEAT QUANTITY DETERMINATION IN LONG DISTANCE GAS TRANSMISSION Georg Strulik, Dipl.-Ing. Thomas Fischer, Dipl.-Ing. GAS QUALITY and QUANTITY CONTROL BEB ERDGAS UND ERDOL GMBH Riethorst 12, 3000 Hannover 51 (Germany)

ABSTRACT

Due to a changed situation, both with imported natural gas and with our own production, it is not always possible to maintain a constant gas quality. Because the possibilities of conditioning and distribution are limited, security of supply must take absolute priority. In this paper, measuring systems are described which are able to determine heat quantities accurately with the help of certain processing routines. These systems will assume ever-greater importance on account of today’s energy prices. Besides the measuring systems, possibilities of measurement recording and processing are described. The following procedures are discussed: – – – –

Determination of characteristic gas quality data. Measurement of gas volume taking gas quality into account (density-correction). Comparison of measurement results with respect to forecasted an actual gas quality. Determination of heat quantities by linking volume and calorific value on various time bases (monthly, daily, hour– Heat quantity determination dependent on gas throughput. – Continuous heat quantity determination through the use of energy computers. INTRODUCTION

Three fourth of the natural gas consumed in Western Europe came from indigenous production, one fourth from imports (Fig. 1). Some 15 % are imported from the USSR and about 9 % from Algeria and Libya.

331

Figure 1. West European Natural Gas Supplies

As can be seen from Fig. 2, the gases are fed into the West European supply system at very different locations. Characteristic of these gases is the wide variation in quality. Upper calorific values range from about 8.5 kWh/m3 to 12.3 kWh/m3. The Federal Republic of Germany has consumed in 1985 about 55×109 m3 natural gas, 30 % from indigenous production and 70 % from imports. Some 30 % are imported from the Netherlands, 25 % from USSR and about 15 % from Norway and Denmark (Fig.3). BEB is the largest gas producer in the Federal Republic of Germany and the second largest gas transmission company. The German Shell and the German Esso (Exxon) companies are the share-holders of BEB. Of the total of about 15 billion m3 (1985) of natural gas, which are handled by a more than 3,000 km long high-pressure pipeline system, some 10 billion m3 stem from our own production (Fig. 4). Five billion m3 are supplied by purchases from the Netherlands and the Norwegian and Danish North Sea sectors. GAS DISTRIBUTION BEB’s gas transport system includes several compressor stations, tie- in and crossover points as well as cavern and pore storage pools. Almost the whole of the pipeline system has been looped, so that high and low-calorific gas can be transported separately (Fig. 5). H gas imports from the Norwegian and Danish North Sea sectors and L gas imports from the Netherlands are fed into these systems. Our own gas fields are mainly located in the South Oldenburg and East Hannover areas. Some of the ’gases produced in these areas are strongly sour with H2 S contents of up to 25 % v or have N2 contents of up to 70 % v. This makes it necessary to operate large-scale gas purification plants. In South Oldenburg, a high-calorific gas is obtained by means of Purisol and Sulfinol purification processes, with some 500,000 tons of elemental sulphur per year being obtained as a by-product. In East Hannover, nitrogen is removed from the gas in a low-temperature plant in order to get a salable L gas of the 1st gas family in accordance with DVGW specification G 260 (ref. 1). Furthermore, high-calorific gases are conditioned with gas having a very low calorific value (Apeldorn, Ho< 4 kWh/m3) so as to obtain an L gas of Slochteren quality. Despite these measures being taken on the supply side, different quality ranges involving changing and fluctuating gas qualities are encountered on the demand side in gas distribution. In addition, some areas show the phenomenon of fluctuating quality ranges.

332

HEAT QUANTITY DETERMINATION IN LONG DISTANCE GAS TRANSMISSION

Figure 2. European Natural Gas Transmission System

Figure 3. Natural Gas Supply of the Federal Republic of Germany

GAS QUALITIES The composition of the natural gas described above differs considerably in many cases in that different percentages of hydrocarbons and inert gas components are found. Up to 20 different gas qualities are fed into the trunkline system. Seasonal variations in the gas composition are due to the provisions contained in the import contracts, which specify a firm offtake pattern over the delivery periods, and to seasonal fluctuations in demand, i.e. periods of high consumption in winter and periods of low consumption in summer. Fig. 6 shows the annual load curves for the upper calorific values in H and L gas and in the low-calorific L gas systems. In the H gas system the calorific value may vary as much as ± 600 Wh/m3 during any one gas business year. In the L gas system, the range is ± 150 Wh/m3 , and in the low-calorific L gas system, ± 350 Wh/m3 .

