Performance of Generating Plant: New Realities, New Needs
A Report of the World Energy Council
AUGUST 2004
Performance of Generating Plant: New Realities, New Needs Copyright 2004 World Energy Council 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, electrostatic, magnetic, mechanical, photocopy, recording or otherwise, without prior permission of the copyright holder. Published August 2004 by: World Energy Council 5th Floor, Regency House 1-4 Warwick Street London W1B 5LT United Kingdom www.worldenergy.org ISBN 0 946121 19 2
ii
ACKNOWLEDGMENTS The WEC Committee on the Performance of Generating Plant is perhaps the best known WEC Committee, both within and outside the WEC family. It is the leading, actionoriented and “hands-on” work programme item reporting to the WEC Programme Committee. As the Chair of the Programme Committee over the past three years, I have been able to follow closely the work of the PGP Committee, and have attended several presentations and events organised by the Committee. These presentations never fail to impress participants by delivering a simple and clear message and demonstrating a whole range of benchmarking options, which can help improve power plant performance, both technical and more recently also economic, without requiring significant investments. The current report is a collection of papers describing some of the Committee’s activities. However, the PGP Committee is not in the business of producing reports. Its focus is on action, and this is clearly reflected in the papers. The most recent Committee initiative the development of the Internet-based global database of power plant performance statistics – is an additional avenue to reach out to utilities and companies around the world and offer them a valued proposition. The real value of collecting, evaluating and exchanging power plant performance data is not always well understood, and the Committee goes a long way towards helping to improve this understanding. I would like to thank all the dedicated members of the PGP Committee, and in particular Dr. Karl Theis, the Committee Chairman who has been leading the Committee in a dynamic and sophisticated manner for the past two years. Last but not least, my special thanks go to Elena Nekhaev, WEC Director of Programmes, for all the support and guidance she has provided.
Norberto de Franco Medeiros, Chairman WEC Programme Committee Rio de Janeiro, August 2004
iii
Officers of the World Energy Council Antonio del Rosario Chair World Energy Council
Norberto de Franco Medeiros Chair Programme Committee
Philip Aiken Vice Chair Sydney 2004
Shige-etsu Miyahara Vice Chair Asia
François Ailleret Chair Studies Committee
Kieran O’Brien Vice Chair Europe
Asger Bundgaard-Jensen Vice Chair Finance
Fred Phaswana Vice Chair Africa
John Derrick Vice Chair North America
Carlos Pierro Vice Chair Latin America/Caribbean
Alioune Fall Vice Chair GEIS Initiative
Gerald Doucet Secretary General
Member Committees of the World Energy Council
Algeria Angola Argentina Australia Austria Bangladesh Belarus Belgium Bolivia Botswana Brazil Bulgaria Cameroon Canada China Congo (Dem. Rep.) Côte d’Ivoire Croatia Czech Republic Denmark Ecuador Egypt (Arab Rep.) El Salvador Estonia Ethiopia Finland France Gabon Georgia Germany Ghana Greece
Guinea Hong Kong, China Hungary Iceland India Indonesia Iran (Islamic Rep.) Ireland Israel Italy Japan Jordan Kenya Korea (Rep.) Latvia Lebanon Libya/GSPLAJ Lithuania Luxembourg Macedonia (Rep.) Mali Mexico Monaco Mongolia Morocco Namibia Nepal Netherlands New Zealand Niger Nigeria Pakistan
iv
Paraguay Peru Philippines Poland Portugal Romania Russian Federation Saudi Arabia Senegal Serbia & Montenegro Singapore Slovakia Slovenia South Africa Spain Sri Lanka Swaziland Sweden Switzerland Syria (Arab Rep.) Taiwan, China Tanzania Thailand Trinidad & Tobago Tunisia Turkey Ukraine United Kingdom United States Uruguay Venezuela Yemen
PGP Committee Membership Chairman:
K Theis
VGB PowerTech e.V (Germany)
Members: Brazil: China: Egypt: France:
S.R. Fernandes Eletronuclear W. Guangyao State Power Corporation M. Sharaf /K. Yassin Egyptian Electricity Authority B. Manoha Electricité de France D. Glorian Consultant Germany: C. Reese RWE Energie AG Hungary: T. Tersztganszky Electricity Licensing India: M. Mohan Central Electricity Authority B. N Ojha NTPC Indonesia: D. Prasetijo PT PLN Iran: Dr A Karbassi Ministry of Power Ireland: G. McMahon/ ESB Power Generation M. Kelly ESB National Grid Italy: L. Salvaderi Consultant Japan: A. Torii/ Toshiba Corporation A. Nakanishi Kyushu Electric Power Co Jordan: H. Khatib Consultant Korea (Republic): Ki-Yoon Lee/Chi-Wan Kim KEPCO Mexico: E. M Vela’zquez Comision Federal de Electricidad Poland: A. Jaworski Elektrociepłownie Warszawskie S.A Romania: I. Marcu Electrocentrale Bucuresti SA Russian Federation: G. Olkhovsky All-Russian Thermal Engineering Inst. Saudi Arabia: A.A. Al-Tuwaijiri Ministry of Industry and Electricity South Africa: V. Micali/T. Moss ESKOM Spain: A. Lopez de Sebastian UNESA Syria M. K. Sheki Ministry of Electricity Switzerland: M. Seifert Gas & Water Industry Association B. Basler Alstom Power Thailand: U. Khunvichai/J Ruangsup-anek Thai Electricity Generating Authority Turkey: M. Kesim/ TEAS-TEK B. Celik Eregil Iron & Steel Works Co. United Kingdom M. Grasby PowerServe Ltd USA: G.M. Curley North American Electric Reliability Council R.R. Richwine Reliability Consultant G.S. Stallard Black & Veatch International Organisations: CIER: J. C Alvarez Salomón IAEA: M. Szikszaine Tabori IBRD: M. Takahashi Corresponding Members: France D. Carbonnier EdF Hong Kong, China: C.P. Yuen China Light & Power Co. Ltd Korea M-J Kim Korea Southern Power Y-K Park Korea Midland Power (KOMIPO) USA S. Hanawalt Calpine Corporation Secretary: E.V. Nekhaev (WEC Director of Programmes)
v
TABLE OF CONTENTS Page ACKNOWLEDGEMENTS WEC OFFICERS AND MEMBER COMMITTEES PGP COMMITTEE MEMBERHIP FOREWORD
INTERNATIONAL AVAILABILITY DATA EXCHANGE FOR THERMAL GENERATING PLANT
iii iv v vii
Section 1
G. S. STALLARD, Black & Veatch (US), R. RICHWINE Consultant (US) V. MICALI ESKOM (South Africa)
THERMAL GENERATING PLANT UNAVAILABILITY FACTORS AND AVAILABILITY STATISTICS
Section 2
M. CURLEY, North American Electric Reliability Council (US)
NUCLEAR GENERATING PLANT UNAVAILABILITY FACTORS AND AVAILABILITY STATISTICS
Section 3
M. SZIKSZAINE TABORI, International Atomic Energy Agency (IAEA)
PERFORMANCE OF HYDRO AND PUMP STORAGE PLANT
Section 4
T. MOSS, Eskom Generation (South Africa)
PROPOSAL OF TECHNICAL, ENVIRONMENTAL AND SOCIOLOGICAL PERFORMANCE INDICATORS FOR RENEWABLE ENERGY SOURCES
Section 5
B. MANOHA, Electricité de France (France) M. HOPPE-KILPPER, Institut für Solare Energieversorgungstechnik (Germany) R. VIGOTTI, ERGA of ENEL (Italy) E, HUGHES, Electric Power Research Institute (USA)
WORKSHOPS AND COMMUNICATIONS CASE STUDIES OF THE MONTH R. R. RICHWINE, Consultant (US)
vi
Section 6
FOREWORD The WEC Committee on the Performance of Generating Plant is a well-established institution in the global power generation community represented by WEC Member Committees all over the world. Since 1974 the Committee has compiled, analysed and published performance data of power generating plants worldwide. I was delighted to join this unique group three years ago and participate in its activities, including benchmarking workshops and presentations, which have revealed to me some new dimensions in the operation of power plants. Availability is a critical indicator for assessing the overall performance of the power plant, both in technical and commercial terms. Moreover, it is a public demonstration of the service the plant provides to its customers. The importance of reliable service should not be underestimated, in particular in the increasingly competitive market environment in which many utilities around the world are operating today. The service provided by the industry – electric power – is not considered to be particularly exciting when things run well and the lights are on. Only when the service is not there, does it become exciting and hits the headlines. Only then do customers begin to understand and appreciate the real value of secure access to electricity and the full extent to which modern society depends on reliable supplies of electric power. The importance of power plant performance is poorly known to industry outsiders, although it is one of the major factors which could have a significant impact on the future of our planet. Analysis of generating plant performance data undertaken by the Committee demonstrates the enormous value of plant availability. It has been estimated that improving the availability of all power plants in the world to the performance levels achieved today by the 25% of best performing plants, is worth a staggering US$80 billion per year. In addition, this improvement in performance would reduce the annual global GHG emissions by 1 billion tonnes CO2 equivalent (i.e. by approximately 4%), along with proportional reduction of other pollutants. This could be achieved using existing best practice technologies at an average benefit to cost ratio of 4 to 1. Case studies from utilities and manufacturers around the world which are included in this report confirm that while some technology enhancements and equipment upgrades will be required, the majority of the improvement will come as a result of addressing human factor issues and power plant management. Moreover, if these “soft” issues are not addressed, new technology plants will be unable to achieve their inherent superior performance potential. Performance improvement of existing power plants is the most cost-effective way to increase the energy producing capabilities of a utility while improving the overall energy efficiency of the industry and producing substantial environmental benefits. To facilitate the international cooperation and information exchange, the PGP Committee has developed the first phase of an Internet based global database of performance indicators for fossil-fired, nuclear, hydro and other renewable power plants. Companies and utilities are invited to join this initiative by registering their data in the database on a fully confidential and anonymous basis. A joint global effort is required to achieve an effective allocation and use of the global energy resources. The WEC and its Committee on the Performance of Generating Plant are taking action to face up to this challenge. Dr. Karl A. Theis Chair Germany, August 2004
vii
Performance of Generating Plant Section 1 INTERNATIONAL AVAILABILITY DATA EXCHANGE FOR THERMAL GENERATING PLANT
G. S. STALLARD Black & Veatch (US) R. RICHWINE Consultant (US) V. MICALI ESKOM (South Africa)
World Energy Council
Work Group Membership: G.S. Stallard (US), Chair K. Yassin (Egypt) D. Glorian (France) C. Reese (Germany) T. Tersztgansky (Hungary) S. Arafin (Indonesia) G. McMahon (Ireland) A. L. de Sebastian (Spain) R. Spiegelberg-Planer (IAEA) until 01.04.2004 M. Szikszaine Tabori (IAEA) from 01.04.2004
Performance of Generating Plant 2004 - Section 1
World Energy Council
Performance of Generating Plant 2004 - Section 1
TABLE OF CONTENTS TABLE OF CONTENTS ..................................................................................................................... 1 INTRODUCTION ................................................................................................................................ 3
Terms of Reference ...............................................................................................................4 INDUSTRY OUTLOOK...................................................................................................................... 5
1.1 US Situation...............................................................................................................5 1.2 The Europe Situation .................................................................................................8 1.2 Africa Situation........................................................................................................11 Introduction ..................................................................................................................... 11 1.2.2 Angola ............................................................................................................. 11 1.2.3 Botswana ......................................................................................................... 11 1.2.4 Kenya............................................................................................................... 11 1.2.5 Namibia ........................................................................................................... 12 1.2.6 Nigeria ............................................................................................................. 12 1.2.7 South Africa..................................................................................................... 12 1.2.8 Sudan ............................................................................................................... 13 1.2.9 Tanzania........................................................................................................... 13 1.2.10 Uganda............................................................................................................. 14 1.2.11 Zambia ............................................................................................................. 14 1.2.12 Zimbabwe ........................................................................................................ 14 VALUE OF PERFORMANCE/AVAILABILITY DATA .............................................................. 15
2.1 The Historic Problem...............................................................................................15 2.2 The Concept of Commercial Availability ...............................................................17 2.3 Commercial Availability Methodologies ................................................................17 2.3.1 Method 1.......................................................................................................... 17 2.3.2 Method 2.......................................................................................................... 18 2.3.3 Method 3.......................................................................................................... 18 2.3.4 Method 4.......................................................................................................... 18 2.3.5 Method 5.......................................................................................................... 19 2.4 Banked Availability Value Method .........................................................................19 2.5 Other Commercial Availability Alternative Measures ............................................21 2.6 Estimating Commercial Availability .......................................................................21 2.6.1 Data Requirements .......................................................................................... 22 2.6.2 Calculation Methods........................................................................................ 22 2.7 Implications of Using Commercial Availability .....................................................23 OTHER ASPECTS OF MEASURING AND OPTIMISING PLANT PERFORMANCE.......... 25
3.1 3.2 3.3
Addressing Risk/Returns of Coal Sourcing on Plant Availability and Production Costs......................................................................................................26 Evaluating Physical Positions of Underlying Assets within Power Trading Framework...............................................................................................................27 Market-Based View of Asset Management.............................................................29
PGP DATA COLLECTION AND ANALYSIS............................................................................... 33
4.1
WEC Goal: A Global Data Collection and Analysis System ..................................35
1
World Energy Council
Performance of Generating Plant 2004 - Section 1
REFERENCES ................................................................................................................................... 36 CASE STUDIES ................................................................................................................................. 37
5.1 5.2 5.3
Use of Conditional Probability Methods to Evaluate Commercial Availability .....37 South Africa’s ESKOM Experience: Commercial Availability Application ..........41 The Italian Power System: Recent Evolution & Issues ...........................................45
APPENDIX 1:
A Demand-related EFOR ..................................................................................... 53
APPENDIX 2:
NERC GADS ......................................................................................................... 56
APPENDIX 3:
WEC PGP COMMITTEE DATA COLLECTION ........................................... 61
1.1 1.2 1.3 1.4
Governing Design Principles............................................................................... 61 The First Phase .................................................................................................... 64 “Peer” Group Data Entry..................................................................................... 66 Single Unit Data Entry ........................................................................................ 67
2
World Energy Council
Performance of Generating Plant 2004 - Section 1
INTRODUCTION During the three years (2001 – 2004) covered in this report, the global power market continued to evolve, however, in a somewhat different direction than in the previous review period. While earlier the evolution predominantly focused on the introduction of competitive and market-oriented frameworks/rules, the more recent trends within the global energy sector are neither clear nor consistent. In fact, following the collapse of Enron and the resulting negative financial impact on independent power production in the US, the migration towards fully deregulated wholesale market in the US has stalled. Today, there is a mixture of traditional utilities, cooperative, and municipal power suppliers; multi-regional independent power producers, and niche players in the US power market. Mechanisms for managing unit selection/bidding, dispatch/commitment, and system reliability issues vary significantly across the regions. Arguably, the US situation has deeply affected power markets and planned or ongoing power industry privatisation initiatives around the world. Generally, power markets in many countries demonstrate the same lack of consistency. An examination of industry trends, actions, and reactions at a higher level, leads to several conclusions. First, capital remains at a premium – hence, getting greatest returns vs. investing long-term in either new or existing plant remain a global concern. Second, fuel selection, energy efficiency and environmental requirements are becoming increasingly interlinked. Third, in most electricity markets, there are commercial entities focused on delivering profits to their shareholders. This creates new expectations and pressures, and even traditional power producers, whether owned by governments or private regulated integrated utilities feel the need to •
address performance within a financial context;
•
manage operational risk, and
•
address environmental/political pressures.
All over the world, plant operators are taking on a broader, business-minded role in today’s energy supply industry (ESI). New industry drivers geared toward profitability, cost control, environmental performance and market economics are shifting the focus away from traditional measures of technical excellence such as availability, reliability, forced outage rate, and heat rate toward the balance sheet bottom line. As the issue of generating plant performance evaluation becomes increasingly complex, the ability to “measure” and analyse this performance is getting even more challenging. There is no clear “right” answer on how to address this issue; different entities, different facilities, different markets, and different obligations will yield different needs. It should not be surprising, therefore, that the ESI seems to be at the crossroad, both in terms of how it measures its performance, but also in terms of data or information necessary to support such measures. From a historical perspective, a plant’s performance was measured within a regulated framework wherein a guarantee to purchase heat and/or power was made. Therefore, performance was largely measured on the ability to deliver the required load first (i.e., the obligation of supply) and second, on the effectiveness of the generating plant to make money. Hence, traditional performance indicators measuring availability, planned outage 3
World Energy Council
Performance of Generating Plant 2004 - Section 1
rates, and unplanned or forced outage rates were regarded as good and adequate measures.
Terms of Reference Prior to 2001, Work Group 6 (WG6) viewed the need to consider and understand the commercial implications of power plant performance and availability as a “new” and separate task, independent of PGP Committee’s traditional role in developing and reporting industry performance statistics and best practices. From 2001, the Work Group shifted its focus toward defining how to integrate fully the needs and realities of commercial performance and availability into the overall PGP Committee’s mission and data collection and analysis work. During its 2001-2004 term, WG6 has also looked at the related issues, in particular at how to promote effectively “International Availability Data Exchange for Thermal Generating Plant” within the realities presented by the increasingly complex and diverse energy markets. WG6’s goal is to improve the value of international power plant availability data exchange in the increasingly competitive global power sector by further evaluating “commercial availability” issues, indicators and definitions, cost comparisons, and benchmarking methodologies. The report will focus on the following: 1. World power industry outlook with respect to the “value” of unit performance/availability. The availability and compatibility of various data sources (including traditional reliability databases such as the North American Electric Reliability Council’s Generating Availability Data System (NERCGADS), the International Atomic Energy Agency’s Power Reactor Information System (IAEA PRIS), Euroelectric’s database, etc., indices, and assessment practices will also be addressed. 2. As power generators strive to understand better what is “best of class” performance, the need to evaluate and benchmark plant performance across industry is growing – in direct opposition to the decreasing willingness to share key data with others in the industry. This has resulted in a gaping hole that currently exists within the data collection and analysis toolsets available worldwide and is the basis for the PGP Committee’s mission to meet that need via new data collection process/database/toolsets that can over time evolve in complexity and power. 3. “Commercial Availability” (CA) concepts are gaining popularity. There is a number of frameworks currently in use in different parts of the world; and examples of these concepts are explored. In addition to that aforementioned database challenge, the second key challenge is the lack of standard means for measuring and applying commercial availability concepts. 4. Other performance measures, including heat rate, capacity, environmental indices, fuel sourcing flexibility, etc. are also vital to unlocking full potential of existing plants’ portfolio.
4
World Energy Council
Performance of Generating Plant 2004 - Section 1
INDUSTRY OUTLOOK 1.1
US Situation
On 14 August 2003, the US witnessed a massive power outage or blackout. Analysis of the situation laid the blame on several factors, including what can be generally described as deficiencies in specific practices, equipment, and human decisions by various organisations. Specific issues the analysis focused on were: •
Failure to understand the inadequacies of the system with respect to voltage instability, vulnerability of specific areas, and operation of the system under inappropriate voltage criteria;
•
Inadequate situational awareness or the inability to recognise or understand the deteriorating condition of the system;
•
Failure to manage tree growth in the transmission right-of-ways;
•
Failure of the interconnected grid reliability organisations to provide adequate real-time diagnostic support.
Recommendations from the assessment were grouped into the following categories: •
Institutional issues related to reliability (14 recommendations), including enforceable standards, strengthening of the institutional framework, assuring participants have met minimum functional requirements, etc.;
•
Strengthening of NERC’s Actions (16 recommendations), including correction of the direct cause of the blackout, strengthening of the compliance programme, better real-time tools, more system protection measures, etc.;
•
Physical and Cyber Security of North American Bulk Power Systems (12 recommendations), including development and implementation of new IT standards, management processes, security control, forensic/diagnostic capabilities, risk and vulnerability assessments, etc.;
•
Canadian Nuclear Power Sector (2 recommendations) focusing on review of operating procedures, training, and installation of backup generation equipment.
Industry players closely monitored the activity to determine if and how the event would impact the industry outlook, as the industry was still recovering from the Enron debacle, the failed California market, and massive overbuild of base-load gas-fired generation capacity. While it is too soon to state with any certainty the overall influence of this event, in general, the results appear to be: •
Reaffirmation of industry regulators (NERC, FERC, etc.) that stronger standards and protocols are required;
•
Stronger focus on system/grid reliability in conjunction with any further market reform.
5
World Energy Council
Performance of Generating Plant 2004 - Section 1
In 2003-2004, there has been significant activity with respect to standard market design (SMD) and the related concept of Regional Transmission Operators (RTOs). The Federal Energy Regulatory Commission (FERC) has been at the centre of the action. In April 2003, FERC published a white paper to clarify its position on the continued wholesale market reform and its drive toward the implementation of standard market design. In this paper, FERC proposed: “The industry has been evolving toward a market-based approach for well over a decade and active longterm wholesale bilateral markets exist in all regions of the country. However, short-term wholesale markets with transparent prices and market structures that will reliably produce just and reasonable prices are not likely to develop without strong Commission action. Wholesale electricity markets do not automatically structure themselves with fair behavioural rules, provide a level playing field for market participants, effectively monitor themselves, check the influence of market power, mitigate prices that are unlawful, or fix themselves when broken. These are the responsibilities of the Commission under current law, and our proposal was made with these responsibilities in mind”.
Furthermore, in the same paper FERC presents its goals and methods it intends to apply under its Final Rule. These include: •
The formation of Regional Transmission Operators (RTOs) and ensuring that all RTOs and independent system operators (ISOs) have good wholesale market rules in place;
•
Require public utilities to join an RTO or ISO;
•
Provide for phased-in implementation and sequencing tailored to each region and allow modifications to benefit customers within each region.
FERC goes on to note: For the basic wholesale market platform, we intend to build upon the existing rules adopted in Order No. 2000 for RTOs by adding features that we have learned are necessary for effective wholesale power markets For example, Order No. 2000 did not include market power mitigation measures and does not prevent flawed market designs. Wholesale electric markets will not be able to deliver full customer benefits in the future without the oversight and transparency that regional independent transmission organisations can provide. Healthy and well-functioning wholesale power markets are central to the national economy, and we believe that regional, independent operation of the transmission system, with proven market rules in place, is the critical platform for the future success of electric markets. Divestiture is not required to achieve independent operation of the transmission system. Companies may remain vertically integrated under an RTO or ISO.
There has also been substantial activity at the state level. Many states are pushing back, especially those that are not currently under RTO’s. Various forms of federal legislation are also in process that could alter the situation significantly. Ideological lines have been drawn around this debate. Those who support FERC’s position insist that it is impossible to have regional power markets and fair access to transmission if each state retains its ability to regulate bundled retail transmission. Their goal is one set of rules for what is now a regional transmission system, with no discrimination in favour of one class of customer. The opposition cites that states should retain jurisdiction over bundled retail transmission because the retail customers have paid
6
World Energy Council
Performance of Generating Plant 2004 - Section 1
for the transmission system, and must continue to have priority access to it; as such, a utility should focus on its own retail load in providing transmission service, as defined by its obligations to state regulators. RTOs are a key element of FERC’s standard market design strategy. An RTO is a largescale (primarily multi-state) electric transmission system operator whose fundamental purpose is to ensure the efficient and reliable delivery of power supplies by removing transmission barriers between power sellers and buyers. The goals for RTOs are significant and include: •
Improved management of congestion in transmission facilities;
•
Elimination of “opportunities” for transmission owners to engage in discriminatory practices;
•
Improved market performance and efficiency;
•
Enhancing the exchanges of power by eliminating "pancaked" rates (multiple "shipping" prices that customers pay when electricity travels through several utilities rather than within one RTO).
RTO’s operations are independent from power market participants, including the buyers and sellers of power. RTOs must be the sole provider of transmission, control transmission facilities in their region and provide transmission service that is reliable, efficient and non-discriminatory. A second, equally “visible” component of FERC’s standard market design is the concept of Locational Marginal Pricing or LMP. LMP is the mechanism preferred by FERC to eliminate “pancaking” of rates and to establish a means for managing transmission congestion. Under LMP, the goal is to send “proper economic signals” that are not present with traditional transmission pricing methods (e.g. congestion costs); transmission customers pay their share of transmission fixed costs in “access charge” (not dependent on usage) and on top of it, they pay a generation “congestion charge.” In simple terms, the RTO calculates prices of energy at the “source” versus the destination (i.e., the sink) and with the difference being the congestion cost. Theoretically, by making prices of energy at source and sink transparent, this allows the consumer to see the “value” the market places on various transmission routes, and encourages behaviour in which power is purchased from other sources on the “right” side of the constraint. As the action continues to unfold, it becomes clear that the US power industry is facing many future uncertainties, while FERC and others regulatory bodies are focusing on: •
The Market – the main component of standard market design, RTOs, LMP, and other aspects of FERC and state regulations to ensure fair, equitable, and transparent market – not the elimination of the market;
•
The need for participants in the market to be “rewarded” or “punished” in economic terms without sacrificing regional grid reliability;
•
Ensuring that the diverse range of power producers, utilities, coops, cogenerators, etc. can co-exist within a viable market environment.
7
World Energy Council
Performance of Generating Plant 2004 - Section 1
WG6 believes that continued presence and importance of the market clearly impacts the goals and roles of the generator and the measures/means in which performance is measured, reported, and ultimately optimised.
1.2
The Europe Situation
In many ways, there is a number of parallels between the US and the European situation. Europe, like the US, witnessed a major power blackout; like in the US, industry participants closely watched and waited to determine if the event would derail their progress toward a fully liberalised European energy market. Finally, like in the US, it seems that the blackout will strengthen Europeans’ resolve to maintain adequate reliability safeguards while still moving ahead toward a fully functional competitive market. The focus of attention amongst European Union legislators in recent years has been the introduction of a fully competitive market in electricity. With the ink barely dry on directives aimed at ensuring choice for all electricity consumers by 2007, the European Commission (EC) has turned its attention to the issues of security of supply and infrastructure. In December 2003, the Commission proposed a legislative package that is intended to complement the opening up of both electricity and gas markets by encouraging investment in infrastructure and preventing the reoccurrence of the blackouts that took place in a few European countries in the summer of 2003. In announcing EC’s proposals, Loyola de Palacio, EC Vice President responsible for energy and transport, said, "This new framework is decisive for reinforcing the European single energy market and preventing Europe having to face a situation like the so-called California experience." Loyola de Palacio made it clear that, in her opinion, the incidents affecting supplies in Europe during the preceding summer had nothing to do with the market opening. At the heart of the new proposals are a set of measures aimed at stepping up investment in power production and strengthening member states' transmission and distribution networks. The European Union is in the process of changing its electricity markets from many, more or less separate monopolistic electricity markets, into one single liberalised market. The implementation of the first liberalisation Directive (96/92/EC) has resulted in fully liberalised electricity markets in some Member States, while there are still markets where not all parts of the liberalisation aspects are in place yet. With the new electricity market Directive (2003/54/EC) the process of integration of the national electricity markets into a single European market (in parallel with the natural gas markets) is reinforced, and in 2007 all customers will have the right to freely choose their suppliers. In fact, the market is undergoing two fundamental changes: one is liberalisation, the other is the integration into a single pan-European market. Both have effects on investments. Liberalisation brings new challenges; integration provides more choice of investments between importing electricity and building own generation capacities. The intentions of the Directives are to establish a competitive market for generation and supply with many unbundled companies, and a separate, naturally monopolistic, regulated sector for transmission and distribution; replacing the pre-liberalisation system where 8
World Energy Council
Performance of Generating Plant 2004 - Section 1
electricity supply in given geographical areas was provided by single vertically integrated companies. This reform is expected to lead to greater economic efficiency and lower prices, yet maintaining the very high security of electricity supply that Europe has enjoyed for many years. Electricity is practically non-substitutable for many end-users, and the penetration of electricity is expected to grow, forming part of the solution for sustainable development. Therefore, the updated Directive establishes an excellent framework for competition in a free pan-European electricity market. In today’s liberalised structure, investments should be triggered by market decisions within a competitive framework and security of electricity supply should be ensured without any distorting interventions. Furthermore, the Directive also provides a list of tools for market surveillance. Unlike legislation governing any other industrial sector, it goes even further, allowing governments to organise direct tendering procedures where judged appropriate to ensure security of supply. The liberalisation of the electricity sector has created a new competitive environment in which the power generation business is fully open to competition, and millions of industrial and household consumers are already entitled to freely choose their power supplier. Some 70% of all consumers in the European Union now have this right of choice. The recent adoption of a new EU "Accelerated Liberalisation Package" will result in full liberalisation of the internal EU electricity and gas markets, in two steps. •
July 2004, all non-household customers (industrial, commercial and professional customers) will be free to choose their supplier;
•
July 2007, at the latest for all households to gain this right.
The acceleration of energy market liberalisation is a key objective for EURELECTRIC, the leading European association of electricity producers. EURELECTRIC’s position is summarised below. •
Ensure convergence between electricity and gas markets. In order to ensure competitive gas prices for gas-fired power generation, the gas market should be opened up to competition in parallel with the electricity market and be accorded the same treatment - in terms of qualitative and quantitative market-opening requirements - as the power sector;
•
The creation of a single EU electricity market & cross-border trade. Opening the individual national electricity markets to competition is a first vital step in the process of creating a single European electricity market;
•
Active Wholesale Markets and Trading. Electricity trading and power exchanges play a key role in the liberalised markets by producing a reliable market price and helping to develop the necessary market liquidity. In liberalised markets electricity can be traded through various arrangements between generators and suppliers, traders and final customers. Bilateral trades and OTC (over-thecounter) trading occupy an important role in the markets. However, trading in organised power exchanges is playing an ever-increasing role. Trading markets can be divided into the following broad product and market categories:
9
World Energy Council
Performance of Generating Plant 2004 - Section 1
o Day-ahead markets or "spot trading": Prices are set for the following day and electricity is physically delivered from seller to buyer. Reliable spot markets are the basis of modern power markets and they set the reference price for other types of contracts. o Physical products: These products allow a market participant to sell or to buy power at a preset price for weeks, months or even years ahead. These contracts can be traded on to other participants, but they always result in physical delivery of electricity. o Financial products: These products allow players to sell or buy power at a preset price for weeks, months or years ahead. However, financial derivatives do not result in physical delivery, but are settled financially between parties. •
Retail markets. The right of a customer to freely choose his/her supplier depends on the degree of market opening applied in a given EU Member State. While large industrial customers are designated as eligible throughout the EU, competitive mass retail markets are now developing in Sweden, Finland, Norway, Denmark, Germany, Austria, the Netherlands and the United Kingdom;
•
Post Liberalisation Activities. Ensuring a sufficient level of investment, properly functioning market, and the identification of the appropriate solutions for fuelmix disclosure and unbundling. Although there are different projections, there seems to be a broad agreement on the need of 500-600 GW new capacity in the time frame 2000 to 2030 for EU-15 only. Depending on the share of electricity production from renewable energy sources and distributed generation, the total investment cost for this new capacity might be in the range of €400 to 700 billion.
Another key European power market participant is the "Union for the Coordination of Transmission of Electricity" (UCTE). UCTE is the association of transmission system operators in continental Europe, whose role is to provide a reliable market base by efficient and secure electric power supply. About 450 million people are supplied with electricity through the networks of the UCTE, bringing the annual electricity consumption to approx. 2300 TWh. The prime objective of UCTE is to provide secure operation of international interconnections in the electricity transmission system of the synchronous area. Main missions of UCTE include: •
Technical and operational co-ordination of interconnection in the UCTE synchronous area;
•
Monitoring and control of the short-term reliability of the system with regard to load, frequency control, stability, etc.;
•
Medium-term adequacy between generation and consumption (3-year power balance forecast);
•
Study and monitor the development of the synchronous area.
As demonstrated by the above discussion, Europe, like the US, remains committed to the continued development of a market environment. At this point, it appears that the “goals” lying before the EU are more reaching than those of the US. As such, it will be imperative for European power generators to be fully capable of managing the complexities presented by a fully liberalised market.
10
World Energy Council
1.2
Performance of Generating Plant 2004 - Section 1
Africa Situation
Introduction The issues below briefly describe the generation industry in the African continent. They also give an indication and ramifications in the trading scenarios. The importance of benchmarking has been clearly demonstrated in South Africa by Eskom in leveraging and asset sweating the generators in improving the availability of plant and avoiding construction of new plant (from a performance perspective). However, the intensive growth in South Africa has necessitated new interventions. Commercial Availability concepts have also been fundamental in Eskom at leveraging not only the technical performance of the generating units but also their ability of adding/destroying value or missing opportunities or not being competitive. 1.2.2 Angola The construction of a 600-MW, two-turbine hydroelectric dam in Angola's northern Uije province was announced by Minister of Energy and Water Botelho de Vasconcelos. The minister stated that the authorities were awaiting the conclusion of a study before work could begin and that construction equipment for the dam was already in the country. Angola seems to be focusing on improving its power sector: it is rehabilitating several dams and pre-completion testing is ongoing at the 520-MW Capanda hydroelectric dam in Malange. 1.2.3 Botswana New interconnection transmission lines along the Southern African Power Pool (SAPP) countries, funded by US$223m from the World Bank will boost Botswana's power sector as it seeks to diversify the source of its supply through the short-term energy market (STEM) electricity trading agreement. STEM commenced trading in 2001 and involves countries including Botswana, Namibia, South Africa, Zambia and Mozambique, along with other countries in the region. Through STEM, Botswana has been able to purchase power cheaper than it previously could on a traditional bilateral power purchase agreement, and ensure reliability of power supply. Power exchange arose because of the distribution of power resources in the southern Africa region: a large reserve of low-cost hydroelectricity in the northern part; and large reserves of cheap coal in South Africa and Zimbabwe. STEM incorporates interconnection and regional power trade and this project is already saving costs and improving the power sector in the countries involved. While there is still a lot of work to be done, such as liberalising the power sector and removing subsidies on electricity, some progress has been made and the SAPP could serve as a model for other regional integrated electricity network programmes in Africa. 1.2.4 Kenya Construction on the Sondu-Miriu hydropower project was set to restart in May 2004. Work was supposed to recommence on 1 April, but a delay has been encountered as some increased costs had been incurred, which will be met by the Kenyan government. The 60 MW dam is expected to be completed in 2006.
11
World Energy Council
Performance of Generating Plant 2004 - Section 1
1.2.5 Namibia Namibia is on a vital search for a reliable energy supply that would ensure that at least half the country's peak energy demand could be met at all times. The government is considering several alternatives to diversifying its power-supply source. Several options being considered include a wind-propelled plant at Lüderitz, solar power and a hydropower plant at Epupa Falls in the Kunene River. A feasibility study on a hydropower plant at the Popa Falls is ongoing. NamPower, the Namibian state utility, has renewed its power-sharing agreement with Eskom, the South African parastatal. The deal means that Namibia will continue to import electricity from South Africa; South Africa supplies its neighbour with up to 240 MW during the Q2-Q3 winter period, although Namibia can export power to South Africa during the rainy season. In addition, Eskom and NamPower will continue to co-operate within the sector in order to improve efficiency. Up to 80% of the country's electricity needs are supplied by Eskom. 1.2.6 Nigeria The privatisation of the National Electric Power Authority (NEPA), the state electricity parastatal, has taken a step forward. NEPA is to be unbundled into 18 separate firms, which will then be privatised. The first of these firms, the Transmission and System Operation Company (Transysco), has now been unbundled at a formal launching of the firm in Abuja, Nigeria. Transysco will be the only transmission company in Nigeria and will operate the country's power grid. The unbundling of Transysco will be followed by changes to the generation and distribution sectors. Six generation companies and 11 distribution firms will be created The unbundling of NEPA has begun despite the country's National Electricity Reform Bill having not yet been approved. NEPA has now officially created Transysco, the company that will operate Nigeria's transmission network, but has done so even though the necessary legal and regulatory framework is not yet in place. The bill has stalled in the National Assembly, as President Olusegun Obasanjo is unwilling to sign the legislation into law as it contains an amendment, made by the National Assembly, which would grant the body the power of approval over nominated commissioners to NEPA. 1.2.7 South Africa The Integrated Energy Plan (IEP) looks to outline the future of the South African energy sector in the long term. Given the strength of the country's coal industry, the IEP envisages that South Africa will continue to be reliant upon coal for the next 20 years. However, a gradual drift towards using cleaner generation methods, such as gas and renewables, is planned, and the Department of Minerals and Energy (DME) is seeking to attract IPPs to the country. From the IEP, 2007 is marked out as the time by which South Africa must bring new generation capacity online, or face a potential shortfall. South Africa is taking steps at the moment to deal with this issue. Energy efficiency programmes are being put in place to reduce demand, while Eskom is considering removing some plants from their mothballed state.
12
World Energy Council
Performance of Generating Plant 2004 - Section 1
The plan now, under the framework put forward by the IEP, is to seek new generation capacity. In order to do so, the DME is looking to secure IPPs (independent power producers) to construct new units. A call has been made to consultants to help in this process. The consultants will also assist in matters such as locating potential sites for IPP plants, as well as the creation of a power pool into which the IPPs can sell their power. Although the IEP does not totally prohibit the utilisation of gas for electricity generation, it does state that it is technically more effective for South Africa if gas is burned at source. In the past, renewables have lost out in South Africa on a cost basis, but this IEP forces them into the country's energy future, despite the increased costs that this involves. This is because the utilisation of renewables clearly brings with it other benefits. These are not only environmental; after all, fossil fuels will run out someday, and the country needs to have the infrastructure in place to replace these in the long term. Eskom has brought one wind farm online already and it is anticipated that another will be up soon, while photovoltaic development is also taking place. The white paper states that renewable energy sources should provide 10,000 MWh within 10 years. The IEP is South Africa's first attempt to provide a comprehensive strategy for the country's energy sector to follow, and appears to be a positive step towards achieving more satisfactory integration in the industry. Introduction of advanced performance measures, such as Banked Availability Value indicators (a measure of Commercial Availability) has enabled Eskom to be one of the leaders in measuring the economic value of its generators. These indicators enabled plant management to understand that the value of availability is not a constant and that a generator could significantly improve its profitability by managing its availability so as to be producing electricity when the market spot prices are at their highest. In other words, a generator’s availability is worth more during certain hours of the year than in others. Consequently, electricity executives would welcome Commercial Availability indicators that would allow them to benchmark the performance of their plant with that of other similar plant around the world without disclosing sensitive commercial information. These new indicators were prototyped (last quarter – 2000) in South Africa and showed success. The project was piloted in 2001 and 2002 enabled further calibration and benchmarking of the indicators. In 2003, they became operationally contractual. 1.2.8 Sudan The government aims at providing electricity to 90% of the population after reportedly securing more than US$2bn from a variety of sources. With peace talks between the government and southern rebels progressing, the government is now looking at the need for economic development. Plans to construct a new 4,000 MW hydropower plant are well under way and theoretically a further 10,000 MW could be added. 1.2.9 Tanzania Tanzania has been granted an International Development Association (IDA) US$43.8m credit by the World Bank to assist the government in implementing emergency measures to avert prolonged shortage of power supply due to the extended drought affecting the
13
World Energy Council
Performance of Generating Plant 2004 - Section 1
country's hydropower system. The IDA credit will enable Tanzania's electricity company, TANESCO, to pay for fuel and energy purchases to prevent load-shedding and power outages in the country over the next eight months. The severe drought in 2003 has significantly reduced water inflows into the country's dams and affected TANESCO's ability to generate electricity from hydropower. The 2003 water inflow to the Mtera reservoir, the most important reservoir in Tanzania's hydropower 1.2.10 Uganda The Ugandan government has closed bidding on the controversial 250-MW Bujagali hydroelectric plant project, to be constructed on the Victoria Nile River near Jinja, 80 km east of the capital, Kampala. The project was widely opposed by environmentalists and lawmakers who claimed it would displace many people and submerge the Bujagali Falls that attract white-water rafting enthusiasts. Construction work on the project was supposed to start in June 2002 by an AES Corp.-led consortium but the company dropped out when the World Bank indefinitely postponed approval of a US$195m loan guarantee for the project. The Ugandan government is reported as claiming that the World Bank and International Finance Corporation are involved in the project and have claimed that the country is facing a power crisis if the dam is not built. Work on the project is scheduled to start next year. 1.2.11 Zambia Zambian Energy Development Ministry has stated that the country’s energy sector needs regulation in order to enable energy firms to make a profit and to allow consumers to pay fair prices. At an Energy Regulation Board (ERB) meeting, the Ministry stated that at present the sector was ‘monopolistic’ in nature. However, no specifics on how the market might be further regulated were put forward. 1.2.12 Zimbabwe The Zimbabwe Electricity Supply Authority (ZESA) has increased the country’s electricity tariffs by 400%. The move appears to have been forced on the utility by the need for ZESA to pay its foreign suppliers for the power that they provide; South Africa’s Eskom and Mozambique’s HCB have recently imposed ultimatums on the utility to settle its debts ZESA has been handicapped by debt as a result of Zimbabwe’s foreign-exchange crunch. ZESA has been keeping afloat by issuing ‘megawatt bills’ but the firm will not be able to sell its debt forever, and therefore has implemented a tariff hike in an attempt to increase its income. However, the scale of the pricing increase could raise domestic tensions within Zimbabwe, and will also hinder the attempts of the Reserve Bank of Zimbabwe (RBZ) to scale back the country’s inflation rate. Iran seems interested in investing in Zimbabwe’s power sector. It is reported that an Iranian delegation has visited Zimbabwe with a view to becoming involved in rehabilitation work at the country’s two major power stations, Kariba and Hwange. ZESA has reportedly offered Sunir, an Iranian firm, free repatriation of investment capital, 100% profit and dividend remittance, and 49% foreign equity ownership in the Kariba plant. It also seems that a Chinese firm, CATIC, may have secured similar contracts to improve the operations of Zimbabwe’s power plants. However, it does seem that Zimbabwe is determined to secure some kind of foreign involvement in its power sector in order to improve performance.
14
World Energy Council
Performance of Generating Plant 2004 - Section 1
VALUE OF PERFORMANCE/AVAILABILITY DATA In recent years the need to develop and use new reliability indices, which more accurately reflect the current market place, has taken on a high degree of urgency. It is brought on by the need of large power consumers for lower electricity prices in order to compete in the global economy. To meet this need electricity generators are being compelled to reduce their costs. Decision-making at all levels is being affected and the old “technical” definitions of reliability are being amended to incorporate economics in order to link better plant performance with the actual cost of electricity supply. Rather than applying traditional measures that are calculated over both demand and non-demand periods, new reliability terms are considering only the hours that the plant would have been dispatched and the financial consequences to the company’s bottom line from the failure to generate during those hours.
2.1
The Historic Problem
Among the traditional measures of plant reliability have been the Equivalent Availability Factor (EAF), the Forced Outage Factor (FOF) and the Equivalent Forced Outage Rate (EFOR) in North America, China and a few other countries and the Unit Capability Factor (UCF) and the Unplanned Capability Loss Factor (UCLF) in Europe and South Africa among others. Those measures that are "factors" (EAF, FOF, UCF, UCLF,) use as their denominator the entire time period being considered (typically one year) without regard to whether or not the unit was required to generate. Therefore, for non-baseloaded units, these factors can lose their relevance (and the more cyclic the demand is, the greater the effect). For example, if a Gas Turbine unit is used exclusively for meeting peak demand periods, it may only be required to generate just a few hundred hours a year. If it were unavailable during 25% of those hours, it would still have high an EAF and UCF and a low FOF and UCLF. If a peaking unit was required to generate 100 hours per year but experienced forced outages during 25 of those demand hours (and no other outages over the 8760 hours in the year), it would still have an EAF and UCF of (8760-25)/8760 x 100 = 99.71% and a FOF and UCLF of (25)/8760 x 100 = 0.29%. Those numbers might look good on paper but the reality is that the unit could only produce 75% of the power required of it. So these factors do not correctly describe the unit's ability to produce its rated capacity when demanded. Of course, for true baseloaded units, such as many nuclear units, which generate every hour they are available, these factors come much closer to depicting the unit's "real" reliability. The terms Forced Outage Rate (FOR) and Equivalent Forced Outage Rate (EFOR) were introduced in an attempt to resolve these difficulties. FOR and EFOR differ only in that EFOR considers the "equivalent" impact that forced deratings have in addition to the full forced outages that is all FOR considers. In this simple example with only full forced outages we will examine the FOR. More detail is included in Appendix 1.
15
World Energy Council
Performance of Generating Plant 2004 - Section 1
The equation for FOR is: FOR = (Forced Outage Hours)/(Forced Outage Hours + Service Hours) X 100
For the example given above the actual service hours are 75 so that the FOR would be: FOR = (25)/(25 + 75) X 100 = 25%
The complement of the FOR might be considered to be the unit's reliability so that Reliability = 100% - 25% = 75%
So it appears that FOR (and EFOR when forced deratings are present) are good measures of a unit's reliability. However, in practice it is extremely unlikely that all of the forced hours that a unit experiences during the course of a year are during its demand period. (In our example all 25 Forced Outage Hours were assumed to occur during the 100 demand hours). Most times a forced outage will also require a few hours to restore the unit to service, during both non-demand periods and some demand periods. In our example the unit might have experienced five forced outages during 25 hours of its demand period (out of 100 hours total demand). However, it is likely that the time to restore the unit to full capability would average more than the five hours each during demand periods. It is much more probable that the total forced outage hours would be several times higher (some previous studies suggest that the average restoration time for a gas turbine forced outage is on the order of 24 hours). Therefore, if we use 24 hours as the average down time, then the total forced outage hours reported would be 5 X 24 = 120 hours. Now the FOR would be: FOR = (120)/(120 + 75) X 100 = 61.5% and the unit's reliability = (100-61.5) = 38.5%,
Both values are obviously unrealistic when attempting to use these statistics to make decisions requiring the expected reliability of units to be used. And yet these values are very close to actual FOR and EFOR statistics being reported for peaking types of generators. This does not mean that FOR and EFOR should not be used, as they are in fact appropriate indicators for baseload or near-baseload types of generating units. However, for cycling or peaking units they are inadequate and new indicators were needed. A few years ago a modification of EFOR was introduced in an attempt 16
World Energy Council
Performance of Generating Plant 2004 - Section 1
to resolve this problem. The term Equivalent Forced Outage Rate-Demand, EFOR(d), was developed which only used that portion of a unit's forced outage (or derating) that occurred during demand periods. As in the earlier example, that would resolve the issue. However, demand periods are not currently part of approximation technique was devised using a MARKOV standard reporting systems, and therefore an approach. Although not perfect, this technique does result in a more accurate calculation of EFOR(d).
2.2
The Concept of Commercial Availability
The term Commercial Availability (CA) emerged in the United Kingdom in the early 1990's following the deregulation of the UK power industry and the introduction of a "market" system. Since a plant's availability only had value to its owner, if it could generate power at a profit, it was only measured during the times the market price was above the plant's variable cost. Initially, CA was not "weighted" and it was assumed that each hour that the unit was economically viable had the same influence on CA. Over time, some users of CA have developed the definition to include the influence of the price/cost gap magnitude so that it could serve as a more accurate indicator of the plant's impact on the company's profitability. (E.g. during hours when the gap is US$20/MWh, the plant's actual availability would have ten times greater influence on the profits than an hour in which the gap was only US$2/MWh). Therefore, CA attempts to measure the actual profit delivered by the plant relative to the potential profit, had been able to deliver every MWh required of it at the actual market price. (Profit here is defined as gross margin, generally the difference between the plant's variable production cost and the market price, or the system marginal cost in the case of regulated companies). Different equations for CA have been independently developed attempting to measure the financial impact of a unit’s unavailability. Some of these are described later in this report.
2.3
Commercial Availability Methodologies
The concept of a Commercial Availability (CA) is gaining popularity despite the fact that there is no specific or accepted definition of the term. Based on a survey of US companies in various locations, the following describes some of the methodologies used by different companies. All of these terms are used to evaluate the “Commercial Availability” (CA) on a historical basis. A discussion of forecasting CA will follow later in this report. 2.3.1 Method 1 The first of the various methods compares the ratio of the actual revenue to potential revenue using the product of generation and market price. In the numerator generation refers to actual production and the denominator generation refers to the requested generation. Requested generation would therefore include actual generation plus any instance of a failed start, forced outage or derating that limited the output of the unit. This method has the advantage of being easy to calculate. However, is does not recognise the impact of fuel price and only considers the potential revenue. Should the owner choose to operate the unit at a loss to avoid a start-up cost, the logic behind the calculation would no longer be fully relevant. 17
World Energy Council
Performance of Generating Plant 2004 - Section 1
2.3.2 Method 2 The next method is based on the ratio of hours the unit was available and when the unit could have been generating at a profit as compared to the total hours that the unit could have been operating at a profit. It is referred to as hours “in the money” versus period hours “in the money”. Again this is a simple calculation but the calculation of hours in the money is very specific to each generator. The method does address the shortcoming of Method 1, since it specifically addresses the marginal cost of operation. However it does not address any difference between available hours and actual generation. In many cases the unit can be available to operate but does not. Though not stated explicitly a derating of the unit is either not addressed, (since the unit is still “available”) or it must be addressed through a calculation of Equivalent Availability (EA). 2.3.3 Method 3 The next methodology attempts to address Method 3 the issues of EA as well as that of fuel Sum (Market Price - Generation Cost) x MW Available prices or marginal costs. The denominator Sum (Market Price - Generation Cost) x Max Capacity is the accumulated sum of the period in Measured hourly when the unit is “in the money” In the money occurs if market price – cost of generation is >0 question of the difference between Market Cost of generation includes fuel, variable O&M, and emissions costs Price and generating costs times the Method 4 maximum capacity of the unit. This Actual Margin difference between market price and Potential Margin generating cost could be referred to as net Measured hourly, margin. This sum is accumulated on an Denominator = Sum of “In the money” hourly generation x hourly market price less marginal fuel cost. hourly basis over the time period in Numerator = Hourly installed capacity x hourly market price less marginal fuel cost question. The numerator sums the product of net margin and the available capacity in each hour.
In the money again refers to the hourly periods in which the cost of generation, including fuel, variable O&M and emissions costs is less than the market price. The data requirements for this method start to become more substantial though this method does address the impacts of availability and marginal costs. 2.3.4 Method 4 Despite the apparently simple description of this method, the calculations are almost as difficult as the previous method. Again data are calculated on an hourly basis and summed over the period in question. Actual Margin refers to the sum of the generation during time periods when the unit was “in the money” times the hourly market price minus the marginal fuel cost. The “Potential Margin” of the denominator is similar though it uses installed capacity in the place of actual generation. Again this method is computationally intensive. Based on the definition provided, it does not address variable O&M or emissions costs. These costs, especially for a peaking unit that incurs a starts-based charge to the O&M costs as a term of an LTSA (Long Term Service Agreement) can be substantial.
18
World Energy Council
Performance of Generating Plant 2004 - Section 1
2.3.5 Method 5 The last of the identified methods of calculating CA is very similar to earlier described methods but is simply expressed in different terms. The ratio of an actual performance over planned (as opposed to actual over capability). This method would allow for Method 5 Actual Generation x (Market Price – Unit Cost) planned outages and an expected number of Planned availability x (Market Price – Unit Cost) forced outages that occur during the Uses the sum of hourly values evaluation period which would have no No penalty for scheduled outages Expected forced outages converted to monthly average such that there impact on the calculation. The planned is no penalty for FOR=plan availability of the denominator is calculated as a monthly average inclusive of planned Availability Ratio 1-EFOR (actual) outage and an expected number/magnitude of 1-EFOR (target) forced outages.
Availability Ratio Though not strictly a method of calculating CA (Commercial Availability), one company chose to use an Availability Ratio. This is the ratio of actual availability to expected availability. In both cases availability is calculated as 1 minus the EFOR (Equivalent Forced Outage Rate) either actual or target. Target EFOR is again based on an annual average EFOR. This method is comparatively simple but does not address issues of marginal fuel price or even market prices. Though the company using this measure was of the opinion this was comparable to a commercial availability ratio it clearly leaves a number of commercial issues unaddressed.
2.4
Banked Availability Value Method
This framework implies that Commercial Availability should reflect the classical relationship between supply and demand with the value of an MWh increasing due to: • An increase in demand, assuming no change in supply; • A decrease in supply with no change in demand; • A combination of the above. The converse is also true and will decrease the value of an MWh. The demand and supply each have a predictable component and a random component. Therefore it is possible, by means of statistical techniques, to forecast the events when the energy will be of value, given the forecast for the demand and supply. This can be seen as a static (mechanistic) approach to define the periods of high worth. In addition, the random behaviour of the demand and supply could significantly change the anticipated worth of the MWh in the short term. The indicator (or family of indicators) should reflect how well the managers of the entity have been able to capitalise on both the long and short-term economic opportunities presented in liberalised markets; i.e. where spot prices continuously respond to the balance between supply and demand. The spot price represents the value (worth) of an MWh to the market during that specific period. 19
World Energy Council
Performance of Generating Plant 2004 - Section 1
If the spot price is greater than the marginal cost of production, then the unit has market value. Under these circumstances the unit is competitive and should be producing to increase shareholder value. If, however the unit is unable to produce (due to planned or unplanned outage or capacity curtailment), then the unit has missed the opportunity to add shareholder value. If the spot price is lower than the marginal cost of production, then the unit is not competitive and has no value in the market. Under these circumstances the unit should not be producing. However, the risk now exists to actually destroy shareholder value by making the unit available and producing at these low prices. The above situations define the four domains of Commercial Availability, namely: • Adding Value; • Destroying Value; • Missing Opportunity; • Not Competitive.
Revenue > Costs Revenue < Costs Dispatched (Actual MW) Adding Value Not Dispatched
Missing Opportuniy
Destroying Value Not Competitive
Commercial Availability Domains This indicator of the Commercial Availability family measures the frequencies in each particular domain for every hourly event. This is weighted by the MW dispatched vs. the Installed Capacity (IC).
Revenue ≥ Costs Revenue < Costs Dispatched (Actual MW) You added value
You destroyed value
You missed the opportunity
You were not competitive
Not Dispatched
Frequency: CA Family of Indicators - VAB To account for the financial impact of the Revenue – Cost relationship, the authors formulated the Banked Availability Value (BAV) family of indicators.
20
World Energy Council
Performance of Generating Plant 2004 - Section 1
This family is illustrated below.
Revenue ≥ Costs Revenue < Costs Dispatched (Actual MW) Earnings x MW
Earnings x MW
Potl Earnings x [Max{IC;D} – MW]
Potl Earnings x [Max{IC;D} – MW]
Not Dispatched
CA Family of Indicators – BAV
2.5
Other Commercial Availability Alternative Measures
As an alternative to calculating Commercial Availability, some companies are choosing to calculate an inverse measure of Commercial Unavailability. This measure simply examines the ratio of the hours the unit could have been operating (was in the money) but was not (available) to the total number of hours in the period in which the unit was “in the money”. This method does not address the magnitude of impact of the market. There is no distinction between being forced off in a market that is US$1.00 over the cost of production versus a market that is US$500 over the cost of production. However, it would allow the user to tie the unavailable hours to specific forced outage and its causes (using a tracking method such as NERC GADS). This would also allow the asset owner to prioritise issues based on a limited evaluation of the commercial impact. The last of the methods identified was very specific to a particular market. This method required the user to track a “balance account” and evaluated a difference in generation based upon actual generation minus the desired output of the unit factored for the period of unit availability (not EFOR) times the difference between the Locational Market Price (LMP) minus the unit cost of production. This method is useful and applicable only in regions which calculate a Locational Marginal Price (LMP). This measure can be either positive (in the case if high availability versus target) or negative when the actual generation does not meet expectations.
2.6
Estimating Commercial Availability
All of the methodologies discussed above are historical measures. They show the achievement (or lack thereof) by the asset relative the specific market. It would also be useful to attempt to forecast commercial availability in order to predict the value of a specific unit. The following discussion focuses on one suggested method for forecasting CA.
21
World Energy Council
Performance of Generating Plant 2004 - Section 1
It is useful at this point to note that the term Commercial Availability (CA) is also sometimes referred to as Financial Availability (FA). The following sections will discuss the data requirements for the suggested method as well as the calculation procedures. 2.6.1 Data Requirements In order to evaluate or estimate the future CA/FA for a specific generating unit, it will be necessary to develop a forecast of the market prices, on an hour-by-hour basis for the region in question. This is required in order to determine the hours during the period in which the unit will be called upon to generate (“in the money”) and the magnitude of that market so as to determine just how far “in the money” the generator would be. This minus the cost of generation reflects the unit margin. A market simulation model is used to project prices and revenue. This simulation will need to include all of the generating units, their operating characteristics (cost, capability, availability) within the region or other area of consideration (i.e. Pool). There are a number of factors the model will need to address or recognise including: •
Transmission Congestion
•
RTO Impacts
•
Hour-by-hour dispatch
Further analysis of all of the issues associated with system or regional demand and price forecasts have been the subject of numerous papers and are not relevant to this discussion. Suffice it to say they can become very complicated and must address the entire region and the economic issues that drive the region and the production within the region. The forecast of future CA/FA will use the output of the market forecast to determine or target the highest value commercial availability opportunities. This requires knowledge of the following: •
Which Units Dispatch During Which Hours
•
Distribution of Energy Prices During Each Hour (market clearing prices or LMP)
•
Distribution of Unit Energy Revenue.
2.6.2 Calculation Methods One of the primary issues associated with forecasting CA/FA is determining the likelihood that the unit is “in the money”, available and generating. This is evaluated by predicting the combined conditional probability of each of these states occurring simultaneously. This conditional probability is the product of the probability of the unit being available when required, starting when called upon and operating for the duration of the period in question. This becomes the estimation of future generation.
22
World Energy Council
Performance of Generating Plant 2004 - Section 1
The next step involves the recognition of prices and dispatch volume. These are the classic price risk and volume risk issues. The calculation method involves simultaneously predicting or evaluating the conditional probability that the unit is generating at the times required time the volume (MW) required at the time and the net margin (market price minus cost of generation). This calculation is integrated over the time domain in question to arrive at the forecast of the future Commercial Availability of the unit.
2.7
Implications of Using Commercial Availability
There will be a wide range of impacts on the way a company evaluates and manages its power plants resulting from the adoption of commercial availability and other tools/processes required to address market dynamics. This also requires a different mindset and approach in applying data and new tools in both day-to-day and performance assessment decisions. Measures and actions must consider ways to quantify and respond to different situations with differing economics. Yet the fundamentals of benchmarking remain relevant, although in new, modified forms. Benchmarking: Selection Over the past few decades benchmarking has become a key tool in most top performing generating companies for performance improvement efforts. A recommended approach is first to identify other "peer" plants whose design and operational characteristics are similar to the unit in question. The WEC has used this advanced statistical technique, simultaneously analysing over 50 plant features, to identify peer units from different parts of the world and then to compare their "traditional" reliability indices. Benchmarking commercial availability will require a new aspect of the plant to be included in the analysis to determine the optimal peer group. That new aspect is an indicator of the plant's economic incentive to generate at different times. Studies have shown that the greater the economic incentive to generate power, the better the plant's reliability can be managed to meet the demand. This implies that a statistic that measures the unit demand must be incorporated into the peer group analysis. Reserve Shutdown Hours (RSH) during different periods during the year may be a useful measure to “add” to the other design and operational features of a plant to optimise the peer selection process, including demand aspects. Benchmarking: Comparisons The actual calculation of commercial availability (whichever definition is finally adopted) is likely to be highly dependent on the precise market price (or marginal cost for a regulated or controlled business environment) per hour (or parts of each hour) which can be matched to the unit's availability in those hours. Since that price (or cost) can and does fluctuate widely over the course of each day, week, month or year, this would require a massive new database containing market cost in order to make the CA calculations. Furthermore, even if such a database was established, the actual CA's will probably not be appropriate to compare since the actual market prices in different regions would be likely to be very different. However, using the concept of conditional probability (CP), it is possible to "benchmark" the conditional probabilities of peer units and then select a goal CP as perhaps the best quartile or best decile or "Optimal Economic Availability" from the CP distributions of 23
World Energy Council
Performance of Generating Plant 2004 - Section 1
the peer units. By combining the goal CP and a unit's unique economics, it is possible to calculate a "goal" for commercial availability objectively without having to create any new data collection processes. Maximising Commercial Availability This focuses on being available to generate when required by the market and when the income and profit potential is highest. Generating units are only maintained and manned to meet market need. The logical reason for this is that stations need not be maintained and manned for periods when they are not required by the market. The daily, weekly, and annual variations in demand for electricity means that it may be possible to reduce generating costs by allowing the units to remain unavailable overnight, at weekends, and for certain parts of the year. When the plant is not required by the market, and although it is technically unavailable, will have no effect on commercial availability. Furthermore, if overhauls, etc., which would affect technical availability can be scheduled for periods when a generating unit is not required by the market, these overhauls can be undertaken without affecting the commercial availability of the unit. In situations where the operation of power plants is driven by power purchase agreements or similar financial vehicles, it will be important to consider implications of capacity payments/obligations. Design New plant design is likely to be affected since the goal is to not to maximise traditional measures of availability or reliability, but to maximise profitability (or minimise cost). One outcome of this different design philosophy will be to reduce the dependency on expensive equipment redundancy and instead install advanced equipment monitoring equipment. Since the objective is to be available "when the plant is needed", being able to better anticipate imminent equipment problems will give needed flexibility to plant management. Furthermore, even if the timing of an event cannot be controlled, the communication of the increased likelihood of an outage will allow others in the organisation (dispatch, trading, marketing, etc.) to take appropriate steps to minimise the financial impact of the outage. Operational "flexibility" also needs to be considered in design. With the addition of advanced control systems and online performance optimisation tools, it is possible to increase the plant's capability to meet demanding load schedules, ramp rates, etc., thereby increasing the potential for sale of additional MWh without compromising plant availability. In addition, since different regions have different economic conditions, the optimal economic design is likely to be different. Other implications It is necessary for the industry to recognise one likely result of using commercial availability in place of the traditional indices, i.e. that these traditional measures (EAF, EFOR, etc.) will almost surely appear worse. Regulatory agencies, financial institutions, insurance providers and even the company's own executives, board members, stockholders and customers must be included in the change process and they must "buy into" the new performance measuring system. Otherwise, they will not be able to understand that although the measures used for monitoring are appearing worse, the company is actually doing better and is delivering a lower cost and more profitable product.
24
World Energy Council
Performance of Generating Plant 2004 - Section 1
OTHER ASPECTS OF MEASURING AND OPTIMISING PLANT PERFORMANCE For more that three decades, the main focus of the PGP Committee’s work was on collection, reporting and analysis of reliability/availability statistics. Over the last nine years, as described in this report, the effort has evolved toward analysis and assessment of the implications of commercial and market forces on how to augment such data collection and analysis practices to provide the industry with additional value. Recently, it has become clear that the PGP Committee, along with the industry, must continue “broadening” its focus to take into consideration other techno-financial aspects of performance as well. This chapter explores the “issue” of performance from the broader context. In some cases, profitability or business “success” is now a function of operations within a dynamic environment – an environment that can cause the “role” of the plant to be redefined from base-load to mid-tier to peaking at different times of year; an environment where the “value” of each MWh can shift by a factor of 10 or even a 100 in a very short period of time; and an environment with many complex often conflicting economic, technical, and environmental objectives. This means that the core performance elements (availability, efficiency, production costs, and unit flexibility) must be tightly coupled to business objectives. Analysis and decisionmaking frameworks must, in turn, be geared toward overall goals and/or specific sources of opportunity presented in the market. This environment is driving the industry to “redefine” its performance: to shift from the traditional technical perspectives to the more global view of efficiently managing all key processes (access to finance, fuels, maintenance, outages, heat rate, etc.) as a collective whole against the backdrop of profitability, environmental stewardship, and risk management. The fundamentals (e.g., availability, efficiency, production costs, and operational flexibility) are key performance elements that directly affect the desired end result (profitability). However, the ability to maximise the end result requires processes or levers that enable operational decision-making based on detailed consideration and evaluation of tradeoffs or options within the context of this end result. This means that targets for fundamental, technical plant conditions are really a function of what the profit opportunities are and what the plant’s role is in capitalising on such opportunities. Obviously, “profits” in reality are a far more complex objective function that could include: •
Profitability/cost management;
•
Obligation of supply;
•
Environmental compliance (under current and future regulatory frameworks);
•
Maximisation of return on capital investment;
•
Optimal strategy for retirement/replacement of generation facilities to address aging, safety, environmental, fuel, or other techno-economic factors.
25
World Energy Council
Performance of Generating Plant 2004 - Section 1
Availability
Efficiency
Prod Costs
Outage Plan
Plant "Role"
Flexibility
Ops Strategy
Opportunity Based
"Planned"
Profitability
3.1
Addressing Risk/Returns of Coal Sourcing on Plant Availability and Production Costs
With the emergence of the competitive power market, generators with older or intermediate-load coal-fired units are becoming increasingly concerned with the need to reduce plant production costs. Often, reducing costs is the only practical way for the unit to be dispatched. Conversely, lower capacity margins within a regional market can require the same units to be capable of maximum generation and high availability during periods of peak demand. With coal often being around 80% of the production cost, reducing fuel costs is a key strategy for reducing generation costs. Yet, while lower quality, cheaper fuels can reduce the bottom line figure for fuel costs, unit capacity, availability, and emissions levels can be negatively impacted. In fact, it is possible, with all costs considered, for a cheaper fuel to be more costly on a total variable cost basis. Understanding plant performance and costs associated with fuel impacts requires very complex analysis. This is due to the fact that the “issues” to be considered range from thermodynamic performance, to maintenance, to availability/capability. Data required for calculating thermal performance can be generally characterised as the plant, unit, system, equipment configuration, performance “curves” and other operational parameters, and physical equipment geometry. These data are then entered into engineering and performance models. For maintenance and availability issues, it is necessary to consider equipment failures, their frequency, costs, time to repair, etc., that occur during operation; for this type of data, historical “maintenance/availability” data sources such as NERC GADS and manufacturer databases have been applied to develop specialised databases.
26
World Energy Council
Performance of Generating Plant 2004 - Section 1
Based on data collected, most often analysis is carried out within detailed techno-financial models such as EPRI’s VISTA. Such a model is capable of predicting plant/unit performance and costs for alternative fuels including consideration of fuel costs, transportation costs, heat rate, auxiliary power, unit capability, availability, maintenance, and waste disposal costs. These values can then be applied within a regional production costing/market simulation model to evaluate how changes in production costs might impact a dispatch plan or a power market bidding strategy. Analysis can be directly applied to the actual fuel sourcing and purchasing process. Economics for various fuel sources will vary according to market prices (fuel and energy on both spot and contract basis), environmental factors (limits, allowance costs, etc.), season, and power demand/characteristics. Hence, results can be applied to: •
Fuel selection;
•
Contract strategy development and negotiations;
•
Dispatch strategy;
•
Maintenance Planning;
•
Capital budgeting.
Opportunity to save money by making more intelligent fuel decisions requires a linkage between the analytics and the market. It is not possible to properly capture value in the fuels/energy markets without the ability to “translate” coal quality/pricing options into discrete generation strategies and actions. Thus, the major change or “result” of this assessment is to drive the analysis from the “back corner” into the day-to-day decisionmaking process. Level of cost savings to be realised is clearly a function of the market, the range of available coal sources and associated quality, and actual characteristics of the physical system and plants/equipment. Opportunity can be quite significant depending on the size of the system, liquidity/structure of the power/energy market, and customer power portfolio. For example, annual savings for a 3-unit coal-fired system Midwest USA (400 MW coal unit and two 130 MW coal units) was recently calculated to be over US$7 million.
3.2
Evaluating Physical Positions of Underlying Assets within Power Trading Framework
Many competitive power markets have associated financial markets for the trading of derivatives and other energy products. Within these markets, many of the energy companies have formed both generation groups and energy trading groups. The introduction of the trading group provides essentially for two separate sources of profit and risk: generation and trading.
27
World Energy Council
Performance of Generating Plant 2004 - Section 1
Historically, the generation portfolio has been largely driven by obligation of supply and by opportunity to purchase incremental power at the system marginal price. Hence, the generator would seek out minimum generation cost scenarios for given level of demand. Traders generally tend to focus on the financial market and generating profits through arbitrage, deal structuring, and leveraging market volatility. Risk is generally managed through hedging, calculating book positions on daily basis, and setting limits for maximum exposure and applying Value-At-Risk (VAR) analysis. In addition to the above, however, it is important to recognise that generation assets actually provide immense value to the trader as each plant is, essentially, a call option for power with a given “price”/reliability signature. Hence, the growing area of interest is how to “connect” generation to trading so that options associated with the physical assets can be properly captured and evaluated within the trading framework – both from the perspective of profit management and risk management. Understanding plant performance and costs associated with power facilities is an important component. “Real-time” (hourly, daily) capture of plant costs (fuel, heat rate, O&M, environmental, etc.), plant capability (capacity, ramping capabilities, etc.), and plant availability (probability of being on-line when demanded at particular MW level) information provide the necessary plant information infrastructure. The probability of a plant being available (or being capable of meeting specific MW demand), can be predicted as a function of plant history, plant condition assessment, historical industry “availability” data sources such as NERC GADS, and specific operations requirements. Analysis within this area is relatively new; to date, most of the analysis has been carried out within GenCo’s/Trading organisations and has not been shared with other industry participants. Analysis focuses on: •
Understanding how fuel/operating options translate to different underlying “financial” options including differential costs, level of power output available, availability, and environmental compliance costs;
•
How the physical plant portfolio can be best “combined” with market options to maximise profits – e.g. when should power being sold or purchased be generated?
•
Understanding the sources and magnitudes of risks associated with the physical power plant portfolio so that the risks can be mitigated/managed.
Analysis can be directly applied in both strategy development and risk management with the “results” of the analysis being utilised within the larger processes of deal structuring, deal capture/position tracking, and risk management. Conversely, the “value” of generation associated with in-system assets can be calculated based on the market position and applied at the plant to coordinate operations/fuels with overall corporate mission. Aligning the interests of the plant operators and traders can be greatly facilitated using appropriate planning and analysis software and solutions. Managing generation assets involves considering all the physical limitations and costs associated with various 28
World Energy Council
Performance of Generating Plant 2004 - Section 1
operating strategies and the recognition/quantification of how such limitations/options are valued in the market. Asset managers must develop strategies to communicate operating constraints to the operations and trading group. In addition to considering operating strategies, asset and portfolio managers continuously evaluate both current and historical performance and determine asset improvements that enhance value.
3.3
Market-Based View of Asset Management
Traditional planning practices can be often lacking when attempting to evaluate opportunities presented by the market. However, it is difficult to distinguish between times when “traditional” thinking continues to provide solid answers and, conversely, times where additional analytical tools are required. A market-driven view of asset management is based on the following five basic concepts: •
Goal is the alignment of assets and operations with corporate (strategic) objectives;
•
Links decision-making and action with information;
•
Risk weighted decision analysis;
•
Multi-year perspective (applying the concept of life cycle cost management);
•
A process, accepted at all levels, that distinguishes between asset ownership, management and operations.
The goals for a particular facility can be quite varied, based on the type of generation entity, the level of competition in its market, perceived strengths/weaknesses, and other corporate goals. Some typical goals for alignment with strategic-level objectives might include: •
Regional market leader;
•
Top 10% in Utility share price;
•
Share Price growth;
•
Increased earnings per share (EPS).
It is not possible for a power plant or even a portfolio of power plants to be “optimised” for all market opportunities. Rather, the fundamental makeup of the facility(ies) (type of equipment, condition, fuel cost, etc.) will limit the options. Hence, rather than focusing on all possible avenues, it is important to consider what the “optimal role” of the plant should be. In other words, to focus on defining and acting based on the expectation as to how the plant should run to maximise its return to the owner. Typical roles might include: •
Base-load plant: o Low cost high volume producer o Generates revenue through MWh sales
•
Intermediate: o Flexible energy resource, quick response and low cost to start or shut down o Generates revenue with sales
29
World Energy Council
Performance of Generating Plant 2004 - Section 1
o Reduces costs by avoiding more expensive energy o Provides system stability and load control •
Peaking: o Always available when needed capacity resource o Avoids high cost capital investments in energy resources o Mitigates system against spikes in peak prices
and to understand whatever limitation exist, such as: •
Physical Limitations (design): o Maximum capacity o Steam generator limits o Turbine/generator limits
•
Transmission access/capacity
•
Legal Restrictions: o Emissions Limits o Air permit requirements
•
Practical Limits o Culture changes o Limits to the rate of acceptance of changes o Labour organisation resistance
•
Capital and Cash flow limitations o Market driven o Regulatory driven
•
Other Unique or Local Circumstances
Encouraging and measuring such behaviour is difficult; many of the tools, practices, and ways of performing day-to-day tasks have been based on historical practices and do not properly address or reinforce the new demands placed on the generator as a function of increased market/financial risk, increased uncertainty, and increased volatility. These “old” systems or ways of doing business actually can serve as barriers to change. Hence, market-based asset management requires to rethink and to redeploy tools and processes that reinforce the “current” business needs. It should be possible to apply the volumes of information available to enhance the understanding of the current situation (now, in short-term, mid-term, and over the longer-term), formulate strategies, and act on such strategies. In simple terms, one must continuously strive to create a culture where the fundamental performance levers are pulled to maximise results. A summary of the high-level levers or value drivers is shown below.
30
World Energy Council
Performance of Generating Plant 2004 - Section 1
As the above listings demonstrate, maximising performance is, indeed, a complex subject.
31
World Energy Council
Performance of Generating Plant 2004 - Section 1
32
World Energy Council
Performance of Generating Plant 2004 - Section 1
PGP DATA COLLECTION AND ANALYSIS During the last 12 years, the PGP Committee has been examining the compatibility of international technical databases, demonstrating the value of benchmarking, and, at present, has investigated the needs for further development of such systems to address the new, more competitive global power industry. The Committee has found that while data currently available continues to be of great importance and has high analytical value, there is greater potential and value in more in-depth, comprehensive data collection efforts. The Committee has therefore been involved in cooperation and discussions with a number of leading organisations throughout the world that collect such data. The Committee has members from: • EURELECTRIC/VGB; • IAEA – International Atomic Energy Agency; • NERC – North American Electric Reliability Council. Discussion above focused on the inability of technical indices to address today’s business and risk environment. The concept of commercial availability or CA was introduced; CA offers the ability to bridge commercial issues and technical issues. This discussion largely ignored, however, a second dimension to the data collection and analysis. This second dimension is the “depth” of the data. In detailed systems such as NERC GADS database, a great amount of detailed information is collected and analysed. Individual failure records are classified as to cause, type (planned, unplanned), as well as by offending system/component. Such information allows one to develop peer system or component-level analyses to help troubleshoot problem areas and to assess better “how” to improve overall performance results. Traditionally, statistics collected and evaluated at the international level have been aggregated by country and, hence, lack specific valuable detail: •
Size, age, fuel type, major equipment configuration/manufacturers, etc.;
•
System- or component-level failure, outage, and other maintenance detail;
•
Linking lost availability to a standard “degree” or measure of need of the unit;
•
Linking of the availability data to measure of urgency.
In many cases, the aggregation of data is a reality imposed by the lack of tools and/or processes to capture, validate, and collect/report data at a lower level. In other cases, such measures were viewed as a means for “protecting” the data. The result was that, limited by such conventional sources of data, many organisations have not been able to track or trend performance and reliability of power plants according to design, operations, or vintage criteria. Without such data, it is not possible to statistically decompose the population into statistically valid peer groups. This limits the accuracy and value of benchmarking activities and prevents the data from being consolidated with that of NERC to provide a truly global source for benchmarking data.
33
World Energy Council
Performance of Generating Plant 2004 - Section 1
At this time, the state of data collection and reporting systems are in a state of flux. Historical tools such as UNIPEDE’s LASCAR (used by PGP to collect/evaluate reliability data for the past 12 years) have been retired; other systems have evolved to take their place such as EURELECTRIC/VGB. However, the scope/focus on the deployment of such systems is far more “narrow” and focused on member issues. National and regional systems such as NERC GADS continue their normal level of activities. Paradoxically, at the same time as data systems and databases have either declined or remained constant, the interest within the industry in applying such data is growing substantially. This appears to be a function of the following: •
Failures in system reliability, including large-scale blackouts have demonstrated need for system reliability;
•
The excitement of the “market” to provide incentives/mechanisms required to create liquid, reliable supply of power has been tempered with the reality of inconsistent transmission capability, realisation that the physical asset (vs. financial market instruments) is fundamental to acceptable industry performance;
•
Lack of capital, limited maintenance due to uncertainty about continued plant viability, growing environmental regulatory pressures, and industry transformations have left many plants in sub-optimal conditions;
•
There is a critical need to be able to place available capital toward areas of greatest need AND identify non-capital intensive performance improvement solutions;
•
Irrespective of actual form of the market, its sheer presence will dictate need for operational strategies, processes, and measures that address the actual business realities;
•
Transparency in market operation and the ability to measure/benchmark performance relative to other industry players is a key strategy for how generators can improve performance;
•
Market dynamics and financial implications on performance make the problem of data collection and analysis that much more challenging.
It will be critical for the industry to develop new, more meaningful indices to evaluate commercial availability issues. As discussed above, the commercial aspects of plant availability and performance are critical even within regulated markets where cost efficiency is not needed to succeed within the power market but, rather, is needed to maximise the utility realised from limited financial resources.
34
World Energy Council
4.1
Performance of Generating Plant 2004 - Section 1
WEC Goal: A Global Data Collection and Analysis System
In working with other organisations in the ESI during the last several years, it has become clear that a “new” solution for collecting, managing, analysing, and reporting performance data and statistics is needed. It has also become clear that the best opportunity for success will lie in the ability to identify a solution that is absent from commercial pressures and capable of evolving over time. Finally, the system needs to take into account a wide range of issues for both developing and developed countries. In the end, it appeared that the best champion for the cause of a new data collection/reporting system was the WEC PGP Committee. In order to take on this task, the PGP Committee enlisted the participation of experts and international organisations to help define the path forward. Workshops and working meetings were held in the US and Europe to address both the data collection requirements from traditional point of view as well as that surrounding commercial availability. PGP Committee’s goal is to create the initial worldwide data collection/reporting system to measure the performance of power plant components, equipment and technologies; this system will be designed to work in conjunction with other existing systems: NERC GADS, EURELECTRIC/VGB, and PRIS. It will also be designed to permit continued evolution of the system capabilities and its ability to interface with other existing data collection systems. A working model of the data collection and reporting process is being tested to collect, compile, and report power plant statistics, is described in Section 2.
35
World Energy Council
Performance of Generating Plant 2004 - Section 1
REFERENCES 1. Richwine, R. R.; et. al.; International Data Exchange within the Global Power Industry – A Critical Activity for the Evolving Competitive Power Market; Houston, Texas. 2. Richwine, R.R.; Curley, G.M.; DellaVilla, S.A.; Lofe, J.J.; Reliability Measures Unreliable…It's Time for a Change; Published in the IGTI's Global Gas Turbine News, 1998 3. Stallard, G.S.; Salvaderi, L.; Richwine, R.R.; Spiegleberg-Planer, R.; Glorian, D.; Corrigal, M.; Micali, V.; Neibo, R.; International Data Exchange Within The Global Power Industry - A Critical Activity for the Evolving Competitive Power Market; Published at the 17th World Energy Council Congress, 1998 4. Stallard, G.S.; Micali, V., deSabastian, A.L.; Salvaderi, L.; Richwine, R.R., Glorian, D.; Heithoff, J.; International Availability Data Exchange for Thermal Generating Plant; published at the 18th WEC Congress, 2001 5. Stallard, G.S.; Richwine, R.R., Salvaderi, L.; Measuring Performance in a CoGeneration Plant, published at Power-Gen Europe Conference, 2000. 6. Electric Power Supply Association, “California: The Real Story, A Situation Analysis”, October 20, 2000, 1401 New York Avenue, NW, 11th Floor, Washington, DC 20005. 7. FEDERAL ENERGY REGULATORY COMMISSION, White Paper -- Wholesale Power Market Platform, April 28, 2003. 8. Eurelectric, “A Competitive Internal Energy Market,” position paper from website. 9. VGB, “Thermal Generating Plant (100 MW+), Availability and Unavailability Factors”, Statistical Data on the Performance of Fossil-Fuelled Power Plants
36
World Energy Council
Performance of Generating Plant 2004 - Section 1
CASE STUDIES 5.1
Use of Conditional Probability Methods to Evaluate Commercial Availability Robert R. Richwine, Chair WG7, Consultant (US)
This is an example of the calculation of Commercial Availability using a random sample of 10 hours during the year. The values are for demonstration purposes only. For simplicity, only full outages are considered. Partial outages would be handled in a similar fashion. In addition the example is for only one of the previously discussed CA definition options (other options would have similar calculations). The intent of this example is to demonstrate how CA could be used and benchmarked using Conditional Probabilities. The following are descriptions of the column headings: 1)
Hour – selected hour during the year; for most cases this will be sequential hours for each day, week, month, etc. For this example 10 random hours from throughout the year are used.
2)
Market Price - US$ per MWhr - (or Marginal Cost for regulated business environment) - this is the price (or cost of the next MW in the dispatch order) you would have to pay to replace the unavailable energy from the unit.
3)
Unit Variable O&M – US$ per MWhr - the unit’s variable cost to produce a MWhr of generation. It includes fuel plus a portion (usually small) of the maintenance budget (that portion that truly varies with the amount of energy produced).
4)
Gross Margin - US$ per MWhr - column 2 minus column 1 – this would represent the contribution to the company’s gross margin (profitability) by this unit in this hour per MW. If zero or below, use zero.
5)
Unit Size – MW the maximum net output capability of the unit.
6)
Potential Gross Margin - US$ - column 4 times column 5 – the sum of this column represents to total potential gross margin for the hours selected.
7)
Available (Yes=1; No=0) – Was the unit available in this hour (for simplicity, no partial availability is considered in this example).
8)
Actual Gross Margin – column 6 times column 7 – the sum of this column represents the total actual gross margin achieved for the hours selected.
9)
Conditional Probability – this column is the goal that is set for the reliability of the unit during this hour (this can be a very complex process if a goal is desired that is optimal from an economic perspective.
10) Goal Gross Margin – column 6 times column 9 – the sum of this column represents the total “goal” gross margin for the hours selected.
37
World Energy Council
Performance of Generating Plant 2004 - Section 1
CALCULATIONS Traditional Availability (%) = total hrs in which the unit was available divided by period hours times 100. Forced Outage Rate (demand) (%) = only for hours when gross margin is positive, total unavailable hours divided by total hours times 100. Commercial Availability (Actual) (%) = Total Actual Gross Margin achieved (column 8) divided by Potential Gross Margin (column 6) times 100. Commercial Availability (Goal) = Total Actual Gross Margin achieved (column 8) divided by Goal Gross Margin (column 10) times 100. Gross Margin Achieved above/below Goal = Total Actual Gross Margin achieved (column 8) minus Total Goal Gross Margin (column 10). EXAMPLE 1 1 Hour
2
3
4
5
6
7
8
9
10
Price (Cost)
Var. O&M
Gross Margin
Unit Size
Potential Gross Margin US$/ MWh
Avail. (Y/N)
Actual Gross Margin US$/ MWh
Conditional Probability
Goal Gross Margin US$/ MWh
3000 0 0 12000 24000 18000 9000 66000
.92 .90 .92 .95 .98 .98 .98 .97 .90 .90
2760 1380 5700 11760 23520 17640 8730 71490
= = = = =
60.00% 28.57% 89.80% 97.27% -US$5490
US$/ MWh
1 2 3 4 5 6 7 8 9 10
20 20 20 20 20 20 20 20 20 20 TOTAL
US$/ MWh
10 5 20 40 80 60 30 30 -
300 300 300 300 300 300 300 300 300 300
3000
30 15 25 40 60 100 80 50 10 20
1500 6000 12000 24000 18000 9000 73500
Traditional Availability Forced Outage Rate (demand) Commercial Availability (Actual) Commercial Availability (Goal) Gross Margin Achieved (above/below goal)
= = = = =
1 1 0 0 1 1 1 1 0 0
6/10 X 100 2/7 X 100 66000/73500 X 100 71490/73500 X 100 US$66000–US$71490
US$
EXAMPLE 2 If the unit in the above example had been able to be returned to service by hour 4, with everything else constant, then: Traditional Availability Forced Outage Rate (demand) Commercial Availability (Actual) Commercial Availability (Goal) Gross Margin Achieved (above/below goal)
= = = = =
38
7/10 X 100 1/7 X 100 72000/73500 X 100 71490/73500 X 100 US$72000-US$71490
= = = = =
70.00% 14.29% 97.96% 97.27% +US$510
World Energy Council
Performance of Generating Plant 2004 - Section 1
EXAMPLE 3 If the unit in the example 1 above had been available in every hour except hour 6, then: Traditional Availability Forced Outage Rate (demand) Commercial Availability (Actual) Commercial Availability (Goal) Gross Margin Achieved (above/below goal)
= = = = =
9/10 X 100 1/7 X 100 49500/73500 X 100 71490/73500 X 100 US$49500-US$71490
= = = = =
90.00% 14.29% 67.35% 97.27% -US$21990
In the above three cases, example 3 gives the best traditional availability numbers but is clearly the worst result from a bottom line, financial perspective. This demonstrates how the traditional reliability measures can differ dramatically from newer, market-based indicators.
39
World Energy Council
Performance of Generating Plant 2004 - Section 1
40
World Energy Council
5.2
Performance of Generating Plant 2004 - Section 1
South Africa’s ESKOM Experience: Commercial Availability Application Vince Micali, Corporate Consultant, Business Sciences, Generation Strategy Dept., ESKOM (South Africa)
Introduction This report is the result from the South African perspective, to provide a short concise introduction the Commercial Availability Indicators produced for the World Energy Council (WEC) PGP Committee. These types of Indicators are presently being used in quite a few countries worldwide. The Short Summary below is an excerpt from the paper already submitted to the Committee. Development Firstly it is necessary to differentiate between concepts and indicators. For instance, the Technical Availability (TA) is a concept and the indicators could be UCF, EAF (defined by UNIPEDE – NERC). Similarly, the Commercial Availability (CA) is also a concept and indicators need to be formulated for quantification.
Commercial Availability (e.g. BAV)
Technical Avail.
Financial Perf.
(e.g. EAF)
(Earnings)
Potential
Actual
Expense
Value
(Inst. Cap.)
(Av. Cap.)
(Cost of Prod)
(Sys. Price)
41
World Energy Council
Performance of Generating Plant 2004 - Section 1
The Framework There are two families of indicators: the Valuable Availability Balance (VAB) which is one set generally used to measure the Trader’s performance and the Banked Availability Value (BAV) utilised at different Boards to measure Commercial Availability (in case of ESKOM) this is reported monthly and contracted on a yearly basis). The mostly utilised family is the BAV and the definitions of its Indicators are given below. Family of Indicators: Banked Availability Value (BAV): Definitions of Computations are given in bold.
Revenue = CoS Dispatched (Added Value) (Actual (MW) Earnings x MW NOT Dispatched
Revenue < CoS (Destroyed Value) Earnings x MW
(Missed Opportunity) (Not Competitive) Ptl Earnings x Ptl Earnings x [Max{IC;Dcl} – MW] [Max{IC;Dcl}-MW]
BAV Indicators calculate the weighted financial (monetary) implication of a unit which can be aggregated at Power Station and Generating Company level for a specified time window. They can be expressed as monetary or % of the absolute total. Example: If a 500 MW unit with a marginal cost (Cost of Sales) of US$40/MWh was sending out 300 MW, for a particular hour, when the system marginal price (spot price for that hour) was US$50/MWh. The Earnings, for that hour, are US$ (50-40) and the profit realised was (50-40) x 300 = US$ 3000 and the potential profit was (50-40) x (500) = US$ 5000. Hence the generating unit could have had a profit of US$5000, but due to bid strategy earned only US$3000. Thus the generating unit had added value (AV) of US$ 3000 and missed opportunity (MO) [(50-40) x (500-300)] of US$ 2000. By definition missed opportunity returns a negative value (reflecting a loss). Therefore: BAV (AV) = 3000 (or 60%); BAV (MO) = -2000 (or -40%) Similarly, if the spot price was US$30/MWh, then the Earnings for that hour would be US$(30-40) and the loss realised by that unit would be (30-40) x (300) = -US$ 3000. Thus the generating unit destroyed value (DV) to the tune of -US$ 3000. Hence, BAV(DV) = -3000 (or -60%). However, the trader did well at not despatching 200 MW for that hour, as the market moved against the unit, and the unit was not competitive (NC) on the market to the tune of (30-40) x (500-300) = -US$ 2000; hence, BAV(NC) = -2000 (or -40%).
42
World Energy Council
Performance of Generating Plant 2004 - Section 1
At times, the declared unit (Dcl) capacity exceeds the installed capacity (IC) value. In that case, the declared (the maximum of the two) is used. These values are measured per hour and aggregated according to needs (e.g. daily/unit, weekly/(units 1 & 3), monthly/Station). CONCLUSION The BAV family of indicators presently constitute a battery of statistics for measuring Commercial Availability and are included in Performance Contracts for the various Generating Companies. For this purpose, it was necessary to develop spreads that assured equity in the setting up of these forward performance contracts. Depending on the trends, a board of executives of a GenCo may decide to concentrate on, say, Added Value and Destroyed Value indicators with equal weights. Another GenCo’s board may include all four indicators with Added Value having the highest weighting factor (e.g. AV: 50%; DV 30%; MO 10% and NC 10%). In Eskom, South Africa, these indicators have been used for the past four years with success to win the confidence of various boards of executives and were implemented in the fashion described above. In conclusion, these indicators performed well as a measure of Commercial Availability and executives were able to quantify the generating units’ exposures and, when necessary, assume hedging positions on these units.
43
World Energy Council
Performance of Generating Plant 2004 - Section 1
44
World Energy Council
5.3
Performance of Generating Plant 2004 - Section 1
The Italian Power System: Recent Evolution & Issues Luigi Salvaderi, Consultant, Fellow IEEE
The Evolution The Italian power system has been gradually deregulated, in accordance with the European Directive 96/92/EC and the new Italian “Electricity Law”, the so-called Bersani Decree 79/99 of 31/3/1999. Italy has established an independent Office of Regulator for Electricity and Gas, according to the law of 14 November 1995, No. 481. The previously monopolistic vertically integrated national utility, ENEL Spa, has been unbundled. In the Power Generation Sector, 15 GW of installed capacity grouped in three GenCos, were transferred from ENEL in November 1992 and sold to companies with mixed Italian /foreign capital. The three GenCos are: 1.
Elettrogen: with installed capacity of 5.4 GW (now 5.8 GW). Shares: 85.3% Endesa (Spain) – 14.7% ASM Brescia (Municipalities of Italy);
2.
Edipower: installed capacity of 7.0 GW. Main shareholder with 40% is Edison-Milano, which in its turn is controlled by the Italenergia Bis (leb) 64%, C Tassara Finanziaria 16%, Edf 18% and Banca Intesa 2%.;
3.
Tirreno Power: 2.4 GW, owned by Electrabel (Belgium) and ACEA (Municipality of Rome).
The competition in the free market began, despite only a few actors on the supply side, all limited to “Big Companies”, with ENEL Production retaining its dominant position. In the Transmission Sector, a new independent Transmission System Operator (TSO), the GRTN, was created and made responsible for Operation and Control of the power grid. GRTN’s shares are controlled by the Treasury, while GRTN also owns the Market Operator. The ownership of the “wires” was left to the previous owners, the major - but not the sole - being ENEL’s daughter Company TERNA. Soon it became obvious that this model had to be reviewed, and finally Parliament approved the reunification of the system operation and ownership in a new Transmission Company, through a recently adopted Bill (n. 251 of 28 October 2003). The shares will be traded on the market, even if public control is anticipated to continue. No company with direct interests in Production, Import, Distribution and Supply of Energy will be allowed to have more than 5% of voting rights in the new Transmission Company by 1 July 2007. This move complies with the framework of the Regulation EC n. 1228/203 of June 2003, on conditions for access to the network for cross-border exchanges in electricity. In June 2004, almost 50% of the shares of Terna, owned by ENEL were placed on the market. In the Distribution Sector and in particular in the Municipalities (MUNIS), where both ENEL and MUNIS have been competing, further rationalisation of grid ownership continued, with a transfer of assets between ENEL-MUNIS according to market
45
World Energy Council
Performance of Generating Plant 2004 - Section 1
conditions. The transaction was carried out in two stages: in 2002, 10 Distribution Companies were sold by ENEL to Municipalies, whilst 12 were then purchased by ENEL. Demand by customers with a current consumption rate >0,1 GWh/yr (in March 2003) corresponding to 170 TWh = 57% of the final consumptions, has decreased.. A further opening to clients with yearly consumption at 0.05 GWh/yr has been proposed in a Bill currently under examination by Parliament. According to the EU Directive 2003/54/EC of 26 June 2003, concerning common rules for the internal market in electricity and repealing Directive 96/92/EC, complete liberalisation for industrial and commercial consumers is expected by 1 July 2004, with a full opening of the market by 1 July 2007. The Italian Electricity market (IPEX) began operating on 31 March 2004. After lengthy discussions, a non-compulsory pool concept was adopted, where bilateral contracts coexist with the exchange. The Market Today There are markets for: i) Energy: Day Ahead & Adjustment, run by the market Operator (GME); ii) System services: Congestion, Operational Reserve, Real Time), run by the ISO. A provisional cap of 500 €/MWh has been enforced. On the Exchange, suppliers are paid different prices for different areas to take into account the existing inter-area congestion, whereas consumers pay a uniform nation-wide price. Even if the model is basically two sided (supply & demand), the demand will be set by the ISO until the end of 2004. A peculiar feature of the Italian market is the Single Buyer (AU), a Company fully owned by the ISO and operational since 1.1.2004. Its- mission is to protect consumers and decrease wholesale energy prices. Its supply sources are:i)
long term contracts with France & Switzerland, in the past affected by ENEL;
ii)
preferential allocation of annual import capacity, with the possibility of contracting foreign suppliers;
iii)
allocation of up to 20% of discounted incentive-based energy;
iv)
physical bilateral contracts with Italian producers, accounting for up to 25% of demand;
v)
financial contracts for differences for base, intermediate and peak capacity, met through auctions;
vi)
the remainder comes from purchases on the IPEX. The AU has been very active and by mid-2004 it has already contracted some 60% of demand for the whole 2004.
Competition in the market begun before the IPEX went “live”, but it was – and to some extent still is unbalanced, due to supply being lower than demand. The market is “short” and the producers/traders are still “price makers”. Italy’s generation mix is predominantly oil/gas based, using relatively old units. Coal use is insignificant, whilst nuclear is not used at all. Hydro resources are underexploited. Import capacity from Central Europe is also limited and is lower than the 10% figure suggested by the European Commission.
46
World Energy Council
Performance of Generating Plant 2004 - Section 1
Consequently, the wholesale price for electricity in Italy is almost double that of in Central Europe. Market liquidity in the first month of IPEX operation was high, at 30% and dominated by trading of marginal energy, since the prevalent concept used for base-load energy is based on bilateral contracting. The average weighted price in April 2004 was 57.55€/MWh. Price tensions were recorded in June. In addition to the transparency of spot prices, financial derivatives are fundamental to ensuring some certainty to those using the market and to encouraging a forward-looking attitude. The Issues The disparity in prices between neighbouring countries explains the maximum load on interconnections at the Northwestern and Northeastern borders. It has also had an impact on the rules governing the allocation of import capacity, including recourse to the Courts, and has resulted in the present pro-rata mechanism. Electricity Production in Italy Fuel Shares (2002) Other Thermal: 4.8% Oil: 26.4%
Hydro: 17.3% Georthermal: 1.8% Wind & PV: 0.6%
Coal & Other Solids: 12.3%
Natural Gas: 36.8%
In 2002, total electricity demand in Italy reached 310.7 TWh and 290.5 TWh were traded. 60% (175 TWh) were sold to the captive market and 40% to the free market (95 TWh + 20.5 TWh of self consumption). The Ministry of Productive Activities (formerly Industry) made some moves to fuel the free market, with the preferential allocation of import capacity. It also reduced prices of the purchase contracts for renewable and other specially treated energy GRTN inherited from ENEL. In 2002, the supply to market, net of self-consumption was: i) import 30 TWh (out of a total of 50.6 TWh); ii) 40 TWh of renewables out of 55 TWh, and iii) 25 TWh purchase from national producers. The high import percentage has raised new concerns following the black out which affected Italy on the morning of Sunday, 28 September 2003 (the import was 6.4 GW, out of total demand of 24.6 GW). A number of inquiries were launched to investigate the cause of the blackout and to suggest remedies. The Union for the Coordination of Transmission of Electricity (UCTE) presented an Interim Report on 27 October and the Swiss Federal Energy Office (UFE) presented a report on 25 November 2003. Common statements by the Italian and French Regulators were released on 1 December and expressed concerns about the UFE Report. They noted a lack of coordination in drawing a conclusion between the UFE and the two Regulators. 47
World Energy Council
Performance of Generating Plant 2004 - Section 1
The lack of information on the status of the Swiss network at the ISOs of the French and Italian interconnected network is an issue. The compliance with the N-1 criteria according to the UCTE standards for all interconnected networks after the first line tripping is also a basic issue. The report by the Italian Regulator, issued on 9 June 2004, particularly stressed that the Swiss electric power companies did not follow the UCTE N-1 rules. It also noted that some Italian generating units trip before the adopted intervention level (47.5 Hz) of the frequency and questioned the overall level of the load shedding mechanisms. A formal inquiry into these and other issues is expected. A report by the Enquiry Commission, launched by the Italian Government, is also forthcoming. The recent blackouts in Europe were a key trigger for the European Commission to launch a package of measures for electricity supply security. They should be in place in time for the Energy Ministers’ Council session on 4 December 2004 and are expected to ensure that the efficient operation of the EU internal electricity market. Supply Security Notwithstanding a theoretically high reserve margin computed with reference to the “installed capacity”, a lot of the installed capacity is not available for operation for various reasons. This results in a tight operational reserve margin. The lack of “cheap power” in recent years has triggered the emergence of a huge number of new market entrants and requests for connection to the National transmission network of more than 100 GW! The industry aims to increase the efficiency of the generation mix to 46% by 2010, compared to 40% at present. So far, 8 GW of re-powering in 10 power plants, 25 new plants, all of them CCGT have been approved, and 11 GW out of the total is to be commissioned in 2008. A very sensitive and so far unresolved political problem, is the balance of power between the regions where the plants will be sited and the Central Administration, which authorities their work. Some perplexities still exist concerning the “right” amount of CCGT, new and transformed, the system could accept with the number of utilisation hours capable of ensuring the required return on investments. A scheme of “capacity payment” has been devised by the Government and its implementation is under way. Renewable Energy In order to support renewables with a marked based approach, the Government ruled that each Producer/Importer must ensure that 2% of its production/import comes from Green Power, certified by the GRTN with related Green Certificates. The Green Certificates, in force since 2002, are coupons, which can be traded on an Exchange, run by the Market Operator since 28 March 2003, or through bilateral contracting Developers/ Producers/Importers. The progressive increase of the renewables share by 0.35 %/yr from 2004 to 2006 is included in the Bill (n. 387. 29 December 2003) and it introduces in the Italian legislation the “Renewable Directive” 2001/77/CE dated 27 September 2001. At the end of 2006 the renewable obligation will therefore reach 3.05%. Further Decrees planned for 31/12/04 and 31/12/07 are expected to increase these percentages for three-year periods 2007-2009 and 2010-2012 respectively. The target of the EU Directive is to achieve by 2010 up to
48
World Energy Council
Performance of Generating Plant 2004 - Section 1
22.1% share of electricity production from renewables in the European Internal Market, a 1997 baseline of 13.9% is used. The share for Italy is 25%, against a 1997 baseline of 16%. Environmental Policies A related issue is the impact of the Kyoto Protocol on industry. The Protocol will enter in force when a minimum of 55 Countries that account for least 55% of the CO2 emission in 1990 have ratified it. After the USA pulled out of the Protocol, 122 countries were left, together accounting for only 44.2% of emissions. Russia’s ratification of the Protocol is still uncertain as political discussions still continue and possible agreements are explored. The European Union has agreed a reduction target of 8%; its 15 Member States (MS) made on 4/3/2002 an internal agreement to distribute the total target in various “burden sharing”. EU and its MS ratified the KP on 2/4/2003. Italy ratified the EU burden sharing agreement with the Bill 120/2002 on 1/6/2002; and has committed to a reduction of 6.5% on the 1990 emission level by 2010. The Kyoto Protocol establishes three mechanisms to allow flexibility in compliance: Emission Trading (ET), a cap and trade mechanism applied to developed countries, based on an “ex ante” allocation of allowances by various Governments to each “eligible” sector with National Allocation Plans (NAPs). The ET Directive covers only 6 GHGs and only some sectors (Combustion Plants, Mineral oil refineries, Coke ovens, Ferrous metals, Cement, Lime, Brick, Pulp & Paper, Glass, Ceramic). It will become compulsory from 2005 and will be implemented in two phases: 2005- 2007 and 2008-2012. While common issues have been already established, the criteria for the precise allocation of allowances to the various sectors by the Governments, which directly impact the cost allocation, and the competitiveness of various sectors- is still an open issue. The two others mechanisms provide flexibility in compliance by introducing, within certain limits, the equivalence of “credits” acquired by project-based solution to the allowances allocated by the Government. The rationale is that both a minor compliance cost and an increase in efficiency will be obtained by a reasonable recourse to such mechanisms. Joint Implementation (JI) will be applied only to countries that under KP have a cap, namely developed countries or to countries with economies in transitions. Since the countries have caps, the mechanism is a zero-sum operation: it is expected that JI will take place especially in Russia, having a great potential for transfer of advanced technologies. Emissions Reductions Certificates (ERCs) will be recognised for reductions against a baseline. Clean Development Mechanism (CDM) allows investors from the developed countries to invest in developing nations, which have no quantitative targets but have ratified the KP. The investors will receive Certified Emission Reduction (CER) credits issued by a UN body. A Directive to link the ET Directive and JI and CDM was approved by the European Parliament in April 2004. The mechanisms are expected to enter in force in 2008, subject to the KP entering into force.