333

Figure 4. Natural Gas Supply of BEB

MEASURING SYSTEMS In BEB’s gas distribution system, gas is delivered via some 220 customer offtake stations, where the volumes are measured. Orifice meters, turbine meters, rotary displacement meters, vortex meters and high-pressure diaphragm meters are used for volume measurement at flowing conditions. The different volume measurement systems are selected according to throughput, measuring range and offtake pattern (ref. 2). The volumes can be corrected to normal conditions (1.01325 bar, 273.15 K) with the following instruments: ideal gas corrector, real gas corrector (PTZ corrector) and density flow meter. The ranges of application are fixed according to measuring pressure and volume throughput. Table 1 shows the systems used for the different measuring ranges. Gas quality measurement in BEB’s system is performed for upper calorific value, Wobbe index, density, CO2 and H2S. These measurements are carried out at representative, strategically located points of the pipeline system. In addition to continuous measurements along the pipelines, gas samples are analyzed in BEB’s laboratory. Normally the upper calorific values are recorded with automatically operating calorimeters. Gas density is measured with density transmitters or balances. Special measuring equipment is used to determine the CO2 and H2 S contents. All other gas components are determined by means of gas chromatography from analyses of samples collected over a month or samples taken individually. PROCESSING OF MEASURED DATA The measured gas quality data are used to determine the compressibility behaviour of the gases and their heat volume (ref. 5). In our supply area, the deviation of real from ideal gas behaviour—the so-called compressibility behaviour—has the following consequences for the determination of the normal volume: For pü=1 bar the correcting quantity is within the accuracy of the K value calculation method (0.25 %) and is therefore neglected (Fig. 7).

334

HEAT QUANTITY DETERMINATION IN LONG DISTANCE GAS TRANSMISSION

Figure 5. BEB’s Transmission System

Figure 6. Annual Load Curve (84/85) of Three Gross Calorific Values

For measuring pressures of 4 bar, the K value influence is around 1 %. Up to this pressure range, ideal gas correctors with subsequent manual K correction can safely be used. At higher measuring pressures, correction systems should be employed which give constant consideration to the compressibility. The K value calibrated into real gas correctors is determined by means of a forecasted gas quality and is checked continuously. Deviations< 0.1 % are tolerated. In the case of greater deviations, the annual volume is

335

Table 1. Correction of Natural Gas/Flow Measurement

corrected retroactively. Table 1 shows the correction systems used by BEB, distinguished by measuring pressures and volume throughput. It can be clearly seen that for higher measuring pressures the K value influence is already largely taken into account when making the corrections. As can be seen from Fig. 8 for high measuring pressures and large volume throughputs, density flow meters are mainly employed, which take the compressibility factor into account directly in the correction process, even in the case of wide variations in gas quality. MEASURING ERRORS It has to be mentioned, however, that the density flow meters currently in use are subject to systematic deviations resulting from the velocity-of-sound effect. The deviations of our sales gas qualities are around 0.2 %. This can be corrected or avoided by means of a calculation procedure or by using density transmitters with low frequencies of oscillation. Another possibility of correcting the velocity-of-sound effect is a recalibration of the transmitters or direct sound velocity measurement. The above-mentioned measuring errors are small, however, in comparison with errors already made when measuring volumes at flowing conditions. In Germany, the calibration error limit when measuring air under atmospheric pressure in the permissible measuring range is ± 1 %. The error may be much greater with measuring conditions involving natural gas. Such errors can, however, be minimized by means of a highpressure calibration of gas meters. A high-pressure calibration has the additional advantage of extending the permissible measuring range, thereby also reducing the minimum flow problem. HEAT VOLUME DETERMINATION Our aim should be—also in a time of fluctuating gas qualities and with a view to the current energy situation —to provide an accurate and economic procedure for heat volume determination of natural gas which is also suitable for the considerable number of small stations (ref. 4).

336

HEAT QUANTITY DETERMINATION IN LONG DISTANCE GAS TRANSMISSION

Figure 7. K-Factor-Influence Taking into Account Different Gas Qualities

In order to avoid a possibly premature installation of expensive calculating equipment in existing stations, BEB has been putting magnetic tape recorders into the gas meter stations since 1978. These cartridge-type units store all measured data of a station needed for invoicing purposes over a period of one month, either in the form of hourly aggregates or hourly averages, and make such data available for subsequent interpretation and correction in a mainframe computer. Apart from the measurement of basic data such as volumes and gross calorific value, as well as analytical data for K value determination, the formation of the heat volumes is also problematical. The results obtained with the existing invoicing method (monthly volume x arithmetic average of monthly calorific values) differ considerably in some cases from heat volume formation based on hourly or daily figures. In view of the prevailing offtake pattern and quality fluctuations, heat volumes based on daily averages of calorific values are almost identical with those based on hourly measurements (Fig. 9, 10, 11, 12). For a number of customers, who conduct synchronous volume and calorific value measurements, BEB has carried out heat volume comparisons on the basis of hourly, daily and monthly calorific values. The daily average of calorific values is of particular importance if stations are not equipped with calorimeters, so that one must rely on the calorific value measurement of an adjacent station (so-called master calorimeter) (ref. 3). Meanwhile the technology of continuous local heat volume formation has been improved to such an extent that systems of this kind are permitted under the applicable standardization regulations and are in use. This method employs on-site heat volume calculators which continuously (i.e. at very small intervals) link the measured volumes with the upper calorific values applicable at the same time. This makes the detection of errors in the case of malfunctions or deviations much more difficult than with “conventional” methods of heat volume formation.