49
World Energy Council
Performance of Generating Plant 2004 - Section 1
A basic issue is the relation between the cost (Euro/Mtonne) of the allowances – and the related penalties for not compliance- which will materialise with the ET and the cost corresponding to the “credits” – in perspective well lower- gained with the JI and CDM. To fulfill the Italian obligations, the Interministerial Technical Committee (CTE) updated in April 2004 a previous Ruling of CIPE n. 123/2002 “Revision of the Guidelines for National Policies and Measures for GHGs Reduction”. Against the 1990 baseline (508 Mt CO2), the burden sharing of –6.5% corresponds to a target of 475 Mt. The Outlook Two scenarios are now considered: Business as Usual (BAU) would by 2010 entail for all sectors an increase up to 607.7 Mt CO2. Reference Scenario 2010: with measures already approved. A first reduction of 37,5 Mt, down to 570,2 Mt CO2. An huge contribution (26 Mt= 69% ) is required from the electricity sector. It is worthwhile to underline that by comparing the Reference and the BAU Scenarios, the reduction required from the sectors under the ET Directive (304.4 Mt in Reference vs. 330.4 Mt in BAU) is 7,9%, higher than the one required from all sectors (6.2%). The reduction required from the thermal power sector (153.5 Mt Reference vs. 179.5 BAU) is 14% almost the double. Criteria for the allowances allocation to the thermoelectric sector have been proposed and are presently being discussed. The allocation mechanism should take into account that the Italian generation system will undergo a considerable restructuring. Different operational ranges have been established for the various technologies. Allocation should be based on future expected emissions, based on the yearly utilisation hours, by technology, according to specific emission rates. A modest recourse to JI&CDM (12 Mt) and a further reduction of 11.2 Mt CO2 from national sinks should reduce the emission to 547 Mt CO2. The remaining gap to the target comes out to be 72 Mt CO2. The strategy of the Italian Government to comply is based on consideration that GHGs control is a “global issue”. Correspondingly a strong focus to extended use of JI& CDM, in addition to ET, is needed. A great concern for the Government and Industry is the impact that the cost for purchasing emissions allowances utilising the ET mechanism could have on the Italian economy: various unit costs, ranging from 10-50 and even up to 100 €/t, are anticipated on which various scenarios can be examined. It is necessary to safeguard the Italian industry in the competitive European Market; the high efficiency and the low carbon intensity of the Italian industry must be taken into account. The marginal cost of internal measures , in term of GNP/CO2 is much higher for Italy than for the other Member States. The Government is correspondingly pushing for an enlarged utilisation of the other two flexible mechanisms (JI and CMD): the same “global target” can be
50
World Energy Council
Performance of Generating Plant 2004 - Section 1
obtained at a much lower compliance cots. In this framework, Italy signed agreements with World Bank on May 2004 for the Community Development Carbon Fund. (CDCF). In the near future many of the present uncertainties will be the focus of the Government and Parliament’s initiatives.
51
World Energy Council
Performance of Generating Plant 2004 - Section 1
52
World Energy Council
Performance of Generating Plant 2004 - Section 1
APPENDIX 1: A Demand-related EFOR An EFOR Equation (or Formula) for Generating Units of any Duty Cycle - The Markov Approach Summary For many system planning and production reliability applications, a standard equation is needed to estimate a demand-related Equivalent Forced Outage Rate (EFOR). A demandrelated EFOR represents the probability that a generating unit will not (or did not) meet its required generation demanded by dispatch. The current IEEE-Standard 762 EFOR equation, also used by the North American Electric Reliability Council (NERC) Generating Availability Data System (GADS), adequately approximates a demand-related EFOR, but only for baseload units. Efforts have been taken in the past (see References) to solve this dilemma. The Markov approaches used in two 1970s technical papers have been combined and refined. The resulting Markov equation is universally suitable to approximate a demand-related EFOR for generating units having any duty cycle. The Markov equation is complex in appearance. However, the Markov approach simply uses the relative historical average forced outage, reserve shutdown and service time (duty) durations to calculate a discount factor, which approximates how much of the reported forced outage time occurred during actual demand conditions. Since this approach provides only an approximation, it should not negatively impact its use as a prediction tool. For applications requiring an exact accounting of demand-related forced outage time (such as contract guarantees), the only perfect method is to record the exact demand time-line with the corresponding outages. This task force endorses the adoption of the Markov approach to estimate demand-related EFOR by IEEE and NERC GADS.
Background EFOR has been in use for modelling generating unit forced outages and deratings in electric utility generating capacity system reliability and production cost evaluations for over 40 years (1). EFOR is also widely used by utility production support staff. Typical applications are goal setting, incentive awards, benchmarking and reliability evaluation. EFOR is defined in the IEEE Standard 762 and in NERC GADS, but with a caution that the definition may not be applicable to non base-load units. There were multiple attempts to modify the definition to suit non base-load units (2, 3, 4, 5).
53
World Energy Council
Performance of Generating Plant 2004 - Section 1
The Markov Approach The basic equation to determine EFOR is as follows: EFOR= (f x FOH) + (EFDH - EFDHRS) x 100% SH + (f x FOH) where:
FOH = EFDH= EFDHRS= SH= f= = r= = T= D= = T+D
(Full) forced outage hours Equivalent forced derated hours Equivalent forced derated hours during reserve shutdown Service hours Discount factor for FOH (1/r + 1/T)/(1/r + 1/T + 1/D) — Reference (4) Average forced outage duration FOH/number of forced outages Average reserve shutdown time Average demand time (duty cycle time) SH/number of successful starts Available hours/number of starts
Advantages of the Markov Approach The following are considered as the advantages of using the above equation to calculate EFOR. 1. Directly provides an approximate demand-related EFOR, which is a popular index for planning, production, and design studies in the U.S. and other countries. 2. Applicable to units with any duty cycles. For truly base-load units (continuous demand) the discount factor would approach 1. 3. It discounts the reported forced outage time for those non demand-related periods when there is little (or no) urgency to repair. The non demand-related periods, by definition, are not applicable to a demand-related EFOR.
Disadvantages of the Markov Approach The following are the disadvantages of the Markov approach. 1. Appears complex. It is more complex that the current IEEE Standard 762 EFOR equation. However, it requires no additional data-reporting burden. Computers handle the calculations. 2. It is an approximation. However, it is a major improvement over the current EFOR equation to estimate a demand related EFOR. It uses average and relative forcedoutage event durations, duty durations, and reserve shutdown event durations to approximate the forced outage discount factor. The only exact way to calculate a demand-related EFOR would be to report the demand time for each generating unit along with the unit’s events. This could be a costly effort and potentially subjective.
54
World Energy Council
Performance of Generating Plant 2004 - Section 1
References [1]
G. Calabrese, “Determination of Reserve Capacity by the Probability Method,” AIEE Transactions, Vol. 69, Part II, 1950.
[2]
P.F. Albrecht, W-D Marsh, F.H. Kindl, “Gas Turbines Require Different Outage Criteria,” Electric World, April 27, 1970, pp. 38–40.
[3]
A.M. Adamson, “Gas Turbine and Diesel Forced Outage Rates and Their Application to Reliability Calculations,” IEEE-ASME Joint Power Generation Conference, September 27–30, 1970.
[4]
IEEE Committee, “A Four-State Model for Estimation of Outage Risk for Units in Peaking Service,” IEEE PES Transactions, March/April 1972.
[5]
M.P. Bhavaraju, J.A. Hyrids, G.A. Nunan, “A Method for Estimating Equivalent Forced Outage Rates of Multiple Peaking Units,” IEEE Transactions on PAS, November/December 1978.
55
World Energy Council
Performance of Generating Plant 2004 - Section 1
APPENDIX 2: NERC GADS North America’s Electric Power Plant Database Improved: Now 98% Compatible with EURELECTRIC/VGB Michael Curley, Manager, GADS, North American Electric Reliability Council For the last three years, GADS Services of the North American Electric Reliability Council (NERC) has worked with the Institute of Electronic and Electrical Engineers (IEEE) to review and improve the IEEE Standard 762, “Definitions for Reporting Electric Generating Unit Reliability, Availability and Productivity.” IEEE 762 is the basis of equations and definitions for the NERC Generating Availability Data System (GADS). With the help of North American, European and South African colleagues, the IEEE 762 has moved closer to that used in other parts of the world. This paper explains several major changes to IEEE 762 to bring about the compatibility. It also mentions some applications that are used by NERC GADS. Background and History Since the early 1960s, North America electric utilities have been collecting equipment failure data on power plant. At first, the only power plants examined were large fossil (400 MW and larger) and nuclear units. In 1979, the NERC was asked to operate and maintain a database of power plant equipment for all technologies and unit sizes from1 MW and larger. With the electric industry experts, the data collection process was expanded and became more comprehensive. More information was collected and processed so that more and better applications could be used to enhance and improve the electric generating units in North America. GADS has provided this service for 22 years and it is the basis of power plant data collection in a number of countries worldwide. As part of that enlargement, there was a need for standardised definitions and equations. NERC worked with the IEEE committees to create a Standard for GADS to use. That task was completed in 1980. The IEEE 762 had a number of definitions and equations that the North American electric industry had been using for years. But at the same time, it also introduced new concepts that have proven to be helpful in examining the performance and productivity of power stations today. One of the problems with IEEE 762 was its focus on certain areas. As mentioned, when IEEE 762 was introduced in 1980, the majority of the units reporting to GADS were large, base-load units. As a result, the equations and definitions of IEEE 762 were focused on units with very little idle time (reserve shutdowns). Peaking and cycling units were treated as if they were base-load. Some equations were applicable to any type of unit operation or technology while others such as forced (unplanned) outage rates distorted the performance of non-load generation. Also, the equations included in the Standard were all time-based. These unweighed equations allowed 500 MW fossil units and 50 MW gas turbines to be equal on the impact of station performance equations. Energy-based (weighted) equations were not mentioned although they were commonly used by many utility owners.
56
World Energy Council
Performance of Generating Plant 2004 - Section 1
What is new with IEEE 762? When the IEEE Standard 762 Review Working Group was created in 2001, it was asked to review the existing Standard and the suggested improvements from GADS Services, and add to the Standard the necessary elements to make it helpful to all electric utility analysts and users. As a result, there are a number of new changes to IEEE 762. Some of these improvements include: 1. Demand-related Equations. Not all generating units operate in base load, there are also units operating on an “as needed” basis and those that move from base load to cycling or peaking operation as they mature with age. It is therefore important to monitor the performance of such units in a way, which will allow accurate and fair evaluation of the unit performance. There are several new equations in IEEE 762 that are applicable to all operating modes. One equation, Equivalent Forced Outage Rate – demand, was developed in the 1970s by a set of mathematicians and engineers to determine the performance of peaking gas turbines and similar machines. It had been used by one of the NERC Regions, Mid-Atlantic Area Council (MAAC or PJM Interconnection) for many years. It analyses the probability of having a forced (or unplanned outage) event at the time when the unit is needed to provide power. This same equation is referred to as “conditional probability” for use in commercial availability work. This equation is presented in the PGP Work Group 6 report at the end of Section 5.1. Another equation was introduced by the Canadian Electricity Association (CEA) around 1980. The results of their work are reflected in an equation entitled Adjusted Utilisation Forced Outage Probability (DAUFOP). DAUFOP is currently used by some generators in Canada. Other equations have been added to the Standard to monitor units, which are not operated in base load. 2. Energy-based Equations. Since 1980, IEEE has recognised time-based equations. These equations treat all units of various sizes on an equal basis. A 500 MW fossil unit is treated in the same way as a 50 MW gas turbine when combining the outage hours or calculating forced outage rates. The new IEEE 762 recognises both time-based (unweighted) and energy-based (weighted) statistics. With weighted equations, the larger units are emphasised more in grouped statistics than are the smaller units in the group. In using the example above, the 500 MW unit would have 10 times the impact on the group statistics as would the 50 MW gas turbine because it is 10 times as large. Energy-based equations used in Europe are now part of the Standard. See item 3. 3. Outside Management Control (OMC) Concepts and Equations. European statistics have taken into consideration certain problems that occur outside the control of plant management for a number of years. These problems are described in the UNIPEDE document “Detailed Descriptions of International Performance Indicators for Fossil-fired Power Plant” under the section “Data Qualification Requirements”,
57
World Energy Council
Performance of Generating Plant 2004 - Section 1
“Clarifying Notes” dated December 1991. These OMC problems include natural disasters, loss of transmission lines, strikes, etc. IEEE 762 and GADS now follow the vast majority of these events outlined by the UNIPEDE document. Some parts are clarified to meet unregulated generator monitoring. UNIPEDE (now Eurelectric) recognises Reserve Shutdown or RS (lack of demand) periods as OMC. RS events are periods of time when dispatch does not need any load and so the generating unit is idle. Also, there are times when the unit is not at full load because of lack of energy needs. Both Eurelectric and IEEE 762 recognise that dispatch requirements are OMC, yet they are still included in many calculations. With the IEEE 762 now recognising OMC events, GADS has begun to calculate Unit Capability Factor (UCF) and Unit Capability Loss Factor (UCLF). These equations have been used in Europe for some time now and they are similar to the IEEE 762 Equivalent Availability Factor (EAF) and Equivalent Unplanned Outage Factor (EUOF), except that the OMC events are removed from the calculations. UCF and UCLF include RS events. More Eurelectric equations will be added to GADS over the next several years.
4. Revision to Forced Outage Equations. The original Forced Outage Rate (FOR) and Equivalent Forced Outage Rate (EFOR) equations considered service hours (SH) as the only times that equipment was in operation and could fail. Base-load units are always in demand and are operated. The Standard did not consider other non-generating yet operating modes. As the IEEE expands to provide performance standards for all technologies and operating modes, it is also important to review non-generating operations of the units. Pumped storage units, combustion turbines, and other technologies have been known to fail while in pumping mode or synchronous condensing operation. Although they are not providing electrical power to the grid, they are still providing a “service.” The new IEEE 762 now considers pumping hours (for pumped storage units) and synchronous condensing hours (VAR supplying) as operating modes and adds the pumping hours and synchronous condensing hours to SH. The new equations are:
forced outage hours FORT = *100 (forced outage hours) + (service hours)+(service hours non-generating mode)
And FOH + ∑ equivalent forced derated hours EFORT = *100 SH+SH non-generating mode+FOH+ ∑ equivalent reserve shut down forced derated hours
58
World Energy Council
Performance of Generating Plant 2004 - Section 1
Again, the IEEE 762 is moving away from the traditional base-load equations to those applicable to cycling and peaking operations. 5. Data Pooling Methodologies Explained There has been a number of ways that the industry averaged plant or peer group statistics. Some for example, have added together 20 forced outage rates and then divided the sum by 20. Others have determined more sophisticated ways to pool statistics. Several years ago, the IEEE Reliability, Risk and Probability Applications Sub-committee (RRPAS) and NERC GADS came up with a better way to pool or average statistics. The method involves summing up each component of the equation in hours and then placing the sums into the equation. The methodology is used for both time-based and energybased equations. Examples of the methodology are shown here for forced outage rates: Time-based (unweighted) n FOHi ∑ *100 FOR = n i =1 ( FOHi + SHi ) ∑ i =1
Energy-based (weighted) n ( FOHi * NMCi ) ∑ *100 WFOR = n i =1 [( FOHi + SHi ) * NMCi ] ∑ i =1
… Where NMC is net maximum capacity (or reference capacity.) IEEE 762 now contains equations for single unit, for pools of units using unweighted calculations and for pools of units using weighted calculations. The RRPAS method of pooling data has been used by NERC GADS for many years. These equations are now part of the IEEE 762 document and are also in the GADS Data Reporting Instructions, Appendix F at http://www.nerc.com/~filez/dri.html.
6. Expansion of Unplanned Outage Extensions Definitions Since 1980, IEEE 762 recognised only extensions to planned outages. However, the North American electric industry also recognised extensions to maintenance outages. The maintenance extensions were then introduced and used in GADS. A combination of IEEE 762 planned and maintenance outages are very close to the Eurelectric unplanned outages. In Eurelectric and Canada, any extension to a planned outage is forced while in GADS, the extensions are not forced but are a part of the original planned or maintenance outage.
59
World Energy Council
Performance of Generating Plant 2004 - Section 1
IEEE 762 and GADS have now modified unplanned extensions to be either part of the original outage or forced outage. The determining factor is whether the cause for the extension is part of the original scope of outage work or it was a new repair/problem discovered during the unplanned outage. If it is new repairs/problems, then they are considered as forced. Otherwise, they are part of the original work and are not forced outage. 7. Common Terms in Line with Industry. IEEE 762 has updated the nomenclature to include maintenance outage, maintenance extensions, demand related equations, weighted and unweighted equations as well as several other words and terms. The Standard continues to use “unplanned” in its definitions too. In Eurelectric, a planned event is one scheduled at least four weeks in advance. In IEEE 762 and GADS, a planned event is “scheduled well in advance” but “well in advanced” does not have a time set on it like Eurelectric. However, it is felt that the unplanned events (forced and maintenance outages and deratings) and planned outages are so close to the Eurelectric definitions that they can be considered equals. 8. Commercial Availability (CA) Introduction. IEEE 762 is introducing the concept of commercial availability. CA is not a universal equation at this point and may never become so. However, the general concept and approaches used in North America, South Africa and other parts of the word are introduced for user consideration and application. These approaches are mentioned in detail in Working Group 6 report in section 5.3 “Commercial Availability.” Conclusion A number of improvements to the data collection analysis procedures have been introduced in North America over the last three years. IEEE Standard 762, “Definitions for Reporting Electric Generating Unit Reliability, Availability and Productivity,” has moved away from focusing on only base-load generating units to all operation modes. Updated, more clearly explained definitions have been added. New equations and instructions for analysing groups or pools of units have been provided. The experiences and needs of the electric industry have been documented and made more uniform. . At the same time, the differences between the definitions and equations of Eurelectric and GADS have been reduced. With this reduction in differences comes a much better compatibility and comparability between the Europe, Africa and North America to exchange experiences and learn from each other. The WEC PGP Committee is planning to finalise within the next year or so the development of its global database of the power plant performance indicators which will allow users analyse data from around the world, whether from Europe, Africa, North America or Asia without concerns about the quality of information. All data will be compatible and collected in a very similar manner. The IEEE Standard 762 “Definitions for Reporting Electric Generating Unit Reliability, Availability and Productivity” is expected to be released before the end of 2004.
60
World Energy Council
Performance of Generating Plant 2004 - Section 1
APPENDIX 3: 1. WEC PGP COMMITTEE DATA COLLECTION During the last several years, it has become clear that a “new” solution for collecting, managing, analysing, and reporting performance data and statistics is needed. It has also become clear that the best opportunity for success will lie in the ability to identify a solution that is absent from commercial pressures and capable of evolving over time. Finally, the system needs to take into account a wide range of issues for both developing and developed countries. In the end, it appeared that the best champion for the cause of a new data collection/reporting system was the WEC PGP Committee. In order to take on this task, the PGP Committee enlisted the participation of experts and international organisations to help define the path forward. Workshops and working meetings were held in the US and Europe to address both the data collection requirements from a traditional point of view as well as that surrounding commercial availability. The PGP Committee is introducing a quicker and easier method for collecting data on electric power plants worldwide. Starting with the 2000-2002 data collection period, countries around the world can report data to the WEC PGP via the Internet. The PGP Committee’s goal is to create the initial worldwide data collection/reporting system to measure the performance of power plant components, equipment and technologies; this system will be designed to work in conjunction with other existing systems: NERC GADS, Eurelectric/VGB, and PRIS. It will also be designed to permit continued evolution of the system capabilities and its ability to interface with other existing data collection systems. 1.1
Governing Design Principles
Integrating commercial aspects of performance into measures, practices, and a data collection system is quite complicated principally because of: 1. A wide variation in market structures worldwide; 2. Differences in “how” profitability can be achieved within these markets, due to differences in forms of contracts, how power plants portfolios are managed, and degree of excess capacity present; 3. The need to be able to “benchmark” results without access to confidential price/cost data. As WG6 continues to explore possible solutions for the collection and application of data benchmarking processes, it will be incumbent on the working group to consider the realities of collecting additional data (and other uses for such data by the party in question), possible forms of measures, and a framework for applying such measures. The ability to address the overall performance and availability “equation” is the principal driver in the system design. To that end, it is recognised that both system capabilities and user acceptance must evolve if the PGP data collection and reporting process is to extend
61
World Energy Council
Performance of Generating Plant 2004 - Section 1
from the realm of “traditional” availability statistics into plant/unit/system/component framework, across generation segments (i.e., from fossil/nuclear to include renewables), and its analytical capabilities to consider commercial availability issues. The system will support levels of data gathering: 1. Detailed – High Level – Plant/Unit/System/Component failure data. The focus is to allow integration from the several existing databases which collect outage information on each component, including the time for repair, cause of failure, impact on the plant system and other important data. From the component outage record, data is compiled into major equipment and plant level performance numbers. Under this system, the data operator can analyse various levels of plant performance and collect data on similar causes of outages between generator technologies. Databases to be initially supported include: • • •
North American Electric Reliability Council’s Generating Availability Data System (NERC GADS); Euroelectic/VGB Data Entry system; IAEA PRIS.
The WEC System will define the methodology and data translation mechanism required to “migrate” data into the system. 2. Mid-level – Unit level Unit level data entry will be entered in the WEC system EITHER via a data entry programme or Excel spreadsheet template; it is also expected that an input file format will be fully documented to allow countries to develop data externally and import it into the programme. 3. Low-level – Aggregate submittal by a utility or country. Data entry for low level data will be available in three forms – data entry within the programme or Excel spreadsheet template, and manually (paper). The system will combine these groups with the results from the high- and mid-level reporting for a more general statistic. Data collected from all countries (either directly via the WEC data entry programme or via import from other databases) is to be housed in a central database at the WEC Secretariat. WEC would oversee all operations and procedures for editing, compiling and disseminating the data. The system must be flexible and user-friendly. Its capabilities should include: 1. Web-based data entry housed on the WEC website (http://www.worldenergy.org/). Ideally, this will be the principal location for both data entry and review of results. Hyperlinks to “help” pages documenting data definitions will be provided to assist in data entry activities.
62
World Energy Council
Performance of Generating Plant 2004 - Section 1
2. Initial database format will be “documented” to allow either ASCII or CSV files to be incorporated into the initial database. This will allow the data to be parsed and imported into the PGP database from either unit or grouped-unit Excel templates. 3. An Excel template for data entry for single units and grouped units; the groupings will be based on historic reporting “groups.” Basic data checking functions will be performed; menus/macros will be developed to guide the user through the data entry process, oversee the data verification process, and “automate” submission of the file to WEC. 4. Data Security, Protocol for Garnering Permission, etc. will be fully implemented to assure the security of data provided; the intention is to develop a database structure, security, data import, programme distribution, and reporting functions consistent with requirements imposed by participants. 5. WEC publication of statistics for this and future reporting periods will include the presentation of traditional indices (i.e., UCLF, UCF, EAF) and will grow over time. Reports required for WEC publications (hard copy) would include: a. UCLF b. UCF c. EAF d. EAUF e. LF f. Unplanned Automatic Grid Separations g. Total Operating Hours h. Thermal Performance. 6. The system will initially possess the capability to filter the data according to the following: a. Country b. Geographic Regions c. Technology Class d. MW Rating e. Fuel Source f. Duty Cycle g. Age h. Any of the above. During system design and planning for the initial implementation, it was important for the PGP Committee to consider the “longer-term” needs of the programme and how it relates to the Committee’s overall mission. Many of the issues/dilemmas cannot be immediately addressed. Only time and continued investigation will allow the Committee to determine how to best develop both the system and its ability to address the commercial implication of power plant performance.
63
World Energy Council
1.2
Performance of Generating Plant 2004 - Section 1
The First Phase
The first phase of the new database development was completed in May 2004. Here are a few highlights of the new website:
This opening screen welcomes the reporter to the website. By clicking on the “login”, the reporter moves to the screen for entering the user name and password as shown below. The WEC guarantees the confidentiality of information submitted and considers all data confidential – whether the information is group statistics or single unit information. The WEC Secretariat will issue the user name and password for data entry. If you have not received your password, please contact Ms. Elena Nekhaev, at
[email protected].
64
World Energy Council
Performance of Generating Plant 2004 - Section 1
After logging into the website, the screen below will come up. This screen provides general instructions for use of the programme. The website is designed to be “user friendly”.
At the top of the screen, in the right-hand corner, there is a “?” box. This box is linked to a help page with definitions or equations used to collect data as defined by the WEC-PGP advisory group. An example of the help screen is shown below.
The general information screen allows the reporter to choose between peer group or single unit data. “Peer Group” is a set of more than one unit in a specific category of units based on unit type (technology: fossil steam, gas turbine, combined cycle, etc), fuel (coal,
65
World Energy Council
Performance of Generating Plant 2004 - Section 1
gas, or oil), and MW size (100-199 MW). “Single unit” allows the user to provide individual unit data for each single unit rather than a group of units. The programme is set up to ensure that all data is entered using the peer group method or the single unit method for a category of units. It is not possible to switch from one group to the other within a group. For example, if peer groups for fossil units are reported, all fossil units have to be reported as peer group units. It is strongly recommended to report the single unit data. 1.3
“Peer” Group Data Entry
Below is an example of the Peer Group choice screen:
Above the “Description” title on the screen, there are buttons to select more Peer Groups than are shown on the screen.
66
World Energy Council
Performance of Generating Plant 2004 - Section 1
After entering the year and the information in the “Group Totals,” an “Entry method” button should be activated. There are two choices for data entry: Energy or Factor, only one or the other can be chosen. Energy allows data entry for the left side of the screen. Factor allows entry for the right side of the screen. If Energy is used on the left side, the right side statistics will be automatically calculated and will appear. If you use Peer Groups, the WEC-PGP recommends you use the Energy (left side). 1.4
Single Unit Data Entry
To enter single, individual unit data, click on the “single unit” link on the General Information screen. The following screen will appear:
The screen is similar to the Peer Group screen, but has more data entry fields for each single unit. This information is necessary to allow the units to be grouped and analysed in more detail then the Peer Groups can be.
Some general design information about each unit, the year of initial operation, fuel, Reference Capacity and type of unit is requested. The design data can be expanded to any degree as requested. The expansion can include information on primary plant equipment such as manufacturers, number of pumps, furnace draft, etc.
67
World Energy Council
Performance of Generating Plant 2004 - Section 1
Once the design is entered, the report leads to the next screen:
As in the Peer Group data entry, the top of the screen is completed and then there is one of two choices: Energy or Factor. Only one can be selected, not both. If Energy on the left side is used, the right side statistics will automatically appear. The Energy and Factor sides can also be expanded in future versions of the programme. At present, the database is available to WEC Members only.
68
Performance of Generating Plant Section 2 THERMAL GENERATING PLANT UNAVAILABILITY FACTORS AND AVAILABILITY STATISTICS
MIKE CURLEY Manager - GADS North American Electric Reliability Council (US)
Work Group Membership: W. Guangyao (China) D. Glorian (France) K. Yassin (Egypt) T. Tersztgansky (Hungary) T. Uchida (Japan) V. Micali (South Africa) G. S. Stallard (US)
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE OF CONTENTS Page INTRODUCTION
3
PGP DATA BASE: THE SCOPE, DEFINITIONS AND TERMINOLOGY……….. 3 1.
The Scope
2.
Definitions and Terminology
BRIEF DESCRIPTION OF THE FIVE PERFORMANCE MONITORED BY EURELECTRIC…………………………………………………...5 1.
Unit Capability Factor (UCF)
2.
Unplanned Capability Loss Factor (UCLF)
3.
Unplanned automatic grid separations per 7 000 operating hours
4.
Successful start-up rate
5.
Industrial Safety Accident Rate
BRIEF DESCRIPTION OF THE INSTALLATION FAMILIES MONITORED…………………………………………………………………………… 5 FIRST RESULTS……………………………………………………………………….. 7 1.
Canada
2.
United States
3.
Europe
4.
Japan
WHAT THE FUTURE HOLDS IN STORE…………………………………………. 10 CONCLUSIONS………………………………………………………………………...10 DATA FROM AROUND THE WORLD……………………………………………...11 Table 1-1:
All Fossil Fuels, Steam Units ≥100 MW
Table 1-2:
Solid Fossil Fuels, Steam Units ≥100 MW
Table 1-3:
Liquid & Gaseous Fossil Fuels, Steam Units ≥100 MW
Table 1-4:
All Fossil Fuels, Steam Units, North America ≥100 MW
Table 1-5:
Solid Fossil Fuels, Steam Units, North America ≥100 MW
World Energy Council
Performance of Generating Plant 2004 – Section 2
Table 1-6:
Liquid & Gaseous Fossil Fuels, Steam Units, North America ≥100 MW
Table 1-7:
All Fossil Fuels, Steam Units, Europe ≥100 MW
Table 1-8:
Solid Fossil Fuels, Steam Units, Europe ≥100 MW
Table 1-9:
Liquid & Gaseous Fossil Fuels, Steam Units, Europe ≥100 MW
Table 1-10:
All Fossil Fuels, Steam Units, Other Countries ≥100 MW
Table 1-11:
Solid Fossil Fuels, Steam Units, Other Countries; ≥100 MW
Table 1-12:
Liquid & Gaseous Fossil Fuels, Steam Units, Other Countries ≥100 MW
Table 1-13:
All Fossil Fuels, Steam Units, Japan ≥100 MW
Table 1-14:
Solid Fossil Fuels, Steam Units, Japan≥ 100 MW
Table 1-15:
Liquid & Gaseous Fossil Fuels, Steam Units, Japan ≥100 MW
Table 1-16:
Combustion Turbines ≥30 MW
Table 1-17:
Hydro Units ≥50 MW
Table 1-18:
Recent Combined Cycle Units ≥100 MW
Table 1-19:
Historical Combined Cycle Units ≥100 MW
Graph 1-1: Unit Capability Factor, Europe, 1990-2003 Graph 1-2: Unplanned Capability Loss Factor, Europe, 1990-2003
2
World Energy Council
Performance of Generating Plant 2004 – Section 2
INTRODUCTION One of the most important criteria for the evaluation of power plant performance is its availability record. The causes of unavailability are thoroughly analysed to identify the areas for performance improvement. The WEC Committee on the Performance of Generating Plant (PGP), formerly a Joint UNIPEDE/WEC Committee, for many years has been collecting statistical data on power plant availability using WEC’s global network of Member Committees. There is no simple way to measure overall plant performance, nor is there a single indicator which could be used for this purpose. Operating conditions vary widely between the countries and regions, and in addition to high reliability, power plants must at the same time achieve a number of other objectives: economic, environmental, societal, etc. These objectives are different for different power plants, and each plant has its own particular aspects to take into account. The increasing competition in the electricity sector has had significant implications for plant operation, and it requires thinking in strategic and economic rather than purely technical terms. This is not always easy for the global community of power plant operators, which is heavily dominated by engineers with a “technical mindset”. The need for the efficient allocation and use of available resources; effective scheduling of plant activities, such as outages and on-line maintenance and greater use of analytical tools to conduct cost/benefit evaluation of proposed activities is changing the industry mindset. This new need, reinforced by the dynamics of the ongoing change, is creating an atmosphere of uncertainty in the market. The uncertainty of meeting the demand for electric power and the shareholders’ profit expectations places additional pressure on power plant operators. The challenge is both to improve the performance of the existing generating plant stock and to build up enough – but not too much - new generation and transmission capacity to meet growth in demand. It is expected that over 700 GW of new power generation capacity will be added worldwide in the next few years. PGP DATA BASE: THE SCOPE, DEFINITIONS AND TERMINOLOGY 1.
The Scope
For more than ten years, power plant availability statistics collected by the WEC PGP Committee were processed and stored using a special software package which was not directly compatible with the majority of most commonly used database applications. During the process of transferring data onto the new PGP database, a fairly large amount of statistical information has been left behind in old format files. Additional information may be interpreted, bearing in mind the wide variety of equipment, and the considerable differences in operation and utilisation, in addition to economic considerations. By August 2004, there was data for over 5,000 unit/years in the database. Not all countries, which participated in previous surveys, have yet been able to enter their data into the new database. As the content of the database grows further, it is expected to become a valid reference source for an availability factor expectation, being particularly useful for countries in the early stages of employing gas turbine plants and combined cycle plants as part of their power systems.
3
World Energy Council
Performance of Generating Plant 2004 – Section 2
WEC data surveys focus on base-load units, since availability and unavailability factors are not suitable for peaking plants. For example, a fossil-fuel plant operating at peak load for a limited number of hours during the year, and spending the rest of time in reserve, excluding planned annual maintenance shutdowns, would show an availability level in the order of 100%, which would not reflect the real situation. Therefore, it was agreed, whenever possible, to exclude this type of installation from the statistics, along with the units whose utilisation factor is less than 40%. 2.
Definitions and Terminology
The calculation methodology and rules introduced for the new database broadly reflect the existing standards and their use should be encouraged within the framework of the WEC survey. The document “Availability and Unavailability Factors Of Thermal Power Plants - Definitions and Methods of Calculation” compiled by the former Joint UNIPEDE/WEC Committee and published in 1991, presents the basic reference terminology and definitions. Some of the original definitions and terminology have been modified. Pursuing the programmes undertaken by WANO (World Association of Nuclear Operators) and IAEA (International Atomic Energy Agency) for nuclear power plant performance monitoring, UNIPEDE has developed a programme for collecting, processing and publishing statistics relating to the performance of conventional fossilfuelled power plants (UNIPEDE, International Union of Producers and Distributors of Electrical Energy, merged with EURELECTRIC in 1999). The following five performance indicators have thus been defined, for international application, for the different areas in which operators must ensure a high degree of vigilance in order to achieve a satisfactory quality of service: •
Unit Capability Factor (UCF);
•
Unplanned Capability Loss Factor (UCLF);
•
Unplanned Automatic Grid Separations per 7000 hours of operation (UAGS 7);
•
Successful Start-up Rate (SSR);
•
Industrial Safety Accident Rate (ISAR).
Using precise terminology and definitions, the programme has been launched in 1994 by UNIPEDE, and opened to all its Members (mostly from Western European countries). This programme was developed to be consistent with similar work carried out within the former Joint UNIPEDE/WEC Committee on the Performance of Thermal Generating Plant (fossil-fuel and nuclear). Integration within the EURELECTRIC system makes it possible to contribute to the activities of the present WEC Committee on the Performance of Generating Plant without requiring collection of additional data. The aim of the programme is to create a high-quality management tool. The included indicators are intended principally for use by operators to monitor their own performance and progress, to set their own challenging goals for improvement, and to gain an additional perspective on performance relative to that of other plants. Thus, the international exchanges will help foster a commitment to emulate the best practices, thereby maintaining the satisfactory level of performance observed.
4
World Energy Council
Performance of Generating Plant 2004 – Section 2
BRIEF DESCRIPTION OF THE FIVE PERFORMANCE MONITORED BY EURELECTRIC 1.
Unit Capability Factor (UCF)
Unit capability factor is the percentage of maximum energy generation that a plant is capable of supplying to the electrical grid, limited only by factors within control of plant management. A high unit capability factor indicates effective plant programmes and practices to minimise unplanned energy losses and to optimise planned outages, maximising available electrical generation. NOTE: Energy Availability Factor (WEC indicator) is defined on the same basis; but EAF is reduced by losses that are not under the control of plant management. The Eurelectric "Therperf” programme is able to produce EAF results, in addition to UCF. 2.
Unplanned Capability Loss Factor (UCLF)
Unplanned capability loss factor is the percentage of maximum energy generation that a plant is not capable of supplying to the electrical grid because of unplanned energy losses (such as unplanned shutdowns, outage extensions or load reductions due to unavailability). Energy losses are considered unplanned if they are not scheduled at least four weeks in advance. A low value for this indicator indicates that important plant equipment is reliably operated and well maintained. 3.
Unplanned automatic grid separations per 7 000 operating hours
This indicator expresses how often a generator is separated from the external grid, in both an unplanned and automatic (manual actions are excluded) manner; it is given as a rate per 7 000 operating hours, thereby taking into account the wide variety of operating regimes. 4.
Successful start-up rate
One of the ways to measure quality of service rendered to an electrical grid - among other indicators - is the promptness with which a unit is connected to the external grid, in line with the grid operator's request. This indicator expresses the level of success in achieving a presence on the grid, at the moment requested, for all start-ups requested on the previous day (immediate start-up for peaking gas turbines). 5.
Industrial Safety Accident Rate
Progress in improving industrial safety performance is monitored by the number of accidents that result in day away from work, days of restricted work or fatalities, per 1,000,000 man- hours worked. BRIEF DESCRIPTION OF THE INSTALLATION FAMILIES MONITORED The five performance indicators are monitored on a unit-by-unit basis, starting from the first full year of commercial service. Data is submitted anonymously using a unit "code", which is known only by the operator who supplies the data. To ensure complete confidentiality (no data can be used for commercial purposes), certain procedures have been defined for the exchange of this information. 5
World Energy Council
Performance of Generating Plant 2004 – Section 2
Three categories of conventional thermal installations are monitored: A - Steam turbines B - Combined cycle, cogeneration C - Combustion turbines. Four types of fuel are monitored: 1 - Coal (excluding lignite and others) 2 - Lignite and others 3 - Liquid fuels 4 - Gaseous fuels. The power rating categories are those recommended by the former Joint UNIPEDE/WEC Committee. Availability and unavailability statistics for steam turbine units are grouped by four basic fuel types as presented in Figure 1. Categories of capacity used in the analysis of steam turbines are shown in Figure 2.
Fossil fuels
Solid fuels
Coal (excl. lignite & others)
Liquid & gaseous fuels
Lignite & others
Liquid fuels
Fig. 1: Steam Turbine - Fuel Types
6
Gaseous fuels
World Energy Council
Performance of Generating Plant 2004 – Section 2
Steam Turbine >100 MW
100 to 199 MW
200 to 399 MW
400 to 599 MW
> 600 MW
200to 299 MW
300 to 399 MW
600 to 799 MW
800 to 999 MW
> 1000 MW
Fig. 2: Steam Turbine - Class of Capacity FIRST RESULTS Whereas in the past, the PGP surveys were triennial, the new Internet-based system will allow the user to enter data and run queries on an ongoing basis, and the WEC Members will be encouraged to update their data annually. Since the 2004 data sample is fairly different from the 2001 one, there are just a few general tables, and a more detailed analysis of the three-year developments and trends is presented for major countries only. 1.
Canada
Canada has been reporting its power plant performance statistics to the WEC-PGP Committee through the NERC GADS programme. There are three utilities in Canada that are part of the NERC GADS system at this present time: Ontario Power Generation, B.C. Hydro, and New Brunswick Electric Power Commission. Canada provided data on 23 base-loaded fossil units, representing 8,890 MW. Thirteen of the 23 units reported in 2002, were in the 400-599 MW range. The average EAF for all Canadian fossil units over the period 2000 to 2002 is 75.7% but with a CUF of 90.8%. There were a surprisingly high number of external problems reported. The average PUF was 11.8%, which is 3.4% higher than the world average (8.4%). The UUF was 13.1%, which is 5.3% higher than the world average (7.8%). The 400-599 MW units reported an EAF of 78.46% with very few external problems (UCF = 78.49%). However, this group of units has had other problems. Almost twice as many outages for this group were unplanned rather than planned events (14.0% UUF and 7.5% PUF). The Canadian Electricity Association (CEA) and NERC GADS are establishing cooperation where the data collected by CEA will be converted into the NERC GADS format. This will facilitate reporting of unit-specific data to the WEC-PGP database by NERC GADS, as it is being done for the units in the United States. Using the unit-by-unit data, it is possible to conduct a more detailed analysis of the fossil units based on age, MW size, operating mode, and many other aspects.
7
World Energy Council
2.
Performance of Generating Plant 2004 – Section 2
United States
The NERC GADS database has shown an increase in participation in the last several years. The reason for a more active participation in GADS is a ruling by the NERC Planning Committee (one of three NERC standing committees) in November 2000. At that time, the NERC PC modified the GADS Data Release Guidelines to state that if a power generator did not report data to GADS, then they would not have access to the GADS software product pc-GAR. Pc-GAR was developed for use in analysing all North American power plants by design, statistics, or performance parameters. Using the software, users can benchmark units, determine peers, examine manufacturer performance, determine the expected performance of units in future years, and many more things. As a result of the Guideline changes, GADS received event and performance records for 4,102 units (648,300 MW) in 2003. This is the highest number of units reported to GADS in its 22 year history. During the 2000-2002 survey period, the United States provided data for 795 base-loaded fossil, 45 combined cycle, and 384 hydro units. The average capacity of these 795 fossil units is 300,463 MW, and 687 units (or 86% of the total) were coal-fired. This 86% represents 246,547 MW of installed capacity. The combined cycle units account for 10,553 MW and the hydro units for 48,130 MW of installed capacity, respectively.
Summary of US units 2000-2002 Average Fossil, All Fossil Fuels Fossil, Solid Fossil Fuels Combined Cycle Hydro All US Base-loaded Units
EAF 84.0 84.4 87.0 81.0 83.8
PUF 8.3 8.1 8.2 16.7 7.0
UUF 7.7 7.5 4.8 2.3 9.2
The average EAF for US plants is the same as in the previous survey (about 83%) and the sums of unplanned and planned outages are the same, too. However, this table demonstrates that unit outages for US fossil units are just about evenly distributed between unplanned and planned events, with planned being slightly higher. This is a decrease in planning of outages from the last survey, which indicated that the US units had twice as many planned outage events as unplanned, similar to the statistics for the combined cycle PUF/UUF ratio above. This increase in the number of unplanned outages for fossil units can be the results of several reasons including: 1) units not allowed to go to outages due to demand for power, and 2) older units were operated for longer periods, resulting in more unplanned events. More and more US companies are joining Independent System Operators (ISO) and become subjected to more strict demand for power. ISO organisations require the cheapest (cost/MW) units to operate first before dispatching other, more costly units. Sometimes there is no permission to take units off line for repairs until it is more convenient to do so. That may be a week or more after the problems were discovered and reported. If a unit needs repairs, prolong operation results in more damage to the equipment and longer repair times. Surprisingly perhaps, the solid fuel (coal) units are slightly better than the liquid and gas units. It is assumed that the coal units are repaired quicker because they are the main source of electric power in the USA. See tables 1-4, 1-5 and 1-6. 8
World Energy Council
Performance of Generating Plant 2004 – Section 2
The US utilities and other companies have constructed a large number of combined cycle units. About 95% of all new construction in the US is done by IPP’s and the majority of these plants are combined cycle. Most of the units are natural gas fired because they considered being environmental friendly. However, in some cases, the demand for natural gas has put a few combined cycle units outside the “low cost” range by the ISO. The resulting action is to operate coal units more (economics). 3.
Europe
Due to the introduction of a new reporting format, the data for Europe is not directly comparable with the previous surveys. Nevertheless, the average EAF calculated for European plants demonstrates an improving trend.
Summary of European Units (All Fossil Fuels, Fossil Steam Units, 100 MW or Larger) EAF 2000-2002
4.
2000
2001
2002
Average
100 to 199 MW
89.45
85.77
83.43
86.22
200 to 299 MW
86.37
90.68
80.40
85.82
300 to 399 MW
87.66
91.18
86.03
88.29
400 to 599 MW
88.93
85.00
76.61
83.51
600 to 799 MW
86.03
94.5
62.10
80.88
Japan
Japan provided an extensive set of data for the 2000-2002 survey (tables 1-13, 1-14 and 1-16) completely following the WEC format, i.e. the same kind of data set provided for the 1997-1999 survey. The available information shows clearly the specific characteristics of operational experience of Japan. The results are country based and presented in separate tables to keep consistency with previous surveys. The set of Japanese base-load fossil-fuel plants represented 148 units and 84,700 MW in the period 2000-2002. The average EAF of these plants is 83.6% (82.9% for the previous survey in 1997-1999), while PUF equals 15.1% (16.2 % previously) and UUF, 1.3% (0.9% previously). As already stated, although the average result for energy availability factor in Japan is in the same order of magnitude as the world sample (83.8 %), the balance between PUF and UUF is again completely different. The Japanese have shown a wonderful example of planning outages and making repairs correctly so that there are few unplanned outages. The fact that the PUF factor for Japan moved from 18.7 % in 1994 to 14.7% in 2002 shows very clearly that a very cautious, but positive policy, has been implemented to achieve better operating optimization in these units. As base-load units, they must operate to the maximum (optimum) use of their capabilities. Therefore, it seems that the extension of such a careful policy, facing a more demanding market, could reveal additional availability improvements.
9
World Energy Council
Performance of Generating Plant 2004 – Section 2
WHAT THE FUTURE HOLDS IN STORE The benefits of the international cross-comparison system henceforth depend - in addition to the current practices described in this report - on the commitment of power plant operators to enhancing them. The underlying goal is to foster international support and participation. Nevertheless, additional factors have to be taken into consideration, as there is a stronger need to reach the global picture of power plant performance, facing the grid and the needs of the users. These factors refer to the different kinds of responsibilities for each type of energy losses: external versus internal (for example, environmental constraints as opposed to equipment reliability and human performance), and technical versus commercial. In addition, the introduction of the concept of commercial availability could help to better address the technical performance of generating plants in the competitive electricity market. The WEC PGP Committee will continue producing statistics that will offer value to all electricity producers worldwide. But this is not all. The WEC PGP Committee has started work to widen the analysis aspect of the WEC PGP database. Within the coming years, the database will be expanded to include data selections based on design and annual performance characteristics for use in benchmarking, reliability determinations, evaluating new and old unit designs as well as other applications for increasing the productivity and reliability of plant equipment. This will follow the example of the NERC GADS software product, pc-GAR. pc-GAR allows the user to compare the performance and design of peer units based on the user’s own selection criteria, not on predetermined criteria by others. The NERC software contains hundreds of design characteristics and 22 years of annual performance records on more than 5,500 generating units. pc-GAR is now used by more than 12 countries, but is limited to comparing US generating units only. It needs to be expanded to include generating units in the world community. This will be accomplished by the WEC PGP Committee work. CONCLUSIONS Key factors influencing plant performance should be identified and analysed to allow a cost/benefit analysis of any activity/programme before its implementation. To analyse plant availability performance, the energy losses/outages should be scrutinised to identify the causes of unplanned or forced energy losses and to reduce the planned energy losses. Reducing planned outages increases the number of operating hours, decreases the planned energy losses and therefore, increases the energy availability factor. Reducing unplanned outages leads to safe and reliable operation, and also reduces energy losses and increases energy availability factor. Access to worldwide statistics on the performance of generating plant will help power plant operators deal with the availability records of their plants in the context of global experience. New software for collecting and new, powerful software for analysing the results are within the scope of the PGP Committee, to bring the world electricity producers closer together in a cooperative manner. The result will be a wonderful exchange of information to better the quality of life for the world community.
10
World Energy Council
Performance of Generating Plant 2004 – Section 2
DATA FROM AROUND THE WORLD TABLE 1-1:
All Fossil Fuels, Fossil Steam Units, ≥100 MW
ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2000 283 40330 10.5 84.1 6.0 132 31668 8.9 82.5 9.4 156 54601 11.5 82.4 4.8 230 115840 10.9 82.2 7.0 160 107355 15.5 77.0 9.3 38 32017 7.6 88.5 3.9 25 27828 15.2 81.1 3.8
2001 287 40640 9.7 80.7 8.3 119 28206 9.9 83.2 5.8 152 53068 11.7 84.4 4.2 196 98502 11.2 84.4 4.4 157 105327 12.2 79.7 6.6 35 29545 7.6 86.4 6.0 27 29928 14.9 82.0 3.1
2002 273 39117 10.3 75.8 11.7 118 28368 7.7 80.7 9.6 145 50536 11.7 78.6 6.8 201 101058 9.7 84.3 6.0 153 102466 10.4 85.1 4.5 38 32150 11.4 82.3 6.4 28 30928 13.0 83.2 3.8
List of countries: Brazil, Canada, Egypt, Japan, United States
11
Average 281 40029 10.2 80.2 8.7 123 29414 8.8 82.1 8.2 151 52735 11.6 81.8 5.3 209 105133 10.6 83.6 5.8 157 105049 12.7 80.6 6.8 37 31237 8.9 85.7 5.4 27 29561 14.4 82.1 3.5
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-2 – Solid Fossil Fuels, Fossil Steam Units, ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2000 239 34068 13.4 78.8 7.8 107 25795 8.8 77.7 13.5 69 24008 11.5 84.1 4.4 166 84341 8.5 83.4 8.1 106 72364 10.6 84.3 5.2 29 24508 10.4 83.7 5.9 16 18828 13.6 80.6 5.8
2001 234 33512 7.0 78.1 14.3 98 23476 8.1 86.3 5.6 68 23622 9.3 88.1 2.9 148 75213 9.5 85.4 5.1 106 72574 8.3 83.7 7.9 24 20255 9.2 85.0 5.7 18 20928 8.2 86.4 5.4
List of countries: Brazil, Canada, Japan, United States
12
2002 234 33752 11.1 76.9 13.5 103 24793 7.1 76.1 16.9 70 24259 12.8 70.2 8.3 161 81697 8.1 83.2 8.6 106 72753 9.3 84.2 6.5 29 24543 9.6 83.9 6.6 19 21928 10.1 83.9 6.1
Average 236 33777 10.5 77.9 11.9 103 24688 8.0 80.0 12.0 69 23963 11.2 80.8 5.2 158 80417 8.7 84.0 7.3 106 72564 9.4 84.1 6.5 27 23102 9.8 84.2 6.1 18 20561 10.6 83.6 5.7
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-3 – Liquid & Gaseous Fossil Fuels, Fossil Steam Units, ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%) 2000 100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2001
44 6262 7.6 87.7 4.8 25 5873 8.9 87.3 5.2 87 30593 11.4 81.6 4.9 64 31499 13.2 80.9 5.9 54 34991 19.2 72.6 11.8 9 7509 6.2 91.0 2.8 9 9000 16.8 81.6 1.7
53 7128 11.0 82.0 5.4 21 4730 11.7 81.7 5.9 84 29446 12.9 82.3 5.0 48 23289 12.9 83.4 3.8 51 32753 15.2 77.3 5.9 11 9290 7.1 86.8 6.1 9 9000 21.6 77.6 0.8
List of countries: Brazil, Canada, Egypt, Japan. United States
13
2002 39 5365 9.6 75.0 10.3 15 3575 8.6 84.2 4.1 75 26277 11.2 83.4 5.9 40 19361 11.2 85.4 3.4 47 29713 11.1 85.6 3.2 9 7607 12.0 81.7 6.3 9 9000 16.0 82.6 1.5
Average 45 6252 9.4 81.6 6.8 20 4726 9.7 84.4 5.1 82 28772 11.8 82.4 5.3 51 24716 12.4 83.2 4.4 51 32486 15.1 78.5 7.0 10 8135 8.4 86.5 5.1 9 9000 18.1 80.6 1.3
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-4 – All Fossil Fuels, Fossil Steam Units, North America ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2000 270 38500 7.4 86.6 5.9 130 31248 7.9 85.4 6.7 94 33101 9.1 82.9 8.0 199 100856 7.8 82.9 9.3 108 74801 14.0 76.4 9.7 34 28601 7.2 87.4 5.4 11 13828 8.7 81.7 9.6
List of countries: Canada & United States
14
2001 265 37625 9.7 83.4 6.9 115 27346 9.9 84.5 5.7 89 31205 10.9 82.7 6.4 165 83518 8.7 83.4 7.9 103 71373 11.8 77.3 10.9 31 26129 5.2 86.8 8.1 11 13828 8.2 83.2 8.6
2002 251 36102 8.6 84.4 7.1 114 27508 6.3 88.1 5.6 82 28673 9.9 83.1 7.1 170 86074 8.2 83.9 7.9 99 68512 6.9 85.5 7.6 34 28734 11.1 80.6 8.3 11 13828 5.1 84.9 10.1
Average 262 37409 8.6 84.8 6.6 120 28701 8.0 86.0 6.0 88 30993 10.0 82.9 7.1 178 90149 8.2 83.4 8.4 103 71562 10.9 79.8 9.4 33 27821 7.8 84.9 7.3 11 13828 7.3 83.3 9.4
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-5 – Solid Fossil Fuels, Fossil Steam Units, North America, ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2000 239 34068 7.1 85.8 7.0 107 25795 6.9 85.6 7.5 67 23308 8.4 85.4 6.2 160 81341 5.5 85.9 8.6 96 65664 8.8 83.9 7.5 29 24508 10.4 83.7 5.9 11 13828 8.7 81.7 9.6
List of countries: Canada & United States
15
2001 230 32930 7.0 85.9 7.0 98 23476 8.1 86.3 5.6 65 22559 9.6 86.4 4.0 142 72213 7.5 83.5 9.0 94 64474 7.3 82.3 10.4 24 20255 9.2 85.0 5.7 11 13828 8.2 83.2 8.6
2002 230 33170 6.6 86.0 7.4 103 24793 4.1 88.4 7.5 67 23196 8.7 86.8 4.5 155 78697 6.3 84.1 9.6 94 64653 7.2 83.5 9.4 29 24543 9.6 83.9 6.6 11 13828 5.1 84.9 10.1
Average 233 33389 6.9 85.9 7.1 103 24688 6.4 86.8 6.9 66 23021 8.9 86.2 4.9 152 77417 6.4 84.5 9.1 95 64930 7.8 83.2 9.1 27 23102 9.8 84.2 6.1 11 13828 7.3 83.3 9.4
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-6 – Liquid & Gaseous Fossil Fuels, Steam Units, North America ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2000 31 4432 7.6 87.0 5.4 23 5453 8.9 85.2 5.9 27 9793 10.6 80.2 9.2 39 19515 10.1 80.0 9.9 12 9137 19.1 69.0 11.9 5 4093 3.9 91.1 5.0
2001 35 4695 11.0 82.1 6.9 17 3870 11.7 82.6 5.7 24 8646 12.9 78.3 8.8 23 11305 9.8 83.3 6.9 9 6899 16.3 72.3 11.4 7 5874 3.1 87.7 9.2
List of countries: Canada & United States
16
2002 21 2932 9.6 83.5 6.9 11 2715 8.6 87.7 3.7 15 5477 10.5 79.1 10.4 15 7377 10.1 83.7 6.2 5 3859 6.6 87.6 5.9 5 4191 11.9 78.9 9.2
Average 29 4020 9.4 84.2 6.4 17 4013 9.7 85.2 5.1 22 7972 11.3 79.2 9.5 26 12732 10.0 82.3 7.7 9 6632 14.0 76.3 9.7 6 4719 6.3 85.9 7.8
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-7 – All Fossil Fuels, Fossil Steam Units, Western Europe ≥100 MW ANNUAL AVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
2000
2001
2002
Average
100 to 199 MW
Count: EAF:
171 89.5
90 85.8
75 83.4
112 86.2
200 to 299 MW
Count: EAF:
116 86.4
70 90.7
62 80.4
83 85.8
300 to 399 MW
Count: EAF:
167 87.7
91 91.2
81 86.0
113 88.3
400 to 599 MW
Count: EAF:
53 88.9
47 85.0
44 76.6
48 83.5
600 to 799 MW
Count: EAF:
36 86.0
24 94.5
22 62.1
27 80.9
800 to 999 MW
Count: EAF:
2 NR
4 NR
5 NR
4 NR
List of countries: Belgium, France, Germany, Italy, Portugal, Spain
17
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-8 – Solid Fossil Fuels, Fossil Steam Units, Western Europe ≥100 MW ANNUAL AVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
2000
2001
2002
Average
100 to 199 MW
Count: EAF:
103 88.4
65 80.1
65 85.1
78 84.5
200 to 299 MW
Count: EAF:
76 83.6
45 89.5
47 84.9
56 86.0
300 to 399 MW
Count: EAF:
90 93.6
82 92.5
72 95.2
81 93.8
400 to 599 MW
Count: EAF:
38 92.3
30 89.4
29 84.8
32 88.8
600 to 799 MW
Count: EAF:
16 NR
20 NR
20 NR
19 NR
800 to 999 MW
Count: EAF:
2 NR
4 NR
5 NR
4 NR
List of countries: Belgium, France, Germany, Italy, Portugal, Spain
18
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-9 – Liquid & Gaseous Fossil Fuels, Steam Units, Western Europe ≥100 MW ANNUAL AVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
2000
2001
2002
Average
100 to 199 MW
Count: EAF:
68 90.2
25 88.6
10 81.8
34 86.9
200 to 299 MW
Count: EAF:
40 88.7
25 91.6
15 77.4
27 85.9
300 to 399 MW
Count: EAF:
77 83.7
9 88.6
9 67.8
32 80.0
400 to 599 MW
Count: EAF:
15 84.5
17 80.6
15 68.4
15 77.8
600 to 799 MW
Count: EAF:
20 86.0
4 94.5
2 62.1
9 80.9
800 to 999 MW
Count: EAF:
NR NR
NR NR
NR NR
NR NR
List of countries: Belgium, France, Germany, Italy, Portugal, Spain
19
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-10 – All Fossil Fuels, Fossil Steam Units, .≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%) 2000 100 to 199 MW
200 to 299 MW
300 to 399 MW
600 to 799 MW
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2001
13 1830 7.3 89.0 3.7 2 420 4.5 91.6 3.9 16 5000 15.2 82.4 2.4 2 1254 1.3 65.0 33.7
22 3015 12.2 78.0 9.8 4 860 13.2 80.7 6.1 17 5363 11.2 85.9 2.9 2 1254 23.6 72.2 4.2
2002 22 3015 15.3 66.3 18.4 4 860 15.0 80.7 4.4 17 5363 19.9 76.2 3.9 2 1254 11.4 87.8 0.8
Average 19 2620 11.6 77.8 10.6 3 713 10.9 84.3 4.8 17 5242 15.4 81.5 3.1 2 1254 12.1 75.0 12.9
List of countries: Brazil, Egypt
TABLE 1-11 – Solid Fossil Fuels, Fossil Steam Units, Other Countries ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%) 2000 100 to 199 MW
300 to 399 MW
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2001
0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0
List of countries: Brazil, Egypt
20
4 582 8.1 70.2 21.6 1 363.0 8.2 90.1 1.7
2002 4 582 6.5 65.9 27.7 1 363.0 47.6 48.6 3.8
Average 3 388 4.9 45.4 16.4 1 242.0 18.6 46.2 1.8
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-12 – Liquid & Gaseous Fossil Fuels, Fossil Steam Units, Other Countries ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%)
100 to 199 MW
200 to 299 MW
300 to 399 MW
600 to 799 MW
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2000 13 1830 7.3 89.0 3.7 2 420 4.6 91.6 3.9 16 5000 15.2 82.4 2.4 2 1254 1.4 65.0 33.7
2001 18 2433 14.3 81.8 3.9 4 860 13.2 80.7 6.1 16 5000 12.7 83.8 3.5 2 1254 23.7 72.2 4.2
List of countries: Brazil, Egypt
21
2002 18 2433 19.8 66.5 13.7 4 860 15.0 80.7 4.4 16 5000 6.1 90.0 3.9 2 1254 11.5 87.8 0.8
Average 16 2232 13.8 79.1 7.1 3 713 10.9 84.3 4.8 16 5000.0 11.3 85.4 3.3 2 1254 12.2 75.0 12.9
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-13 – All Fossil Fuels, Fossil Steam Units, Japan ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%) 2000 100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2001
0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 46 16500 17.1 82 0.9 31 14984 14.5 84.0 1.5 50 31300 18.1 81.1 0.8 4 3416 8.5 90.8 0.7 14 14000 16.4 82.1 1.5
22
0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 46 16500 13.4 85.8 0.8 31 14984 16.5 83.0 0.5 52 32700 13.1 85.3 1.6 4 3416 15.0 85.0 0.0 16 16100 14.8 83.5 1.7
2002 0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 46 16500 16.3 81.6 2.1 31 14984 11.0 88.3 0.7 52 32700 15.0 83.5 1.5 4 3416 12.3 87.3 0.4 17 17100 16.4 81.8 1.8
Average 0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 46.0 16500.0 15.6 83.1 1.3 31 14984 14.0 85.1 0.9 51 32233 15.4 83.3 1.3 4 3416 11.9 87.7 0.4 16 15733 15.9 82.5 1.7
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-14 – Solid Fossil Fuels, Fossil Steam Units, Japan ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%) 2000 100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2001
0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 2 700 17.7 81.4 0.9 6 3000 14.6 82.4 3.0 10 6700 14.2 85.2 0.6 0 0 0.0 0.0 0.0 5 5000 18.5 79.5 2.0
23
0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 2 700 8.7 89.3 2.0 6 3000 16.6 83.0 0.4 12 8100 10.4 86.6 3.0 0 0 0.0 0.0 0.0 7 7100 8.3 89.6 2.1
2002 0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 2 700 21.1 58.5 20.4 6 3000 13.5 86.1 0.4 12 8100 13.6 85.6 0.8 0 0 0.0 0.0 0.0 8 8100 15.1 82.9 2.0
Average 0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 2 700 15.8 76.4 7.8 6 3000 14.9 83.8 1.3 11 7633 12.7 85.8 1.5 0.0 0.0 0.0 0.0 0.0 7 6733 13.6 84.4 2.0
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-15 – Liquid & Gas Fossil Fuels, Fossil Steam Units, Japan ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED STEAM TURBINE UNITS MORE THAN ONE YEAR OLD (%) 2000 100 to 199 MW
200 to 299 MW
300 to 399 MW
400 to 599 MW
600 to 799 MW
800 to 999 MW
1000 MW & UP
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2001
0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 44 15800 17.1 82 0.9 25 11984 14.4 84.5 1.1 40 24600 19.3 79.9 0.8 4 3416 8.5 90.8 0.7 9 9000 15.2 83.5 1.3
List of countries: Japan
24
0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 44 15800 13.6 85.7 0.7 25 11984 16.5 83.0 0.5 40 24600 13.9 85.0 1.1 4 3416 15.0 85.0 0.0 9 9000 20.0 78.7 1.2
2002 0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 44 15800 15.9 82.7 1.4 25 11984 10.4 88.8 0.8 40 24600 15.6 82.7 1.7 4 3416 12.3 87.3 0.4 9 9000 17.5 80.8 1.7
Average 0 0 0.0 0.0 0.0 0 0 0.0 0.0 0.0 44 15800 15.5 83.5 1.0 25 11984 13.8 85.4 0.8 40 24600 16.3 82.5 1.2 4 3416 11.9 87.7 0.4 9 9000 17.6 81.0 1.4
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-16 – Combustion Turbines, World, ≥30 MW ANNUAL UNAVAILABILITY OF BASE-LOADED COMBUSTION TURBINES MORE THAN ONE YEAR OLD (%) 2000 30 to 49 MW
50 to 74 MW
75 to 99 MW
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2001
17 625 8.3 92.9 2.2 8 456 3.7 93.2 3.1 0 0 0.0 0.0 0.0
18 659 94.9 2.1 1 69 18.9 80.1 1.0 0 0 0.0 0.0 0.0
2002 38 1561 3.5 89.4 3.9 0 0 0.0 0.0 0.0 1 81 0.0 95.9 4.1
Average 24 948 5.9 92.4 2.7 5 263 11 87 2 1 81 0.0 95.9 4.1
2002 198 16425 9.6 88.3 1.2 96 27790 23.3 80.1 0.8
Average 295 24501 15.2 85.1 1.4 105 25998 21.9 80.4 1.7
TABLE 1-17 – Hydro Units, World, ≥50 MW ANNUAL UNAVAILABILITY OF HYDRO UNITS (ALL OPERATING MODES) MORE THAN ONE YEAR OLD (%)
50 to 149 MW
150 and above MW
Count: MW: PUF: EAF: UUF: Count: MW: PUF: EAF: UUF:
2000 346 28831 25.2 78.2 1.7 111 25384 9.8 87.6 2.3
25
2001 340 28248 10.7 88.6 1.2 109 24821 32.7 73.4 1.8
World Energy Council
Performance of Generating Plant 2004 – Section 2
TABLE 1-18 – Recent Combined Cycle Units, World, ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED COMBINED CYCLE UNITS MORE THAN ONE YEAR OLD (%) 2000 100 to 199 MW Unit-years: 73 MW: 10035 Average MW/unit: 137 PUF: 10.39 EAF: 86.8 UUF: 2.28 200 to 299 MW Unit-years: 51 MW: 12262 Average MW/unit: 240 PUF: 12.19 EAF: 85.08 UUF: 2.73 300 MW & UP Unit-years: 24 MW: 11112 Average MW/unit: 463 PUF: 9.1 EAF: 90.66 UUF: 1.19
2001 71 9867 139 9.77 85.55 4.35 54 13004 241 8.48 88.46 4.94 30 13636 455 20 83.2 1.03
2002 75 10321 138 9.63 85.47 3.1 55 13308 242 7.24 89.06 3.3 34 15616 459 12.12 87.45 2.43
Average 73 10074 138 9.9 85.9 3.2 53 12858 241 9.3 87.5 3.7 29 13455 459 13.7 87.1 1.6
TABLE 1-19 – Historical Combined Cycle Units, World, ≥100 MW ANNUAL UNAVAILABILITY OF BASE-LOADED COMBINED CYCLE UNITS MORE THAN ONE YEAR OLD (%) 1991-1993 100 to 199 MW Unit-years: Average MW/unit: PUF: EAF: UUF: 200 to 299 MW Unit-years: Average MW/unit: PUF: EAF: UUF: 300 MW & UP Unit-years: Average MW/unit: PUF: EAF: UUF:
26
1994-1996 148 134 14.4 84.2 1.4 30 245 9.5 84.7 5.8 12 355 9.9 78.0 12.2
220 141 12.5 85.9 1.6 95 236 10.8 86.6 2.6 56 393 8.5 88.3 3.3
World Energy Council
Performance of Generating Plant 2004 – Section 2
GRAPH 1-1 – UNIT CAPABILITY FACTOR, WESTERN EUROPE
GRAPH 1-2 – UNPLANNED CAPABILITY LOSS FACTOR, WESTERN EUROPE
27
Performance of Generating Plant Section 3 NUCLEAR GENERATING PLANT UNAVAILABILITY FACTORS AND AVAILABILITY STATISTICS
MARIANNA SZIKSZAINE TABORI International Atomic Energy Agency (IAEA)
Work Group Membership: M. Szikszaine Tabori (IAEA) S. R. Fernandes (Brazil) W. Guangyao (China) D. Glorian (France) K. Yassin (Egypt) A. R. Karbassi (Iran) A. Torii (Japan) T. Uchida (Japan) K. Yoon Lee (Rep. of Korea) A. A. Al Tuwaijri (Saudi Arabia) V. Micali (South Africa) J. Ruansup-anek (Thailand)
World Energy Council
Performance of Generating Plant 2004 – Section 3
TABLE OF CONTENTS Page
1.
Nuclear Power Generating Units
3
1.1
Nuclear Power Information at the IAEA
3
1.2
Status of Nuclear Power Worldwide [2]
4
1.3
Worldwide Energy Availability and Unavailability
6
1.3.1
Approach Used for the Availability Analysis
6
1.3.2
Sustained Improvement in Plant Performance Worldwide
7
2.
Conclusions
11
3.
List of Countries Providing Information on Nuclear Power to IAEA-PRIS
13
REFERENCES
13
LIST OF TABLES Table 1:
Nuclear Power Reactors in Operation and Under Construction in the World (as of 31 December 2003)
5
LIST OF FIGURES Fig. 1:
World Energy Availability Factors
7
Fig. 2:
Distribution of Reactors with High Availability Factor
8
Fig. 3:
Distribution of Units According to Energy Availability Factor in 2002
8
Fig. 4:
Average Availability and Unavailability Factors (1990-92, 1993-95, 1996-98, 1999-2001)
9
Fig. 5:
Energy Availability Factors by Reactor Type (1999-2001)
10
1
World Energy Council
Performance of Generating Plant 2004 – Section 3
Page ANNEX 1 APPENDIX N1 -
STATUS OF NUCLEAR POWER PLANTS WORLDWIDE (AS OF 31 DECEMBER 2003)
A1-1
APPENDIX N2 -
NUMBER OF NUCLEAR POWER PLANTS BY REACTOR TYPE AND REGION (AS OF 31 DECEMBER 2003)
A1-2
APPENDIX N3 -
NUMBER OF NUCLEAR POWER PLANTS BY AGE (AS OF 31 DECEMBER 2003)
A1-3
APPENDIX N4 -
DISTRIBUTION OF NUCLEAR POWER PLANTS BY AGE AND TYPE (AS OF 31 DECEMBER 2003) DISTRIBUTION OF NUCLEAR POWER PLANTS BY AGE AND REGION (AS OF 31 DECEMBER 2003)
A1-4
APPENDIX N6 -
NUCLEAR SHARE OF ELECTRICITY GENERATION IN 2003
A1-6
APPENDIX N7 -
ANNUAL WORLD ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS
A1-7
APPENDIX N8 -
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS PWR
A1-8
APPENDIX N9 -
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS BWR
A1-9
APPENDIX N10 -
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS PHWR
A1-10
APPENDIX N11 -
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS WWER
A1-11
APPENDIX N12 -
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS LWGR
A1-12
APPENDIX N13 -
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS BY AGE
A1-13
APPENDIX N14 -
ENERGY AVAILABILITY BY REGION NORTH AMERICA
A1-14
APPENDIX N15 -
ENERGY AVAILABILITY BY REGION WESTERN EUROPE
A1-14
APPENDIX N16 -
ENERGY AVAILABILITY BY REGION EASTERN EUROPE
A1-15
APPENDIX N17-
ENERGY AVAILABILITY BY REGION FAR EAST
A1-15
APPENDIX N18 -
ENERGY AVAILABILITY BY REGION MIDDLE EAST AND SOUTH ASIA
A1-16
APPENDIX N19 -
ENERGY AVAILABILITY BY REGION LATIN AMERICA
A1-16
APPENDIX N20-
ENERGY AVAILABILITY BY REGION AFRICA
A1-17
APPENDIX N5 -
2
A1-5
World Energy Council
Performance of Generating Plant 2004 – Section 3
1. Nuclear Power Generating Units The statistics presented in this section are based on data collected by the International Atomic Energy Agency (IAEA) for its Power Reactor Information System (PRIS).
1.1 Nuclear Power Information at the IAEA The Power Reactor Information System (PRIS) Information and data on nuclear power reactors have been collected by the IAEA since its establishment. PRIS covers two kinds of data: general and design information on power reactors, and data on operating experience with nuclear power plants. General and design information covers data on all reactors that are in operation, under construction, or shutdown in the world. Operating experience data covers operating reactors and historical data on shutdown reactors since the beginning of commercial operation. PRIS contains the largest amount of worldwide statistical information on operating experience. Although there are other data banks in existence, which could sometimes be referred to, the IAEA PRIS is considered the most complete and authoritative source of statistical data on the subject area. PRIS makes it easy to identify individual units with their main characteristics, and to determine nuclear power development status and trends worldwide, in regions or in individual countries. Since 1990, the IAEA has compiled information, available but spread over a large number of documents, on additional technical characteristics, covering items related to the mode of plant operation, safety characteristics, safety features, existence of a safety analysis report and of emergency plans, plant environment, etc. This additional information on plant characteristics, which provides a better overview of the plant design and mode of operation is being implemented in PRIS. Uses of PRIS The IAEA’s PRIS can be used to assess nuclear power performance through many different indicators and outage causes in a systematic and homogeneous manner. The data contained in the system is useful for identifying problem areas and overall trends. The amount of operating experience data available on the system allows a statistical analysis to be performed. Currently, the principal nuclear power performance indicators cover: plant availability and unavailability; planned and unplanned outages; nuclear safety related events; unavailability of safety systems and support functions; worker safety related events; radiation exposure; fuel reliability; and volume of radioactive waste. Among those, PRIS provides information on energy availability factors, planned unavailability factors, unplanned unavailability factors due to causes in the plant and external to the plant, load factors, operating factors, unit capability factors, unplanned capability loss factors, and recently, the number of scrams. Statistics on planned and unplanned outages, including causes, are also available. It is important not to give priority to a single performance indicator, as this could distort the overall findings. Performance indicators are a tool to help identify problem areas, where improvements are necessary, but they do not provide either the root cause or the solutions.
3
World Energy Council
Performance of Generating Plant 2004 – Section 3
PRIS generates many products to the IAEA Member States and international organisations: such as MicroPRIS, PRIS-PC (front-end tool interface with on-line connection to PRIS through the Internet), PRIS on CD-ROM, and through a public Internet site at the address: http://www.iaea.or.at/programmes/a2/. Currently, these products are distributed to more than 700 organisations. In addition the IAEA Secretariat handles daily a considerable number of ad-hoc requests on nuclear power plants information and statistics.
1.2 Status of Nuclear Power Worldwide A total of 439 nuclear power plants were operating around the world at the end of 2003, according to data reported to the IAEA’s Power Reactor Information System. The plants had a total net installed capacity of 361 GW(e). Also during 2003, two nuclear power plants representing 1625 MW(e) net electric capacity were connected to the grid, one in China, and the second one in Korea. Additionally, construction of one new nuclear reactor in India started in 2003, bringing the total number of nuclear reactors reported as being under construction to 31, out of which 18 are located in Asia, where both population and economic growth are high, as is per capita energy consumption. Nuclear accounts for approximately 16% of the electricity produced in the world, which translates into one in every 6 light bulbs in the world today being lit by nuclear power. It is important to note that nuclear power reduces the emissions of Greenhouse Gases from the power sector by 500-600 Mt carbon (C) per year, equivalent to 17-20% of the total, depending on the assumption. The ten countries with the highest reliance on nuclear power in 2003 were: Lithuania, 79.9%; France, 77.7%; Slovak Republic, 57.4%; Belgium, 55.5%; Sweden, 49.6%; Ukraine, 45.9%; Slovenia 40.5%; Republic of Korea, 40.0%; Switzerland, 39.7%; and Bulgaria, 37.7%. In total, 16 countries relied upon nuclear power plants to supply at least a quarter of their total electricity needs. Worldwide in 2003, total nuclear generated electricity increased to 2524.03 TWh. Cumulative worldwide operating experience from civil nuclear power reactors at the end of 2003 exceeded 11,143 reactor-years. In North America, where 120 reactors supply about 20% of electricity in the United States and 13% in Canada, the number of operating reactors has increased due to recommissioning two delayed reactors in Canada in 2003. In Western Europe, with 141 reactors, overall capacity has declined slightly. In Eastern Europe with 68 reactors, a few partially built plants are likely to be completed, while aging units are being shut down. Only in the Middle East, Far East and South Asia, with a total of 104 reactors at present, are there clearly pronounced plans for expanding nuclear power, particularly in China, India, the Republic of Korea and Japan. For new plants, there are a number of economic challenges. Market de-regulation and privatisation have changed the strategy. Operators are no longer guaranteed cost recovery over a long period of regulated rates and private investors expect a high, rapid and secure return on investments. Nevertheless, the prospects for nuclear power in the near future might change. Some US utilities have showed intentions for building new plants, but currently ongoing licensing renewals for operating plants will add 20 more years and produce 50% more electricity than was allowed under the original license. The US National Energy Policy drafted in 2001 recommends “the expansion of nuclear energy as
4
World Energy Council
Performance of Generating Plant 2004 – Section 3
a major component of national energy policy”. The Industry’s Vision outlines an approach to meet future energy demand by adding 50GWe of new nuclear generating capacity by 2020. The Department of Energy (DOE) and industry’s joint “ Nuclear Power 2010 Program” focuses on exploring sites for new construction, demonstrating a new regulatory process, and implementing strategies to enhance the business case for building new plants. China is the most recent developing country to adopt nuclear power. It is currently operating 9 units and will have 17 units by 2010. India currently operates 14 Nuclear Power Plants (NPPs) and is constructing 8 more units, including the Fast Breeder Test Reactor. In Europe the Finnish utility TVO has decided to build a 5th NPP as an addition on the existing Olkiluoto site. It is a 1600Mwe EPR from a French German consortium. Table 1: Nuclear Power Reactors in Operation and Under Construction in the World (as of 31 December 2003) Country
Reactors in Operation
Reactors under Construction
Nuclear Electricity Supplied in 2000
No. of Units
Total MW (e)
No. of Units
Total MW (e)
TW (e).h
% of Total
2 1 7 2 4 16 8 6
935 376 5760 1901 2722 11323 5977 3548
1
692
3
2610
7.03 1.82 44.61 13.34 16.04 70.29 41.59 25.87
8.59 35.48 55.46 3.65 37.71 12.53 2.18 31.09
1
1040 21.82 420.70 157.44 11.01 16.37
27.32 77.68 28.10 32.69 3.30
230.80 123.28 14.30 10.51 3.80 1.81 4.54 138.39 12.66 17.86 4.96 59.36 65.50 25.93 85.31 76.70 763.74
25.01 40.01 79.89 5.23 4.48 2.37 9.33 16.54 6.05 57.35 40.45 23.64 49.62 39.73 23.70 45.93 19.86
Argentina Armenia Belgium Brazil Bulgaria Canada China Czech R. DPR Korea Finland France Germany Hungary India Iran Japan Korea RP Lithuania Mexico Netherlands Pakistan Romania Russian Fed. S. Africa Slovak R Slovenia Spain Sweden Switzerland UK Ukraine USA
4 59 18 4 14
2656 63363 20643 1755 2550
53 19 2 2 1 2 1 30 2 6 1 9 11 5 27 13 104
44139 15850 2370 1310 449 425 655 20793 1800 2442 656 7584 9451 3220 12052 11207 98298
Total
439
361094
8 2 3 1
3622 2111 3696 960
1 3
655 2825
2
776
4
3800
31
25387
2524.03
Total Operating Experience to 31 Dec. 2003 Years Months
50 36 191 25 129 486 39 74 0 99 1346 648 74 223 0 1123 220 36 23 59 35 7 761 38 100 22 219 311 143 1329 279 2871
7 3 7 3 2 11 1 10 0 4 2 0 2 5 0 7 8 6 11 0 10 6 4 3 6 3 2 1 10 8 10 8
11143
5
Note: The total includes the following data in Taiwan, China: • 6 units, 4884 MW (e) in operation; 2 units, 2660 MW (e) under construction; • 37.37 TW (e).h of nuclear electricity generation, representing 21.5% of the total electricity generated there; • 134 years 1 month of total operating experience.
5
World Energy Council
Performance of Generating Plant 2004 – Section 3
1.3 Worldwide Energy Availability and Unavailability 1.3.1 Approach Used for the Availability Analysis The statistics on nuclear power plants cover non-prototype units in operation or shutdown at the end of 2003, i.e. plants with more than 100 MW(e). The information reported to the IAEA Power Reactor Information System up to end of December 2003 formed the basis for this survey. Although, energy availability and unavailability factors are presented in this report for the year 2002, data analysis for nuclear power plants is made for the three-year period (19992001) to be consistent with other statistics in this report. The basic performance indicators for this study are the Energy Availability Factor (EAF) and the Planned (PUF) and Unplanned Energy Unavailability Factors (UUF). The Energy Availability Factor is the ratio of the actual energy generation (net) in a given period, expressed as a percentage of the maximum energy that could have been produced during that period by continuous operation at the reference capacity. Energy losses are planned if they are scheduled at least four weeks in advance. Planned energy losses are considered to be under plant management control and include losses due to planned outages for refuelling, maintenance, testing, etc. Unplanned energy losses include losses due to unplanned outages for maintenance, testing, repair, etc. and due to causes beyond the control of management (external causes). The availability and unavailability data of nuclear power plants are presented here according to their type, region and age. The following groups have been established for observing trends on energy availability and unavailability factors: Selection of Regions (R): World: North America: Latin America: Western Europe:
all units in the world Canada and the USA Argentina, Brazil and Mexico Belgium, Finland, France, Germany, Netherlands, Spain, Sweden, Switzerland and the United Kingdom Eastern Europe: Armenia, Bulgaria, Czech Republic, Hungary, Kazakhstan, Lithuania, Russian Federation, Romania, Slovak Republic, Slovenia and Ukraine Far East: China, Korea Rep., DPR Korea and Japan Africa: South Africa Middle East and South Asia: India, Iran and Pakistan This classification is in accordance with the International Atomic Energy Agency grouping of countries and geographical areas [3]. Period of Observation: The period of observation covers: P = 1:average over the years 1993, 1994, 1995 P = 2: average over the years 1996, 1997, 1998 P = 3:average over the years 1999, 2000, 2001. P = 4:average since commercial operation and up to 2002 (lifetime). 6
World Energy Council
Performance of Generating Plant 2004 – Section 3
The period of observation P = 1, 2 and 3 are used to compare trends on availability and unavailability presented in the two previous WEC Triennial Reports, published in 1998 and 2001, with the current three-year analysis. In an attempt to avoid conclusions based on individual results, samples with less than 10 units are not considered. The availability and unavailability statistics include only nonprototype units, i.e., units with capacity bigger than 100MW(e). Also, the Gas Coded Graphite Modulated Reactor (GCR) and Advanced Graphite Modulated Reactor (AGR) type of reactors are not considered in the analysis. 1.3.2 Sustained Improvement in Plant Performance Worldwide Worldwide Results There has been a steady improvement in the World Energy Availability Factor as shown in Figure 1. The EAF grew from about 73% in 1992 to 83.3% in 2001, and achieved 83.7% in 2002.
100 95 90 85 80 75
73.1
74.4
75.4
76.7
79.1
78.1
77.1
1996
1997
80.8
82.1
83.3
83.7
2001
2002
70 65 60 55 50 1992
1993
1994
1995
1998
1999
2000
Fig. 1: World Energy Availability Factors The number of plants demonstrating higher energy availability factors (greater than 75%) has also increased. In 2002, 51 out of 439 operating nuclear power plants presented an energy availability factor between 70 and 79% and 328 plants presented an EAF higher than 80%. (Figure 2).
7
World Energy Council
Performance of Generating Plant 2004 – Section 3
100 90 80 70 60 50 40 30 20 10 0 1992
1993
1994
1995
1996
1997
1998
70-79%
80-89%
1999
2000
2001
2002
>90%
Fig. 2: Distribution of Reactors with High Availability Factor
120 98
100 84
(%)
80
70 62
60 49 40 15
20
<3 0%
0
2 0
3
10
1
0
30 -3 5% 35 -4 0% 40 -4 5% 45 -5 0% 50 -5 5% 55 -6 0% 60 -6 5% 65 -7 0% 70 -7 5% 75 -8 0% 80 -8 5% 85 -9 0% 90 -9 5%
2
19
Fig. 3: Distribution of Units According to Energy Availability Factor in 2002 The cumulative world energy availability factor since the beginning of commercial operation and up to 2002 for non-prototype reactors, is 76.1%, while the planned energy unavailability factor (PUF) is 16%. There has been a steady decrease in both planned and unplanned energy unavailability factors over the last years indicating continuing improvement in plant maintenance management (Figure 4). The average planned energy unavailability factor decreased continuously from about 16.1% in the period 1993-1995 to 14.5% in 1996-1998 and achieved 12% in 1999-2001. The improvement in the unplanned energy unavailability factor (UUF) was also significant. Although the cumulative average UUF (since the beginning of commercial operation) is 8, it decreased from 7% in the period 1993 – 1995 to 6.5% in 1996 – 1998 to 4% in the last three-year period (1999-2001) and achieved 3.4% in 2002.
8
World Energy Council
55%
60%
65%
30
70%
25
Planned EUF (%)
Performance of Generating Plant 2004 – Section 3
75% 20 1990-1992
80%
1993-1995 1996-1998
15 85% 2002 10
1999-2001
90%
5 0
3
6
9
12
Unplanned EUF (%)
Fig. 4: Average Availability and Unavailability Factors (1990-92, 1993-95, 1996-98, 1999-2001) These values demonstrate the results of the efforts made by the nuclear industry for the safe and reliable operation of nuclear power plants. These improvements also reflect the impact of de-regulation and privatisation of the electricity market which have affected all electricity producers, but have also contributed to optimising operation of nuclear power plants in all its aspects: safety, reliability and economics. However, the sustainable improvements in energy availability can also be attributed to a process of learning from experience. A number of initiatives taken by the Member States and the Agency have a significant role in the achievement of sustained improvement in the overall performance of the plants. Survey by Reactor Type A survey by reactor type shows that there is considerable increase in the availability of PWR and PHWR units. The PWR units have improved the energy availability factor from 76.1% in 1992 to 86.5% in 2002. The three-year average also increased considerably from 78.5% between 1993-1995, to 81% between 1996-1998 and achieved 85% in 19992001. In 2002, the average EAF of all operating PWR units was 86.5%. For PWR units, the three-year average of planned energy unavailability factor (PUF) was 10% in the period 1999-2001, while the unplanned energy unavailability factor (UUF) was 4%. In 2002, 85 out of 207 PWR units presented EAF higher than 90%. The PHWR units also increased the energy availability factor from an average of 70%, in 1993-1995, to 81% in 1999-2001. Since 1997, the EAF has continuously recovered and increased. In 2002, the average EUF of all operating BWR units was 84% and 9 out of 29 units presented EAF higher than 90%. For PHWR units, the three-year average of planned energy unavailability factor (PUF) was 12% in the period 1999-2001 and the unplanned energy unavailability factor (UUF) was 6%, while UUF decreased to 4% in 2002.
9
World Energy Council
Performance of Generating Plant 2004 – Section 3
Many PWR units have optimised the frequency of refuelling outages. They now operate with longer fuel cycles. Others have implemented an outage strategy, which also enables a shorter duration of refuelling outages. Some of them perform refuelling outages in less than two weeks, while others in more than a month. The IAEA has also assisted its Member States in exchanging information on good practices for outage optimisation, improving nuclear power plant performance and other activities, which have contributed to a reduction in outage duration. Since 1992, the average energy availability factor of BWR units has varied from 71.5% in 1992 to 82.7% in 2002. The last three-year (1999-2001) average is 85%, but the EAF decreased to 82.7% in 2002, due to the unplanned energy unavailability factor increasing from 3% to 6% in 2002. The WWER units presented a cumulative EAF of 70% up to 2002. Nevertheless, the yearly values have been below this reference until 2000. The last three-year average (1999-2001) was 72% and the EAF increased to 77.3% in 2002. The availability of WWER units, as well as the LWGR (or RBMK), shows the impact of longer planned outages for refurbishment and backfittings since 1992. Note that the Finnish units of WWER type included in this group have always presented very high EAF. They have also demonstrated some of the lowest planned and unplanned energy unavailability factors in the world. The LWGR units show an increase in the availability (as an average of energy availability) of about 60% in the period between 1996-1998 rising to about 63% in 19992001. The yearly values have been below the cumulative EAF 67%. Planned unavailability factor of about 29% and unplanned unavailability factor for the last three years is 2.4%, which is below the cumulative value 4.1% Figure 5 presents the average energy availability values for the period 1999-2001 by reactor type. The energy unavailability factors by reactor type are presented in Appendices N8 to N12. 90
85
84
85 81
76
80
70
70
62
EAF (%)
60 50 40 30 20 10 0 ABWR
AGR
BWR
LWGR
PHWR
PWR
WWER
Fig. 5: Energy Availability Factors by Reactor Type (1999-2001)
10
World Energy Council
Performance of Generating Plant 2004 – Section 3
Survey by Region The analysis of energy availability and unavailability values by region since the beginning of commercial operation (lifetime) and in the last years (1992-2002) is presented in Appendices N14 to N20. The results show an increase in the yearly average of energy availability factor since 1992 in almost all world regions. In particular in North America the yearly EAF increased from 74% (1992) to 90% (2002). This increase is mainly due to the US units (74.7% in 1992 to 90.4% in 2002), which have improved considerably their performance in the last ten years, and to a lesser extent the Canadian units, which have increased the availability in the last three years. In Western Europe, the yearly EAF has also increased since 1992, although at a lower rate, from 74% in 1992 to 84% in 2002. This could be attributed to the uncertainties given by different countries’ energy policies in the region. In Eastern Europe, where the majority of units are of WWER and LWGR type, with the exception of the PWR unit operating in Slovenia, the yearly EAF has been between 66% to 74% in the last ten years. The plants in Latin America have also improved the EAF, although some variations and low values were experienced in some years, plants have recovered in the last years. The units in operation in the Far East present almost constantly high values of EAF. The EAF varies from 76% (1992) to 81% (2002). World regional analysis is difficult because operating plants in such large regions are often of different types; operate in different countries and under different economic and energy market conditions. More in-depth analysis should consider smaller regions or countries and other criteria used in benchmarking analysis, for instance. Survey by Age The calculation of energy availability factor by age of reactor since the beginning of commercial operation up to 2002 shows that units have a higher energy availability factor in the first year of commercial operation, which decreases in the subsequent years and achieves higher values again after the 5th year of commercial operation. The value is almost constant for some years. For reactors with more than 25 years of commercial operation the values of energy availability factor increases. The results should be analysed in detail to identify the role of learning from experience of older plants and other factors such as upgrade of capacity.
2. Conclusions In terms of energy availability, the performance of worldwide nuclear power plants has been steadily improving over the last several years. They have achieved high performance standard, the average availability has increased significantly, and in many countries, longer refuelling cycles were implemented and significant reductions in the duration of refuelling outages, number of scrams and number of significant events were achieved. All these are positive aspects, which play a role when assessing the future of the current operating nuclear power plants in the competitive environment and on planning additional capacity to the electricity grids.
11
World Energy Council
Performance of Generating Plant 2004 – Section 3
The average energy availability factor for nuclear power plants in the world has steadily increased from about 74.4% in 1992 to the current value, which is above 83.7%, with many utilities achieving significantly higher values. The worldwide results present a steady decrease in both planned and unplanned energy unavailability factors over the last years, indicating a continuing improvement in plant maintenance management. Therefore, it provides a good indication of improvements in plant reliability and safety in the last five years. The results by reactor type show that there is a considerable increase in the availability of PWR and PHWR, mainly due to a decrease of the planned and unplanned unavailability in the last three years. These results confirm the improvements in the worldwide energy availability factor, when compared with the results presented in the last surveys. Nevertheless, the determinant factors on a regional basis depend on the energy and economic situation, on the regulatory framework of the countries and, worldwide, the quality of the operators, more than the plant location. Many utilities are changing or are going to change their management strategies and address aspects, which require more attention to all areas. Some aspects have been identified as major recent industry transformations contributing to cost reduction: process changes (refuelling and other planned outage reductions), re-engineering and an aggressive self-assessment and capital investments. The implementation of these strategies, together with the learning from experience, contributed among others to: the decrease of unplanned outages; better outage management strategies which resulted in a reduction of the planned outage time (planned maintenance); effective failure prevention programmes including effective maintenance and on-line preventive maintenance, which help identify in advance equipment degradation indication due to ageing, etc. Nuclear plant operators are achieving high availability through integrated programmes, where international co-operation is playing a key role. The IAEA activities, which include nuclear power plant performance assessment and feedback, outage optimisation and effective quality management, are important examples of international co-operation to improve the performance of operating nuclear power plants. The World Association of Nuclear Operators (WANO) also plays a role in maximising the safety and reliability of the operation of nuclear plants, by exchanging information and encouraging communication of experience. Technological maturity, economic competitiveness and financing arrangements for new plants are key factors in decision making. Continued vigilance in nuclear power plant operation, and the enhancement of safety culture and international co-operation are highly important in preserving the potential of nuclear power to contribute to future energy strategies.
12
World Energy Council
Performance of Generating Plant 2004 – Section 3
3. List of Countries Providing Information on Nuclear Power to IAEA-PRIS Nuclear Power Plants Argentina - Armenia - Belgium - Brazil - Bulgaria - Canada - China - Czech Republic - DPR Korea - Finland - France - Germany - Hungary - India - Italy - Iran - Japan Kazakhstan - Korea, Rep. - Lithuania - Mexico - Netherlands - Pakistan - Romania Russian Federation - Slovakia - Slovenia - South Africa - Spain - Sweden Switzerland - Ukraine - United Kingdom of Great Britain and Northern Ireland United States of America.
REFERENCES
[1] Availability and unavailability factors of thermal power plants – Definitions and methods of calculation, UNIPEDE, March 1991 (Ref. 02011 REN91WE) [2] Nuclear Power Information at the IAEA, R. Spiegelberg-Planer, International Atomic Energy Agency, proceedings of the Workshop on Nuclear Reaction Data and Nuclear Reactors Physics, 16 February to 12 March 2004, International Centre for Theoretical Physics, Trieste, Italy. [3] IAEA Reference Data SeriesNo2, Nuclear Power Reactors in the World, International Atomic Energy Agency, Vienna, Austria, 2003. [4] Operating Experience with Nuclear Power Stations in Member States in 2002, International Atomic Energy Agency, Vienna, 2003, STI/PUB/1105. [5] Unavailability Factors of Thermal Generating plant and Availability Statistics 1995, D. Glorian, EdF, France and R. Spiegelberg-Planer, International Atomic Energy Agency, 16th World Energy Council Congress, Tokyo, Japan, 1995. [6] Unavailability Factors of Thermal Generating plant and Availability Statistics 1998, D. Glorian, EdF, France and R. Spiegelberg-Planer, International Atomic Energy Agency, 17th World Energy Council Congress, Houston, USA, 1998. [7] Exchange of Availability/Performance Data on Gas Turbine and Combined Cycle Generating Plant, 15th World Energy Congress, Madrid, Spain, 1992.
13
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N1 - STATUS OF NUCLEAR POWER PLANTS WORLDWIDE (AS OF 31 DECEMBER 2003) Nuclear Electricity
Total Operating Experience
Supplied in 2003 % of TW(e).h Total
to 31 Dec. 2003
Reactors in Operation Reactors under Construction Country No of
Total
No of
Total
Units
MW(e)
Units
MW(e)
2
935
8.59
50
7
ARMENIA
1
376
1.82 35.48
36
3
BELGIUM
7
5760
44.61 55.46
191
7
BRAZIL
2
1901
13.34
3.65
25
3
BULGARIA
4
2722
16.04 37.71
129
2
70.29 12.53
486
11
16
11323
CHINA
8
5977
CZECH R.
6
3548
DPR KOREA
3
1
692
2610
7.03
Months
ARGENTINA
CANADA
1
Years
41.59
2.18
39
1
25.87 31.09
74
10
0
0
1040
FINLAND
4
2656
21.82 27.32
99
4
FRANCE
59
63363
420.70 77.68
1346
2
GERMANY
18
20643
157.44 28.10
648
0
HUNGARY
4
1755
14
2550
8 2
2111
JAPAN
53
44139
3
KOREA RP
19
15850
1
LITHUANIA
2
2370
MEXICO
2
1360
10.51
NETHERLANDS
1
449
3.80
PAKISTAN
2
425
1.81
ROMANIA
1
655
1
655
4.54
30
20793
3
2825
2
1800
SLOVAK R
6
2442
SLOVENIA
1
656
SPAIN
9 11
INDIA
74
2
3.30
223
5
0
0
3696
230.08 25.01
1123
7
960
123.28 40.01
220
8
14.30 79.89
36
6
5.23
23
11
4.48
59
0
2.37
35
10
9.33
7
6
138.39 16.54
761
4
6.05
38
3
17.86 57.35
100
6
4.96 40.45
22
3
7584
59.36 23.64
219
2
9451
65.50 49.62
311
1 10
IRAN
RUSSIA S. AFRICA
SWEDEN SWITZERLAND
11.01 32.69 3622
16.37
12.66 2
776
5
3220
25.93 39.73
143
UK
27
12052
85.31 23.70
1329
8
UKRAINE
13
11207
76.70 45.93
279
10
USA
104
98298
763.74 19.86
2871
8
Total
439
361094
11143
5
4
31
3800
28387
2524.03
Note: The total includes the following data in Taiwan, China: — 6 units, 4884 MW(e) in operation; 2 units, 2600 MW(e) under construction; — 37.37 TW(e).h of nuclear electricity generation, representing 21.5% of the total electricity generated there; — 134 years 1 month of total operating experience.
A1-1
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N2 - NUMBER OF NUCLEAR POWER PLANTS BY REACTOR TYPE AND REGION (AS OF 31 DECEMBER 2003)
160
140
120
100
80
60
40
20
0 Africa
Eastern Europe
Western Europe
48
2
2
1
92
2
1
Latin America
WWER PWR PHWR
2
69
46
1
16
6
13
35
31
12
GCR 1
FBR
AGR
Far East
17
LWGR
BWR
Middle East and South Asia
North America
2
1 20
1
14 2
ABWR
A1-2
2
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
Age (years)
APPENDIX N3 - NUMBER OF NUCLEAR POWER PLANTS BY AGE (AS OF 31 DECEMBER 2003)
3
44 43 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
1 2 2 1 3 7 5 11 10 15 23 11 16 14 15 7 21 22 19 22 33 32 24 22 14 11 10 4 6 9 6 4 6 3 4 4 6 3 6 2 0
5
10
15
20
No. of Reactors
A1-3
25
30
35
World Energy Council
APPENDIX N4 -
Performance of Generating Plant 2004 – Section 3, Annex 1
DISTRIBUTION OF NUCLEAR POWER PLANTS BY AGE AND TYPE (AS OF 31 DECEMBER 2003)
140
120
100
FBR
Number of Reactors
GCR AGR 80
LWGR WWER PWR
60
PHWR BWR ABWR
40
20
0
0-5
6 - 10
11 - 15
16 -20
1
FBR
21 - 25
26 - 30
1
1
> 30
12
GCR AGR
4
6
LWGR
1
5
4
7
4
WWER
5
2
4
22
12
3
2
PWR
11
14
24
66
45
40
13
PHWR
8
4
6
10
6
2
2
BWR
1
5
6
24
16
22
16
ABWR
2
A1-4
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N5 - DISTRIBUTION OF NUCLEAR POWER PLANTS BY AGE AND REGION (AS OF 31 DECEMBER 2003)
160
140
Number of Reactors
120
100
80
60
40
20
0 Latin Am erica
Middle East and South
North Am erica 3
21
1
4
32
14
19
22
2
26
51
1
2
30
31
1
2
7
6
1
6
1
5
Eastern Europe
Far East
26 - 30
8
6
21 - 25
8
15
21
14
2
11 - 15
23
15
6 - 10
3
16
0-5
5
9
Africa
1
> 30
16 -20
2
A1-5
Western Europe
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N6 - NUCLEAR SHARE OF ELECTRICITY GENERATION IN 2003(%)
0
10
20
30
40
50
60
LITHUANIA FRANCE 57 56 50 46
SLOVENIA KOREA, REP. SWITZERLAND
40 40 40 38
(%)
BULGARIA ARMENIA HUNGARY CZECH REP.
36 33 31 28
GERMANY FINLAND JAPAN UNITED KINGDOM
27 25 24
SPAIN USA RUSSIA CANADA
24 20 17 13
ROMANIA ARGENTINA SOUTH AFRICA
PAKISTAN CHINA
80 80 78
SLOVAKIA BELGIUM SWEDEN UKRAINE
MEXICO NETHERLANDS BRAZIL INDIA
70
9 9 6 5 5 4 3 2 2
A1-6
90
World Energy Council
APPENDIX N7 -
Performance of Generating Plant 2004 – Section 3, Annex 1
ANNUAL WORLD ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS Energy Availability Factor
10 0 95 90
(%)
85
81
79 76
80 75
76
70 65 60 55
19 82 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02
50
Y ear ly
P = 1
P = 2
P = 3
Li f et i me
Planned Energy Unvailability Factor Yearly
P=1
P=2
P=3
Lifetim e
30 25
(%)
20
16
16
14
15
12
10 5
19 82 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02
0
Unplanned Energy Unavailability Factor Yearly
P=1
P=2
P=3
Lifetim e
14 12 8
8 6
6
6 4
4 2 0
19 82 19 83 19 84 19 85 19 86 19 87 19 88 19 89 19 90 19 91 19 92 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02
(%)
10
A1-7
World Energy Council
APPENDIX N8 -
Performance of Generating Plant 2004 – Section 3, Annex 1
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS FOR PWR
100 95 90 85 80 75 70 65 60 55 50
85 81
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
77
1992
(%)
PWR - Energy Availability Factors
Lifetime
PWR - Planned Energy Unavailability Factors 30 25
(%)
20 15
14
14
11
10
10
5
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
0
Lifetim e
PWR - Unplanned Energy Unavailability Factors 14 12
8
7
6
6
6 4
4 2
Yearly
P=1
P=2
A1-8
P=3
Lifetim e
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
0 1992
(%)
10
World Energy Council
APPENDIX N9 -
Performance of Generating Plant 2004 – Section 3, Annex 1
ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS BWR
100 95 90 85 80 75 70 65 60 55 50
85 80
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1994
1993
1995
75
74
1992
(%)
BWR - Energy Availability Factors
Lifetim e
BWR - Planned Energy Unavailability Factors 30 25
(%)
20
18
17
15
15
11 10 5
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
0
Lifetim e
BWR - Unplanned Energy Unavailability Factors 14 12
8
7
7
6 4
4
3
2
Yearly
P=1
P=2
A1-9
P=3
Lifetim e
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
0 1992
(%)
10
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N10 - ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS PHWR
100 95 90 85 80 75 70 65 60 55 50
81
P=1
P=2
P=3
2002
2001
2000
1999
1996
1995
1994
1993
Yearly
1998
70
70
1997
73
1992
(%)
PHWR - Energy Availability Factors
Lifetim e
16 13
12
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
6
1993
20 18 16 14 12 10 8 6 4 2 0
1992
(%)
PHWR - Unplanned Energy Unavailability Factors
Lifetim e
PHWR - Planned Energy Unavailability Factors 30 25
15 10
13
12
11
12
5
Yearly
P=1
P=2
A1-10
P=3
Lifetim e
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
0 1992
(%)
20
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N11 - ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS WWER
100 95 90 85 80 75 70 65 60 55 50
72
70
67.5
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
64
1992
(%)
WWER - Energy Availability Factors
Lifetim e
WWER - Planned Energy Unavailability Factors 30 25
25
25
22
21
(%)
20 15 10 5
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
0
Lifetim e
8 5.5
4
Yearly
P=1
P=2
A1-11
P=3
Lifetim e
2002
2001
2000
1999
1998
1997
1996
1995
1994
2 1993
20 18 16 14 12 10 8 6 4 2 0
1992
(%)
WWER - Unplanned Energy Unavailability Factors
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N12 - ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS LWGR
100 95 90 85 80 75 70 65 60 55 50
67
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
63
60
58
1992
(%)
LWGR - Energy Availability Factors
Lifetim e
LWGR - Planned Energy Unavailability Factors 40
36
35 29
30 (%)
25
29
25
20 15 10 5
Yearly
P=1
P=2
P=3
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
0
Lifetim e
Yearly
P=1
P=2
A1-12
P=3
Lifetim e
2002
2001
2000
1999
2.4 1998
1996
2.7 1995
1994
3.6
1997
4.1
1993
20 18 16 14 12 10 8 6 4 2 0
1992
(%)
LWGR - Unplanned Energy Unavailability Factors
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N13 - ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS BY AGE Energy Availability Factor by Age
100 95 90
EAF (%)
85 80 75 70 65 60 55
0 1 2 3 4 5 6 7 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
50
Planned Energy Unavailability Factor by Age
30 25
EAF (%)
20 15 10 5
0 1 2 3 4 5 6 7 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
0
Unplanned Energy Unavailability Factor by Age
30 25
15 10 5 0 0 1 2 3 4 5 6 7 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
EAF (%)
20
A1-13
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N14 - ENERGY AVAILABILITY BY REGION NORTH AMERICA North America 100 86
90
(%)
80
74
73
1992
1993
77
78
88
89
90
2001
2002
84
84
80
77 72
70 60 50 40 30 1994 1995
1996
1997 1998
1999 2000
APPENDIX N15 - ENERGY AVAILABILITY BY REGION WESTERN EUROPE Western Europe 100 90
(%)
80
74
79
80
80
83
84
83
82
83
70 60 50 40 30 1992 1993
1994
1995 1996
A1-14
1997
1998 1999
2000
2001 2002
World Energy Council
APPENDIX N16 -
Performance of Generating Plant 2004 – Section 3, Annex 1
ENERGY AVAILABILITY BY REGION EASTERN EUROPE Eastern Europe
100 90
(%)
80 70
66
64
61
61
65
67
64
65
68
70
74
60 50 40 30 1992 1993
1994
1995 1996
1997
1998 1999
2000
2001 2002
APPENDIX N17- ENERGY AVAILABILITY BY REGION FAR EAST Far East 100 90
(%)
80
76
78
76
80
81
1995
1996
83
83
82
83
83
81
70 60 50 40 30 1992
1993 1994
1997 1998
A1-15
1999
2000 2001
2002
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N18 - ENERGY AVAILABILITY BY REGION MIDDLE EAST AND SOUTH ASIA Middle East and South Asia 100 90
(%)
80
78
1999
2000
75
78
68
70 60
76
60
56
52
50
1995
1996
46
50
39
40 30 1992
1993
1994
1997
1998
2001
2002
APPENDIX N19 - ENERGY AVAILABILITY BY REGION LATIN AMERICA
Latin America 100 90
83 78
(%)
80 70
68
69
71
72
1995
1996
82
80
80
67
63
60 50 40 30 1992
1993
1994
1997
A1-16
1998
1999
2000
2001
2002
World Energy Council
Performance of Generating Plant 2004 – Section 3, Annex 1
APPENDIX N20- ENERGY AVAILABILITY BY REGION AFRICA Africa 100 90
85
80 (%)
70 70 60
79
79
81 76
73 69
62
58 45
50 40 30 1992
1993
1994
1995
1996
1997
A1-17
1998
1999
2000
2001
2002
Performance of Generating Plant Section 4 PERFORMANCE OF HYDRO AND PUMP STORAGE PLANT
TERRY MOSS Eskom Generation (South Africa)
Work Group Membership: T. Moss (South Africa), Chair R. S. Chadha (India) R. N. Misra (India) C. Cakmak (Turkey) G. Varol (Turkey)
World Energy Council
Performance of Generating Plant 2004 – Section 4
PERFORMANCE OF HYDRO AND PUMP STORAGE PLANT The Largest Source of Renewable Energy Hydropower is the world’s largest source of renewable energy used for power generation; it accounts for 17% of the world’s electricity production. Hydro resources are widely spread and used around the world, and more than 150 countries use hydropower for electricity generation. The main remaining hydropower potential development exists in developing countries in Asia, South America and Africa. For example, Europe has developed 75% of its economic hydro potential, while Africa only about 7%. Hydropower technologies are reliable, mature and efficient. The energy conversion efficiency can reach well over 90%. Hydropower can be used to provide different services: from base-load supply to peak-load and system-backup services from hydro storage schemes. By replacing outdated or worn-out equipment within existing hydropower infrastructure, service life of hydro power plants can be extended by 30-50 years. Storage of electric power is one of the major challenges facing further development of many environmentally sound renewable energy technologies. Hydropower storage schemes, such as pumped storage, are uniquely efficient and suitable to support many intermittent renewable technologies. Given its versatile range of applications, hydropower is widely used to provide grid stability and essential ancillary services, e.g. black start-up, frequency control and flexible reactive loading. In the majority of power systems, the short-term power demand can reach 200% of the system average value. Hydro storage plants can be deployed at short notice to cover the peak demand, and this gives the plant operator a considerable competitive advantage. In a market environment, electricity prices can change both significantly and quickly during the day. Hydro peaking plants offer a unique flexibility, which enables utilities to follow and quickly respond to changes in demand. There is an enormous potential for further development of all types of hydropower, both large and small. Nevertheless, the most efficient and least costly way of increasing hydropower’s contribution is improving the performance of existing schemes.