337

Figure 8. Number of Employed Corrector-Systems Taking into Account Pressure and Volume and Flow

Figure 9. Load Curve and Heat Quantity Deviation of Local Public Utility

MINIMIZATION OF MEASURING ERRORS In view of energy costs, the minimization of measuring errors is gaining ever greater importance. Whenever possible, balancing of supplies and offtakes is a suitable method to detect measurement errors, although this is restricted to areas with an identical quality. As in many cases suitable balancing areas are not available and balancing measurements are often too expensive, the single measurement must be subjected to a suitable checking procedure. Most of the volume meter runs in BEB’s system have therefore been equipped with a standby run. By means of suitable piping, the main meter run and the standby run can be connected

338

HEAT QUANTITY DETERMINATION IN LONG DISTANCE GAS TRANSMISSION

Figure 10. Load Curve and Heat Quantity Deviation of a Transmission Company

Figure 11. Load Curve and Heat Quantity Deviation of an Industrial Consumer

in series. In order to avoid continuous operation of the standby meter runs and to have them ready for use in the case of malfunctions, such series connections have so far in most cases been made only two or three times a year. This results only in random results depending on the meter loads prevailing at the respective times. Major deviations can be detected in this way. Also corrector checks at certain intervals belong in this category. It is very difficult to determine a drifting of gas meters or correctors, and even impossible to record the time of mal-functioning. For this reason the German gas industry has for some time been testing or applying continuously operating checking systems. In most cases they consist of two different metering systems (e.g. vortex meter and turbine meter) which are connected in series. If a calculated deviation is exceeded, an alarm gives the information, that the measurement equipment should be tested imediately. For stations with small throughputs a standby meter run is not required. If the throughput is, however, equal to or greater than 150,000 m3/h (Vn), it is advisable for economic reasons to provide also for independent operation of the meter runs in the case of a malfunction. In additon to continuous ΔVn onitoring of the meters, BEB is currently making continuous corrector checks on a test basis in small metering facilities. The meter runs are equipped with additional pressure and

339

Figure 12. Load Curve and Heat Quantity Deviation of a Power Station

temperature transmitters, whose measuring results are recorded on magnetic tape as hourly averages together with the volumes at flowing conditions. After exchanging the magnetic tape cartridges, the hourly volumes at flowing conditions are calculated by the central data processing function, taking into account the K value, and are compared with the normal volume per hour from the corrector, which is also stored on a recording medium. In this way it is possible to reduce off-normal periods to one hour and detect any drift at a very early stage. As soon as such a trend has been determined, a meter check is made on site. Such a system does not require any routine checks. This applies in particular to the combined operation of computers in gas offtake stations with a central computer, periodic calling up of measured data from the stations via acoustic coupler and remote data transmission, as well as automatic preparation of measured values in the offtake stations with the aid of small computers to the greatest possible extent. However, gas offtake regulation by the customer continues to be a problem. When determining actual capacity offtakes, an exact heat volume dispatching is of crucial importance, in particular in view of the very high capacity charges. Customers without calorific value metering can regulate their offtakes only by means of a pre-set average calorific value. In view of the necessary safety margins it is not possible to make economically optimum use of the contractually provided volumes, even when an arithmetic average of the upper calorific values is used for invoicing purposes. An accurate heat volume regulation is only possible if the periods for heat volume formation are shorter. This calls for investments, however, in calorific value metering, small computers and the relevant software. Such regulating procedures, e.g. also by process computer, are only economically justified in the case of large throughputs. SUMMARY Summarizing it can be said that the various different calibrating standard legally prescribed in Europe specify only minimum accuracies for metering equipment. However, mainly because of the cost of energy, both the gas industry and the customers are interested in arriving at much more accurate measuring results. At the same time it must be ensured that the steadily increasing number of data can be provided at ever shorter intervals so as to be able to meet the contractual payment dates and avoid interest-related losses.