Hydro & Pump Storage Plant Work Group within the PGP Committee The 2002-2004 three-year work programme of the PGP Committee saw the incorporation of Work Group 5 on the Performance of Hydro & Pump Storage Plant into the full threeyear cycle for the first time. The terms of reference and a work plan for the ensuing three years were formally adopted at the 2001 Congress in Buenos Aires. A commitment was made in the work plan to ensure that plant performance data on Hydro & Pump Storage Plant could be collected for input to the PGP triennial report on Plant Performance. This turned out to be a significant challenge, as most of the PGP member utilities have been accustomed to focusing their benchmarking and plant performance data collection activities on thermal and nuclear plants. The collection of data amongst the "hydro fraternity" is in no way as prevalent as within thermal and nuclear “communities”. Nonetheless the UNIPEDE suite of plant performance indicators (as prepared by the Hydrostep Working Group) was put forward as an instrument for both recording and benchmarking of Hydro & Pump Storage Plant performance.
1
World Energy Council
Performance of Generating Plant 2004 – Section 4
There was little response from the industry and this combined with the fact that the NERC suite of Plant Performance Indicators, and those of UNIPEDE, were not directly comparable led to a change in focus for Work Group 5 mid way through the 2002-2004 work cycle. Efforts were refocused on two major initiatives. Firstly, support was given to Work Groups 2 and 3 in their effort to develop a common database, which would allow input of information, using either time-based or energybased indicators, thus permitting the direct comparison of information produced from data collected using either the NERC or UNIPEDE suites of indicators. The second major initiative was to endeavour to reach a significant number of owners and operators of Hydro & Pump Storage Plants worldwide. To achieve this the Chairman of Work Group 5 drafted a Cooperation Agreement between WEC and the International Hydropower Association (IHA). It was agreed that the IHA would assist in canvassing its member utilities and encourage them to submit data for the WEC - PGP triennial report. The Co-operation Agreement was signed by the WEC Secretary General and the President of the IHA in May 2004. Action was taken in June whereby the IHA would approach its utility members with a request under the signature of the IHA President to contribute data to the WEC-PGP "Sydney Report". However, to ensure confidentiality and prevent unauthorised access to the database, inputting of data into the database requires a unique user-id and a password. For previous WEC PGP Surveys, power plant performance data was collected at a national level and submitted as peer groups by the Member Committee. During the initial phase of the new database development, the same principle was applied providing a single point of access per country through the WEC Member Committee. However, additional user-id’s and passwords for individual users can be issued on request. As the PGP Committee moves towards collection of unitspecific data, the database will be adapted further to accommodate an increasing number of data reporters, but to ensure system and data security and confidentiality, the system access will always remain password-protected. This security barrier adds an extra degree of difficulty to accessing the database, and it's impact on the number of utilities willing and able to contribute to the 2004 report remains to be seen. The database section of hydro power plants currently has data for over 3,500 units, mainly from Canada and the United States. Work Group 5 also participated in National Seminar and Round Table events. These events in the past used to be focused on fossil fuel-fired power plants, but significant advances were made during the 2002-2004 work programme to introduce and increase both Hydro & Pump Storage and Renewable Energy (Work Group 4) participation. This was usually well received by the host organisation and the participants, and continued future participation is strongly recommended for future events. One final area of Work Group 5 activity focused on establishing working relations with other global organisations with an interest in the Hydro & Pump Storage field, including VGB PowerTech, EPRI, the North American Pump Storage Users Committee and the IHA. Extended co-operation with these and other organisations during the forthcoming three-year work programme is seen as essential in terms of gleaning maximum benefits from the available global resource and expertise base.
2
Performance of Generating Plant Section 5 PROPOSAL OF TECHNICAL, ENVIRONMENTAL AND SOCIOLOGICAL PERFORMANCE INDICATORS FOR RENEWABLE ENERGY SOURCES
BRUNO MANOHA Electricité de France (France) MARTIN HOPPE-KILPPER Institut für Solare Energieversorgungstechnik (ISET) (Germany) ROBERTO VIGOTTI ERGA of ENEL (Italy) EVAN HUGHES Electric Power Research Institute (USA) RUGGERO BERTANI International Geothermal Association
Work Group Members: B. Manoha (France) U. Said Said (Egypt) M. Hoppe-Kilpper (Germany) M. H. Abdurrachman (Indonesia) R. Vigotti (Italy) T. Moss (South Africa) E. Hughes (US) R. Bertani (IGA)
World Energy Council
Performance of Generating Plant 2004 - Section 5
SUMMARY The work presented here has been performed within a Work Group of the World Energy Council's (WEC) Committee on the Performance of Generating Plant (PGP). Its objective is to define performance indicators for generating plants using Renewable Energy Sources (RES). A first phase of the work (1999-2001) has proposed technical performance indicators, as it has been done in the past by WEC and UNIPEDE for nuclear and fossil-fired power plants. These technical performance indicators have been proposed for wind, photovoltaic, biomass and geothermal energy, and were presented in a report presented at the 18th WEC Congress, Buenos Aires, October 2001. The present second phase (2002-2004) aims at extending the work to proposals of environmental and sociological indicators, as RES are considered particularly beneficial in terms of sustainable development. As for the first phase, leading experts in RES, working with international organisations (IEA, Eurelectric, IGA, WEC, etc.), have participated in the work in order to develop standards. Sample examples on existing installations are given in this report to demonstrate the magnitude of the performance indicators. We hope that these indicators will be useful to all experts, organisations and countries willing to develop renewable energy. In the future, through benchmarking, cross-comparisons, intercomparisons, and eventually databases, it is hoped that these indicators will help compare the respective benefits and deficiencies of various technologies, analyse and identify their weaknesses, and thus lead to improvements in their performance, and finally contribute to a more efficient and rapid development of RES.
RESUMÉ Le présent travail a été réalisé dans le cadre du Conseil Mondial de l’Energie (CME) et de son Comité sur les performances des centrales de production électrique (PGP). Son objectif est de définir des indicateurs de performance pour les filières utilisant des sources d’énergie renouvelable (ENR). Une première phase (1999-2001) a consisté à proposer des indicateurs techniques, comme cela a été fait par le passé pour les centrales nucléaires et thermiques classiques par le CME et l’UNIPEDE. Ces indicateurs techniques ont été proposés pour l’éolien, le photovoltaïque, la biomasse et la géothermie, et le travail a été consigné dans un rapport présenté au 18ème Congrès du Conseil Mondial de l’Energie à Buenos Aires en Octobre 2001. La seconde phase présentée ici (2002-2004) a pour principal objectif d’étendre l’étude à des indicateurs environnementaux et sociologiques, les ENR ayant pour caractéristique de participer fortement à ce qu’on a coutume d’appeler le développement durable. Comme pour la première phase, d’éminents spécialistes ont travaillé à la définition de ces indicateurs, en liaison avec divers organismes internationaux reconnus en la matière (IEA, Eurelectric, IGA, CME, etc.), afin que les indicateurs proposés puissent être considérés comme aussi standard que possible. Des exemples sont présentés de manière à fournir de premiers ordres de grandeur de ces indicateurs de performances.
Nous espérons que ces indicateurs s’avéreront très utiles pour toutes les personnes, organismes et pays souhaitant développer les énergies renouvelables. A l’avenir, ils devraient aider, à travers des travaux d’inter comparaisons, de benchmarkings, voire la mise en place éventuelle de bases de données, à comparer et positionner les différentes techniques utilisées, permettant ainsi de mieux identifier leurs faiblesses éventuelles, d’améliorer leurs performances et d’aider, de façon générale, au développement des énergies renouvelables.
1
World Energy Council
Performance of Generating Plant 2004 - Section 5
Acknowledgements This report has been made possible through the work of international experts who have generously volunteered their time and expertise to this project. We are grateful to the four sub-group leaders Messrs. Martin Hoppe-Kilpper, Roberto Vigotti, Evan Hughes and Ruggero Bertani, the ISET team and the members of the International Geothermal Association, for their valuable contributions. We also wish to thank all the participants listed in Annex 1, and especially Mrs Elena Nekhaev, WEC Director of Programmes, for their efficient support.
2
World Energy Council
Performance of Generating Plant 2004 - Section 5
TABLE OF CONTENTS Page SUMMARY / RESUMÉ
1
ACKNOWLEDGEMENTS
2
1.
Introduction and Background
5
2.
Objectives and Scope of the Work
6
2.1
Objectives of the Work Group
6
2.2
Expected Benefits of the study
6
2.3
Scope of the Work
7
2.4
The Four Sub-groups and the Participants
7
3.
8
4.
Overview of the Present Development and Future Expansion of Renewable Energy Review of Technical Performance Indicators Proposed in Phase 1
4.1
Wind Energy
11
4.2
Solar PV Energy
16
4.3
Biomass Energy
19
4.4
Geothermal Energy
22
5.
Proposal of Environmental Indicators
25
5.1
General Environmental Indicators
25
5.2
Specific Indicators for Wind, PV, Biomass and Geothermal Energies
32
6.
Proposal of Sociological Indicators
35
7.
Conclusion and Recommendations For The Future
39
8.
General References
40
9
LIST OF FIGURES Fig. 1:
Measured Power Curve of Turbine with Pitch Control and Variable Speed (a) and Stall Control and Constant Speed (b).
12
Fig. 2:
Wind Speed – Frequency and Weibull Distribution
13
Fig. 3:
Distribution of Wind Energy Based on Energetically Weighted Frequency Distribution of the Wind Direction in 12 Sectors
13
Fig. 4:
Gross Wind Energy Supply by Years (Germany 1993-1999)
14
Fig. 5:
Monthly Capacity Factor of the Wind Turbines in the Scientific Measurement and Evaluation Program (WMEP) of ISET, Germany, 1990-1999
14
3
World Energy Council
Performance of Generating Plant 2004 - Section 5
Page ANNEX 1:
Contributors to WEC PGP Work Group - Phase 2
41
ANNEX 2:
Renewable Energy in the World
42
ANNEX 3:
Synthesis Table on the Main Advantages and Drawbacks of the Various RES Technologies
51
ANNEX 4:
List of Proposed RES Indicators
52
ANNEX 5:
Commercial Availability
54
4
World Energy Council
1.
Performance of Generating Plant 2004 - Section 5
Introduction and Background
The mission of the World Energy Council is “ to promote the sustainable supply and use of energy for the greatest benefit of all”. This encompasses, among other things: • • • •
The conservation of energy resources; The protection of the environment; Strategic future planning; The assistance to developing countries to help them meet their energy needs.
In the electricity sector, this can also mean the energy demand management, and the development of renewable energy. The WEC Committee for the Performance of Generating Plant (PGP) was established 30 years ago to enable the countries and electricity producers to evaluate the respective performances of plants, detect their weaknesses, and gain experience from successful performance improvement efforts of other producers. Renewable energy is developing rapidly, and it was felt that the time was ripe to begin the same work on this form of energy as well. After a first attempt focused on wind energy only, presented at the 17th WEC Congress in Houston in 1998 (1), a specific Work Group was established in 1999, and the results of the first phase of its work, focused on technical performance indicators for RES, were presented during the 18th WEC Congress in Buenos Aires, in October 2001 (2). These results are reviewed later in this document. Because of the specific characteristics and widely perceived advantages of renewable energy, especially in terms of “sustainable development”, it was decided to extend the work to cover environmental and sociological indicators. To propose such indicators is the main task of phase 2, presented in this report. Economical indicators have not been proposed and included at this stage of the work, because it seemed a very difficult task taking into account the variety of circumstances, the rapid development of RES and issues of “confidentiality’. However some general economical indications, figures and comments are given in the report (5.1.4 and 8.), in order to give a more complete overview of the issues facing RES. In this report, Chapter 4 reviews the technical performance indicators proposed during phase 1. Chapters 5 and 6 present the proposed environmental and sociological indicators defined during this phase 2. The complete list of proposed indicators is presented in Annex 4.
5
World Energy Council
Performance of Generating Plant 2004 - Section 5
2.
Objectives and Scope of the Work
2.1
Objectives of the Work Group
The objective of the Work Group is to provide information and enable benchmarking for generating plants using renewable energy resources, in order to help improve the efficiency of the systems and the design of new projects, and enable potential project participants to evaluate and compare them in terms of their respective performances. indicators are necessary. The first phase of the work group has resulted in a proposal for technical indicators. Environmental and sociological indicators are proposed in the present report (phase 2 of the work). Economic indicators could be developed in a future phase. At this stage, these indicators are only proposed concepts, which will still have to be more precisely defined and compared with existing or emerging norms and standards. The general objective of this work is to promote efficient development of Renewable Energy around the world, and to improve the performance of generating plants utilizing them.
2.2
Expected Benefits of the Study
Improvements in power plant performance brings many direct benefits, including: • • •
Increased generating capacity; Fewer (and shorter) outages; Better power plant economics.
These direct benefits produce equally important secondary benefits: improved confidence in RES technologies, more effective use of existing generating capacity, reduced or deferred need for construction of new generating capacity, and lower overall generation costs. Another benefit can be achieved through an improved management process. Mediocre management often is the reason for poor performance of generating plants and their inability to realise their inherently superior performance potential. Moreover, better awareness and knowledge of the environmental and sociological advantages of RES can also improve public acceptability of the systems. Conversely, an early identification of potential problems (for example, the potentially high mortality of bird populations near wind turbine plants) can help minimize and mitigate compensation claims and thus ensure the development of renewable energy resources in a sustainable manner. The development of performance indicators is the first step in a process of creating a large database, which would enable power plant operators to compare their own plant performances with others, and make improvements.
6
World Energy Council
2.3
Performance of Generating Plant 2004 - Section 5
Scope of the Work
The following conditions were agreed for the first phase: • • • •
2.4
Work limited to grid-connected power plants ; Focus on wind, solar PV (photovoltaic), biomass and geothermal energy; Work divided into “ sub-groups”, each with a particular expertise in one type of RES (wind, solar PV, biomass or geothermal); Each sub-group chaired by a leading expert, to ensure the results to be accepted as consistent with “standards” accepted by the international community (in particular IEA, Eurelectric, IGA, WEC, etc.), and to avoid duplication of efforts.
The Four Sub-Groups and the Participants
The following experts chaired the subgroups: • • • •
Wind: Martin Hoppe-Kilpper from ISET (Germany), responsible in particular for the German “250 MW wind programme”; Solar / PV: Roberto Vigotti from ERGA of ENEL (Italy), also Chairman of RES working groups at the International Energy Agency (IEA) and Eurelectric; Biomass: Evan Hughes from EPRI (USA), biomass specialist at the US Electric Power Research Institute; Geothermal Energy: Ruggero Bertani from Italy, Executive Director of the International Geothermal Association (IGA).
The task of each "RES leader" was to propose specific indicators for the specific RES sourced by the expert group, to validate them in the respective RES community, and to find a few power plants which would allow testing and provide initial, actual values of these indicators (with the help of some experts previously involved with the questionnaire sent at the beginning of phase 1). During phase 1, about 60 potential participants from 20 countries were contacted and many of them accepted to participate and contribute to the project. Special effort was made to find experts from all over the world, well known in the RES community, and involved in international organisations, to ensure that the results of the Work Group would be widely accepted. Annex 1 gives the list of the main contributors to this research.
7
World Energy Council
3.
Performance of Generating Plant 2004 - Section 5
Overview of the Present Development and Future Expansion of Renewable Energy
In order to understand the development of renewable energy, it is essential to consider the existing electricity demand. Today, nearly two billion people in the world have no access to electricity (and another billion has less than five hours of electricity a day). The gap between industrial and developing countries is increasing dramatically. A US – EIA study predicts that in the next 20 years, the global demand for electricity will increase by 54% (world average), with a growth of 91% in developing countries. Moreover, environmental concerns (above all, the increasing concentration of greenhouse gases in the atmosphere) require drastic changes in the behaviour of both developed and developing countries to ensure a transition towards “sustainable development”, which in particular includes the development of renewable energy. RES are not the only solution, but they are an important part of the “energy mix. Annex 2 illustrates a few significant figures and general information concerning electricity in the world by the end of 2002, including contribution and development of renewable energy sources. The figures highlight the need to develop clean and renewable energy, in order to: • • • •
Provide electricity to developing countries, especially in the remote areas and in the areas where wind, sun, biomass and geothermal energy are abundant; Protect the environment, especially in terms of low emissions of pollutants, and in Particular greenhouse gases, which have the potential to lead to climate change; Conserve fossil energy resources; Contribute to security of energy supply.
In this respect, it would seem reasonable to expect that the “developed countries”, which in total are the largest electricity consumers and the largest polluters in the world, would take on a responsibility to be at the forefront in developing clean and economic energy resources. Although RES are presently marginal, in relative terms, in world energy production, except for biomass and hydropower (which is not included here and accounts today for about 90% of the RES electricity production), their development is important especially for wind and solar / PV, whose average annual growth rate is about 25 to 30%. USA, Japan and Europe (in particular Germany, Spain, Italy and Denmark) are the leading countries, but developing countries, such as India, Mexico and Brazil are also making an important effort in the development of RES. Finally, Annex 3 provides an overview of the main advantages and drawbacks of the various types of renewable energies, showing that their optimal use depends on many factors and must be carefully examined and compared to other energies, in order to maximise their contribution to sustainable energy future.
8
World Energy Council
4.
Performance of Generating Plant 2004 - Section 5
Review of Technical Performance Indicators Proposed in Phase 1
This section reviews the results of the first phase of the project, in which technical performance indicators were proposed. For details, see the WEC PGP report 2001(2). The following table summarises these main indicators.
Technical Performance Indicator
Plant size (MW) Expected life time Capacity Factor (%) Load Factor (%) Specific Energy Production (kWh/m2) Equivalent Full Load Hours (hours) Availability Factor (%) Reference Yield (hours/day) Array Yield (hours/day) Final Yield (hours/day) Performance Ratio (%) Efficiency (higher heat value) Fuel Moisture Annual energy sent out (MWh/year)
Ease of measurement / Comparability estimate between sites
Type of RES All All Wind, Biomass, Geothermal Geothermal Wind
Easy Rather difficult Easy
OK OK OK
Easy Easy
OK OK
All
Easy
OK
Wind, Biomass, Geothermal PV
Easy
OK
Rather difficult
OK
PV
Rather difficult
OK
PV PV Biomass
Rather difficult Rather difficult Easy
OK OK More or less
Biomass Biomass
Easy Easy
More or less More or less
A few details (definitions and sample values) are reviewed and repeated in the following for each RES (except for plant size and expected lifetime, general indicators easily understandable), in order to present a full list of proposed indicators in the present report. General comments: • As for environmental and sociological indicators proposed in the future chapters, these first definitions will have to be very precisely defined in a future phase, so that they can be very clearly understood and used, and be “officially” considered as real standards. • Along with these indicators, general information on the RES plants will have to be provided in order to be able to compare and make useful statistics in the future. In particular, the following data will be necessary when introduced in the future data bases:
9
World Energy Council
-
Performance of Generating Plant 2004 - Section 5
Age; Vintage; Type of technology (type of PV cells, of wind turbine wings, of biomass fuel, of geothermal technique); Cycle of operation.
10
World Energy Council
4.1
Performance of Generating Plant 2004 - Section 5
Wind Energy
The performance of wind farms is very much dependant not only on the wind turbines themselves (rated power, rotor diameter and hub height), but also, and mainly, on the wind conditions on the site. Therefore, along with the following performance indicators, data on the wind regime must also be provided in order to be able to compare the performances of various wind farms. We can mention that for wind energy, statistics already begin to exist in various countries, especially in Europe with the “Wind Stats” monthly publication (5). The following technical performance indicators have been proposed for wind energy: 4.1.1 Definitions of the Proposed Performance Indicators Electricity Production • • • •
Total energy production [kWh] delivered to the grid at the connecting point during the monitoring period, usually one year; Specific energy production [kWh/m2]. This parameter is strongly dependent on the site. For example, an average value for three blades of a 600 kW wind turbine in Italy is ∼0,87 [MWh/y/m2]; Equivalent full load hours i.e. the annual energy production in relation to the rated power of the turbine, in hours; Capacity factor1 i.e. the ratio of the total actual energy production during one year over the theoretical, potential energy production (rated power x 8760 hours), dimensionless.
Note: 1 “capability factor” is also used by some experts in place of “capacity factor” Technical Availability This term seems to be more effective than “absolute availability” as no wind measurements are necessary. The following definitions are based on the ‘definition of Terms in the Energy Industry’ by VDEW (Germany). • • • •
'Nominal period' 1 is the complete period covered by the report, usually one year; 'Period of non-availability' 2 is the period during which a plant is not functioning. This can be scheduled (maintenance) or unscheduled (malfunction, failure, network problem); 'Technical availability' 3 is the period of availability over the nominal period, in percentage form; 'Average technical non-availability' divides the total period of non-availability by the number of considered turbines.
Note: the following equivalent terms are used by other experts: 1 “relevant time period” in place of 'Nominal period' 2 “Unavailability Factor” (UAF) in place of 'Period of non-availability' 3 “ Availability Factor” (AF = 1-UAF) in place of 'Technical availability'
11
World Energy Council
Performance of Generating Plant 2004 - Section 5
4.1.2 Examples of Values of the Performance Indicators The following examples are given by ISET (Germany) – see report phase 1 for additional details -
1,2
1,2
1
1 power output [P/Pr]
power output [P/Pr]
Power Curve [ P=f(v): In order to make power output of different turbine models comparable, their performance should be measured and documented under uniform international standards (IEC) by accredited laboratories.
0,8 0,6 0,4 0,2
0,8 0,6 0,4 0,2
0
0
0
5
10
15
20
25
0
5
10
15
20
25
wind speed [m/s]
wind speed [m/s]
a) b) Fig. 1: Measured Power Curve of Turbine with Pitch Control and Variable Speed (a) and Stall Control and Constant Speed (b). Note: these curves are usually only related to “standard conditions” and not applicable with precision to the various installations on site Wind Regime at Given Locations Wind measurements should be carried out with a minimum sampling rate of 1 Hz (BINmethod) and documented with regard to international standards (IEC) using the following parameters: • • • •
Mean wind speed; Weibull distribution; Directional distribution of wind energy; Turbulence intensity (ratio between standard deviation and mean value).
12
World Energy Council
Performance of Generating Plant 2004 - Section 5
Weibull parameter: A = 8,2 m/s ; k = 2,2
measured in 30 m height
25% rel. frequency
frequency
20%
Weibull distribution
15%
10%
5%
0% 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
wind speed in m/s
Fig. 2: Wind Speed – Frequency and Weibull Distribution
0 330
345
15
30
315
45
300
60
285
75
270
90
255
105
240
120
225
135 210
195
165
150
180
Fig. 3: Distribution of Wind Energy Based on Energetically Weighted Frequency Distribution of the Wind Direction in 12 Sectors Note: It can also be useful to use two independent parameters : i) frequency of the speed in the various sectors; ii) average speed for each sector, instead of only one parameter which combines the two effects Expected Value of Annual Electricity Production In terms of technical availability, modern plants regularly achieve, on average, values between 98 and 99 percent. However, turbine models, which are in the introductory phase, in the first operational years, achieve values of 95%. As an example, some results from Germany: the coastal locations provide on average approximately 2,200 –2,400 full-load hours, whereas inland sites achieve an average of approximately 1,300-1,500 full-load hours. Figure 4 presents gross energy supply for the years 1993-1998. This period is meteorologically too short for deriving reliable long-term trends. Nevertheless, a significant result obtained from these figures is that the gross available wind energy of a calendar year can fluctuate, by up to 15 percent, from the long-term average value.
13
20
World Energy Council
250
Performance of Generating Plant 2004 - Section 5
W /m ² c o a stlin e , isla n ds (1 9 9 3 -'9 9 : 17 6 W /m ² ) Hig h la n d s ( 1 9 9 3 -'99 : 1 0 6 W /m ² ) No rth e rn lo wla n d s ( 1 9 9 3- '9 9 : 8 7 W /m ²)
209
No rth e rn lo wla n d s, wo o de d (1 9 9 3 - '9 9 : 6 4 W /m ² )
200 183
179
173 182 159
14 7
150 125 119
113
111 102
107
92
100
87
84
98
76
81
84
69
74
71
70
70
50
61 52
50
1995
1996
0 1993
1994
1997
1998
1999
Fig. 4: Gross Wind Energy Supply by Years (Germany 1993-1999) In order to ascertain the statistical expectation values for the wind power production in individual months, the energy delivery data from the plants recorded in Germany is used and the “capacity factor” is presented in Fig. 5. The actual monthly energy delivery achieved by all observed plants is thereby related to the theoretically maximum total monthly production with continuous full load, whereby 125,000 monthly energy delivery reports from the years 1990-1998 were available. The average values of the individual months (and therefore the expectation values) range from 15% in July, to 32% in January. According to the general weather conditions and the annual available wind, fluctuations can be noted which range from the minimum value in August 1997 with 7% capacity factor, to a maximum value of 45% in January 1993. 60
%
50
Capacity Factor
M axim um Values 1990-99
40
30 M ean Values 1990-99
20
10
M inim um Values 1990-99
0 Jan
Feb
M ar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Fig. 5: Monthly Capacity Factor of the Wind Turbines in the Scientific Measurement and Evaluation Program (WMEP) of ISET, Germany, 1990-1999
14
World Energy Council
Performance of Generating Plant 2004 - Section 5
4.1.3 Additional Possible Indicators Related to “Weather Variability Control” The performances of wind and PV energies largely depend on the weather, and an accurate forecast of the wind and sun conditions are of prime importance for an efficient and economically viable use of these energies. The idea is then to try to define indicators related to the forecast of the weather conditions. Two kinds of indicators can then be defined: one related to the estimate of the electricity production, the other one to the prediction of the money value (one day, one week, one month or one year in advance). These indicators are still first ideas and have to be further investigated by specialists, but they could be the following:
Effective kWh (or money) produced during day d (week w, month m, year y)
Fd (%) =
_____________________________________________________________________________________________
(Fw, Fm, Fy) before
kWh (or money) forecasted one day (one week, one month, one year)
15
World Energy Council
4.2
Performance of Generating Plant 2004 - Section 5
Solar PV Energy
As for wind energy, the performances of photovoltaic energy is of course very much dependant on the weather (sun) conditions, and therefore on the site and region where the plant is installed. For example, if northern Europe has an average of only 700 hours per year of “usable” sun, southern Europe can count on at least twice this figure. Apart from the indicator proposed in chapter 4.1.3 above, related to the production forecast, the following technical performance indicators have been proposed for solar / PV energy: 4.2.1 Definitions of the Proposed Performance Indicators: •
Reference yield, YR The Reference yield YR is the daily (monthly or annual) in-plane irradiation ES,A divided by the STC reference in-plane irradiance GSTC ( = 1 kW/m²) YR =
ES, A [kWh/(m 2 ⋅ d)] GSTC [kW/m 2 ]
It has the value h/d and can be considered as the number of hours per day during which the solar radiation would be at reference irradiance level, in order to contribute the same energy incident as was monitored. •
Array yield, YA The array yield YA is the daily (monthly or annual) array energy output EA,d per kWp of installed PV array power Po: EA, d [kWh/d] YA = Po [kWp] It has the dimension kWh/(d · kWp) and can be considered as the number of hours of array operation per day at Po, which would give the same energy output as the recorded integral value for that day (month or year).
•
Final yield, Yf The final yield Yf is the daily (monthly or annual) plant useful energy output Euse per kWp of installed PV array power Po: Yf =
Euse [kWh/d] Po [kWp]
It has the dimension kWh/(d · kWp) and can also be considered as the number of hours of plant operation per day at Po, which would give the same energy output as the recorded integral value for that day (month or year). •
Performance Ratio, PR
16
World Energy Council
Performance of Generating Plant 2004 - Section 5
The Performance Ratio, PR indicates the overall effect of losses on the array’s rated output due to array temperature, incomplete utilisation of the irradiation, and system component inefficiencies or failures. It is given by the ratio: PR =
Yf YR
Note: the performance ratio can also be defined as [ PV module efficiency x inverter efficiency ] in order to make a fair comparison. Remark: all these performance indicators are difficult to measure (they depend in particular on the position of the sun) and require specific measurement instruments.
4.2.2 Examples of Values of the Performance Indicators a)
The following tables give typical values of the proposed indicators, obtained from a sample of various plants (central and roof top) in Italy :
Serre PV Central Station (3.3 MWp)
1995
1996
1997
1998
1999
Reference Yield, YR [kWh/m²] [h/d]
1674 4.58
1641 4.50
1784 4.89
1704 4.67
1723 4.72
Performance Ratio, PR [%]
48.0
70.4
68.7
68.7
68.6
b)
PV roof top plant
Bovisa (3 kWp) 1999
Cesi (3 kWp) 1999 Cagliari (3 kWp) 1999
Reference Yield, YR [h/d]
3.75
3.32
4.03
Final Yield, Yf [h/d]
2.44
2.35
3.15
Performance Ratio, PR [%]
65.0
70.9
78.2
The proposed PV performance indicators have also been applied to the Saijyo Project in Japan, which is the first and biggest project, with 1000 kW output, which started operation in 1980 and ended in 1998 (T. Kaneda, 2001):
17
World Energy Council
Performance of Generating Plant 2004 - Section 5
Year
2000
Reference yield YR h/d Array Yield YA h/d Final Yield Yf h/d Performance Ratio PR %
3.84 3.56 3.32 86.00
18
World Energy Council
4.3
Performance of Generating Plant 2004 - Section 5
Biomass Energy
Biomass can be divided into four sub-categories, which correspond to the definitions used by the International Energy Agency (IEA) (2): • • • •
Wood, logging and agricultural residue, animal dung; Solid industrial waste; Solid municipal waste; Biogas.
In order to make fair comparisons between biomass plants, one should first specify to which biomass category the plant belongs. The following table presents the performance indicators proposed for biomass energy, with their definitions and examples of typical values obtained in the USA (68 US power plants giving typical, low and high values – for more details about the categories of the corresponding US biomass plants, please refer to the 2001 Report).
19
World Energy Council
Technical Indicator (1)
Performance of Generating Plant 2004 - Section 5
Units
1
Size
2
Generation (based on # 1, 3, 4)
3
Capacity (2) (annual)
4
Hours per year equivalent of the capacity factor (if run (3) at full power)
5
Efficiency (higher heat value)
6
Heat Rate (SI units)
7
HHV Higher heating value (at moisture & (4) ash free)
MJ/kg
8
Fuel ash content on dry basis
%
9
Dry Higher heating value (at ash given, but no moisture)
MJ/kg
10
Fuel moisture (5)
11
Wet Higher heating value (at moisture given, typical 2% ash)
MWe GWh/year
factor
hours
MJ/kWh
Definition
Plant size in net power output
Typical Value
Low Value
High Value
25.00
8.00
65.00
153.30
35.04
455.52
0.700
0.500
0.800
6132.00
4380.00
7008.00
0.201
0.155
0.263
18.00
13.50
23.00
Heat content (HHV basis) if there were no moisture and no ash
19.70
18.60
20.90
Fraction of the total fuel input by weight (dry, not as-received, basis) that is ash
2%
1%
10%
Total electricity (net) generation per year Total net electricity generation per year divided by plant size in net MW times the 8760 hrs in a year Number of hours that would give the net annual generation if all operation were at full net size of plant Net output in heat unit equivalent of the electricity divided by the fuel heat input using the higher heating value of the fuel Fuel heat input per net electricity output
Heat content (HHV basis) if 19.3 17.7 19.5 there were no moisture, at ash (2% ash) (10% sh) (1% ash) given
%
Moisture (H2O) as a fraction of the total weight of fuel (wet, not dry basis)
MJ/kg
HHV at typical 2% ash content, at moisture given
30%
10%
60%
13.5 7.7 17.3 (30% (60% (10% moisture) moisture) moisture)
Some other more «economical» indicators can also be proposed for US biomass power plants : 12
Capital cost
US$/kW
Total cost to build the power plant divided by the net output in kW
1 400.00
1 000.00
2 000.00
13
Fuel cost
US$/GJ
Cost of fuel as received per unit weight divided by heat content of fuel (HHV basis) as received per unit weight
1.42
0.95
2.37
14
Fraction of capital cost to maintain (repair) plant each year
%/year
Annual cost to run the plant, but not including fuel cost and not including employee cost, divided by the cost to build the plant
4.00
3.00
5.00
15
Number of employees
# per 20 MWe
Number of full-time operating and supervising staff members per 20 MW of power plant net size
20.00
16.00
30.00
20
World Energy Council
Performance of Generating Plant 2004 - Section 5
The following comments clarify the values of the indicators obtained from the US sample mentioned in the table above: (1) Size : in general, biomass plants tend to be small rather than large. They can be as low as a few hundred kWe, but the more realistic ranges from a few MW to 25 or 40 MW. 65 MW is rather large. (2) Capacity factor: this can reach 90%. (3) Hours per year equivalent of the capacity factor (if run at full power): it can reach 7800 hours/year. (4) LHV (Lower Heating Value, or lower calorific value) is also used, especially in gasification. Usually LHV is 94% of HHV for biomass, so efficiency on LHV basis is 6% higher than on HHV basis. (5) Fuel moisture: general typical values range from 5-10% to 50%. Anything over 50% moisture creates too many problems. Bagasse for example has traditionally been used at 50% moisture content. (6) Availability and Unavailability Factors could also be proposed as performance indicators, similar to what has been done for « classical » fossil fired power plants.
21
World Energy Council
4.4
Performance of Generating Plant 2004 - Section 5
Geothermal Energy
The following technical performance indicators have been proposed for geothermal energy: 4.4.1 Definitions of the Proposed Performance Indicators Capacity Factor =
Load Factor =
Total MWh generated in period x 100 Installed Capacity (MWe) x period (hours) Total MWh generated in period x 100 Maximum Load (MWe) x period (hours)
Availability Factor = Total hours of operation of plant during the period x 100 Total length of period (hours) The unavailability (%) of the plant (100-availability factor) is split into two categories: • •
Planned outage - An outage scheduled well in advance (at least two weeks) of the actual outage; Forced outage - Unplanned outage that requires the plant to be taken out of service immediately or before the next planned outage.
Both the Capacity and the Load Factor are needed to describe the technical performance of the plant. Other indicators which describe partial performance and operational conditions of the plant are listed in the 2001 report (2).
22
World Energy Council
4.4.2
Performance of Generating Plant 2004 - Section 5
Examples from Actual Geothermal Plants
The following table shows data collected from typical power plants: one 60 MW and 20 MW standard Italian plants, and one 50 MW Japanese plant. Italian Plant #1 (1999)
Italian Plant #2 (1999)
Japanese Plant 1/4/97-31/3/98
Installed Capacity
60 MW
20 MW
50 MW
Maximum Load
55 MW
17 MW
48.3 MW
Annual Produced Electricity
462,8 MWh
142,2 MWh
361,7 MWh
Hours of Operation of the Plant
8748 hrs
8483 hrs
8112 hrs
Capacity Factor
88.1 %
81.2 %
82.6 %
Load Factor
96.1 %
95.5 %
85.5 %
Availability Factor
99.9 %
96.8 %
92.6 %
The average values for the same reference periods for some tens of Italian and Japanese plants are shown in the following table: Factors
Japan
Italy
Capacity Factor
75.6 %
75.1 %
Load Factor
84.2 %
89.3 %
Availability Factor
92.1 %
92.1 %
23
World Energy Council
Performance of Generating Plant 2004 - Section 5
24
World Energy Council
5.
Performance of Generating Plant 2004 - Section 5
Proposal of Environmental Indicators
The environmental indicators proposed below can be divided into two categories: • General indicators applicable to all kinds of RES generating plants; • Specific indicators, which are only applicable to specific types of RES. For each indicator the following is included: • Its definition; • Examples of values found in the literature or provided by experts in the domain, to give an order of magnitude and focus on narrower range of values; • Comments.
5.1
General Environmental Indicators
Four general environmental indicators are proposed. Three of them can be rather easily measured and considered as “standard” indicators: • • •
Contribution to the reduction of greenhouse gas emissions (tonnes CO2/MW/y; Pollutant emissions during the life cycle (g / kWh); Land area required for production of 1 GW (km2).
The fourth one “External environmental cost” is more difficult to estimate and it can create controversy. It should be evaluated in all cases where it is possible. 5.1.1 Contribution to the Reduction of Greenhouse Gas Emissions (t/MW/y) Definition: AvCO2 (t/MW/y) = Avoided CO2 emissions (in metric tonnes per MW per year), compared to what would have been emitted by a new plant built in the region, given the same annual production (in kWh), using as fuel the most likely future fuel choice, or by the plant most likely to be displaced by the new RES facility (usually the oldest plant scheduled for retirement). Sample values: Wind *Wind Power Denmark: avoided CO2 = 2000 t/MW/year
*US AWEA: 750kW wind turbine avoids 1500t CO2/year
Solar / PV
Biomass
Geothermal
Must take into account CO2 content of electricity used for manufacture of cells (manufacture nearly needs as much energy as will be produced during 2 to 3 years – Lifetime of a cell ~ 20 to 25 years ). The amount of CO2 will depend on the fuel mix of the country where the PV cells are manufactured.
When managed in a sustainable cycle (energy crops, replanting harvested areas, etc.), biopower generation can be viewed as a way to recycle carbon, and can be considered a carbon – neutral power generation option.
ORMAT: 700MW geothermal avoids 13 million tonnes of CO2 in 20 years (i.e. 930t/MW/year)
25
World Energy Council
Performance of Generating Plant 2004 - Section 5
Total European wind In Paris, France, 106 m2 of roof-PV cells avoid emissions farms (16300 MW) of 22t CO2 per year avoid 30 million tonnes CO2/year (source ADEME France)
Roughly speaking, 1 ton of wood avoids 1 ton of fossil fuel CO2 emission (see comments below)
Italy: in 2010, 2500MW wind should avoid 2MtCO2 (i.e. ~800t/MW/yr) (based on emissions of modern CCGT).
A 60 years old US oak and maple forest stocks 2 tonnes C / ha / year
New wind capacity 5.5 TWh.
Comments: • The reference technology displaced must be quoted (coal, oil, gas). The avoided CO2 must be compared with either the displaced power (plants that are retired as new RES plant comes on-line) or plants that would have been constructed to meet future demand. First of all, the values of the “avoided CO2” should be related to a “reference-technology”, for instance oil fired steam generating unit or the CCGT. The fuel type and the efficiency of the units can be used as “benchmarks” to make the comparisons meaningful and comparable among various sources. •
•
Obviously the indicator of avoided CO2 can sometimes be difficult to use for comparison. In France for example, where about 75% of electricity is produced by nuclear plants, there is practically no avoided CO2 where compared to nuclear generation. As shown in this case, the avoided CO2 emissions are not a performance indicator in itself, but depends on the country context. The same applies in countries where hydropower is dominant; For wind, the above definition probably needs certain refinement because the CO2 abatement depends on the wind conditions and hence on the wind plant factor over the year. The best way to express the CO2 abatement is to put it in terms of tonnes of CO2 avoided per MWh (or per MW/y) of generated energy in a given country;
•
For PV, there can be a big difference between developed and developing countries: the electricity needed to manufacture the cells is consumed in the country of manufacture whereas the electricity produced by the cells is often provided to populations which would not have access to electricity at all. It is essentially a transfer of power and of benefit to poor populations;
•
For biomass, when compared to coal combustion in a pulverised coal power plant at 10,000 Btu/kWh heat rate (10.5 MJ/kWh or 34% efficiency at higher calorific value heat rate), avoided emissions of carbon from the fossil fuel used for planting, fertilising, harvesting and transporting biomass are estimated to be in the range from 3% to 10%. 1 tonne wood for 1 tonne CO2 (not C) emissions is about right, assuming that wood and fossil fuel have the same heat value, which is not the case. The avoided emissions should be estimated on the basis of the same amount of heat released in the boiler, assuming that all other factors are the same;
•
This indicator can in some cases be dependent on a required power guarantee (especially for wind farms);
26
World Energy Council
•
Performance of Generating Plant 2004 - Section 5
The specific costs of avoiding CO2 emissions could eventually be a complementary indicator. The problem is that it would be rather difficult to make regional comparisons. Moreover, it would not allow to compare, for one type of RES (wind, PV etc) the different existing technologies. Anyway such costs can be easily calculated, in a given region, from the production cost of the RES plant multiplied by the quantity of CO2 that would have been emitted by the displaced technology (coal, oil, gas etc).
5.1.2 Pollutant Emissions During the Life Cycle (g / kWh): Definition:
QCO2, QSOx, QNOx (g/kWh) = Quantities of CO2, SOx and NOx emitted per kWh during the life cycle of the plant Sample values: Type of pollutant
Wind
Solar / PV
Biomass
Geothermal
(energy crops current practice) CO2 ( g/kWh)
7–9
98 - 167
17 - 27
79 (source IEA) 122 (source IGA)
- Newest flash for steam plants: 1 lb CO2/MWhe (source ORMAT) SO2 ( g/kWh)
0.02 – 0.09
0.2 – 0.34
0.07 – 0.16
0.02
NOx ( g/kWh)
0.02 – 0.06
0.18 – 0.30
1.1 – 2.5
0.28
Source: IEA, Benign Energy?, The Environmental Implication of Renewables, OECD Paris, 1998
Comments: • For wind and PV, these emissions can be considered as close to zero during the production phase; the emissions mentioned in the above table are mainly due to the construction and decommissioning phases (PV modules in particular); •
The same remark applies to SOx and NOx in geothermal energy: SOx and NOx are not present in the geothermal fluid; only minor quantities are released by the diesel engines during the drilling activity, but this is insignificant and not relevant during the production life of the plant;
•
As for CO2 emissions in geothermal plants, it should be noticed that CO2 is NOT produced by the human activities, but by NATURAL processes, by deep chemical reactions. It is naturally released to the atmosphere in all geothermal/volcanic areas. The natural CO2 soil degassing from a standard geothermal area is of the same order of magnitude as of the gas emitted from a geothermal power plant. Geothermal energy
27
World Energy Council
Performance of Generating Plant 2004 - Section 5
does not create new CO2 molecules, but simply concentrates at the chimney the natural emission from deep underground layers; •
In non-quantitative terms biomass can be considered better than coal, but not better than natural gas, except for the greenhouse gas CO2, where biomass is better than both natural gas and coal. The biomass emissions can be estimated at about 1/30 - 1/10 of the coal emissions; or compared to natural gas, this becomes 1/10 -1/3 because coal emits about 0.25 tonne-C/MWh and natural gas about 0.10 tonne-C/MWh in gas turbine-combined cycle;
•
For biomass, specific environmental problems (emissions, but also possibly noise) can arise during the production of the plant, due to the daily transportation of the fuel (wood, logging, industrial, municipal or agricultural wastes) from the source to the plant;
•
Additional emissions like particulates and heavy metals could have been added as general environmental indicators, but it was decided to exclude them at this stage (except, partly, for biomass – see 5.2.3) because they are more difficult to measure. Particulate matter (or dust) is regularly measured at power plants and also, certain heavy metals are attracting increased attention, especially in the US, where the EPA is expected to announce soon measures for the reduction of these emissions. They should not be omitted.
5.1.3 Land Area Required for Production of 1 GW (km2): Definition: A (km2) = Necessary area for exploitation of a plant producing 1 GW of electric power . Sample values: Land area for a
Wind
Solar / PV
Biomass
Geothermal
100 km2
30 km2
5 000 km2
200 km2
1 GW plant
Sources : World Energy Council and US DOE (IGA for geothermal energy)
Comments: • For a wind farm, the area can also be used for agricultural purposes (or other uses), which can be an advantage . It is then necessary to specify the corresponding land use; •
For a wind farm, the “strict” area required for the installation of the turbines is of about 1 km2 for 1 GW. But a very large upwind area (land in front of the wind turbine) and wake area (land downwind of the turbine) is required to ensure the smooth flow of air. This land is virtually “sterilised” and has no other use and therefore should be counted in if it cannot be used for other purposes (like agricultural purposes for example – see above);
•
For biomass, the area very much depends on the type of biomass (agricultural, industrial waste, etc). EPRI (USA) estimates 800 acres per MW or 0.8 million acres per GW which is about 3200 ha / GW. This is based on 5 dry short tonnes per acre per year and 10,000 Btu/kWh (10 dry metric tonnes per ha per year and 10.5 MJ/kWh higher calorific value basis) and capacity factor about 75%;
28
World Energy Council
•
Performance of Generating Plant 2004 - Section 5
For geothermal energy, two plants can generate about 500 MW by being fed from the same surface (a site of 10 x 10 km of land) with two reservoirs at different depths (like in Larderello, where a shallow and a deep reservoir are being used from the same geothermal area). Another water-dominated reservoir (Berlin, El Salvador) extends over a land area of about 4 km2 for 50 MW. The proposed value is 200 km2 per 1 GW, even if the area will not be totally occupied by wells, piping and services road: it is fully compatible with other human or natural activities. The effectively occupied area from well-pads, pipelines, plants and other auxiliary equipments could be 1/101/20 of the quoted number (source IGA).
29
World Energy Council
5.1.4
Performance of Generating Plant 2004 - Section 5
External (environmental) cost (%):
Definition:
Ext (%) =
External unit cost _______________________________
( per MWh )
Unit production cost
- External cost = cost due to environmental and health impacts, not included in the producer financial cost ( cf. ExternE European Program) - Production cost includes investment amortisation, exploitation, fuel and R&D
Sample values: Type of RES External cost per MWh*
Average production cost per MWh ** Corresponding Ext (%)
* ** ***
Wind Solar / PV 0.5 to 2.6 € PV connected to the grid ~ 1.4 to 95 3.3 € 95 (only one study) ~ 45 €
~ 650 €
1.1 to 5.8 %
0.2 to 0.5 %
Biomass Geothermal ~ 2.0 to 50 € 95 0.2 – 0.5 € *** (highly dependant on the type of biomass). ~ 128 € 30 – 100 € *** 1.6 to 39 %
0.2 to 16%
Source : ExternE European project Source : Eurostat Source: International Geothermal Association (IGA)
Comments: • For general information, the following table gives the average investment, production and external costs of the various types of electrical power plants:
30
World Energy Council
Type of plant
Performance of Generating Plant 2004 - Section 5
* Average investment cost ( € / kW)
* Production cost (including exploitation, investment amortization, fuel and R&D)
** External costs ( €95 / MWh) (bracket depending on the countries)
% external costs / production costs
• 0.2 – 17 % (0% discount rate) • 8 – 65 % (3% discount rate) 58 – 300 %
( € / MWh)
Nuclear
1730
29
Coal (CFB)
1273
33
• 0.05 – 4.8 (0% discount rate) • 2.3 – 18.8 (3% discount rate) 19 – 99
Gas (CCGT)
488
35
7 - 31
20 – 89 %
40
0.04 – 6.03
0.1 – 15 %
0.5 – 2.6
1.1 – 5.8 %
Hydro large scale Hydro small scale
1250
Wind (on shore)
1000
45
8000
650
1200
128
1.4 – 3.3 (only one study) 2.0 – 50
2400/1800/1400***
55/45/37 ***
0.2 – 0.5 ***
Solar PV connected Biomass Geothermal
grid-
0.2 – 0.5 % 1.6 – 39 % 4 – 14 % ***
* **
Source : Eurostat (Les Echos Group), March 2003 Source : European Commission ExternE 1999
***
Source : IGA (International Geothermal Association) for the three standard unit size 15/30/55 MW; 0.2 – 0.5 €/MWh external cost includes H2S abatement system and silencers for noise reduction
•
As a matter of fact, this indicator is essentially proposed for future consideration, as it is, on one hand difficult to estimate (and can be subject to misinterpretations, confusion and controversies), and on the other hand difficult to apply from one plant to another, as it heavily depends on the individual country circumstances (where production and investment costs can be very different) and on the various subsidies given to the RES in various countries. It has been kept at this stage because it reflects the fact that the real cost of the electricity must take into account all costs, and in particular the impact on environment and human health, which is one of the main reasons why RES can have large advantages compared with the conventional plants;
•
Both the cost and the percentage should be given, although both can sometimes be confusing. The cost gives an idea of the additional cost due to environmental and health effects, and the percentage shows the part of these effects on the overall cost;
31
World Energy Council
•
Performance of Generating Plant 2004 - Section 5
For biomass, EPRI (USA) usually does not include R&D in the USUS$/MWh total cost; EPRI quotes a cost range for biomass from US$20/MWh for co-firing with coal to US$90/MWh from current technology “stoker boiler of 50 MWe size”, but US$125/MWh can be the cost if fuel is expensive.
5.1.5 Other Possible General Indicators: Further general environmental indicators could be possibly considered in the future: •
Duration of operation to produce the energy consumed for its manufacturing (particularly adapted for PV modules); Time of return / Reversibility / man-hours to return to original state; Emissions from back-up power.
• •
It was decided not to go further in the elaboration of these indicators at this time, because they are generally difficult to estimate, heavily dependant on the site, and also difficult to use to make comparisons between plants.
5.2
Specific Environmental Indicators for Wind, PV, Biomass and Geothermal Energies
5.2.1 Specific Environmental Indicators for Wind Energy Four specific environmental indicators are proposed for wind farms: • • • •
Visual effects / Landscape protection distance; Noise from wind turbines; Bird fatalities; Shadow casting.
A. Visual effects / Landscape Protection Distance (m): Definition
Comments and sample values
dmin (m) = Minimum distance away from nearby dwellings
France recommends a minimum distance of 500 m
It is difficult to define a standard indicator, as the visual effect depends very much on the type of landscape, and is to a large extent subjective. Most often a case by case approach is followed. B. Noise from Wind Turbines (dB): Definition
Comments and sample values
• S f = Maximum noise (dB) at the foot of the wind turbines • S 500 = Maximum noise (dB) 500 m away from the wind turbines • S st = Maximum noise (dB) at standard distance H + D/2 according to norm IEC 61400-11
32
- EDF recommends S 500 lower than 35 dB - Nordex 2.5 MW = 102.8 dB(A) at hub for a 8m/s wind - For most turbines, S f < 55 dB - WT 2 MW : S 500 = 39.7 dB and S 250 = 48.2 dB (source Systèmes Solaires)
World Energy Council
Performance of Generating Plant 2004 - Section 5
Note: The important point is the differential increase in respect to existing noises. In many countries, especially in Europe, specific laws impose maximum increase of 5dB(A) during the day and of 3dB(A) during the night. C. Bird Fatalities (n/y): Definition
Sample values
Nb = Number of birds killed per wind turbine per year
• • • •
0.34 Tarifa Spain 1.34 Blyth Mouth UK 0.4 to 1.3 average Europe 5.2 average California
Comments: • The number of birds killed depends primarily on different site conditions; • This number depends on the measures taken to minimise the fatalities, namely, colour of poles, bird repelling measures, etc. D. Shadow casting (h/y): Definition
Comments and sample values
Nsc = Number of hours per year when the proximate dwellings suffer of shadow casting from the wind turbines
France : maximum 5 to 12 hours per year (source ADEME)
E. Other Possible Environmental Indicators: Additional environmental indicators could be proposed such as interference with radio, TV, and microwave transmission. Also, specifically for offshore wind farms, the effects on navigation, fishermen, aquatic fauna & flora could be proposed. As they strongly depend on the specific site circumstances, and seem difficult to quantify, they have not been further investigated at this stage of the study. 5.2.2 Specific Environmental Indicators for Photovoltaic Energy Quantities of toxic materials in cells and batteries (g/Wp): Qtox (g/Wp) = Quantities of toxic materials contained in the cells and batteries, that will have to be recycled or disposed of after the lifetime of the cells and batteries (Cd, etc.)
As already stated (5.1), the duration of operation to produce the energy consumed for the manufacturing of the PV modules could be added. It has not been proposed at this stage because, on one hand the PV cells very often are not used in the country where they have been manufactured, and on the other hand the operation of the PV plant can be very
33
World Energy Council
Performance of Generating Plant 2004 - Section 5
different from one place to another, and therefore very difficult to compare. This could be further investigated and proposed in a future phase. 5.2.3 Specific Environmental Indicators for Biomass Energy Emissions during the life cycle (g / kWh):
Two specific environmental indicators are proposed for biomass: Q ash (g/kWh) = Quantities of ash emitted, with their composition (Se, Pb, As, B …?) cf. in particular agricultural residues, wood wastes, animal wastes, energy crops. QCH4 (g/kWh) = Quantities of CH4 emitted from landfills (decomposition of biomass material) or decomposing animal manure (land-applied or left uncovered in a lagoon) Comments:
•
Bio-mass is a large and complex subject. Environmental and other impacts need to be carefully examined case by case; it is difficult to establish a few general criteria. For example, wood-processing wastes could be different from lumberyards’ waste, agricultural waste, forest detritus, or energy farm trees. Impacts could be different in wetlands, desert or arid areas, forest lands, prairies, etc. The type of biomass and conditions of use must be described carefully in detail to make useful comparisons;
•
Apart from key benefits such as low GHG (greenhouse) emissions, biomass has other positive environmental effects for which there are no easy standard environmental indicators: elimination of wastes and/or odours, preservation of the landscape, soil conservation, agricultural reconversion, eventual positive effect on biodiversity, etc. These could be considered further in a future phase.
5.2.4
Specific Environmental Indicators for Geothermal Energy
H2S emissions during the life cycle of the plant (g / kWh) Definition
Comments and sample values *
QH2S (g/kWh) = Emissions of H2S during the life of the plant, in g per kWh
- 1 g H2S/kWh - H2S is 1% in volume of the CO2 released: this is a rough world-wide average
* Source: International Geothermal Association (IGA)
34
World Energy Council
Performance of Generating Plant 2004 - Section 5
6.
Proposal of Sociological Indicators
6.1
Jobs Created by the Plant (n / MW)
Definition: Nj = Number of jobs (direct / indirect) created by a 1 MW power plant for the different steps : manufacture, installation, operation and maintenance Sample values: Wind: • A European study of 1999 assumes that 17 job–years of employment are created for every MW of wind energy capacity manufactured, and a further 5 job – years for the installation of every MW, bringing the total to 22 job – years; • WindPower Denmark : 50,000 jobs created in the wind industry by the end of 2001 throughout the world; • The latest update of Wind Force 12 (EWEA, 2003a) suggests that the feasible number of jobs created in the wind industry worldwide by 2020 will be 1.8 million; • In Canada, 10,000MW in 2010 would create 80,000 to 160,000 jobs; • In UK (source DTI), RES should have created between 17,000 and 35,000 jobs in 2020 (8,000 in 2003). Solar / PV: • Estimate by CEA-France 2003 : 20 jobs per produced MW, 30 jobs per consumed MW; • 6,400 jobs in 1997 (70% = Germany + USA + Japan); • SEIA (Solar Energy Industry Association – USA) 3,800 jobs created for every US$ 100 million of PV cell sales; • cf. SEIA & DOE 2001 20,000 employed in the PV industry in the United States. Biomass: • France 4,5 direct jobs created for 1,000 tep (tonnes equivalent petroleum) produced or distributed. • USA EPRI gives an average value of about 20 full-time operating and supervising staff members for a 20 MW biomass power plant. Altogether (including operation, maintenance, truck drivers, etc …), a total of 1.6 jobs per MW is estimated by EPRI in the USA. Comments: • Note that subsidies in some countries (for wind and PV in particular) can distort the data and make meaningful comparisons impossible. •
A distinction must be made between job-years for manufacturing and installation on the one hand, and jobs for O&M on the other hand (wind and PV in particular), especially when the manufacturer is not in the country of installation.
35
World Energy Council
•
Performance of Generating Plant 2004 - Section 5
The number of jobs is not always related to the number of MW. For example, for geothermal plants, there is no difference related to the unit size. For each standard Unit (15-30-55 MW) there is a direct O&M personnel of about 30. The indirect personnel is very difficult to estimate: taking into account the construction phase of each component, drilling of wells, building the plant, and the resource assessment researcher, we can easily account for 100/200 jobs related to each geothermal unit, even if for a limited time.
36
World Energy Council
6.2
Performance of Generating Plant 2004 - Section 5
Providing Access to Electricity
Definition: Na = Number of households / total number of people having access to electricity Produced by a 1 MW plant, and who would not have this access if another type of plant were to be built (grid connected). Sample values: Wind : • WindPower Denmark: 1MW wind energy provides electricity to 500 to 800 households in Europe; • Edens Italy: a wind plant rated 1MW can provide electricity for 1000 houses (without heating); • France: 35 MW Eole-RES = annual electricity consumption (without heating) of 45,000 people. Geothermal : • 30-40 million people now having access to geothermal electricity worldwide. Comments: • This study focuses only on grid-connected plants. In developing countries, isolated wind turbines or PV modules may bring power to people who may not, otherwise, have access to electricity at all without them.
6.3
Industrial Safety Accident Rate
Definition: SAR = number of accidents for all utility personnel permanently assigned to the plant (contractor personnel not included), that result in one or more days away from work (excluding the day of the accident) or one or more days of restricted work (excluding the day of the accident), or fatalities, per 1,000,000 man-hours worked. Sample values: Geothermal: • IGA data on a best-practice case study: from a geothermal plant utility company in the first eighth months of 2003 there were approximately 15 accidents per 1,000,000 man-hours worked. Comments: • This indicator is already widely used for other conventional types of power plants (nuclear, fossil-fired, etc.); • The purpose of this indicator is to monitor progress in improving industrial safety performance for all utility personnel permanently assigned to the utility’s staff;
37
World Energy Council
•
Performance of Generating Plant 2004 - Section 5
This indicator was chosen as the personnel safety indicator over other indicators, such as injury rate or severity rate, because the criteria are clearly defined, utilities currently collect this data, and the data are the least subjective.
38
World Energy Council
7.
Performance of Generating Plant 2004 - Section 5
Conclusion and Recommendations for the Future
It is hoped that the present research work will be useful for the development of Renewable Energy. It can be noted that nearly all performance indicators proposed in this report (technical, environmental and sociological) can be used for all types of power plants and not only RES. The next phase of this research work could consist in further validation, inter comparisons and benchmarking projects using these RES performance indicators. As a matter of fact, some of the proposed definitions probably need to be more precise and refer to existing norms and standards. Other additional performance indicators could also be considered, especially for wind and PV, such as the quality (losses) of the connection to the grid, and the efficiency due to the variability of the natural conditions (“availability when needed”). Moreover, the results of the work will become really useful when databases will be set up by volunteer organisations able to create and maintain them. This would allow continuous progress report, in particular in developing countries. Various installations could also be compared and contacts established with other organisations and similar plants, so that experience can be compared, eventual difficulties detected and, finally, performance improved. The following complementary tasks could be performed in the future, along with complementary benchmarking and setting up of databases: •
Definition of economic performance indicators : the main problem for many RES (in particular PV) is their high investment cost, which often restricts their development (a WEC Task Force on Renewables has been recently formed by WEC to focus on financial and economic aspects of RES). As far as economic and commercial aspects are concerned, let us mention a very interesting and innovative work done by ESKOM, South Africa, about what they have called “Commercial Availability”. This work, although basically performed for “classical” power plants (hydro in particular) could very usefully be applied to RES, especially wind and PV. This work is not described here (see reference (6) and short summary and example in annex 5), but the idea is to define indicators reflecting the relationship between supply and demand, and the opportunity to dispatch power depending on the related costs and revenues;
•
Apart from hydro power, already being examined within WEC PGP Work Group 5, the performance of other renewable energy technologies such as solar thermodynamic, heat pumps, fuel cells, biogas, etc., could also be investigated in the future.
39
World Energy Council
8.
Performance of Generating Plant 2004 - Section 5
General References Reports:
(1) - LEGERTON M L, ADAMANTIADES A G, ANCONA D - Exchange of availability / performance data and information on renewable energy plant : wind power plants - 17th WEC Congress, Houston Sept. 1998
(2) - MANOHA B., VIGOTTI R., BERTANI R., HUGUES E., HOPPE-KILLPER M. –
Performances of renewable energy generating plants – Phase 1 – 18th WEC Congress, Buenos Aires, Oct. 2001
(3) - La production d’électricité d’origine renouvelable dans le monde – Cinquième inventaire - Edition 2003 - Observ’ER - EDF - Systèmes Solaires Paris – Déc 2003
(4) - Survey of Energy Resources - World Energy Council – 1998 (5) - Wind Stats monthly publication – UK (6) - Micali V., Statham B., Moss T. – ESKOM – Commercial Availability & Availability Earnings Ratio, Proposing Paper, WEC - PGP Meeting, Rio de Janeiro, April 2000
General Web sites: World Energy Council: UNIPEDE / EURELECTRIC: U.S. Department of Energy / Energy Efficiency and Renewable Energy Network(EREN): International Energy Agency (IEA) / Renewable Energy Working Party: National Renewable Energy Laboratory (US): Systèmes Solaires / Observ’ER : CADDET Renewable Energy Newsletter : WREN - World Renewable Energy Network: Sustainable Energy and Development: Renewable Energy DataBase: Sustainable Energy and Development:
www.worldenergy.org www.eurelectric.org www.eren.doe.gov www.iea.org/techno www.nrel.gov www.systemes-solaires.com www.caddett-ee.org www.wrenuk.co.uk http://solstice.crest.org www.osti.gov/html/eren/eren.html www.eeca.govt.nz http://solstice.crest.org
Annex 2 provides additional web sites specifically for wind, solar, biomass and geothermal energies.
40
World Energy Council
Performance of Generating Plant 2004 - Section 5
ANNEX 1 Contributors to WEC PGP Work Group On RES Performances – Phase 2 Surname
First Name
WEC WG Members:
Position Organisation * = Sub-group leaders
Town ** =Chairman Pisa
Bertani *
Ruggero
Executive Director
Hoppe-Kilpper *
Martin
Head of Division Institut für Solare Kassel Information and Energieversorgungstechnik e.V. Energy (ISET) Economy
Hughes *
Evan
Manoha **
Bruno
Vigotti *
Roberto
Environment Managerial Adviser Renewable Energy Unit Manager
International Geothermal Association
Country
ITALY GERMANY
EPRI
Palo Alto
USA
Electricité de France, Division R&D
Chatou
France
ERGA of ENEL Business Development
Milan
ITALY
WEC members: Glorian
Daniel
Senior Advisor to the Director
EDF - Division Recherche et Développement,
Saint-Denis
FRANCE
Virkkala Nekhaev
Elena
Programmes Manager
World Energy Council
London
UK
Theis
Karl
Director
VGB Power Tech
Essen
GERMANY
Specialists who contributed to phase 2: Adamantiades
Achilles G.
Director of Engineering
Infrastructure Capital Group,LLC
Washington
USA
Beslin
Guy
Research engineer
Electricité de France
Chatou
FRANCE
Bronicki
Lucien Y.
Chairman
ORMAT Industries Ltd
Yavne
ISRAEL
Dal Pane
Enzo
Edens S.p.A. / Edison
Bologna
ITALY
Kaneda
Takeshi
Tokyo
JAPAN
Li
Albert
Project Manager Mitsubishi Research Institute Inc., Senior engineer China Light & Power International
Kowloon
HONG-KONG
Millborrow
David
Consultant
Wind Stats
Lewes
ENGLAND
Pineau
Dominique
Research engineer
Electricité de France
Chatou
FRANCE
Rosillo-Calle
Frank
Research Fellow King's College London
London
ENGLAND
Salvaderi
Luigi
Consultant on power systems
Rome
ITALY
41
World Energy Council
Performance of Generating Plant 2004 - Section 5
ANNEX 2 Renewable Energy in the World Electricity in the World: The following figures show the electricity production in the world in 2002 (Observ’ER Systèmes Solaires (3)): • Electricity consumption per inhabitant: 444 kWh / inhabitant in South Asia ~ 15 000 kWh / inhabitant in North America; • Total electricity production: 16 127 TWh; • Composition of the various energies for electricity production:
Source
Fossil Hydro (incl. Pump storage) Nuclear Biomass Geothermal Wind Solar • •
Electricity Average production 2002 Share % 2002 annual growth (TWh) 1993 – 2002 (%) 10 484.0 65.0 % 3.4 % 2 643.0 16.4 % 1.0 % 2 720.0 16.9 % 2.4 % 175.1 1.1 % 4.9 % 49.3 0.31 % 2.5 % 53.6 0.33 % 29.2 % 1.95 0.01 % 13.5 %
Annual growth 2001 – 2002 (%) 3.3 % 1.8 % 2.1 % 3.2 % 1.2 % 35.1 % 20.1 %
Total renewable sources: 2 932 TWh, representing 18.1 % of the total electricity production. Among them, 90% is hydro (2643 TWh). Average annual growth rate of renewable energies: 2.8 % between 1993 and 2002 (larger for wind and solar than for biomass and geothermal, which represent the largest area of RES). The average growth rate between 2001 and 2002 is 2.9 %.
Renewable Energy in the World: The following table details the main figures of the development of renewable energy at the end of 2002.
42
World Energy Council
Performance of Generating Plant 2004 - Section 3
WIND
Installed power
Forecast 2010
31 412 MW
90 000 MW (65 000 MW for Europe)
(39 294 MW in 2003)
SOLAR / PV
1 246 MWp
11 300 MWp
43
(5 GWp Japan + 3 GWp Europe + 2.14 GWp USA)
Average cost / kWh
Top countries
4 to 7 US Germany cents Spain USA Denmark (7 948 MW in (25,1 % in 3.0 TWh 2003) 2003) Italy 6 868 MW
535 MWp cells & modules produced
175.1 TWh 2.6 Gtep average produced in scenario 2025 2002
GEOTHERMAL 8 356 MW between 21 GW and 32 GW (49.3 TWh produced in (depending on scenarios) 2002)
Annual growth rate (%)
382 Mwe between 2000 and 2002
35.1 %
19.8 %
19.4 TWh 8.4 TWh 8.1 TWh 4.9 TWh India 1.4 TWh
up to Japan 433 MWp ~1 US $ Germany 189 MWp USA 171 MWp Australia 34 MWp Italy 24 MWp Netherlands 21 MWp
3.2 % (4.9 % mean annual 19932002)
~ 3 to 7 USA US cents Japan Germany Finland Brazil Canada
1.2 % (2.5 % mean annual 19932002)
3 to 5.5 USA 2 018 Mwe US cents Philippines 1 834 Mwe Mexico 953 Mwe Italy 862 Mwe Indonesia 797 Mwe
43
69.2 TWh 14.1 TWh 11.7 TWh 10.4 TWh 9.5 TWh 7.0 TWh
Observations
Wind farms : Buena Vista (USA) = 183 MW / Koudia (Morocco) = 50 MW / Helgoland (offshore project Germany) = 1200 MW Turbine powers : Mini 500 - 750 kW (onshore) / Average 1297kW (installed in 2002 Germany) / Max 3500kW (onshore) / Prototype 4.5MW (offshore) Examples : Carrisa Plains (USA) = 5 200 kWp / Napoli (Italy) = 3 300 kWp / Saijo (Japan) = 1 000 kWp / Munich (Germany) = 1 000 kWp / Toledo (Spain) = 1 000 kWp
Various types of biomass : Wood energy / Industrial wastes / Domestic wastes / Landfill biogas Size : 2 to 100 MW (average ~20MW)
Various types of technologies : - Aquifers (180 to 350 °C) - Binary (90 to 150 °C) - Fractured deep rocks
Performance of Generating Plant 2004 – Section 5
BIOMASS
Installed in 2002
World Energy Council
Renewable Energies in the World by the end of 2002:
World Energy Council
Performance of Generating Plant 2004 - Section 5
Evolution of the RES Production of Electricity: The following tables show the evolution, between 1993 and 2002, of the production of electricity from renewable energy sources in the world and in the main producing regions, i.e. USA and Western Europe:
Wind Electricity Production (TWh) 60
Western Europe 50
USA Total World
TWh
40 30 20 10 0 1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
Year
Solar Electricity Production (TWh) (Heliothermodynamics + PV)
2.5 Western Europe
2
USA Total World
TWh
1.5 1 0.5 0 1993
1994
1995
1996
1997
1998
Year
44
1999
2000
2001
2002
World Energy Council
Performance of Generating Plant 2004 - Section 5
Biomass Electricity Production (TWh) Western Europe USA Total World
200 180 160 140
TWh
120 100 80 60 40 20 0 1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
Year
Geothermal Electricity Production (TWh) Western Europe
60
USA Total World
50
TWh
40
30
20
10
0
1993
1994
1995
1996
1997
1998
Year
45
1999
2000
2001
2002
World Energy Council
Performance of Generating Plant 2004 - Section 5
Main RES Producing Countries: The following histograms show the main electricity producing countries in 2002 for wind, solar (thermodynamics + photovoltaic), biomass and geothermal energies:
Wind Production in 2002 (TWh)
20 18 16 14 12 10 8 6 4 2
Rest of the World
Netherlands
United Kingdom
Italy
India
Denmark
USA
Spain
Germany
0
Solar Production in 2002 (TWh) (Heliothermodynamics + Photovoltaic) 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1
46
Rest of the world
Spain
Netherlands
Mexico
Italy
Australia
India
Germany
Japan
United States
0
World Energy Council
Performance of Generating Plant 2004 - Section 5
Biomass Production in 2002 (TWh) 80 70 60 50 40 30 20 10
Rest of the World
Netherlands
United Kingdom
Canada
Brazil
Finland
Germany
Japan
USA
0
Geothermal Production in 2002 (TWh) 16 14 12 10 8 6 4 2
47
Rest of the world
Indonesia
New Zeland
Japan
Italy
Mexico
Philippines
USA
0
World Energy Council
Performance of Generating Plant 2004 - Section 5
Among the national and international standards and directives proposed for the development of RES, we can note the following: • In the USA, the Federal Energy Regulation Commission imposes a requirement of at least 5% RES of total additional capacity to the electricity producers (excluding large hydro > 10MW); • In the European Union, a project directive states that by 2010, 12% of the energy consumption should be provided by RES (wind, solar, biomass geothermal and tidal energy). By category, the targets can be assumed as:
Biomass Small hydro Wind Solar (thermal) Photovoltaic Geothermal electric thermal Others
1995 44.8 Mtoe 9.3 GW 3.5 GW 6.5 Mm2 0.075 GW 0.5 GWe 0.75 GWth 0
1
2010 135 Mtoe 13.8 GW 40 GW1 100 Mm2 3 GW 1 GWe 2.5 GWth 1 GW
Considering the present trend, this target could be increased up to 65 GW in Europe in 2010
48
World Energy Council
Performance of Generating Plant 2004 - Section 5
SPECIFIC WEB SITES: The following web sites for wind, solar, biomass and geothermal energies are given below: Specific web sites for wind: European Wind Energy Association (EWEA): American Wind Energy Association (AWEA): IEA programme on Wind turbines: Windenergie Report Deutschland 2003: Budesverband WindEnergie 2002: Wind Power monthly: Wind Stats Newsletter: Specific web sites for solar / PV: European Solar Energy Association: Solar Energy Industry Association: Utility Photovoltaic Group: IEA programme on Solar Power and Chemical Energy Systems (Solar PACES): IEA programme on Solar Heating and Cooling:
www.ewea.org www.igc.apc.org/awea/ www.afm.dtu.dk/wind/iea/ www.iset.uni-kassel.de http://wind-energie.de www.windpower-monthly.com www.gridwise.com/windstats/
www.eurosolar.org www.seia.org www.upvg.org www.demon.co.uk/tfc/SolarPACES.html www.iea-shc.org
Specific web sites for biomass: Links to bioenergy resources online: www.esd.ornl.gov/bfdp/inforesr.html European Biomass Association: www.ecop.ucl.ac.be American Bioenergy Association: www.biomass.org IEA Bioenergy Department: www.ieabioenergy.com FAO: www.fao.org/forestry/ National BioEnergy Industries Association (US): www.bioenergy.org IEA programme on Bioenergy: www.forestresearch.cri.nz/ieabioenery/ Specific web sites for geothermal energy: Geo-Heat Center: Geothermal Database and Publications: Geothermal Education Office: Geothermal Energy Technology: Heat Pumps:
www.oit.edu/~geoheat/ www.smu.edu/~geothermal.htm http://geothermal.marin.org/ www.doe.gov/get/getright.html www.geo-journal.stockton.edu vulcan.geo-phys.stockton.edu www.heatpumpcentre.org/home.htm www.sb.luth.se/vatten/projects/iea/ www.earthenergy.co.uk http://earthenergy.ca/ghg.html www.ghpc.org/ http://doegeothermal.inel.gov/heatpumps.html www.demon.co.uk/geosci/earthen.html www.geoexchange.org www.igshpa.okstate.edu
49
World Energy Council
Performance of Generating Plant 2004 - Section 5
50
Performance of Generating Plant 2004 - Section 5
ANNEX 3 Main Advantages and Drawbacks of RES Technologies WIND - Inexhaustible - No emissions
MAIN ADVANTAGES
- “Local” energy
SOLAR PV
BIOMASS *
- Inexhaustible
- Cheap energy
- Ideal for many remote regions
- Large variety of fuel types
- No polluting emissions
- “Local” energy
- “Local” energy
- Wastes elimination
- Low maintenance
- Agricultural reconversion
- Discontinuous
- Discontinuous
- Collection and transport
- Visual impacts
- Recycling of toxic materials
GEOTHERMAL - Clean, “local” and cheap energy
World Energy Council
World Energy Council
HYDRO - Clean and cheap energy - Instantaneous energy
- Independent on climate or season
- Storage of energy
- High capacity factor
MAIN DRAWBACKS
- Limited zones
- Displacement of population - Impact on river ecosystems
- Air pollution
- Noise
- Discontinuous - Grid connection
- Danger to birds
- Season or climate dependant
- Still very expensive - Competition with other uses * Very much dependant on the type of biomass
51
- Risks of conflicts for share between countries
Performance of Generating Plant 2004 – Section 5
51
- Variety of direct uses
World Energy Council
Performance of Generating Plant 2004 - Section 5
ANNEX 4 Complete List of Proposed RES Indicators Performance Indicator
Type of RES
Easiness to measure / estimate
Comparability from one site to another
All All All Wind, Biomass, Geothermal Geothermal Wind Wind, Biomass, Geothermal Wind, PV PV PV PV PV Biomass
Easy Rather difficult Easy Easy
OK OK OK OK
Easy Easy Easy
OK OK OK
Technical indicators Plant size (MW)
Expected life time Equivalent Full Load Hours (hours) Capacity Factor (%)
Load Factor (%) Specific Energy Production (kWh/m2) Availability Factor (%)
Weather Variability Control Reference Yield (hours/day) Array Yield (hours/day) Final Yield (hours/day) Performance Ratio (%) Efficiency (higher heat value) Fuel Moisture Annual energy sent out (MWh/year) Environmental indicators Duration of operation to produce the energy consumed for its manufacturing (years) Man-hours to return to original state (hours/kW) Avoided CO2 (t/MW/y) Q emissions CO2 during production and during the whole life cycle (g/kWh) Q emissions SO2 during production and during the whole life cycle (g/kWh) Q emissions NOx during production and during the whole life cycle (g/kWh) Land area required for 1 GW (km2) External cost (in €/kWh, or in % of production cost)
52
Rather difficult OK Rather difficult OK Rather difficult OK Rather difficult OK Rather difficult OK Easy Rather difficult
Biomass Biomass
Easy Easy
Rather difficult Rather difficult
All
Rather easy
Rather difficult
All
Rather easy
OK
All
Easy
Rather difficult
All
Rather easy
OK
All
Rather easy
OK
All
Rather easy
OK
All All
Easy Very difficult
OK Difficult
World Energy Council
Performance of Generating Plant 2004 - Section 5
Qtox toxic materials contained in the cells & batteries Qash quantities of ashes emitted (g/kWh) QCH4 quandities of CH4 emitted (g/kWh) QH2S quantities of H2S emitted (g/kWh) Dmin Visual / landscape protection distance (m) Sf Sound at foot of turbine (dB) St Sound of turbine (dB) at standard distance H+D/2 (according to norm IEC 61400-11) S500 Sound 500m away from turbine (dB) Birds killed (no/year) Shadow casting (hours/year) Sociological indicators No Jobs created direct (full time employees) and indirect (n/MW) Na People access to electricity (n/MW) SAR number of accidents / 1,000,000 man-hours
PV
Difficult
OK
Biomass Biomass
Rather difficult Easy
OK OK
Geothermal Wind Wind Wind
Wind Wind Wind
All All All
Easy OK Rather difficult Rather difficult Easy Easy
OK OK
Rather difficult Rather difficult Difficult Difficult
OK Rather difficult
Rather difficult Rather difficult Easy Easy
Rather difficult OK
Note: along with these indicators, general information on the RES plants has to be provided in order to be able to compare and make useful statistics in the future. In particular, the following data will be necessary when introduced in the future databases: • Age; • Vintage; • Type of technology (type of PV cells, of wind turbine wings, of biomass fuel, of geothermal technique); • Cycle of operation.
53
World Energy Council
Performance of Generating Plant 2004 - Section 5
ANNEX 5 Commercial Availability (From Micali V., Statham B., Moss T. – ESKOM – Commercial Availability & Availability Earnings Ratio, Proposing Paper, WEC - PGP Meeting, Rio de Janeiro, April 2000) The work performed by ESKOM, South Africa, concerning “commercial availability”, is described in the above mentioned report. We just show hereafter the basic ideas and an example of results. The principle could be very usefully applied to RES, and especially to wind and PV power plants. It is based on the definition of indicators reflecting the classical relationship between supply and demand, and the opportunity to dispatch power depending on the related costs and revenues. Four domains of Commercial Availability are defined in the following table: REVENUE > COST
REVENUE < COST
DISPATCHED (Actual MW)
Adding Value
Destroying Value
NOT DISPATCHED
Missing Opportunity
Not Competitive
On this basis, the following indicators are then defined: • •
The Valuable Availability Balance (VAB) family of indicators measures the frequencies in each particular domain for every hourly event; The Banked Availability Value (BAV) family of indicators accounts for the financial impact of the Revenue – Cost relationship.
An example of results for two power stations A and B is given hereafter (BAV normalised in %) Banked Availability Value (BAV) (Normalised in %)
-7.7%
90.4%
Station B
-1.9%
-2.7%
67.4%
Station A
-28.9%
-0.9% -40.0%
-20.0%
0.0%
Adding Value
20.0%
Not Competitive
54
40.0%
Missing Opportunity
60.0%
Destroying Value
80.0%
100.0%
Performance of Generating Plant Section 6 WORKSHOPS AND COMMUNICATIONS CASE STUDIES OF THE MONTH (CASOMs)
Compiled and Edited By
R. R. RICHWINE Chair, Work Group: Workshops and Communications Consultant (US)
World Energy Council
Performance of Generating Plant 2004 – Section 6
Work Group Membership: R. Richwine (US) C. W. Kim (Korea) V. Micali (South Africa) G. S. Stallard (US) E. Nekhaev (WEC)
ii
World Energy Council
Performance of Generating Plant 2004 – Section 6
TABLE OF CONTENTS Page INTRODUCTION…………………………………………………………………… vi CASOM 1: HIGH IMPACT LOW PROBABILITY (HILP) REDUCTION PROGRAMME Vincent Ryan – Electricity Supply Board of Ireland (ESB)……… …………………...… 1 CASOM 2: PEAK SEASON RELIABILITY Robert R. Richwine - Consultant…………………………………………………………. 5 CASOM 3: AN APPLICATION OF BENCHMARKING TO ESKOM’S 90:7:3 PROGRAMME Vincent Micali - Eskom, Generation, Energy Management Department……………….. 7 CASOM 4: DESIGN OR MANAGEMENT - WHICH INFLUENCES YOUR PLANT’S RELIABILITY MOST? Robert R. Richwine - Consultant………………………………………………………... 15 CASOM 5: GENERATING UNIT AVAILABILITY FOLLOWING PLANNED OUTAGES Robert R. Richwine - Consultant………………………………………………………... 17 CASOM 6: OPTIMUM ECONOMIC AVAILABILITY Robert R. Richwine Consultant……………………………………………………….….19 CASOM 7: PEER UNIT BENCHMARKING ASSESSING FACTORS AFFECTING AVAILABILITY Robert R. Richwine - Consultant Michael Curley - Manager, Generating Availability Data System, North American Electric Reliability Council……………………………………………………………... 23 CASOM 8: PREDICTING UNIT RELIABILITY Robert R. Richwine - Consultant………………………………………………………... 27 CASOM 9: AVAILABILITY IMPACT OF FLUE GAS DESULPHURISATION SYSTEMS Robert R. Richwine - Consultant………………………………………………………... 31 CASOM 10: ESTIMATING NEW TECHNOLOGY RELIABILITY Robert R. Richwine - Consultant………………………………………………………... 33 CASOM 11: RELIABILITY VERSUS DEMAND Robert R. Richwine - Consultant……………………………………………………….. 35 CASOM 12: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 1 – AWARENESS Robert R. Richwine - Consultant………………………………………………………... 39
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 13: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 2 – IDENTIFICATION Robert R. Richwine - Consultant……………………………………………………….. 45 CASOM 14: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 3 – EVALUATION Robert R. Richwine - Consultant………………………………………………………... 53 CASOM 15: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 4 – IMPLEMENTATION Robert R. Richwine - Consultant……………………………..…………………………. 61 CASOM 16: ARE RELIABILITY MEASURES UNRELIABLE? PART 1 Robert R. Richwine - Consultant……………………………………………………….. 67 CASOM 17: ARE RELIABILITY MEASURES UNRELIABLE? PART 2 – USING COMMERCIAL AVAILABILITY Robert R. Richwine - Consultant…………………………….………………………….. 71 CASOM 18: THE RELATIONSHIP BETWEEN SCHEDULED MAINTENANCE AND FORCED OUTAGES AND ITS ECONOMIC IMPACT ON SELECTING AVAILABILITY GOALS Robert R. Richwine - Consultant……………….………………………………….……. 75 CASOM 19: STATISTICS OR UNDERSTANDING: WHICH ONE DO YOU BELIEVE? Robert R. Richwine - Consultant………………………………………………………... 79 CASOM 20: PEER SELECTION FOR BENCHMARKING - DOES IT MAKE A DIFFERENCE? Robert R. Richwine - Consultant………………………………………………………... 81 CASOM 21: ESTIMATING A GENERATING PLANT’S FUTURE MAINTENANCE COST Robert R. Richwine - Consultant………………………………………………………... 83
CASOM 22: AGING OR VINTAGE WHICH IS MOST RESPONSIBLE FOR DIFFERENCES IN BOILER TUBE LEAK RATES?: PART 1 Robert R. Richwine - Consultant………………………………………………………... 89
iv
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 23: BOILER TUBE LEAK STUDY: PART 2 Robert R. Richwine - Consultant………………………………………………………...95 CASOM 24: CHANGING GAS TURBINE DESIGN REQUIREMENTS Robert R. Richwine - Consultant………………………………………………………... 99 CASOM 25: THE FUTURE IS NOT WHAT IT USED TO BE Robert R. Richwine - Consultant……………………………….…………………….... 103
v
World Energy Council
Performance of Generating Plant 2004 – Section 6
INTRODUCTION The main objective of the PGP Committee’s Work Group 7 is to develop the Committee’s communication activities to ensure wide international dissemination of the results of its work. This includes the organisation of workshops and other events to present and discuss the numerous useful concepts developed by the Committee and to promote their wide application by the global electric power sector. It also aims to develop the use of IT and the Internet for communication purposes, including the introduction of the “Case Study of the Month” initiative on the WEC’s Global Energy Information System at www.worldenergy.org. Case Study of the Month Concept On the first day of each month, an automated e-mail announcement is sent to the WEC distribution list with an abstract of the current month’s case study and a link to GEIS (through the PGP “home page” directly to the case study). On the website, one will find the following: • • •
The "Case Study of the Month" - a brief summary (2-3 pages) of the impact that the study had on decisions made; A link to the detailed report (if it is available on the Internet) or to the submitting organisation's website (for ordering the detailed report); A link to a PGP archive of previous "Case Studies of the Month" and/or other areas of the WEC/PGP website (announcements of upcoming workshops, meetings, etc.).
Case studies demonstrate the ways reliability data has been used by different organisations to make better decisions that resulted in actual improvements and cost savings. A collection of the published Case Studies is included in this report. Workshops and Presentations The second component of the Work Group 7 activities is Workshops and Presentations. During the 2002-2004 work cycle, the Committee organised and/or participated in the following events: 2002: 17-18 April, Shanghai: 9-11 October, Berlin: 2003: 24-25 March, Amsterdam: 16-19 June, Atlanta: 9 September, Kiev: 15-17 September, Copenhagen: 2004: 15 April, Rome:
PGP Workshop “Reliability Measurement, Analysis & Benchmarking” VGB Annual Conference “Operational Outages for Power Generation” Conference ASME Conference WEC Executive Assembly, PGP Workshop VGB Congress “Power Plants 2003” Seminar & Round Table “Improving Reliability of Generating Plant: The Real Value”
vi
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 1: HIGH IMPACT LOW PROBABILITY (HILP) REDUCTION PROGRAMME Vincent Ryan – Electricity Supply Board of Ireland (ESB) BACKGROUND Within ESB a High Impact Low Probability Event has become known as a HILP. By definition HILPs occur infrequently but have a high impact in terms of outage duration and/or cost. In 1992, ESB was an island utility of approximately 4000 MW, with no interconnection to other systems. Seven large generating units ranging in capacity from 250 MW to 305 MW made up almost 50% of the generating capacity. A one-year outage on one of these units could reduce the system availability by 6% to 7.5%. The occurrence of a HILP on one of these units could cause cancellation of a planned outage or threaten the ability to provide a reliable service. Prior to 1992, ESB had experienced a considerable number of HILPs. In one year, the impact on availability was 4%. The need to take action to reduce HILPs was recognised, but the task was deemed daunting. When a transformer on one of the large units failed in 1992 (resulting in a 9 month outage) it was decided to "grasp the nettle" and take action. DATA COLLECTION AND ANALYSIS ESB commissioned NERC (North American Electric Reliability Council) to create a database of HILPs from their GADS database. NERC produced a list of events which resulted in outages of over 1000 hours from their database of over 5000 units. The database identified: • • • • •
The type of forced outage; Start and end dates/ times; Duration of event; Cause code; and A short description of what caused the event.
From this database ESB produced the above list of causes of HILP with durations of over 3000 hours. This was used to prioritise the risks to be addressed. The study indicated that the average HILP frequency in the NERC database was one HILP (of 3000 hours outage duration) every 70 unit-years. ESB subsequently defined a HILP as: • • •
An unplanned outage of 3 months on any unit; A significant environmental incident; An incident which could cause loss of income greater than 10% for the particular generating station.
Using the database, and adding the experience of ESB specialists, a list of 80 potential causes of HILPs was drawn up. These were then prioritised and an action programme drawn up.
1
World Energy Council
Performance of Generating Plant 2004 – Section 6
APPLICATIONS AND PROCESS A failure Modes and Effects Analysis was drawn up for each HILP by the appropriate ESB plant specialist. Only those modes which could potentially cause a HILP were listed. For each mode, a list of actions was drawn up which could: • • •
Reduce the probability (Prevention); Provide early warning (Detection); Reduce impact or duration of outage (Mitigation).
Some of the actions listed required large expenditure (e.g. purchase costly spares). Other actions could be easily implemented (e.g. revise operating instructions, install an alarm). All actions were divided into the following 3 categories: • • •
Short Term - Those which could be implemented immediately; Medium Term - Those which could be implemented within a year; Long Term – Those which required considerable resources and could not be done immediately.
SUCCESS STORY On 16 March 1994, a sacrificial anode on a condenser became detached from the tube plate in the condenser water box. As it had been fixed through the tube plate, it left a hole about 12 mm (half inch) in diameter and resulted in a large quantity of seawater entering the hot well. If the estuarine water had entered the boiler, a long outage could have resulted with possible need for a chemical cleaning costing in excess of US$1 million. A "Massive condenser leak" had been identified as a potential HILP. A condensate conductivity trip had been installed under the heading of "detection" during the HILPs
2
World Energy Council
Performance of Generating Plant 2004 – Section 6
programme. The action taken (under the heading "mitigation") was to trip the extraction pump. The unit was taken off load before a large influx of contaminated condensate could enter the boiler. A prolonged outage and possible need for acid wash were avoided.
References: Gerard Caffrey; Thomas Hanson, Electricity Supply Board, Dublin, Ireland and Ronald J. Niebo; Michael Curley, North American Electric Reliability Council, Princeton, New Jersey, USA, Minimising High Impact Low-Probability Forced Outages, published PWRVol. 28 1995, Joint Generation Conference, Volume 3 ASME 1995.
3
World Energy Council
Performance of Generating Plant 2004 – Section 6
4
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 2: PEAK SEASON RELIABILITY Robert R. Richwine Consultant BACKGROUND The Reliability Engineering Department at Southern Company, one of the largest U.S. investor-owned electric utilities, has the responsibility of forecasting the future reliabilities of each of its generating units for use in the Planning Department's generation expansion planning models. This forecast was typically a single value for each unit for the entire year. ISSUE RECOGNITION Southern Company plant managers mentioned to Reliability engineers that their plant's’ goals for reliability were higher during periods of peak demand than for other times during the year. The Reliability engineers understood that the economic optimal reserve margin (the excess amount of installed generating capacity required to insure costeffective levels of customer service reliability) determined by System Planning's expansion-planning models were highly dependent on the plant's "peak period reliability". Therefore, if the plants were in fact more reliable during these peak periods, the system's reserve margin could be lowered without reducing customer service reliability below the economic optimal (defined as the point where the incremental cost of further increasing reliability is equal to the incremental value the customer receives as a result of increased reliability of service coming from the increased capacity). DATA COLLECTION Reliability data for each plant for "peak periods" and non-peak periods was collected for each plant for the previous five years, an easy task using the North American Electric Reliability Council's (NERC) Generating Availability Data System (GADS) programme. DATA ANALYSIS Comparison's was made between the plants' reliabilities during the peak season versus other times of the year. Statistical analysis was performed that indicated a very high probability that the plants were in fact exhibiting higher reliabilities during peak periods rather than simply a random variation. Therefore, forecasts of plant reliability incorporating seasonal variations could be made with a high degree of confidence. RESULTS APPLICATIONS The Reliability Engineering Department began supplying the System Planning Department with two sets of reliability forecasts, one for the peak season and one for the non-peak season.
5
World Energy Council
Performance of Generating Plant 2004 – Section 6
EFFECTS The economic optimal reserve margin was reduced by one full percentage point. For Southern Company's 30,000+ MW of capacity at the time, this amounted to over 300 MW of peaking capacity that did not need to be built without lowering customer service reliability. COST SAVINGS The avoided 300+ MW of peaking capacity represented a cost savings of ~US$100 million. As a final note, this study was extended to include the entire North American industry, with comparable results.
References: Lofe, J.J., Richwine, R.R., Decreasing System Peak Reserve Margin Requirements. Lofe, J.J., Bell, F.J., Curley, G.M., Seasonal Performance Trends; published by NERC.
6
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 3: AN APPLICATION OF BENCHMARKING TO ESKOM’S 90:7:3 PROGRAMME Vincent Micali Eskom, Generation, Energy Management Department PREAMBLE There are essentially three tenets in a production business: the availability of the product, its reliability and good value for money. These are not mutually exclusive but highly interlinked. They do however have a serial relationship in the product acquisition: a product must essentially be available first (or perceived to be, as in the money markets), and then reliable, for the price that one is prepared to pay. These tenets form the ingredients and posture for benchmarking (i.e. one can now start to compare). For comparative purposes, then the business ought to measure and perform accordingly to targets that are set and, then, once achieved, high performance and productivity ought to be sustained. This paper was produced in response to the World Energy Council (WEC) request (Ref. 1). As a practical example of how a benchmarking process may be utilised, the principles adopted by South Africa’s electric utility, Eskom, in its “90:7:3” generation plant availability improvement initiative are reviewed. Further details are available in the 1998 WEC PTGP Monograph (Ref. 2). The figures refer to the utility values respectively for Unit Capability Factor (UCF), Planned Capability Loss Factor (PCLF), and Unplanned Capability Loss Factor (UCLF), the indicators being those as defined by UNIPEDE. Benchmarking performance against International Best Quartile (IBQ) values enabled positioning of the business amongst top international peer utilities. The databases that furnish input for such positioning are most important, their accuracy (e.g. definitions adhered to by all parties, data consistency), appropriateness/relevance to the situation and ease/speed of access being critical factors. After scrutiny of various databases in the international arena, due to the unfortunate fact that they are mutually exclusive (e.g. forced outages are measured differently), a decision had to be made in terms of selecting one. In this case the UNIPEDE database was used. One then needs to position the business for optimal performance. Care has to be taken in the definitions of the Key Performance Indicators (KPIs) and whether these match the strategies of the business (Ref. 2 and Ref. 3). The following is an outline regarding the importance of, and linkages between, benchmarking, performance and its sustainability. STRATEGIES Various performance improvement strategies may be adopted, depending on the resources available to effect improvement and the time desired in which to make the changes as shown in Figure 1. The “step change” approach (Strategy E in Fig 1) was adopted by Eskom as a strategy i.e. the immediate application of intensive resources. This essentially meant that a certain result (target for a KPI) had to be reached in time T* with those resources. For other organisations that, for instance, had to achieve a certain result (KPI level) by a specific time T*, it might have been appropriate, due to constraints, to utilise
7
World Energy Council
Performance of Generating Plant 2004 – Section 6
other resource level applications at different times (see strategies A to D). Combinational hybrids such as D up to point of intersection and then B are also possible (fear of downtrend at time of intersection). I decided to name this family of strategies the STATHAM’s curves, in honour of Brian A. Statham, who was the original founder of these concepts in the early nineties and is regarded as one of the best strategists in the business. A: B: C: D: E:
Slow incremental – allows resources accumulation Hybrid incremental, as in “A” with rapid accumulation Constant rate Hybrid, medium resources & breather period Step Change, resource intensive
E
T* Figure 1: Performance Improvement Strategies CONCEPT OF RISK AND LINK TO REMUNERATION A statistical approach was adopted that included the concept of risk in achieving targets (Ref. 3). For example, the term “target” implied that there was only a 20% chance of bettering the particular figure (see Fig. below). The “target” and its risk were relative to the expected (or most likely) value to occur. Before embarking on setting up benchmarks, “targets” and sustainability indices, it is important to understand the relevant KPIs’ behaviour in the business. The sustainability index measures the level at which the value of the KPI becomes significantly risky and the desired result may not be achieved. A method of doing this is to study the statistical distribution properties. For instance, a KPI might be Normally (Gaussian) or Lognormally distributed. This kind of finding is not a trivial one. It permits the business to be edged towards its “true” risks.
8
Performance of Generating Plant 2004 – Section 6
Probability Density Function
World Energy Council
20% KPI
Target
Figure 2: KPI Gaussian Distribution
Probability Density Function
As an example, if a KPI (say Profits) is assumed to be Normally distributed while, in fact, it is Weibull on its expectation, then the concept of the 20th percentile might actually be at the mode (apex) of the Weibull (see Fig. below). This would be edging the business on the low side of the KPI (Profits) i.e. there is no stretch. Subsequently, there could be questions on why the business has a high frequency of outcomes on the target (too little stretch), whilst it was edged on the 20% risk. Cumulative Distribution Functions (CDFs) are often used in this respect with percentiles of importance (e.g. Quartiles, Deciles).
Normal
Weibull
20% Target
KPI (Profits)
Figure 3: KPI (Profits) Distributions An annual KPI Workshop served to bring together various parties to agree mutually on the figures to be achieved. These figures were linked to personal performance contracts and remuneration schemes to provide the necessary incentive to meet targets (Ref. 4).
9
World Energy Council
Performance of Generating Plant 2004 – Section 6
RELATIONSHIP BETWEEN BENCHMARKING, TARGETS AND SUSTAINABILITY OF PERFORMANCE Figure 4 shows risk progression towards a benchmark (A), with (D) being the minimum values required to achieve sustainability. (B) and (C) show target and expected values progression respectively. This risk progression assumes a constant rate of resource allocation (see strategy C in Fig. 1), a normal KPI distribution and Fig 4 shows the KPI to be at risk at each point in time. There are two spaces along which a KPI may move in time, one space being along a growth model (C) and the other along a static (or stationary) model (A). Whichever the model, the sustainability index, e.g. (D) in the growth model, ensures that the strategic directive is maintained (Ref. 5). This is a Quality Assurance technique.
A: Benchmark (e.g. IBQ) B: Target Progression C: Expected values D: Sustainability Figure 4: Risk Progression Towards a Benchmark
10
World Energy Council
Performance of Generating Plant 2004 – Section 6
PERFORMANCE PROGRESS A progression from Benchmarking is the setting of milestones toward goals for the KPIs. One way of achieving this, is through a mechanism that sets “targets” with its corresponding risks on a particular KPI. These “targets” are associated with other values that assess the volatility of the KPI.
For instance, if the desired KPI is the UNIPEDE UCF, then let us assume that after modelling and forecasting the expected value is 89%. As “target” is defined as a 20% Chance of Bettering (CoB) its corresponding value, from the expected value and the distributional model (or behavioural pattern) of the UCF, the “target” would be set at 90%. The other associated values (with respect to the expected value, in terms of the terminology used in the case study) would be: UCF Ceiling: 91% (5% CoB), UCF Kickin: 87% (80% CoB) and UCF Baseline: 85% (95% CoB) (Refs 2, 3 and 4). Figure 5 is a diagram of the original 1994 strategy with target figures to improve UCF from 80% at the beginning of 1995 to 90% by 1999. It also gives a listing of actual performance achieved. Although the original plan was to achieve 90% by 1999, this figure was reached by 1996, with ten-percentage points improvement being made from 1994 and 1996. An all time annual high of 91,5% was recorded in 1997.
Actual was higher at: 1995: 84,3% 1996: 90,6% 1997: 91,5%
Target of 90% UCF had 20% chance of being achieved
Figure 5: Original UCF Strategy and Actual Performance
11
World Energy Council
Performance of Generating Plant 2004 – Section 6
EXECUTIVES DRIVING CHANGE The change dynamics incorporated executive management setting demanding expectations and enforcing these with consequences, driving the initiative with passion and zest (see Fig. 6). This is in line with international experience, which shows that highperforming organisations must be led from the top down.
Figure 6: Executives Driving Change FEEDBACK MECHANISMS Individuals requiring feedback or knowledge of result to perform at high levels are most important (e.g. business reviews, reporting of results, posters, link to personal performance contracts). Figure 7 shows a typical example of this process, namely the causes of unavailability for a particular period.
12
World Energy Council
Performance of Generating Plant 2004 – Section 6
Generator Stator Boiler Tube Leaks Milling Plant Outage Slip Combustion Unit Electrical Coal Plant
Figure 7: Components of a 3,2% UCLF IMPORTANT POINTS Eskom’s experience or lessons learned can be summarised as follows: • • • • •
Participants should understand the concept of risk in achieving targets; Databases which are accurate, appropriate/relevant to the situation and which have easy/speedy access are most important to benchmarking processes to improve business positioning and performance; Executives must drive the change process from the top down; Consider moving from IBQ (e.g. UNIPEDE) to a World Best Quartile (WBQ), which would encompass benchmarking against an amalgamation of appropriate and relevant world databases; Such processes are not perfect, and consideration needs to be given to moving towards maximising “commercial” availability as opposed to “technical” availability on which the above example has been based. One needs to avoid driving “technical” availability past the point of diminishing returns.