340

HEAT QUANTITY DETERMINATION IN LONG DISTANCE GAS TRANSMISSION

In this connection, systems meeting the requirement of reducing the costs of local meter checks in the stations, thereby making energy measurement more economic, will gain more and more importance in the near future. REFERENCES CITED 1 2 3 4 5

DVGW G 260/1 Gasbeschaffenheit DVGW G 492/II Groβ-Gasmessung DVGW G 686 Thermische Gasabrechnung G. Strulik, “gwf-gas/erdgas” 124/83, Vol. 3, p. 165 pp. PTB, Richtlinie G 9

STATUS OF FEDERAL REGULATION OF NATURAL GAS AND ITS IMPLICATIONS FOR FIELD APPLICATIONS OF ENERGY MEASUREMENT Paul D.Hubbard Special Assistant to the Director* Office of Pipeline and Producer Regulation Federal Energy Regulatory Commission Washington, D.C.

Introduction Changes in the federal regulation of natural gas can have implications for energy measurement. A particularly significant example relates to the procedures for determining the quantity of energy units involved in wellhead sales and other “first sales” of natural gas. That feature was the subject of the Federal Energy Regulatory Commission’s so-called “BTU Rule,” which was issued to implement the wellhead pricing provisions of the Natural Gas Policy Act of 1978. This paper reviews some of the technical aspects of the BTU Rule and discusses the expected future implications of federal regulation on natural gas energy measurement. I. The Natural Gas Policy Act of 1978 A. Reasons for NGPA 1. Gas shortages of early and mid-1970s 2. Price differences between interstate and intrastate markets B. Effects of NGPA 1. Merging of interstate and intrastate markets 2. Implications for energy measurement

342 STATUS OF FEDERAL REGULATION OF NATURAL GAS AND ITS IMPLICATIONS FOR FIELD APPLICATIONS OF ENERGY MEASUREMENT

II. Wellhead Pricing Approaches A. The former Federal Power Commission (FPC) volume-referenced prices 1. Development of “base” prices 2. Method of adjusting for variances in heating value B. NGPA energy-referenced prices 1. Carry-over of former prices set by FPC 2. Comparisons with oil and other energy sources 3. Incentive prices III. The “BTU Rule” A. Effects of “Standard Conditions” for determining heating value 1. Imputed 820 pounds of water vapor per million cubic feet of total gas 2. Physical limit on guantity of water vapor 3. Contract limit of 7 pounds of total water per million cubic feet of gas B. “As-delivered” basis for determining water vapor content and resulting guantity of BTUs IV. Concerns over effects of BTU Rule A. Rising gas prices in late 1970s and early 1980s B. Overall potential dollar impact of BTU Rule V. Court review of BTU Rule A. D. C. Circuit Court decision 1. Recognition of scientific accuracy 2. Interpretation of Congressional intent B. U.S. Supreme Court denial of review C. Reguirement of refunds

*Currently (May 1986) Deputy Director in an acting capacity.

343

VI. Implications of Future Regulation for Natural Gas Energy Measurement A. Decreasing amounts of price-regulated gas B. Relationship between ceiling prices and market prices Conclusion The controversy surrounding the so-called “BTU Rule” has been resolved by the U.S. Courts. Perhaps a lesson to be learned from this controversy is that, even when a U.S. Court may be impressed by scientific accuracy, the court has the “last word” because it interprets the law. This does not mean, however, that technical accuracy should not be strived for in natural gas energy measurement. What it does mean is that, when the scientific community has imput in the formulation of proposed legislation and regulations, technical advise should be clear and unbiased. If this approach is followed, scientific accuracy could be fully considered and accounted for. The overall result should be that scientific accuracy and all other relevant considerations are properly balanced in laws and regulations. With the passage of time, an increasing portion of natural gas will not be governed by existing price regulations. Additionally, some of the Maximum Lawful Prices under the Natural Gas Policy Act are above current market clearing prices and some contract prices. This relationship should continue for at least as long as oil prices remain low and gas inventories remain high. Where federal price ceilings do not exist, or where price ceilings are above market prices, contracting parties are free to adopt mutually acceptable provisions to accurately account for energy measurement. Such provisions would, however, have to be coupled with pricing provisions which recognize and adjust to market competition. The logical implication is that improved accuracy in natural gas energy measurement will be increasingly important.

Note: All of the comments by the author are offered as general information for discussion purposes. They are not presented as an official view of the Federal Energy Regulatory Commission. Neither are they intended as a specific recommendation on any matter or issue.

344 STATUS OF FEDERAL REGULATION OF NATURAL GAS AND ITS IMPLICATIONS FOR FIELD APPLICATIONS OF ENERGY MEASUREMENT

Ref: Adapted from Katz, Cornell, Kobayashi , Poettmann, Vary, Elenbaas and Weinaug, Handbook of Natural Gas Engineering (McGraw-Hill Book Company, Inc., New York, Toronto, London, 1959). Reprinted with permission © 1959 McGraw-Hill Book Company, Inc. GPO 879 235