13
World Energy Council
Performance of Generating Plant 2004 – Section 6
References: Corrigall M.R.V., Crookes B.T., Micali V, Contribution Towards PTGP WG4 “International Data Exchange” Paper, In response to Rome meeting presentation, February 1998. Corrigall M.R.V., Crookes B.T., The Impact of Government and Institutional Policy Changes on the Southern African Electricity Supply Industry. 17th WEC Congress, Houston, Texas, September 1998.
Micali V., Statham B.A., Appropriate Statistical Techniques for Rational Target Setting, WEC Regional Energy Forum, Cape Town, October 1994. Micali V., Jacobs N.B., Development of Production and Supply Agreements and Performance Contracts, Generation Group Procedure GGP0422, Eskom, Johannesburg, February 1996. Campbell I.G., Sustainability Index – Managers Manual, Doc No. 7741 A 162 S, Eskom, Johannesburg, February 1998.
14
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 4: DESIGN OR MANAGEMENT - WHICH INFLUENCES YOUR PLANT’S RELIABILITY MOST? Robert R. Richwine Consultant BACKGROUND There are many occasions when a plant’s future reliability must be estimated. A generation executive setting goals; a developer compiling a pro-forma or a planner committing to a construction program; a trader/marketer trying to manage risks; a bank engineer providing advice on the commercial feasibility of a new project; an insurer setting premiums. These and others use reliability estimates on a routine basis. However, many times the importance of plant management is not fully appreciated. Instead, it is often treated as a commodity; something that can be taken for granted; i.e. one plant management team will get the same results as any other team. People who work closely with power plants know this is not true, but how can we prove it? DATA We began by compiling reliability data from the North American Electric Reliability Council’s (NERC) Generating Availability Data System, which contains unit specific reliability and design data on over 5000 units. Using an advanced statistical method we evaluated over 50 design and operational characteristics of more than 1700 fossil steam units to find out which plant features were the most significant in determining the optimal peer selection criteria. Based on this analysis we were able to divide the fossil steam units population into 19 separate groups (the result of this process was enlightening itself since the typical ways of dividing the population, size and fuel type, were not the most statistically significant, while other characteristics such as criticality, vintage and duty cycle were much more important. For more details on this study, request reference 1). ANALYSIS Using the 19 peer groups we plotted each group’s probability distribution of their unit’s Equivalent Forced Outage Rate (EFOR). We then selected the best performers (top tenth percent) within each peer group (one data point per group) and plotted the resulting probability distribution. We also plotted the probability distribution of the difference between the worst performers (bottom tenth percentile) and the best performers from each group. Comparing the spread of the two distributions we were able to clearly see that there was a much greater spread between the best and worst performers within each peer group (average was 20 percentage points) than between the best performers across all the groups (average was 4 percentage points). Based on this result, we estimated that design and operational differences between power plants could only account for approximately 20-25 % of the variation in reliability. The other 75-80 % must come from another source, which we can generally classify as management.
15
World Energy Council
Performance of Generating Plant 2004 – Section 6
CONCLUSIONS Although there will always be unique differences between any two power plants that cannot be easily quantified, we believe that the statistical evidence clearly points to the quality of management as the key factor in determining the plant’s performance. A superior management philosophy and personnel will achieve consistently top performance, even with a relatively poor design or difficult operating mode; whereas a weak O&M program will have poor results, even with a superior design. “What you do with what you get is much more important than what you get to begin with”. RECOMMENDATION All of the organisations that need to forecast a plant’s future performance will need to devote much more time and energy to evaluating the quality of the plant’s O&M programme and management methods. Much more analytical research is required in order to identify the characteristics of superior management and quantify their impact on plant performance.
References: Niebo, R.J., Richwine, R.R, Peer Unit Benchmarking: Assessing Factors Affecting Availability. Niebo, R.J., Richwine, R.R., Curley, G.M., Jenkins, A.K., Establishing Realistic Availability Goals Using Statistical Benchmarking Techniques. Richwine, R.R., Jenkins, A.K., Optimizing O&M Costs to Maximize Profitability.
16
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 5: GENERATING UNIT AVAILABILITY FOLLOWING PLANNED OUTAGES Robert R. Richwine Consultant BACKGROUND For a number of years there has been speculation that a generating unit’s reliability is lower than normal during the period of time immediately following its return to service after a planned overhaul. Earlier studies had been done that gave indications that this was true, so the North American Electric Reliability Council’s (NERC) Generating Availability Trend Evaluation (GATE) Working Group undertook a project to study the daily and hourly performance of units during their first week of service following a planned outage. The purpose was multi-fold: • • •
To determine the likelihood of having a forced outage and the day it is most likely to happen; To identify the components most often responsible for the forced outages; and To determine if there is a relationship between forced outage occurrences and the reasons for the preceding planned outages.
DATA AND ANALYSIS Base-load coal-fired steam units 400 MW and larger in capacity were included in the study sample. NERC provided data from its Generating Availability Data System (GADS) for the analysis. The study approach required building timelines for each individual unit between long planned outages (those lasting more than one week) and then analysing the forced outages that occurred during those intervals, especially during the first week after the planned outage. RESULTS SUMMARY • • • •
• •
There is a 47% chance that a unit will have its first forced outage on the day it returns to service following a long planned outage. If it doesn’t, there is only a 5% chance it will have one on the second day; The probability of having at least one forced outage in the week following a long planned outage is 63%; Forced Outage Factor (FOF), the ratio of forced outage hours to period hours, decreases at a significant rate with each successive day during the first week; The major contributors to forced outage hours during the first week are boiler tube leaks and turbine vibration. The next most significant contributors are boiler control systems and turbine overspeed test failures. A turbine problem is the most likely cause if the first outage occurs on the first day of service; About 60% of forced outages that occur in the first week, regardless of cause, last less than one day; Other results are available in the referenced study.
17
World Energy Council
Performance of Generating Plant 2004 – Section 6
CONCLUSIONS Recognising that generating units have often experienced a higher rate of forced outages when returning to service from a long planned outage can be a first step to developing action plans to improve their performance. Identifying components that have been the leading causes of these outages will help power plant staff focus their resources toward those pieces of equipment that have the greatest potential for improvement. But even if no additional improvement is cost-effective, planners, dispatchers and trader/marketers can incorporate this knowledge into their daily decisions and thereby minimise the negative economic impact that affects the corporate bottom line due to their plant’s lower reliability following planned outages.
References: Corio, M.R., President, Applied Economic Research Company, Inc., Mills, J.B., Senior Reliability Engineer, Southern Company Services, Inc., Costantini, L.P., Director, North American Electric Reliability Council, Generating Unit Availability Following Planned Outages.
18
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 6: OPTIMUM ECONOMIC AVAILABILITY Robert R. Richwine Consultant For each generating plant operating in its own individual business environment there is a point beyond which increasing Availability is not cost-effective. This point at which the minimum cost of an additional increase in Availability is equal to the value of that increase, or the point of diminishing returns, is called the plant’s “Optimum Economic Availability” (OEA). The first step in estimating a plant’s OEA is to determine the plant’s Cost versus Availability curve (Figure 1). Cost
$
AVAILABILITY
100%
4
Figure 1 This curve is the minimum cost required to achieve and maintain different levels of Availability. (For this discussion I will only be addressing non-fuel Operations and Maintenance spending, including refurbishment capital, once the plant has been built. Design costs optimisation will be left for another case study). Developing this Cost curve is a very difficult task since consistent cost data is difficult to obtain and there is a very wide “scatter” to the data points. Figure 2 was created from cost data reported by United States Electric Utilities to the Federal Energy Regulatory Commission (FERC) and from Availability data reported to the North American Electric Reliability Council (NERC) through its Generating Availability Data System (GADS). Cost
$
AVAILABILITY
100% 8
Figure 2
19
World Energy Council
Performance of Generating Plant 2004 – Section 6
As is clearly evident, the data points resemble a “shotgun” pattern so that any linear regression would have a very poor correlation coefficient. However, if we apply an advanced statistical technique known as “frontier analysis” applied to data from similar plants, we get the curve shown in Figure 3. Cost
$
AVAILABILITY
100% 9
Figure 3
Those plants lying on or close to the frontier curve are achieving the highest levels of availability for various levels of spending, undoubtedly through the application of “industry best practices” Operations and Maintenance techniques. One of the most valuable insights you will get is when you compare your plant against its frontier plants. If your plant is in the “interior” (above the frontier curve) you know that it could either increase its availability without increasing its costs or decrease its spending without lowering its availability. After all other similar plants are doing so! Why not yours? Of course in benchmarking both cost and availability it is vital to select as appropriate a peer group as possible. Our studies have revealed that other design and operational factors besides the traditional size and fuel type are often much more statistically important. For fossil steam units supercritical vs. sub critical or base-load vs. cycling or vintage are often far more significant. For a detailed discussion see Ref. 1, 2, 3. Other difficulties lie in the need to account for differences in labour rates and productivity, material costs and tax and environmental requirements as well as monetary exchange rates (and sometimes government subsidies) when benchmarking using international data. So we see that our plant should be on the “frontier”. But where on the frontier is optimal? To answer that question we need to plot the Worth of Availability curve (Figure 4). Worth
$
AVAILABILITY
100% 5
Figure 4
20
World Energy Council
Performance of Generating Plant 2004 – Section 6
This curve describes how much the plant is worth to your company at different levels of Availability. Although it is not a simple task to estimate these worth values, many companies that have done so and, after distributing the data throughout the generation organisation, have received great value resulting from their staff’s understanding the value of their plant’s performance. By subtracting the minimum Cost curve from the Worth curve, we get the “Net Value” curve (Figure 5).
Net Value (Worth Minus Cost)
$
X
OEA
X
AVAILABILITY
100% 6
Figure 5 The point where that curve is at its maximum value is the plant’s Optimum Economic Availability (OEA). Therefore, for this plant at this time a higher Availability goal would not be cost-effective. In the future as the unit ages and/or new, more efficient O&M techniques are implemented the frontier Cost curve will change, so the plant’s OEA will change. And it is likely that economic conditions will change so that the Worth will change, with the OEA changing also. We can see, then, that the plant’s OEA is a dynamic goal, changing to meet the changing technical, operational and economic environment that we are constantly facing. Today’s evolving market-driven business environment brings a host of new challenges to the already difficult job of managing a modern power plant. We are now asking generation management, from the executive down to the individuals on the plant floor, to focus on economic results as well as maintaining superior technical skills (actually superior financial results start with technical know how and is supplemented with economic insight). Therefore, it will become increasingly vital to the continued survival of our companies that we use modern tools and goal systems to send the proper signals to our generation personnel. They will be the ones who will have to make better “economic” decisions that will determine if we win or lose in the marketplace. We can and should benchmark our plant’s Optimum Economic Availability to objectively set economically rational expectations. However, we must also realise that the OEA will only be achieved when many employees are able to make better day-to-day operations, maintenance and capital investment decisions that support the company’s long-term financial goals.
21
World Energy Council
Performance of Generating Plant 2004 – Section 6
References: Glorian, D., Niebo, R., Salvaderi, L., Unit Benchmarking Demonstrates Compatibility of European and North American Data Collection Systems, 1995 WEC Congress. Niebo, R., Richwine, R., Curley, M., Jenkins, K., Establishing Realistic Availability Goals Using Statistical Benchmarking Techniques, 1995 International Joint Power Generation Conference. Richwine, R., Niebo, R., Peer Unit Benchmarking: Assessing Factors Affecting Availability, 1992 International Joint Power Generation Conference.
22
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 7: PEER UNIT BENCHMARKING ASSESSING FACTORS AFFECTING AVAILABILITY Robert R. Richwine, Consultant and Michael Curley, Manager, Generating Availability Data System, North American Electric Reliability Council
For decades generating companies have been comparing their plants performance against other plants in order to: • • • • • •
Set realistic performance goals; Identify opportunities for improvement; Give advance warning of threats; Set appropriate incentives; Trade knowledge and experiences with their peers (and sometimes to brag about their successes); and Quantify and manage performance risks (a growing vital action in an increasingly competitive business environment).
During most of this time it was assumed that the best selection criterion for the peer group was obvious (fuel and size range for fossil steam units, BWR/PWR/CANDU for nuclear, etc.). Recent studies, however, have raised serious questions about the appropriateness of this selection procedure. Other factors such as criticality, duty cycle, vintage, etc. have been identified as being far more important than fuel or size. Clearly, if the peer selection process is faulty, the entire benchmarking exercise will give misleading results. This case study will discuss an advanced selection process developed by the authors that statistically determines the best select criteria for any individual unit. When you begin your peer selection process one goal is have a population that is large enough for statistical validity and the larger the better. We use the North American Electric Reliability Council’s (NERC) Generating Availability Data System’s (GADS) database which contains unit specific design and performance data on over 5000 individual units, with data extending back to 1982 (data is also available back to the 1960’s from a predecessor system that evolved into the GADS program). Although the data is unit specific, it remains confidential through safeguards developed and administered by NERC. Equally important as a large population is to ensure that the peer units selected have as close a match in design and operating characteristics as possible to your unit. However, if we were to require “exact” matches in all of the design and operational factors that have been identified as important, we will often end up with a peer group of “zero”. Therefore, we must find a way to balance the need for a large population with the need for “exact” match. We achieve this by simultaneously statistically analysing over 50 (for fossil steam units) design and operating characteristics of the unit being benchmarked to find the most statistically important features using techniques that compare the entire distribution of candidate unit’s reliabilities. Once the most important characteristic is identified, we then analyse the remaining characteristics to find the next most important and continue until either there are no more statistically important features or the population is too small for
23
World Energy Council
Performance of Generating Plant 2004 – Section 6
statistically validity. Some of the characteristics for fossil steam units we consider are: • Vintage; • Age; • Criticality; • Fuel firing system; • Boiler circulation type; • Boiler draft type (pressurised vs. balanced vs. converted); • Turbine manufacturer; • Boiler manufacturer; • Unit size; • Reheats; • Generator manufacturer; • Condenser cooling water type; • Duty cycle; • Fuel etc. After having done numerous benchmarking studies for utilities both in North America and around the world, we have been surprised at the number of times that characteristics which we previously assumed were key such as size and fuel type were much less important than others not previously considered. Although each unit must be individually analysed to find its proper peer group, there are some new features that often are identified by the process as most important. The following diagram indicates one common result of the peer selection process.
All Fossil Units
CRITICALITY
Super
Sub
VINTAGE
<1972
MODE OF OPERATION Cycling
Baseload
Size Draft Type Fuel
Boiler Mfr. Draft Type Size
<1972
24
World Energy Council
Performance of Generating Plant 2004 – Section 6
We can see that the most important factor was criticality. Then for the supercritical branch, the second most important factor was vintage (early, pre-1972, vintage supercritical units have much lower reliability whereas recent vintage supercritical units have performed much better as learning curve theory predicts). After the second branch on the supercritical side, there are no longer enough units to be able to continue the process. However, we can see that both fuel and size ranges are less important than vintage for supercritical units. Also, it would be inappropriate to include sub critical units in the peer group. On the sub critical side we find that the second most important factor is “Duty Cycle”. Therefore, including base load units in the population for benchmarking a cycling unit would be improper, and may result in unrealistic expectations for reliability being imposed on the plant management. Because there are many more sub critical units in the total population, we can continue the analysis and will finally find fuel type and size becoming relevant factors in addition to others. In performing a benchmarking analysis the key activity of peer group selection is often overlooked by assuming traditional design and operating characteristics are adequate (fuel type and size for fossil steam units, etc.). However, we have found from our benchmarking analyses of hundreds of units from around the world that other factors are often much more statistically important. In benchmarking it is vital that the most similar units be selected for comparison to the candidate unit or the results will be invalid.
References: Establishing Availability Goals using Benchmarking Techniques, IJPGC Conference, 1995. Peer Unit Benchmarking, IJPGC Conference, 1992. Unit Benchmarking Demonstrates compatibility of European and North American Data Collection Systems, WEC Congress, 1995.
25
World Energy Council
Performance of Generating Plant 2004 – Section 6
26
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 8: PREDICTING UNIT RELIABILITY Robert R. Richwine Consultant When I headed the Reliability Engineering Department at a large (30,000 + MW) electric utility, we had the responsibility of annually estimating future reliabilities of over 100 generating units for the Planning Department to use in expansion models and production cost models (we generally used Equivalent Forced Outage Rate (EFOR) as a measure of reliability). One day the Director of Planning asked me: “How accurate are your estimates? If they predict too low a reliability (high EFOR) we may be building unnecessary plants, but if your predictions are too high (low EFOR) we could be putting people in the dark due to insufficient generation.” I replied that we had been making these estimates for many years and that I thought they were pretty accurate. But I really didn’t know. Therefore we decided to investigate by comparing the actual EFOR’s for the most recent year against the predictions we had made in the previous year (this is a practical example of Statistical Process Control). The frequency distribution in Figure 1 below is the result.
EFORACTUAL - EFORPREDICTED
6
Figure 1 We were terrible! Not only did we have a very wide variation in the results, but also more significantly we had a bias of almost 4 percentage points! That meant that we had been dramatically understating the system’s reliability, and that we were probably planning too much (4 percentage points of 30,000 MW is 1200 MW) new generation. The reason for these high EFOR estimates was that during this period the company’s plants were consistently improving their reliabilities, but we were not adequately incorporating this into our forecasting methods. We had been using historic statistical averages without incorporating new information (aggressive and successful improvement programmes) into
27
World Energy Council
Performance of Generating Plant 2004 – Section 6
our forecasts. We realised that the past was important but that we needed to consider more factors. The basic principle we followed was that if we could understand and quantify the relationship between past conditions and the resulting past reliability (which factors influenced each unit’s reliability and to what degree) and then anticipate what conditions the unit would likely encounter in the future, we could develop an equation to more accurately predict the unit’s future reliability. BASIC PRINCIPLE Past_Conditions ~ Future Conditions Past Reliability Future Reliability We began a project using multi-variable linear regressions to try to develop this new prediction equation (today we might use neural network techniques). After evaluating numerous possible influencing factors, we found that the most important were: • • • • •
The previous year’s reliability (no surprise – this is what we had been doing); Previous year’s duty cycle; Current year Operations and Maintenance activities including outage time and spending; Previous year’s O & M spending (there was a lagging effect – up to two years); Certain design characteristics such as fuel and major equipment manufactures.
After we developed the new prediction equation, we applied it to the original prediction year (which was not used in the development of the equation) and subtracted the new predictions from the actual EFOR’s for that year. The results are shown in the dashed line frequency distribution in Figure 2 below (the solid line is the original distribution). EFORACTUAL - EFORPREDICTED
1 0
Figure 2
Clearly our new equation was much better! We had substantially reduced the variability, but much more importantly, the new process had almost completely eliminated the bias. Now our Planning Department could develop their generation plan and estimate production cost with a much higher degree of confidence. A summary of this and other similar techniques can be found in references 1 and 2. 28
World Energy Council
Performance of Generating Plant 2004 – Section 6
References: Predicting Unit Availability: A Top-Down Perspective, published by the North American Electric Reliability Council’s (NERC) Generating Availability Trend Evaluation (GATE) Working Group, 1991. Predicting Unit Reliability, published by NERC’s GATE Working Group, 1995.
29
World Energy Council
Performance of Generating Plant 2004 – Section 6
30
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 9: AVAILABILITY IMPACT OF FLUE GAS DESULPHURISATION SYSTEMS Robert R. Richwine Consultant In recent years many North American electric utilities have been faced with the likelihood of adding flue gas desulphurisation (FGD) systems to their coal-fired units. Of immediate concern was the impact that these systems might have on the unit’s availability and reliability. Therefore the North American Electric Reliability Council’s (NERC) Generating Availability Trend Evaluation (GATE) Working Group assembled a task force of electric utility experts familiar with the design and operation of FGD systems to perform an in-depth assessment using the detailed data base contained in NERC’s Generating Availability Data System (GADS) as well as information available from federal agencies, industry organizations and internal data from the task force’s companies. The results of this study are expected to help generating unit designers, utility planners and production staffs, and others considering FGD installations. DATA AND ANALYSIS Using the unit specific design, performance and availability/reliability data included in NERC’s GADS database for 111 FGD systems (49,796 MW of scrubbed capacity) the task force was able to compare the impact on availability due to FGD systems for a wide variety of situations; e.g. included in initial design versus retrofitted later, early versus more recent vintage, inclusion of spare modules, etc. In addition the task force was able to examine the range of impacts on individual and not simply averages, leading to a much greater understanding of cost-effective ways to minimise the potential negative impact on performance. EXECUTIVE SUMMARY OF FINDINGS The overriding finding reached by the Task Force is that the performance of FGD systems has improved rapidly since the first FGD systems were installed. FGD system’s impact on unit Equivalent Unavailability Factor (EUF) and unit Equivalent Forced Outage Rate (EFOR) was much less than previously believed. Many reasons were found to explain this improvement. Two primary reasons are that 1) FGD technology matured through the cooperation of manufactures and utilities, and 2) utilities gained experience and knowledge in the operation and maintenance of the FGD cycle. Several of the reports specific findings are described in the following: •
•
FGD systems had minimal effect on unit availability. In 70% of FGD installations, the FGD contributed 1 percentage point or less to the total EUF experienced by the unit. At an additional 12 %, FGD contributed between 1% and 2% of total EUF and at only 7 % (eight units) was EUF increased by more than 4 percentage points (the mean was 1.35% but the median was only 0.31%, indicating that a few bad actors skewed the average substantially). Similarly, the FGD systems contributed 0.25 percentage points or less to unit EFOR in over 67 % of the units, and only 7 % of the FGD systems increased EFOR by more than 2 percentage points (again the mean was 0.45% while the median was only 0.06%); The reduced performance that occurred in FGD systems was primarily due to damage 31
World Energy Council
• •
• • •
• • •
•
Performance of Generating Plant 2004 – Section 6
to the stacks, plugging of the mist eliminators, and repairs to the ductwork and absorber towers; No statistically significant performance difference was found in the unit EUFs or EFORs between units where the FGD system was part of the original design versus those where the FGD system was retrofit at a later date; The inclusion of a spare scrubber module was the only design characteristic that proved to be statistically significant in reducing availability losses. Approximately 30 % of the units equipped with spare modules experienced no change in unit EUF due to the FGD system; In most cases no difference was found in the current performance of the earliest and latest designs as most initial problems were eventually corrected. However, some problems with the earlier designs still exist and continue to degrade performance; No conclusive difference was found between units equipped with or without flue gas bypass systems; No statistically significant performance difference was found in the unit EUFs or EFORs between FGD systems equipped or not equipped with flue gas reheat capability. One anomaly was noted: the few FGD systems that had direct-combustion reheaters did have significant impacts on unit EUFs and EFORs; Capacity losses due to the operation of the FGD system at time of peak averaged 1.74% of total unit capacity; the median was 1.40%. At the time of peak just over 70% of the FGD systems used 2% or less of the unit’s capacity; Station Service (house load or internal station electricity requirements) annual energy requirements for the FGD systems averaged 1.67% of the unit’s capacity (median = 1.28%). Just over 70% of the FGD systems had requirements of 2% or less; Typical manpower requirements for the operation and maintenance of FGD systems were 4.3 persons per 100 MW of generating capacity, divided into 2.8 for operations and 1.5 for maintenance. Separate FGD engineering and chemical laboratory staffing was negligible; Units with reduced performance have higher operating and maintenance (O&M) costs. Where design or operating problems were corrected, O&M costs were reduced.
References: Impact of FGD Systems, North American Electric Reliability Council publication, 1991. Flue Gas Desulfurization System Impact on Availability – 1989-1991 Performance Update, International Joint Power Generation Conference, Atlanta, Georgia, October 1822, 1992.
32
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 10: ESTIMATING NEW TECHNOLOGY RELIABILITY Robert R. Richwine Consultant
Our Case Study “Predicting Unit Reliability” primarily dealt with forecasting performance of mature technologies. This Case Study addresses another requirement we often face which is to predict the performance of new, advanced technologies after they have gone through their initial “break-in” phases. Almost all new technologies experience significant “learning” as the manufacturers, Architect-Engineers, operators and service providers gain first hand experience once a new type plant has gone into service. They try to incorporate that knowledge back into the design, as well as modifying the plant’s Operations and Maintenance programmes. This can be especially true when the original design criteria sought to maximise the efficiency of the plant and/or minimise the initial capital cost, criteria which are almost always in conflict with high reliability (the real goal, of course, should be to minimise the TOTAL of the plant’s performance cost, Operations and Maintenance cost plus design costs – see our Case Study Optimum Economic Availability). As an example, in a paper published at ASME Turbo Expo, June 2001, New Orleans, La., USA, the authors, Axel von Rappard, Consultant, and Sal Della Villa, President of Strategic Power Systems, Charlotte, North Carolina, USA, reported data from 47 F-class gas turbines, collected in 2000. Average availability for these units over a five-year period was 87%, compared to 94% for mature models over the same time period. The paper went on to discuss many of the efforts being made to correct the problems currently being encountered that should lead to improved performance of future F-class machines. What then should we anticipate will be the average reliability for these future machines when the technology reaches maturity? To gain some insight we can look to two earlier studies published in 1985 which analysed the maturing reliability of what was then a relatively new technology that had received unfavourable early reviews in its reliability, i.e. supercritical fossil steam units (Ref 1, 2). In this study the authors applied learning curve theory to the early supercritical vintage data to try to estimate what reliability the technology might ultimately be capable of achieving (of course not every operator will achieve this potential – see our case study Design or Management – Which Influences Your Plant’s Reliability Most). One of the major findings of the study of 77 early vintage supercritical units (Ref 1) was that the reliability of supercritical units was improving, but at a decreasing rate with respect to the year of initial operation (as predicted by learning curve theory). In fact the use of learning curve forecasting methodology resulted in a reliability forecast that was 55% less than the simple average of the actual early reliability statistics (Figure 1 below). As an equally interesting footnote, actual reliability of the most recent vintage supercritical units has been even better than originally forecast by learning curve theory (but still within the range of uncertainty of that forecast).
33
World Energy Council
Performance of Generating Plant 2004 – Section 6
So what should we expect for this new generation of advanced gas turbines? Certainly, if manufacturers, owners, operators and service providers continue to work together as they have in the past, we should expect a continuous improvement in reliability. An improvement, which can be, predicted with reasonable accuracy from the application of learning curve theory to early reliability data. We should look at the new data when it is published to see if our expectations are realised.
References: Richwine, R.R.& Lofe, J.J., Prediction of Equivalent Forced Outage Rates for Future Supercritical and Subcritical Electric Generating Units Using Learning Curve Theory, 1985. Curley, G.M., A View of Supercritical Fossil Steam Generator’s Performance – Past, Present and Future. Figure 1
O u ta g e R a te s v e r s u s Y e a r o f In itia l O p e r a tio n
1 2
34
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 11: RELIABILITY VERSUS DEMAND Robert R. Richwine Consultant BACKGROUND One key issue which often creates disputes (sometimes heated) between plant operators and its dispatch organisation is the impact on plant reliability resulting from the way in which the plant is dispatched. Most will agree that the “easiest” duty cycle that a fossil steam plant can follow is one where the unit is started, ramped slowly up to full load, and then left alone for extended periods of time. Likewise, the most severe duty, from a reliability perspective, is one in which the plant is frequently being started and shutdown. In fact many generating companies have placed constraints on the dispatch organisation, requiring a minimum period of downtime before a unit can be restarted following a shutdown in an attempt to minimise these cycles. A similar duty cycle that gives almost as much concern to plant operators is when a unit that has been designed for base load duty becomes “load-following” or is constantly being ramped up and down between its minimum and maximum rating. Although this duty cycle may not be as harmful as a full shutdown/start-up cycle, there are seldom many restraints placed on the dispatch organisation for this type duty beyond the plant’s prescribed ramp rate. Therefore, many units are required to operate almost continuously between overhauls (or forced outages) but seldom at a steady state load level. Several generation managers complained of this problem and the fact that their performance goals did not adequately reflect the impact this duty cycle had on their plant’s reliability (or Operations and Maintenance costs either, but that’s another study). My Reliability Engineering Department decided to undertake an analysis comparing generating plant’s reliability versus various load-following duty cycles to try to quantify the impact. ANALYSIS The first step was to identify the design and operational characteristics of the peer group we wanted to study. Because most of our company’s plants were coal-fired, we wanted only coal-fired units. From our other benchmarking studies, we decided to include only subcritical units. In addition we only wanted units that were in operation most of the time when not on planned maintenance or forced outages. Therefore, we selected units that had low Reserve Shutdown Hours (RSH) per year (RSH are hours during which the unit is available, but not dispatched for economic reasons). The next issue was to determine which plant performance factors we should compare. For the demand factor we decided to use the term Output Factor (OF). OF is the ratio of the plant’s actual generation divided by the possible generation if the plant had been dispatched at 100% load during every hour that it was actually in service (this differs from Capacity Factor (CF) in that CF’s denominator is the possible generation if the plant had been dispatched every hour during the study period, not just the service hours). Therefore, a low OF will indicate a plant that is ramped frequently between its minimum and maximum rating, whereas a high OF will be for a plant that is operated close to its maximum most of the time. This combined with plants having the select criteria of low RSH gave us the demand factor we wanted.
35
World Energy Council
Performance of Generating Plant 2004 – Section 6
For the reliability factor we first considered the Equivalent Forced Outage Rate (EFOR), a measure of reliability during the time the unit was actually required to generate. However, we realized that since EFOR included the equivalent impact of plant partial forced outages it would also influence the OF variable (the plant could not generate at 100% even if was needed). Therefore, we decided to use the Forced Outage Rate (FOR), which only includes full forced outages and would give us truly independent variables. RESULTS After obtaining the required data from the North American Electric Reliability Council’s (NERC) Generating Availability Data System (GADS) we plotted the regression line of FOR vs. OF as shown in the figure below.
RELIABILITY vs. DEMAND
NERC GADS
36
World Energy Council
Performance of Generating Plant 2004 – Section 6
Much to our surprise the regression line was horizontal! This seemed to indicate that there was no correlation between Forced Outage Rate and Output Factor. However, we have learned that when there is an apparent disagreement between the results of a statistical analysis and plant operator’s opinions, we should dig a little deeper before claiming that our analysis is superior to their first hand experience. Therefore, we divided the FOR into its two constituent pieces: Failure Rate and Repair Time. When we plotted Failure Rate against Output Factor we found a statistically significant relationship with the lowest OF plants having failure rates 2.5 times greater than plants with the highest Output Factors. However, when we plotted Repair Time against Output Factor we again got a highly statistically significant correlation but with a reverse slope such that plants with high OF had Repair Times 2.5 times greater than plants with low OF’s. These opposing trends had the effect of cancelling out each other so that the resulting FOR curve was horizontal. So we concluded that both first hand plant operator’s experience as well as our statistical analysis yielded valuable insights into the issue. And the results of this analysis have helped both the generation and dispatch organizations better understand the effects of their actions on plant reliability. NEW QUESTIONS But these results have led to further questions: • •
Why do low Output Factor machines have much higher Failure Rates? Why do high Output Factor machines have much longer Repair Times?
It has been speculated that for plants with low OF’s, the failure modes they experience tend to be control problems as the plants are constantly experiencing transient conditions, and are much different from the failure modes for high OF plants which would only come off line for major equipment problems with long repair times. In addition these control problems may tend to be “trips” from which the plant can often recover in a short amount of time. In addition high OF plants are obviously the most economic plants to operate so that often times the dispatch organisation will try to get a plant to “hang on until the weekend” or another lower cost time period. The result of this can be additional damage to the plant, often requiring longer time to repair than if the plant had declared a forced outage immediately upon discovering the problem (it may still be the economic choice, but clearly the plant’s reliability targets would be negatively impacted. Additional analysis will be required to determine if the above ideas are truly the root causes of this relationship between reliability and demand or if some other reasons are more important. References: Establishing Availability Goals using Benchmarking Techniques, IJPGC Conference, 1995. Peer Unit Benchmarking: Assessing Factors Affecting Availability, IJPGC Conference, 1992. The Importance of Data Recording and Analysis in Managing Availability Improvement, WEC Performance of Generating Plant Conference, 1985.
37
World Energy Council
Performance of Generating Plant 2004 – Section 6
38
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 12: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 1 – AWARENESS Robert R. Richwine Consultant US$80 Billion per year PLUS 1 Billion tonnes of CO2 Reduction This is the potential positive economic and environmental impact that the World Energy Council estimates would result from closing the gap between the average performance currently being achieved by the worldwide fleet of generating plants and the level of performance that “the best class” plants are achieving. It has also been proven that the greatest performance gains for the least cost are obtained through better management of the generating plants (see our Case Study “Design or Management – Which Influences your Plant’s Reliability Most?”). This month’s case study will begin a four part series that will demonstrate how top performing generating companies collect, share and analyse performance data and have achieved improved the performance of their plants as a result. During many of the workshops conducted by the World Energy Council’s (WEC) Performance of Generating Plant (PGP) Committee, we have been presented with the results of numerous successful Performance Improvement Programmes implemented by companies from many regions throughout the world. While each programme has been unique it its details, we have observed that most have a few key steps in common. Also, and what is pertinent to these monthly Case Studies, these common steps all emphasise how performance data contributes to improved generating plant performance. This month’s Case Study will describe in general the four steps common to successful Performance Improvement Programs: 1) Awareness, 2) Identification, 3) Evaluation, and 4) Implementation. In addition it will examine in detail how performance data are used in Step 1, Awareness. The following three monthly case studies will examine data usage in Steps 2, 3 and 4. COMMON STEPS IN SUCCESSFUL PERFORMANCE IMPROVEMENT PROGRAMMES Step 1 – Awareness The first step in successful Performance Improvement Programmes is to make company executives, managers, and generation staff aware of the potential for improvement that exists at each of the company’s plants and the importance of achieving that potential. Later in this case study two of the most important of the many possible actions that can be taken to enhance the awareness of that potential will be described in more detail:
39
World Energy Council
Performance of Generating Plant 2004 – Section 6
1) Benchmarking of current plant performance; 2) Economic Value of Improved Performance. Step 2 - Identification The ability to identify a wide range of “best practice” options aimed at performance improvement is an increasingly important requirement in managing power plants in today’s cost-conscious business environment. In the past it was perhaps adequate to simply identify the “best” technical option, obtain financing for that option and then implement that option at the lowest possible cost. However, in today’s increasingly market-based business environment, an entire range of viable options must be identified, so that the most cost-effective option can be determined. The effective use of every available channel to identify those viable options, from original equipment manufactures to consultants recommendations, to local plant staff insights (perhaps the most valuable of all) is necessary to be able to determine the best use of the company’s available (but limited) resources. Next month’s case study will describe in more detail some of these “identification” techniques being used by ”best in class” companies. Step 3 – Evaluation After all viable options have been identified (Step 2), their cost and technical impacts on plant performance must be estimated and combined with the worth of unit improvement (Step 1) so that each option can be economically evaluated. All financially justifiable projects can then be prioritised across unit, plant and company levels to ensure that the most cost-effective options are chosen. Decision support tools that can automate this activity are also needed to expedite this process and provide the necessary documentation quickly and cost-effectively. A future case study will describe this process in more depth. Step 4 – Implementation After reviewing the output of the Evaluation process (Step 3) the most cost effective set of options must be chosen for funding. Final decisions must also include each candidate project’s intangible aspects as well as its economic aspects. Intangibles historically have included factors such as employee moral, corporate image, safety, customer satisfaction and environmental impact (although environmental impact is increasingly able to be quantified in monetary terms so that projects with high positive environmental impact will have higher benefit to cost ratios and are more likely to be chosen for implementation). Financing for those projects that are finally chosen must be arranged and the projects designed and installed. Performance goals for the plants where the projects are installed must be set giving proper consideration to the expected performance improvement stated in the prioritisation analysis. Finally, the actual results achieved after the project has been installed should be calculated and compared to the expected impact. This should be done so those successful projects can be repeated at other plants and unsuccessful projects rejected or modified to increase their chances of success elsewhere. This post installation comparison of expected verses actual results will then be fed back into the awareness and identification steps and thereby closes the loop on the process of continuing improvement.
40
World Energy Council
Performance of Generating Plant 2004 – Section 6
AWARENESS The first step in successful Performance Improvement Programmes is to create awareness in the company’s generation executives and staff of the opportunity for improvement in their plant’s technical performance and the economic and environmental benefits resulting from that improvement. Accurate and consistent data plays a vital role in creating that awareness. Although data is used in numerous ways in the awareness step, this discussion will focus on two of the most important: 1) Benchmarking of current plant performance and 2) Economic Value of Improved Performance. BENCHMARKING Previous case studies have described examples of benchmarking and how benchmarking has helped generating companies: •
•
•
•
• •
Set realistic, achievable goals - In setting performance goals it is important to have an aggressive goal that causes the plant management to strive to improve, yet one that is possible to achieve. It is also important that this goal-setting process be objective and easy for everyone to understand so that the results are accepted as reasonable plant expectations. Comparing the plant’s historical performance to the performance of its peers will help in selecting these goals; Identify opportunities for improvement - By analysing plant system and component data from similar (peer) plants, we can find areas within the plant that we should focus on that have the most potential for cost-effective improvement projects. This will be further discussed in next month’s case study when we focus on Step 2) Identification; Give advance warning of potential problems – Analysing data from other similar (peer) plants can also alert us to problems that they have faced that we may encounter in the future. Our Case Study 1 describes one way in which a progressive generating company used peer data to proactively develop cost-effective actions in order to prevent, detect or minimise the consequences of High Impact-Low Probability (HILP) events; Trade knowledge and experiences with peers – By identifying other plants with similar designs and modes of operation to our plants, we can begin to directly communicate with them in order to share our experiences and knowledge. Adapting their successful programmes to our unique situation will give allow us to “get down the learning curve” much quicker; Determine appropriate incentives - Often, plant management is given financial or other types of incentives based on their plant’s performance. Benchmarking can help to quantify the appropriate incentives to offer; Quantify and manage performance risks - As we move into a more competitive business environment, it will be necessary to identify, quantify and manage risks, instead of simply avoiding risks. The insight and data coming from benchmarking will be invaluable in attaining the necessary proficiency in risk management (this is a very complex area that merits a case study of its own which we will be publishing later in the year).
41
World Energy Council
Performance of Generating Plant 2004 – Section 6
ECONOMIC VALUE OF IMPROVED PERFORMANCE One of the most important activities in any awareness campaign is to estimate the short and long term value the company receives from the improvement in each aspect of a unit’s performance; i.e. availability, reliability (forced outage rate), efficiency (heat rate), auxiliary equipment power requirements, capacity, etc. In addition, as market-based emissions trading processes evolve, it will become possible to forecast the financial impact to a unit’s bottom-line profitability due to its environmental performance. Understanding the value to the company has proven highly motivating to the generation staff in developing better day-to-day decisions process. This worth data will also be used during Step 3) Evaluation, when we will seek to justify and prioritise the many improvement projects competing for the plant’s limited resources. Value of improved performance of the company’s generating plants is derived from one or more of the following areas: •
•
•
•
Reduced fuel costs – Improving the efficiency of an individual plant means the plant uses less fuel, resulting in an obvious savings. Also, improved reliability of a more efficient plant will allow that plant to replace the generation that would otherwise have to be generated from a less efficient plant. Therefore, the total fuel cost for the company would be less. In addition the emissions in both of the above cases would be less in order to meet the same customer demand for electricity; Deferred new plant construction costs – Increasing the reliability of the company’s plants has the effect of delaying new capacity construction required to meet increasing demand since the current fleet is able to generate more electricity. At one company whose installed capacity was in excess of 30,000 MW, a one percentage point increase in reliability allowed the deferral of over 300 MW of new capacity for a documented savings of over US$100,000,000 dollars; Reduced reserve margin criteria – When the reliability of a company’s generating plant fleet is low, the expansion planning organisation often must increase the reserve margin in order to compensate in order to provide a reasonable level of customer service reliability. At one company whose Equivalent Availability was only 50%, the planning reserve margin was 100%! After implementing on an aggressive Availability Improvement Program, they were able to raise their availability to over 80% and were able to reduce their reserve margin to approximately 25%. Another large company raised their availability from 68% to 92+% and was able to lower their reserve margin from 40+% to 13%. These savings are carried forward indefinitely into the future if the company is able to sustain the higher levels of generating plant reliability. Higher customer service reliability – In many countries around the world, low levels of customer service reliability translate directly into low economic growth. Giving current and potential future customers the confidence that the electricity supply is reliable will help to ensure long-term economic prosperity for the company and the country. While this area is much more difficult to quantify than the first three (above), it can be higher than the other three combined where customers are experiencing a high incident of electricity service interruptions.
42
World Energy Council
Performance of Generating Plant 2004 – Section 6
While the forecasts of the economic value of improved performance are not easy to make, it is important to make the estimates and then publicise the results throughout the company. At one company where a formal Availability Improvement was successfully implemented, one plant staff executive commented “ The Availability Improvement Programme gave an insight into the value of availability, changing the plant staff’s perception of generation and demonstrated the value of a reliable operating plant”. CONCLUSIONS Several important conclusions can be reached based on the experience of generating companies worldwide that have improved their generating plant’s performance: • • • • • • • • •
Performance improvement can be achieved; Performance improvement requires a systematic, comprehensive and continuous programme with strong commitment from executive management; A strong data collection and analysis programme is a vital element in any successful Performance Improvement Programme; Some equipment replacement, refurbishment or upgrades will be required; The majority of the focus should be on improving management practices; The addition of advanced technology power plants combined with improvement in management practices are complementary pieces of the total improvement puzzle; New advanced technology plants will only reach their full inherent design potential if the most effective management practices are applied; Performance improvement of existing power plants is a proven, cost-effective way to increase the energy producing capabilities of a utility while producing substantial environmental benefits; A successful Performance Improvement Programme starts with creating the awareness in the minds of generation executives and staff of the potential for improvement and its value to the company and the country.
CASE STUDY 13 Case study 13 focuses on the second step in the performance improvement process: Identification. It describes techniques for identifying options that might be considered for evaluation and implementation (steps 3 & 4) with the emphasis on how performance data can be used to help in task.
43
World Energy Council
Performance of Generating Plant 2004 – Section 6
44
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 13: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 2 – IDENTIFICATION Robert R. Richwine Consultant This month’s case study will continue a four part series that will demonstrate how top performing generating companies collect, share and analyse performance data and have achieved improved performance of their plants as a result. Last month’s case study described in general the four steps common to successful Performance Improvement Programmes: and examined in detail how performance data are used in Step 1, Awareness. This month we will focus on Step 2, Identification. The following two monthly case studies will examine data usage in Step 3, Evaluation and Step 4 Implementation. COMMON STEPS IN SUCCESSFUL PERFORMANCE IMPROVEMENT PROGRAMMES Step 1 - Awareness The first step in successful Performance Improvement Programmes is to make company executives, managers, and generation staff aware of the potential for improvement that exists at each of the company’s plants and the importance of achieving that potential. In the previous case study, two of the most important of the many possible actions that can be taken to enhance the awareness of that potential will be described in more detail: 1) Benchmarking of current plant performance; 2) Economic Value of Improved Performance. Step 2 - Identification The ability to identify a wide range of “best practice” options aimed at performance improvement is an increasingly important requirement in managing power plants in today’s cost-conscious business environment. In the past it was perhaps adequate to simply identify the “best” technical option, obtain financing for that option and then implement that option at the lowest possible cost. However, in today’s increasingly market-based business environment, an entire range of viable options must be identified, so that the most cost-effective option can be determined. The effective use of every available channel to identify those viable options, from original equipment manufactures to consultants recommendations, to local plant staff insights (perhaps the most valuable of all) is necessary to be able to determine the best use of the company’s available (but limited) resources. This month’s case study will describe in more detail some of these “identification” techniques being used by top performing companies.
45
World Energy Council
Performance of Generating Plant 2004 – Section 6
Step 3 – Evaluation After all viable options have been identified (Step 2) their cost and technical impacts on plant performance must be estimated and combined with the worth of unit improvement (Step 1) so that each option can be economically evaluated. All financially justifiable projects can then be prioritised across unit, plant and company levels to ensure that the most cost-effective options are chosen. Decision support tools that can automate this activity are also needed to expedite this process and provide the necessary documentation quickly and cost-effectively. A future month’s case study will describe this process in more depth. Step 4 – Implementation After reviewing the output of the Evaluation process (Step 3) the most cost effective set of options must be chosen for funding. Final decisions must also include each candidate project’s intangible aspects as well as its economic aspects. Intangibles historically have included factors such as employee moral, corporate image, safety, customer satisfaction and environmental impact (although environmental impact is increasingly able to be quantified in monetary terms so that projects with high positive environmental impact will have higher benefit to cost ratios and are more likely to be chosen for implementation). Financing for those projects that are finally chosen must be arranged and the projects designed and installed. Performance goals for the plants where the projects are installed must be set giving proper consideration to the expected performance improvement stated in the prioritisation analysis. Finally, the actual results achieved after the project has been installed should be calculated and compared to the expected impact. This should be done so those successful projects can be repeated at other plants and unsuccessful projects rejected or modified to increase their chances of success elsewhere. This post installation comparison of expected verses actual results will then be fed back into the awareness and identification steps and thereby closes the loop on the process of continuing improvement. STEP 1 - AWARENESS STEP 2 – IDENTIFICATION There are many ways in which historical plant data can be of invaluable assistance in identifying trouble spots and viable options for improving a plant’s performance. The following are some of the techniques used by top performing generating companies: • • • • • • • •
Performance Reporting; Benchmarking Analysis; Computer Modelling; Root Cause Analysis; Coal Quality Evaluation; Plant Examinations; High Impact, Low Probability (HILP) Event Reduction; Trend Analysis.
46
World Energy Council
Performance of Generating Plant 2004 – Section 6
PERFORMANCE REPORTING In addition to the reporting of basic indices such as Availability, Reliability, Efficiency and Environmental performance for the plant as a whole, many companies will report the plant’s systems, sub-systems and equipment trends in order to detect changes and focus attention on areas with the greatest potential for improvement. The following table is a simple example of one type of report for a group of similar plants:
Unplanned Unavailability (%) by plant System Plant Boiler Tubes
TurboGenerator
Feed Water
Fuel Burn
Boiler Other
Total
“A”
2.32
0.23
0.76
0.14
1.14
0.35
4.94
“B”
0.73
0.14
0.00
0.04
0.00
0.00
0.91
“C”
2.26
0.0
0.00
0.01
0.01
0.99
3.27
“D”
0.97
0.02
3.80
0.01
1.07
0.42
6.29
By compiling this data over time you can detect trends in areas that warrant closer examination.
BENCHMARKING ANALYSIS In addition to the table shown above a benchmarking analysis can be performed, comparing the plants’ unplanned unavailability against its peers both at the overall plant level as well as the system (or sub-system, equipment or component) level. Now we can add a line or lines to the above table that show the performance of the peer plants at the average, best quartile, best decile, etc. levels in order to determine where to focus our efforts.
47
World Energy Council
Performance of Generating Plant 2004 – Section 6
Unplanned Unavailability (%) by plant System (With peer plant’s results added)
Plant
Boiler Tubes
TurboGenerator
Feed Fuel Boiler Other Water Burn
Total
“A”
2.32
0.23
0.76
0.14
1.14
0.35
4.94
“B”
0.73
0.14
1.56
0.04
0.00
0.00
2.47
“C”
2.26
0.0
0.00
0.01
0.01
0.99
3.27
“D”
0.97
0.02
3.80
0.01
1.07
0.42
6.29
“Peers” (Average)
3.92
0.55
1.15
0.44
1.56
1.88
9.50
“Peers” 2.56 (Best quartile)
0.32
0.81
0.33
1.01
1.55
6.58
“Peers” 1.18 (Best decile)
0.12
0.23
0.11
0.50
0.77
2.91
By examining this data we can clearly see that the area with the most potential for costeffective improvement is plant “D” in the feed water system (no surprise since it’s the largest unavailability area anyway). However, it may come as a surprise that the area with the next greatest potential (as compared against the best performing peers) would be at plant “B” in the feed water system (the values shown here are for demonstration purposes only and are not actual values for any particular plant or peer groupings of plants). Some generating companies have also set up colour charts of the trends of their plants, which have proved helpful in focusing on areas with greatest potential for improvement. A future case study will explore benchmarking at the plant system and equipment levels in more detail. COMPUTER MODELLING This technique is often valuable in initial design configuration decision-making as well as to identify areas of an existing plant where cost-effective improvements are likely. Several commercially availability computer programmes are available that require: • A functional description of the plant’s equipment configuration (e.g. a reliability block diagram) showing the amount of equipment redundancy and the percentage that the plant’s capacity is reduced when one or more components are out of service; • Failure rate and repair time for each of the components;
48
World Energy Council
Performance of Generating Plant 2004 – Section 6
•
Storage time available (if any) from the onset of the equipment failure until the plant must be removed from service or its capacity reduced; • The amount of annual Planned Outage Hours. Output of the programme is the expected availability and reliability of the plant and the amount of time that the plant is expected to be at various levels of capacity. In addition a ranking of the systems, subsystems and components is given based on their contribution to the plant’s unavailability. Numerous technologies have been modelled including Nuclear, Fossil-Steam, Gas Turbines, Combined Cycles, and Hydro among others. From the output of the programme the engineers can pinpoint areas of the plant where they can increase its performance at the least cost. ROOT CAUSE ANALYSIS This systematic technique involves an in-depth investigation of the underlying causes of in-service failure of the plant’s equipment and is extremely beneficial when formulating options for preventing future similar occurrences. The methodology depends on accurate data to determine the true cause of the failures and has been successfully applied at numerous companies around the world. COAL QUALITY EVALUATION Often times a coal-fired plant is designed with the expectation that it will burn coal of a certain quality but later is required to burn the “lowest-priced” coal available. As a result its performance (availability, efficiency, costs, environmental) is degraded. This technique uses a computer programme that models the plant’s performance for the design coal quality and then calibrates and validates itself for the actual fuel being burned. It can then evaluate proposed new coals of differing qualities to determine the “total” cost of these fuels including the operational costs in addition to the delivered cost. The tool has been used in many applications including evaluations of various levels of “washing” the coal, expected benefits from “additives”, determination of appropriate contractual penalties placed on suppliers who deliver coal outside the specified quality ranges, etc. The following figure shows how one coal purchasing decision at one company was changed as a result of using the model.
49
World Energy Council
Performance of Generating Plant 2004 – Section 6
As can be seen if the purchasing company only considered the annual delivered cost, coal “A” would be the least expensive by US$4,200,000 per year. However, when the operating cost is included (cost impact on O&M, availability, efficiency, etc.) then coal “B” becomes the least expensive by a total of US$6,700,000 per year. PLANT EXAMINATIONS One of the most effective techniques that have been employed is to utilise teams of experts to review the operations, maintenance and management practices at a company’s plants. This team, varying from 4 to 8 people depending on the situation, is charged with comparing what is being done at the plant with what, from their expertise, should be done (this is not the same as a traditional plant ‘audit” where the focus is on comparing what the plant is doing against what their procedures say they should be doing). Each examiner spends several days with their counterpart at the plant. Typically one reviews overall management, one reviews operations, one reviews maintenance and one reviews engineering, although occasionally a specialist in an area of known problems will be used (such as water chemistry, coal-handling, etc.). There are two keys to successfully using this concept: 1) The personnel conducting the examination must be the best in their respective fields and be recognized as the best by their plant counterpart; 2) The results of the examination must be used to help the plant improve it performance and not as a method to find someone to blame for past failures. And the plant staff must be convinced that the results will be used as such. A future case study will describe this process in more detail.
50
World Energy Council
Performance of Generating Plant 2004 – Section 6
HIGH IMPACT LOW PROBABILITY (HILP) EVENT REDUCTIONS At any power plant there are many potential catastrophic (high impact) events that can cause extended forced outages. Although taken individually no one event is very likely to occur (low probability), when added together the results can add substantially to any plant’s unavailability. Good management, however, using data from other plants can often find cost-effective ways to prevent, detect or mitigate the effects of these events. For a detailed summary of one company’s successful HILP reduction programme see our Case Study 1. TREND ANALYSIS Knowing what has happened in the past is very important, but it is not sufficient to be content with that knowledge. Rather we must apply that knowledge of past plant performance to help us gain a better understanding of what the future will be. To do that, trend analysis of performance data can yield revealing information about the likelihood of future events and allow us to make better decisions. Our monthly case studies have previously summarised several of these trend analyses including: • •
• • • •
Peak Season Reliability – This study demonstrated how many plants achieve a higher level of reliability during periods of maximum demand when they are most valuable; Design or Management – Which influences your plant’s reliability most? – This study demonstrated that between 75%-80% of a plant’s reliability were due to its management practices with only 20%-25% due to design or mode of operations differences; Generating Unit Availability Following Planned Outages – This summary described the results of a study showing that a plant’s reliability is often much worse in the hours and days immediately following a planned overhaul; Availability Impact of Flue Gas Desulphurisation Systems – This summarized a study assessing the impact of FGD systems on plant performance; Estimating New Technology Reliability – This study described how “learning curve theory” has been used to predict new technology’s future performance; Reliability versus Demand – This study explored the relationship between Output Factor (how hard the plant is running when it is running) and its reliability.
CONCLUSION From these many and varied examples of ways performance data has been effectively used by generating companies (and these are by no means all of the ways data has been used) it should be apparent that this data forms the backbone of any successful Performance Improvement Program, especially in Step 2 – Identification that was this month’s focus. CASE STUDY 14 Case study 14 describes in detail the third step in the performance improvement process: Evaluation. It describes ways of economically justifying and prioritising the options identified is Step 2 Identification using performance data.
51
World Energy Council
Performance of Generating Plant 2004 – Section 6
References: Richwine, R.R. The Importance of Data Recording and Analysis in Managing Availability Improvement, presented at the World Energy Council Conference.
52
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 14: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 3 – EVALUATION Robert R. Richwine Consultant This month’s case study will continue a four part series that will demonstrate how top performing generating companies collect, share and analyse performance data and have achieved improved performance of their plants as a result. COMMON STEPS IN SUCCESSFUL PERFORMANCE IMPROVEMENT PROGRAMMES Step 1- Awareness The first step in successful Performance Improvement Programmes is to make company executives, managers, and generation staff aware of the potential for improvement that exists at each of the company’s plants and the importance of achieving that potential. Two of the most important of the many possible actions that can be taken to enhance the awareness of that potential will be described in more detail: 1) Benchmarking of current plant performance; 2) Economic Value of Improved Performance. Step 2 - Identification The ability to identify a wide range of “best practice” options aimed at performance improvement is an increasingly important requirement in managing power plants in today’s cost-conscious business environment. In the past it was perhaps adequate to simply identify the “best” technical option, obtain financing for that option and then implement that option at the lowest possible cost. However, in today’s increasingly market-based business environment, an entire range of viable options must be identified, so that the most cost-effective option can be determined. The effective use of every available channel to identify those viable options, from original equipment manufactures to consultants recommendations, to local plant staff insights (perhaps the most valuable of all) is necessary to be able to determine the best use of the company’s available (but limited) resources. Step 3 – Evaluation After all viable options have been identified (Step 2) their cost and technical impacts on plant performance must be estimated and combined with the worth of unit improvement (Step 1) so that each option can be economically evaluated. All financially justifiable projects can then be prioritised across unit, plant and company levels to ensure that the most cost-effective options are chosen. Decision support tools that can automate this activity are also needed to expedite this process and provide the necessary documentation quickly and cost-effectively. This month’s case study will describe this process in more depth. 53
World Energy Council
Performance of Generating Plant 2004 – Section 6
Step 4 – Implementation After reviewing the output of the Evaluation process (Step 3) the most cost effective set of options must be chosen for funding. Final decisions must also include each candidate project’s intangible aspects as well as its economic aspects. Intangibles historically have included factors such as employee moral, corporate image, safety, customer satisfaction and environmental impact (although environmental impact is increasingly able to be quantified in monetary terms so that projects with high positive environmental impact will have higher benefit to cost ratios and are more likely to be chosen for implementation). Financing for those projects that are finally chosen must be arranged and the projects designed and installed. Performance goals for the plants where the projects are installed must be set giving proper consideration to the expected performance improvement stated in the prioritisation analysis. Finally, the actual results achieved after the project has been installed should be calculated and compared to the expected impact. This should be done so those successful projects can be repeated at other plants and unsuccessful projects rejected or modified to increase their chances of success elsewhere. This post installation comparison of expected versus actual results will then be fed back into the awareness and identification steps and thereby closes the loop on the process of continuing improvement.
STEP 1 - AWARENESS (For details of Step 1 – AWARENESS see CASOM 12) STEP 2 – IDENTIFICATION (For details of Step 2 – IDENTIFICATION see CASOM 13) STEP 3 – EVALUATION
Following the benchmarking activity and the estimating of the economic value of performance improvement described in the Awareness step and the identification of options for improved performance discussed in the Identification step it is now time to determine which options offer the most benefits for the least cost. This month we will examine the process for deciding which options should be selected for two types of projects: 1) Major capital projects and 2) day-to-day Operations and Maintenance projects/decisions. MAJOR CAPITAL PROJECTS These type projects are typically evaluated once a year during the annual budget process and selected projects are often implemented during major overhauls when the unit is out of service at planned times. There are two key components in these evaluations: • Justification; • Prioritisation.
54
World Energy Council
Performance of Generating Plant 2004 – Section 6
JUSTIFICATION The first hurdle that any proposed project should have to overcome is to answer the question “Is this a good project to implement; Yes or No?” There are many standard economic evaluation methods that can help answer this question including Life Cycle Benefit to Cost Ratio, Net Present Value, Internal Rate of Return, and Payback Period as well as several others. Most of these use the same three elements in various ways: • Impact; • Value; • Cost. For example, the equation for the Benefit to Cost Ratio is BCR = (Impact X Value)/Cost = Benefits/Cost Whereas the equation for Net Present Value is NPV = (Impact X Value) – Cost = Benefits – Cost (These are of course very simplified versions of these equations. A more complex methodology is required to include other factors such as the time value of money, etc. These more complex methods are covered in detail in the WEC workshops). IMPACT The impact that a proposed improvement project is expected to have on future plant performance is the area in the evaluation step where historical performance data is used most extensively. In estimating impact we must “predict the future” in various performance areas expected to be influenced by the project such as: • Reliability; • Availability; • Efficiency; • Auxiliary Power Requirements; • Operations and Maintenance Cost; • Capacity; • Environmental; • Other (voltage control, ramp rate, etc.); • Intangibles. In forecasting the degree of these impacts we must estimate what will happen if we don’t implement the project versus what will happen if we do implement the project. The difference between these estimates will be the impact we use in the evaluation. In making these estimates we will rely heavily on both our own plant historical data as well as data from industry sources for other similar plants. But we cannot be satisfied with only the historical data. Rather, we must use that data to understand what has happened in the past and why, but then project that into the future. This key concept can be expressed in the following relationship: Past conditions gave past results as future conditions will give future results.
55
World Energy Council
Performance of Generating Plant 2004 – Section 6
Therefore the better we understand the relationships between historic plant performance and the conditions that gave rise to that performance, the better we will be able to accurately estimate the future performance of our plant, with and without the proposed improvement project. Many of our previously published case studies give specific studies describing these cause/effect relationships and especially how data was used to increase our understanding and application of them. We should also note that individual impacts can be either positive or negative. For example, we might be evaluating a project where we are installing an extra Boiler Feed pump to improve reliability (positive reliability impact) but we will have to increase the annual Maintenance budget (negative O&M cost impact) to properly maintain it. It is important in include all impacts in the evaluation, both positive and negative, in order to develop an objective, credible evaluation. VALUE After the impact have been estimated we can then multiply each impact times its respective economic value to get the benefits for that area and then sum all the benefits to get the total benefits expected to be realised if we implement the project: BENEFITS = SUM (Impact X Value) for each impact area COST The final element is to estimate the capital cost of the project. Typically this will be the budget cost necessary to implement the improvement project and include cost such as materials, engineering, installation, overheads and removal costs (if any). It only includes the initial cost of the project and not any subsequent O&M cost (those will be included as negative impacts in the O&M portion of the benefits). Simple evaluations might only include the “overnight” capital costs; however, more accurate analysis will include the finance charges associated with the capital cost (for example the Cost might be calculated as the Present Value of Revenue Required (PVRR) for the project). INTANGIBLES Whenever possible an attempt should be made to quantify all impacts in monetary terms. However, many projects have other, non-quantifiable, impacts that should be considered such as morale, safety, company image and environmental (in some countries of the world many of the environmental impacts resulting from improvement projects can be quantified in monetary term and can be included in the tangible benefits; in other countries they cannot be quantified in monetary terms but should be fully described and included among the intangibles). When the decision maker is considering the project for implementation he/she will now have the advantage of having all relevant information needed for the justification of the project.
56
World Energy Council
Performance of Generating Plant 2004 – Section 6
PRIORITISATION The second and more difficult hurdle that a proposed improvement project must overcome is to be selected for installation from among the many competing projects at its plant as well as competing projects from other plants within the company and sometimes even with competing capital projects from other divisions within the company (e.g. transmission, distribution, facilities, etc.). It must be recognised that there are only limited resources available within any company at a particular time and it is vital that the projects selected will make the best use of these resources (money, time, manpower, etc.) In order to make this optimal selection a different question must be asked. During Justification we asked: “Is this a good project – Yes or No?” Now we must answer a more difficult question: “If the company doesn’t have the necessary resources to implement all justified projects this year, what will hurt the least to delay until a later date”. One technique that can help answer this question is to calculate the total benefits that will not be realised if the project is delayed until the next window of opportunity (typically one year; on occasion, however, projects can only be performed in conjunction with a turbine overhaul – therefore, the delay might be 5 or more years). Then a calculation can be made that estimates the financial consequences of the delay. For a one-year’s delay this would be the difference between the company’s “cost of capital” and the forecast “inflation rate” of the project’s cost times the current “overnight” capital cost of the project. This means there will actually be a financial savings realised if the project is delayed. Therefore, the ratio of lost benefits to financial savings is a measure of the economic impact due to delaying the project. If that ratio is greater than 1.0 then the project should be implemented as soon as possible and the higher the ratio, the more important economically it will be to implement the project immediately. As a simple example, consider a project to replace a pump that is nearing the end of its life: The analysis might show a high life cycle benefit to cost ratio (eventually the pump will be completely worn out and will have to be replaced if the plant is to keep operating). However, the risk incurred by delaying the replacement by one year may be evaluated as low so that the lost benefits would also be low, relative to the life cycle value. Therefore, the ratio would be low, causing it to be ranked below other projects that having greater economic necessity for immediate implementation (a higher ratio). This “Delay Benefit to Cost Ratio” ranking of all projects being considered can then be used to decide which projects will have to wait until additional resources are available. These techniques for evaluating and prioritising capital projects are also covered in more detail during our WEC workshops.
57
World Energy Council
Performance of Generating Plant 2004 – Section 6
OPERATIONS AND MAINTENANCE DECISIONS Every day every generating plant staff around the world is faced with making numerous Operations and Maintenance decisions that can significantly influence the technical and economic performance of the plant and its company. Although individually each decision may only affect performance slightly, the cumulative effect of thousands of these decisions every year often determines whether the plant is a top performing plant. As we have shown in our Case Study: Design or Management- Which influences your plant’s reliability most? Superior management leads to superior day-to-day decisions which results in superior performance. Furthermore, it was concluded that management can account for up to 80 percent of the gap between a top performing plant and a poor performing one, whereas design or even mode of operation may only account for as little as 20 percent. Therefore, if your benchmarking analysis reveals that your plant has an opportunity for performance and economic improvement, an in-depth review of management methods might offer substantial improvements for relatively little cost compared to major equipment replacements. To assist their management in making better O&M decisions some companies have developed and are using advanced decision-support computer tools that combine nearterm forecasts of the value of performance improvement with the technical consequences of a variety of options when facing day-to-day O&M decisions. Since the value of performance improvement can vary dramatically from one hour to the next depending on fluctuations in demand and supply, the economic optimal solution is not always obvious. As an example, at one company a mid-merit plant had an average value of availability of US$40,000 per day during the non-peak season, but US$160,000 during the peak season! And some peak season days could be over US$320,000 (days when supply is lower than normal and demand is higher than normal)! Therefore, it is obvious that during the peak season plant management can justify spending 4-8 times more money to avoid an outage (or reduce the duration of an outage) than for the same event during the non-peak season. Plants used primarily for peaking generation will have an even greater variation (a factor of 40 and above have been observed) and even true base loaded units have shown a 1.5-3 variation in value. By combining the forecast values of plant performance with the technical consequences of viable O&M best practice options the plant staff can determine which option is the economic optimal in the thousands of individual O&M decisions they are required to make every year. SUMMARY This month we have focused on how to economically evaluate capital improvement projects using justification and prioritization techniques. In addition we have discussed how similar techniques are used when facing the difficult day-to-day Operations and Maintenance decisions every plant faces thousands of times each year. By using historic performance data to understand the relationship between past conditions and past performance we can be more accurate in predicting what the future results will be for various options we are considering. Those predictions, combined with the forecasts of economic value of performance improvement will allow our plant’s staff and corporate
58
World Energy Council
Performance of Generating Plant 2004 – Section 6
executives to make better decisions that will lead to best-in-class technical and economic performance. CASE STUDY 15 Case study 15 describes in detail the fourth step in the performance improvement process: Implementation. It describes the process of deciding which projects to approve for installation using the results from Step 3 but also including other less tangible aspects of each proposed project. In addition it examines how the plant goal setting process is affected by which projects are chosen. It also focuses on techniques for comparing the actual results achieved from each implemented project against the expected results used in the evaluations (Step 3) and feeds these results back into Step 1 (Awareness), Step 2 (Identification) and step 3 (Evaluation), thus closing the loop in the Performance Improvement Process.
59
World Energy Council
Performance of Generating Plant 2004 – Section 6
60
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 15: PERFORMANCE DATA TO PERFORMANCE IMPROVEMENT: ANSWERING THE US$80 BILLION PER YEAR QUESTION STEP 4 – IMPLEMENTATION Robert R. Richwine Consultant This month’s case study will conclude a four part series that has demonstrated how topperforming generating companies collect, share and analyse performance data and achieve improved performance of their plants as a result. This month we will consider details in Step 4, Implementation. All of these steps are covered in greater depth in the workshops the WEC presents at locations around the world. For information about our workshops see the section at the end of this Case Study. COMMON STEPS IN SUCCESSFUL PERFORMANCE IMPROVEMENT PROGRAMMES Step 1 - Awareness The first step in successful Performance Improvement Programmes is to make company executives, managers, and generation staff aware of the potential for improvement that exists at each of the company’s plants and the importance of achieving that potential. Two of the most important of the many possible actions that can be taken to enhance the awareness of that potential will be described in more detail: 1) Benchmarking of current plant performance; 2) Economic Value of Improved Performance. Step 2 - Identification The ability to identify a wide range of “best practice” options aimed at performance improvement is an increasingly important requirement in managing power plants in today’s cost-conscious business environment. In the past it was perhaps adequate to simply identify the “best” technical option, obtain financing for that option and then implement that option at the lowest possible cost. However, in today’s increasingly market-based business environment, an entire range of viable options must be identified, so that the most cost-effective option can be determined. The effective use of every available channel to identify those viable options, from original equipment manufacturers to consultants recommendations, to local plant staff insights (perhaps the most valuable of all) is necessary to be able to determine the best use of the company’s available (but limited) resources. Step 3 – Evaluation After all viable options have been identified (Step 2) their cost and technical impacts on plant performance must be estimated and combined with the worth of unit improvement (Step 1) so that each option can be economically evaluated. All financially justifiable projects can then be prioritised across unit, plant and company levels to ensure that the most cost-effective options are chosen. Decision support tools that can automate this 61
World Energy Council
Performance of Generating Plant 2004 – Section 6
activity are also needed to expedite this process and provide the necessary documentation quickly and cost-effectively. This month’s case study will describe this process in more depth. Step 4) Implementation After reviewing the output of the Evaluation process (Step 3) the most cost effective set of options must be chosen for funding. Final decisions must also include each candidate project’s intangible aspects as well as its economic aspects. Intangibles historically have included factors such as employee moral, corporate image, safety, customer satisfaction and environmental impact (although environmental impact is increasingly able to be quantified in monetary terms so that projects with high positive environmental impact will have higher benefit to cost ratios and are more likely to be chosen for implementation). Financing for those projects that are finally chosen must be arranged and the projects designed and installed. Performance goals for the plants where the projects are installed must be set giving proper consideration to the expected performance improvement stated in the prioritisation analysis. Finally, the actual results achieved after the project has been installed should be calculated and compared to the expected impact. This should be done so those successful projects can be repeated at other plants and unsuccessful projects rejected or modified to increase their chances of success elsewhere. This post installation comparison of expected verses actual results will then be fed back into the awareness and identification steps and thereby closes the loop on the process of continuing improvement. STEP 1 - AWARENESS (For details of Step 1 – AWARENESS see CASOM 12) STEP 2 – IDENTIFICATION (For details of Step 2 – IDENTIFICATION see CASOM 13) STEP 3 – EVALUATION (For details of Step 3 – EVALUATION see CASOM 14) STEP 4 – IMPLEMENTATION Following the first three steps, AWARENESS, IDENTIFICATION and EVALUATION, the final step in successful Performance Improvement Programmes is IMPLEMENTATION. Included in this step are the following activities: • • • • •
Selecting the most cost-effective improvement options; Arranging financing for the options chosen; Setting performance goals based on the options chosen; Comparing the results achieved to the expected results; Feeding back these results into Steps 1, 2 & 3.
62
World Energy Council
Performance of Generating Plant 2004 – Section 6
SELECTING Many times the total resources required for implementing all justified projects (Step 3) will be greater than are available in any particular year. Therefore, the executive decisionmaker must select from among many “good” projects that group of projects, which will make best use of the company’s limited resources (money, manpower, time, etc). The prioritised list of justified projects compiled in Step 3, Evaluation, can be extremely valuable in assisting him in making those difficult choices. However, that list should only be seen as a starting point. The decision-maker must also incorporate his assessment of the value of the “intangible” impacts of each project so that his choices reflect the combination of the tangible plus intangible value of the proposed projects. Because the key concept in achieving and sustaining performance improvement is making the “best use of available resources”, using these evaluations derived from performance data is a vital part of the process. FINANCING Another important effect that having these objective, defendable project evaluations has had is that several companies have actually been able to obtain additional resources to implement justified projects that otherwise would have been delayed until a future date. Those responsible for obtaining financing for improvement projects have used these evaluations to help convince lending institutions that their loans will be safe and that the systematic process that the company uses ensures that the most cost-effective projects will be done first. GOALS Setting proper goals is one of the most vital actions that must be accomplished if the plant is to achieve its objective of improving performance. Setting the goals too high without providing the necessary resources to accomplish them creates a defeated attitude in the plant staff even before they can begin. Using the benchmarking results discussed in Step 1, Awareness, and combined with the forecasts of performance value (Step 1) can help determine the plant’s “Optimum Economic Performance” (OEA). This process will also estimate the money required to achieve and maintain the plant’s OEA. If there is insufficient funding provided to achieve this OEA goal (each plant’s unique point of diminishing returns) then with the actual funding provided, the plant’s “Achievable Economic Performance” can be calculated and used for the performance goal. In this way there is direct linkage between the knowledge at the plant level (proposed improvement projects and their cost, impacts and value to the company), decisions made at the company executive level as to how much money is available and which projects are ultimately selected, and back to the plant’s performance goals objectively determined by which projects are implemented. These linkages will help to ensure that all decisions made throughout the generation organisation will be made with consistent overall objectives. COMPARISONS All too often when the decisions are made to implement selected projects, the evaluations are forgotten and assumed to be of little additional value. Nothing could be further from the truth! A process should be set up that compares the “expected” cost, impact and value
63
World Energy Council
Performance of Generating Plant 2004 – Section 6
of each project (as predicted in Step 3 – Evaluation) against the “actual” cost, impact and value after the project has implemented. An example of the value in comparing predictions against actual results can be found in our Case Study – Predicting Unit Reliability Comparing what actually happened to what was predicted to happen is not done to lay blame for projects that failed to live up to their predictions or even to reward those who advocated successful projects, but rather to learn from both successful and unsuccessful projects in order to improve the process. By doing this we can more confidently assess the potential for the implementation at other plants of successful projects and eliminate or modify unsuccessful ones. FEEDBACK After all of the other activities in the four steps (Awareness, Identification, Evaluation, and Implementation) have been completed, results should be fed back into each of the steps. Obviously if the implemented projects were successful, a new benchmarking analysis should confirm the improvement. Also, the new performance data will be incorporated into the database and used to identify viable options for consideration at other plants. More accurate evaluations will be able to be performed, improving the chance that in the future the available resources will be spent in the most cost-effective manner. So we see that a successful Performance Improvement Programme is a continuous process as depicted in the figure below:
i tif n tio ca
Performance Improvement
tio n Pr io r it iz a
Im pl
em en ta
tio n
Aw a
en Id
re ne ss
Common Elements In Successful Performance Improvement Programs
64
World Energy Council
Performance of Generating Plant 2004 – Section 6
CONCLUSIONS This month’s Case Study concludes the four part series on successful Performance Improvement Programmes covering Awareness, Identification, Evaluation, and Implementation. Companies that have applied the process detailed in this series have made substantial progress in improving and sustaining the performance of their power plants. Of course, this process will not guarantee that your plants will achieve the same high levels of performance, but it will give you a blueprint for success. It is also my personal observation that the exact same programme is not necessarily the optimal one for every country. Each country faces its own set of challenges and opportunities so that a successful programme must be tailored to meet these unique conditions. Furthermore, a successful programme is one that is created and supported from within each generation organization and not imposed from outside. However, the general framework as discussed in the last four Case Studies is an excellent starting point for any organisation seeking to improve their plant’s performance.
References: Richwine, R. R., The Importance of Data Recording and Analysis in Managing Availability Improvement; World Energy Council Conference, Rome, Italy. Neibo, R. J., Richwine, R. R., Managing Generating Unit Availability. Caffrey, G., Richwine, R. R., Prioritizing Expenditures for Availability Improvement. Richwine, R. R., Tutorial: Applying Economic Processes for Day-to-Day Plant Decisions, WEC Workshop, Prague, Czech Republic. Richwine, R. R., Optimum Economic Performance: Reducing Costs and Improving Performance of Nuclear Power Plants. Richwine, R. R., Performance Improvement in Coal-fired Power Stations Neibo, R. J., Richwine, R. R., Peer Unit Benchmarking: Assessing Factors Affecting Availability Curley, G. M., Neibo, R. N., Richwine, R. R., Jenkins, A. K., Establishing Realistic Availability Goals using Statistical Benchmarking Techniques. Glorian, D., Neibo, R. J., Salvaderi, L., Unit Benchmarking Demonstrates Compatibility of European and North American Data Collection Systems, WEC Congress.
65
World Energy Council
Performance of Generating Plant 2004 – Section 6
66
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 16: ARE RELIABILITY MEASURES UNRELIABLE? PART 1 Robert R. Richwine Consultant INTRODUCTION For over 30 years concerns have been raised periodically about the definitions of traditional measures and indices of power plant reliability. While these concerns are most pronounced in peaking and cycling technologies (especially Gas Turbine and Combined Cycle), base-load technologies are affected even more. Although there have been numerous attempts to devise more suitable indices, consensus has not yet been reached. In recent years the need to develop and use new reliability indices, which more accurately reflect the current market place, has taken on a high degree of urgency. It is brought on by the need of large power consumers for lower electricity prices in order to compete in the global economy. To meet this need electricity generators are being compelled to reduce their costs. Decision-making at all levels is being affected and the old “technical” definitions of reliability should be amended to incorporate economics in order to link better plant performance with the actual cost of electricity supply. Instead of measures that are calculated over both demand and non-demand periods, new reliability terms should consider only the hours that the plant would have been dispatched and the financial consequences to the company’s bottom line from the failure to generate during those hours. This month’s case study is the first in a two part series that examines in detail the traditional reliability measures to demonstrate their shortcomings in actual application. Next month we will review a new metric, Commercial Availability, where it is being used and consider some of the implications arising from its usage. THE HISTORIC PROBLEM Among the traditional measures of plant reliability have been the Equivalent Availability Factor (EAF), the Forced Outage Factor (FOF) and the Equivalent Forced Outage Rate (EFOR) in North America, China and some other countries and the Unit Capability Factor (UCF) and the Unplanned Capability Loss Factor (UCLF) in Europe and South Africa among others. Those measures that are “factors” (EAF, FOF, UCF, UCLF,) use as their denominator the entire time period being considered (typically one year) without regard to whether or not the unit was required to generate. Therefore, for non-baseloaded units, these factors can lose their relevance (and the more cyclic the demand is, the greater the effect). So if a Gas Turbine unit is used exclusively for meeting peak demand periods it may only be required to generate just a few hundred hours a year. If it were unavailable during 25% of those hours it would still have high an EAF and UCF and a low FOF and UCLF. For example if a peaking unit was required to generate 100 hours per year but experienced forced outages during 25 of those demand hours (and no other outages over the 8760 hours in the year) it would still have a EAF and UCF of (8760-25)/8760 x 100 = 99.71%
67
World Energy Council
Performance of Generating Plant 2004 – Section 6
and a FOF and UCLF of (25)/8760 x 100 = 0.29%. Those numbers might look good on paper but the reality is that the unit could only produce 75% of the power required of it. So these factors don’t come close to describing the unit’s ability to produce its rated capacity when demanded. Of course for true baseloaded units such as many nuclear units who generate every hour they are available, these factors come much closer to depicting the unit’s “real” reliability. The terms Forced Outage Rate (FOR) and Equivalent Forced Outage Rate (EFOR) were introduced in an attempt to resolve these difficulties (FOR and EFOR differ only in that EFOR considers the “equivalent” impact that forced deratings have in addition to the full forced outages that is all FOR considers. In this simple example with only full forced outages we will examine the FOR). The equation for FOR is: FOR = (Forced Outage Hours)/(Forced Outage Hours + Service Hours) X 100 For the example given above the actual service hours are 75 so that the FOR would be: FOR = (25)/(25 + 75) X 100 = 25% The complement of the FOR might be considered to be the unit’s reliability so that Reliability = 100% - 25% = 75% So it appears that FOR (and EFOR when forced deratings are present) are good measures of a unit’s reliability. However, in actual practice it is extremely unlikely that all of the forced hours that a unit experiences during the course of a year are during its demand period. (In our example all 25 Forced Outage Hours were assumed to occur during the 100 demand hours). Most times a forced outage will have some of the hours required to restore the unit to service occur during non-demand periods and some during demand periods. In our example the unit might have experienced five forced outage during 25 hours of its demand period (out of 100 hours total demand). However, it is likely that the time to restore to unit to full capability would average more than the five hours each during demand periods. It is much more probable that the total forced outage hours would be several times higher (some previous studies suggest that the average restoration time for a gas turbine forced outage is on the order of 24 hours). Therefore, if we use 24 hours as the average down time then the total forced outage hours reported would be 5 X 24 = 120 Hours. Now the FOR would be: FOR = (120)/(120 + 75) X 100 = 61.5% and the unit’s reliability = (100-61.5) = 38.5%, both of which are obviously unrealistic when attempting to use these statistics to make decisions requiring the expected reliability of units to be used. And yet these values are very close to actual FOR and EFOR statistics being reported for peaking types of generators! Does that mean that FOR and EFOR should not be used? Absolutely not! They are in fact reasonable indicators for base load or near base load types of generating units. However, for cycling or peaking units they are inadequate and new metrics were needed.
68
World Energy Council
Performance of Generating Plant 2004 – Section 6
A few years ago a modification of EFOR was introduced to attempt to resolve this problem. The term Equivalent Forced Outage Rate-Demand, EFOR(d), was developed which only used that portion of a unit’s forced outage (or derating) that occurred during demand periods. As we saw in the earlier example, that would resolve the issue nicely. However, demand periods are not currently part of standard reporting systems so that an approximation technique was devised using a MARKOV approach. Although not perfect, this technique does result in a more accurate calculation of EFOR(d). See Ref. 3 for a more detailed discussion of this technique. THE NEW PROBLEM As the industry moves into a more competitive market-based business environment, the reliability indicators must be able to have a direct linkage between a plant’s performance and the corporate or portfolio cost and/or profit of electricity. They should incorporate the large (often very large) variation in the value of a unit’s reliability. For example, in previous efforts the author has made to quantify the value of availability improvement of an individual unit within a large company, he has found that even for efficient coal-fired units there can be a factor of 100 between its value during a low demand period compared to its value during a high demand period and other generators are having unexpected forced outages. Even nuclear units have a significant variability in their value during different times. This variation will inevitably result in different economically optimal decisions at different times. As an example of this the following example was developed from actual data at one large generating company using value-based availability to measure its plants’ performance: On a Tuesday morning a small boiler tube leak was detected at one of its large efficient coal-fired units that ran close to base load. Two options were identified: 1) Remove the unit from service immediately and repair the tube as quickly as possible so as to minimize the downtime of the unit; 2) Continue to operate the unit until the weekend when the demand is lower and the cost impact of the unit’s unavailability is less per hour. However, the unit would be exposed to a risk of a longer duration outage due to possible additional tube damage. When the plant staff evaluated these options for an event that occurred during the nonpeak season, it was found that the cost was minimised by choosing Option 1. This was because the differential between the weekday and weekend-day cost per hour of this unit’s unavailability at this time was not enough to offset the increased risk of a longer outage if the unit was operated until the weekend. Total Cost – Non-Peak Season 1) Option 1 - US$115,000; 2) Option 2 - US$184,000.
69
World Energy Council
Performance of Generating Plant 2004 – Section 6
This choice also had the effect of minimising the forced outage hours resulting from this event. However, when the exact event occurred during the company’s peak season the cost were: Total Cost- Peak Season 1) Option 1 - US$354,000; 2) Option 2 - US$265,000. In this case, Option 2 is clearly the best economic choice, even though it had the effect of reducing the plant’s availability beyond that of Option 1. As we can see, often the objective of maximising a unit’s “technical” performance (in this case availability) brings it into direct conflict with the company’s goal of minimising cost and/or maximising profitability. This example is only one of thousands of decisions that a power plant staff must make every year and illustrates the vital importance in developing performance metrics that establish direct linkages between the plant’s goals and the company’s overall financial objectives. In this way we will be encouraging the plant staff to make optimal economic decisions, not just ones that maximise their technical goals. None of the traditional statistics such as EAF, UCF, FOF, UCLF, FOR, EFOR or even EFOR(d) adequately make this linkage between technical and economic goals. However, some companies have begun using a new measurement technique called Commercial Availability that promises to do exactly that!
References: Richwine, R. R., Curley, G. M., DellaVilla, S. A., Lofe, J. J., Reliability Measures Unreliable…It’s Time for a Change; Published in the IGTI’s Global Gas Turbine News, 1998. Stallard, G. S., Salvaderi, L., Richwine, R.R., Spiegleberg-Planer, R., Glorian, D., Corrigal, M., Micali, V., Neibo, R., International Data Exchange Within The Global Power Industry – A Critical Activity for the Evolving Competitive Power Market, Published at the 17th World Energy Council Congress, 1998. Stallard, G. S., Micali, V., deSabastian, A. L., Salvaderi, L., Richwine, R. R., Glorian, D., Heithoff, J., International Availability Data Exchange for Thermal Generating Plant, published at the 18th WEC Congress, 2001. Stallard, G.S., Richwine, R.R., Salvaderi, L., Measuring Performance in a Co-Generation Plant, published at Power-Gen Europe Conference, 2000.
70
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 17: ARE RELIABILITY MEASURES UNRELIABLE? PART 2 – USING COMMERCIAL AVAILABILITY Robert R. Richwine Consultant INTRODUCTION Last Month we examined why traditional measures of a power plant’s availability/reliability (EAF, FOF, UCF, UCLF and EFOR) were inadequate in today’s increasingly competitive business environment. We also saw that a new measure, EFOR(demand), while an improvement, did not totally address the problem. This month we will review a new measure, Commercial Availability (CA), which has begun to be used by generating companies around the world. This statistic attempts to measure the impact a plant’s availability has on the company’s cost of generating electricity (and its profitability when the company operates in a market-type environment). We will also discuss some of the implications that result from the adoption of Commercial Availability as the primary availability measure. BACKGROUND The term Commercial Availability originated in the United Kingdom in the early 1990’s following the deregulation of its power industry into a “market” system. Since a plant’s availability only had value to its company if it could generate power at a profit, its availability was only measured during the times the market price was above the plant’s variable cost. Initially CA was not “weighted” with respect to the magnitude of this gap so that each hour that the unit was economically viable had the same influence on CA. Over time some users of CA have evolved the term to include the influence of the price/cost gap magnitude so that it can be a more accurate indicator of the plant’s impact on the company’s profitability. (E.g. during hours when the gap is US$20/MW-HR the plant’s actual availability would have ten times the influence as an hour in which the gap was only US$2/MW-HR). Therefore, CA attempts to measure the actual profit delivered by the plant relative to its potential profit if it had been able to deliver every MW-HR required of it at the actual market price (profit here is defined as gross margin, generally the difference between the plant’s variable production cost and the market price, or the system marginal cost in the case of regulated companies). Although numerous companies in many countries have begun using Commercial Availability as one of their primary measures of availability, there is currently no standard definition for its calculation. In fact, at a recent meeting of an industry group, a survey of those attending revealed that about 1/3 of the companies represented were using Commercial Availability at some level, but none calculated CA in exactly the same way. Clearly the industry is in great need of a standard definition. The Performance of Generating Plant (PGP) Committee of the WEC has had a Working Group exploring this issue as well as other questions related to the increasingly widespread use of Commercial Availability (Ref 3).
71
World Energy Council
Performance of Generating Plant 2004 – Section 6
IMPLICATIONS OF USING COMMERCIAL AVAILABILITY There will be a wide range of impacts on the way a company evaluates and manages its power plants resulting from the adoption of Commercial Availability and other tools/processes required to address market dynamics. This requires a different mindset and approach in applying data and new tools in both day-to-day and performance assessment decisions. Measures and actions must consider ways quantify and respond to different situations with differing economics. Yet the fundamentals of benchmarking remain relevant, although in new, modified forms as discussed below. •
Benchmarking – selection - Over the past few decades benchmarking has become a key tool in most top performing generating companies performance improvement efforts. Good technique is to first identify other “peer” plants whose design and operational characteristics are similar to the one we wish to benchmark. The WEC has used this advanced statistical technique, simultaneously analyzing over 50 plant features, to identify peer units from different parts of the world and then to compare their “traditional” reliability indices. Benchmarking Commercial Availability will require a new aspect of the plant to be included in the analysis to determine the optimal peer group. That new aspect is some indicator of the plant’s economic incentive to generate at different times. Since the greater the economic incentive to generate is, the better the plant’s reliability can be managed to meet the demand, then we will need a statistic that measures the unit demand and incorporate that into the peer group analysis. Some preliminary work has been done analysing a measure called Reserve Shutdown Hours (RSH) during different periods of the year (RSH is Hours the unit is available, but not required to dispatch). By adding this measure to the other design and operational features of a plant, we should be able to optimise the peer selection process, including demand aspects.
•
Benchmarking – comparisons - After we have selected the best peer group for our benchmarking analysis, what will we compare? The actual calculation of Commercial Availability (whichever definition is finally adopted) is likely to be highly dependent on the precise market price (or marginal cost for a regulated or controlled business environment) that exists per hour (or parts of each hour) and matched against the unit’s availability in those hours. Since that price (or cost) can and does fluctuate widely over the course of each day, week, month or year we would have to create a massive new database containing market cost in order to make the CA calculations. Furthermore, even if we did create such a database the actual CAs will probably not be appropriate to compare since the actual market prices in different regions would be likely to be very different. The WEC Working Group 6 is exploring this issue (see Ref. 3 for a discussion on their work to date). One of the approaches being considered is to calculate a term called Conditional Probability (CP). CP represents the likelihood (that’s the probability part) that the unit can deliver the requested amount of energy during a specified time period corresponding to that unit’s demand profile (that’s the conditional part) as indicated by its Reserve Shutdown Hours trends. CP, then, would be similar to the Equivalent Forced Outage Rate (demand) statistic but would likely be different during different demand periods. So what we would be doing would be to “benchmark” Conditional Probabilities of peer units and then select a goal CP as perhaps the best quartile or best decile or “Optimal Economic Availability” from the CP distributions of the peer units. Combining the goal CP and our unit’s unique economics we can then arrive at a “goal” Commercial Availability objectively
72
World Energy Council
Performance of Generating Plant 2004 – Section 6
without having to create any new data collection processes. •
Maximising Commercial Availability - This focuses ones attention on being available to generate when required by the market and when the income and profit potential is highest. Generating units are only maintained and manned to meet market need. The logical converse of this is that stations need not be maintained and manned for periods when they are not required by the market. The daily, weekly, and annual variations in demand for electricity means that it may be possible to reduce generating costs by allowing the units to remain unavailable overnight, at weekends, and for certain parts of the year. The plant is not required by the market and although technically unavailable, such periods have no effect on Commercial Availability. Furthermore, if overhauls, etc., which would affect technical availability can be scheduled for periods when a generating unit is not required by the market, these overhauls can be undertaken without affecting the Commercial Availability of the unit. We also will need to consider implications of capacity payments/obligations as they become more important elements in a deregulated market structure.
•
Design – New plant design is likely to be affected since we are no longer concerned with maximising traditional measures of availability or reliability, but in maximising profitability (or minimizing cost). One outcome of this different design philosophy will be to reduce the dependency on expensive equipment redundancy and instead install advanced equipment monitoring equipment. Since we are only interested in being available “when the plant is needed”, being able to better anticipate imminent equipment problems will give needed flexibility to plant management. Furthermore, even if we cannot control the timing of the event, communication of the increased likelihood of an outage will allow others in the organization (dispatch, trading, marketing, etc.) to take appropriate steps to minimize the financial impact of the outage. Operational “flexibility” also needs to be considered in design. With the addition of advanced control systems and online performance optimisation tools it is possible increase the plant’s capability to meet demanding load schedules, ramp rates, etc., thereby increasing the potential for sale of additional MW-HRS without compromising plant availability. In addition, since different regions have different economic conditions, the optimal economic design is likely to be different.
•
Other implications – There will be many other implications associated with adopting Commercial Availability including modifying the overall goals system for the plant to include the financial impacts of other performance parameters such as efficiency, Operations and Maintenance costs and environmental impacts, etc. Decision analysis tools using information scattered throughout the organisation are needed that combine the technical consequences of various courses of action with their economic impact on the corporate bottom line to give the decision maker all relevant information they need to make the best decision. Finally, it is necessary for the industry to recognise one likely result of using Commercial Availability in place of the traditional indices; that is, these traditional measures will almost surely appear worse. Regulatory agencies, financial institutions, insurance carriers and even the company’s own executives, board members, stockholders and customers must be included in the change process and “buy into” the new metric. Otherwise, how can we expect them to believe that although the measures they are used to monitoring are appearing worse, the company is actually doing better and delivering a lower cost and more profitable product?
73
World Energy Council
Performance of Generating Plant 2004 – Section 6
CONCLUSION For many years the electric generating industry has been aware that traditional measures or plant reliability need improvement, especially for cycling and peaking types of technologies. However, it has usually remained of academic interest to those of us closely involved in Reliability Engineering. However, new times are requiring new, more appropriate measures that link technical performance with financial results. The catalyst for this new interest in reliability measures is the evolving market-based business environment brought on by the need of our customers for lower electricity prices to help them meet the demands of the competitive global economy. Commercial Availability is one measure that has evolved to meet that need and has been successfully adopted by numerous companies around the world. For many years the World Energy Council’s Performance of Generating Plant Committee has had a Working Group investigating this and other issues relating to the increasingly competitive business environment generating companies around the world are facing. Their work is ongoing (Ref 2, 3) and we hope it will lead the industry to a more uniform application of this and other new measures of plant performance.
References: Richwine, R. R., Curley, G. M., DellaVilla, S. A., Lofe, J. J., Reliability Measures Unreliable…It’s Time for a Change, Published in the IGTI’s Global Gas Turbine News, 1998. Stallard, G. S., Salvaderi, L., Richwine, R.R., Spiegleberg-Planer, R., Glorian, D., Corrigal, M., Micali, V., Neibo, R., International Data Exchange Within The Global Power Industry – A Critical Activity for the Evolving Competitive Power Market, Published at the 17th World Energy Council Congress, 1998. Stallard, G. S., Micali, V., deSabastian, A. L., Salvaderi, L., Richwine, R. R., Glorian, D., Heithoff, J., International Availability Data Exchange for Thermal Generating Plant, published at the 18th WEC Congress, 2001. Stallard, G.S., Richwine, R.R., Salvaderi, L., Measuring Performance in a Co-Generation Plant, published at Power-Gen Europe Conference, 2000.
74
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 18: THE RELATIONSHIP BETWEEN SCHEDULED MAINTENANCE AND FORCED OUTAGES AND ITS ECONOMIC IMPACT ON SELECTING AVAILABILITY GOALS Robert R. Richwine Consultant BACKGROUND I prepared this study over 20 years ago at a time when generating companies, regulatory bodies and governmental agencies were considering using the term Equivalent Availability Factor (EAF) as a simple, measurable index for evaluating power plant reliability, even to the point of using it to determine rewards for exceeding or penalties for failing to reach some previously set value (Equivalent Availability Factor is similar to the term Energy Availability Factor used in Europe and some other countries). This study pointed out certain fallacies in equating increased availability with increased productivity (as measured by decreased production cost) without adequate attention given to how that increased availability is achieved. Since this study has not been updated to account for the recent general performance improvement of the industry as a whole, the actual algorithms derived may not be applicable today. However, I believe that the general trend observed in the analysis holds true and especially that the economic issue being addressed is still valid. INTRODUCTION The techniques for estimating a generating unit’s Optimum Economic Availability (OEA) have focused on comparing the unit’s unique “frontier” cost per percent improvement to its marginal value. While it has been generally accepted that the amount of unplanned outages a plant experiences is, in part, a function of the amount of scheduled maintenance, that relationship has not been well defined. This study explores that relationship and examines its economic impact on selecting availability goals. ANALYSIS Using data from the Edison Electric Institute’ database (a predecessor of the current GADS system administered by the North American Electric Reliability Council) for large coal-fired units from 1971-1978, a linear relationship between planned and unplanned unavailability was derived. The correlation coefficient was -0.895 indicating a strong relationship. Because the data is now over 25 years old and the industry availability has improved substantially since that time, the constant in the equation is no longer accurate. However, I believe that the slope of the curve remains close to the original calculated value of -0.5897 This means that for every additional percentage point that we increase our scheduled outages (the total of planned and maintenance outages and equivalent deratings) the forced outages and equivalent deratings decrease by 0.5897 percentage points.
75
World Energy Council
Performance of Generating Plant 2004 – Section 6
MAXIMISING AVAILABILITY As we can easily see, in order to maximise availability using the equation derived from this data we should minimise the amount of scheduled outages since for every percentage point we reduce our planned outages we could expect an increase in our unit’s availability of 0.4103 percentage points (1-0.5897) (Obviously this applies only within the limits of the data; outside of those limits the relationship is probably not linear). If maximising availability were the task of a generating company then this approach would be optimal. However, we are not simply charged with maximising availability, but rather with minimising the cost of producing and delivering power to meet our customers demand (or maximising profit in the case of deregulated systems). Although maximising availability has usually been considered to be synonymous with minimising cost, the following analysis will demonstrate that this assumption is not necessarily valid. MINIMISING COST It has long been recognized that an hour or day or percent of scheduled outage is normally much less costly (in terms of replacement energy cost) than an hour or day or percent of forced outage, since a generating company attempts to schedule its scheduled outages during periods of minimum demand. For many years some generating companies have been calculating these hourly, daily, weekly and seasonal costs. An analysis of these costs has shown that for large base-loaded coal-fired units the “average” ratio of forced outage cost to scheduled outage cost varies between 3.0 and 4.5 (during unusually “tight” conditions between supply and demand this ratio can be much higher and the volatility can be much larger during peak seasons). Also, this cost ratio for more cyclic and especially peaking units is much greater (at least one and sometimes two orders of magnitude greater). Therefore, when we calculate cost of outages and use the lower ratio for large coal units of 3.0, we in effect change the slope of the equation (in monetary terms, not percent) by a factor of 3.0 so that it is now: Slope = 3.0 X -0.5897 = -1.7691 This means that for every dollar we “save” by reducing our scheduled maintenance, we can expect to “pay” an average of 1.7691 extra dollars in increased forced outages. So if our goal is to minimise total cost we should obviously increase our planned and maintenance outages so as to reduce our unplanned outage cost, regardless of the effect that might have on our availability. NEW MEASURE Some companies that have recognized this problem of using EAF to evaluate a plant’s performance have begun using a term called Commercial Availability (CA) which attempts to address the issue by “weighting” the traditional availability measures by some indicator of marginal or market cost. While this holds great promise to properly measure the true economic impact of a plant’s outages on the company’s financial performance, no single definition has been adopted by the industry as yet and there are other implications that have not been fully resolved.
76
World Energy Council
Performance of Generating Plant 2004 – Section 6
CONCLUSIONS Although this study is over 20 years old and the performance of North American power plants have improved substantially since that time, the general results of this study are still applicable today. That is that increasing availability is not necessarily synonymous with decreased cost (or increased profit) and that such an assumption by generating companies, regulatory bodies or governmental agencies may lead to decisions which are not in the best interest of the consumer. While this study was not done to downgrade increased availability as a desirable goal, it does demonstrate that the manner in which that goal is reached must be evaluated very carefully. To simply use the traditional measures of availability for judging performance can be misleading and possibly counterproductive to the true goal of minimising cost (or maximising profitability).
References: Richwine, R.R., Correlation of the Effect of Planned Maintenance on Unplanned Outages: The Economic Impact on Selecting Availability Goals, 1981, Portland Ore.
77
World Energy Council
Performance of Generating Plant 2004 – Section 6
78
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 19: STATISTICS OR UNDERSTANDING: WHICH ONE DO YOU BELIEVE? Robert R. Richwine Consultant This month’s case study is a little different. There are no derived or hypothetical equations or in depth general analysis. Rather, it is a story of how one young engineer with access to his company’s power plant performance data undertook a study that changed the maintenance philosophy of his company, eventually helping to increase the plant’s availability and saving millions of dollars. BACKGROUND Several years ago I led a consulting effort to create a Performance Improvement Programme at a large Northeast U.S. generating company. During the time I spent at their offices, a young reliability engineer asked me “Bob, where do you get your ideas from to perform the statistical analyses you’ve been showing us? I’ve been placed in charge of collecting our plant’s reliability data and compiling summaries for top management but I think we could do more analysis and bring more value to the company.” I replied that I was sure he was right about bringing more value by studying the data (see all of our previous WEC case studies published in the last 18 months) and that the way I got most of my ideas was from spending time at my company’s power plants talking with the plant staff about the issues that are of most concern to them. They generally have deep insight into those issues but often don’t have the time to explore them in detail. The next time I visited his company the reliability engineer told me that he had taken my advice and visited the plants and talked with the staff about their problems. He told me that a recurring issue identified at several of the plants was: • • • •
During major planned overhauls no outside support was allowed (budget constraints); Even with overtime the local plant staff was insufficient for the increased workload; In order to meet the planned outage schedule routine maintenance at the other units at that plant was often skipped; These other units suffered from an increase in forced outages as a result.
He then told me that he had set up a study comparing the reliabilities of the units at each plant during times other units were on major planned outages to their reliabilities when no other units were on planned outages. He then told me that the results of his statistical analysis showed no correlation and asked me what he should do. I asked him if he had more confidence in his statistical analysis or the understanding of the plant staff gained through years of actually living with the issue. He replied “Well, since you put it that way, I might question my analysis.” We then discussed his assumptions and decided that he should modify his original hypothesis to compare the reliabilities with a “lagging” effect of 1 – 3 months (this tests the idea that the units at the plant other than the one on a major overhaul would not begin to “feel” the effects of insufficient maintenance for some period of time).
79
World Energy Council
Performance of Generating Plant 2004 – Section 6
When I returned he showed me the results of his revised analysis, which showed a substantial reduction in the reliabilities of the “other” units 2-3 months following the overhaul of another unit at the same plant. By combining the results of his analysis with the economic impact of the reduced reliabilities, he was able to demonstrate conclusively to his company’s executive management that forbidding external support during a major overall was “false economy”, costing them much more in plant unreliability than they were saving in maintenance costs. A new policy allowing increased external support during major overhauls was then implemented, helping to lead to a substantial increase in their power plant reliability. CONCLUSIONS The purpose of this month’s case study was to demonstrate how generating companies have studied their own unique issues using their internal plant performance data (no national databases were required in this case, although the company did collect their data in a standard format for national reporting requirements which made the analysis much easier). Furthermore, substantial cost savings have been realised stemming from changes in Operations and Maintenance policies as a result of these studies. Another purpose is to encourage data analysts to get out to their power plants and talk to those plant staff members who have to live with their plants every day and find out what their issues are. They may not have the time to study the issues in detail or have the necessary statistical tools or expertise to use them or even the ability to express them in ways that the analyst can fully comprehend, but they will have a deep understanding that only comes from first hand knowledge accumulated over years of experience. The final purpose is to warn data analysts that if your statistical results don’t conform with the understanding expressed by your plant staff, I’d take another look at the analysis before I would challenge their knowledge. As we found in the preceding story, their understanding was the one to believe.
80
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 20: PEER SELECTION FOR BENCHMARKINGDOES IT MAKE A DIFFERENCE? Robert R. Richwine Consultant In a previous case study we discussed the importance of properly analyzing design and operational characteristics to select the best peer group for comparison in any benchmarking study. In the year since that study was published we have been asked one question several times: “Does it really make a difference?” In other words would we have different expectations for our plants if we used the traditional select criteria of fuel type and size range (for fossil steam plants) rather than the more statistically accurate criteria found using the technique discussed in the case study? To answer this question I have reviewed the results of several benchmarking studies we have done over the years using the advanced statistical technique and compared the resulting reliability distributions to the results using traditional criteria. In one early benchmarking analysis (in fact the client was the one for whom the technique was originally developed) we found that the single most important design feature affecting reliability was “criticality”, while the second most important feature was “vintage”. The following table compares the Equivalent Forced Outage Rate (EFOR) distribution for the early vintage supercritical units against more recent vintage plants: SUPERCRITICAL TECHNOLOGY EARLY VINTAGE EFOR (mean) EFOR (median) EFOR (best quartile)
RECENT VINTAGE
15.60% 12.17% 8.14%
9.68% 8.08% 5.47%
Since at these points in the reliability distribution (mean, medial and best quartile) the recent vintage plants show a substantially better EFOR than the early vintage plants, it would be highly inappropriate to include these recent vintage plants in any peer population if the unit we are benchmarking was an early vintage plant, since it would be compared against recent vintage plants that had clearly benefited from the “learning curve” of the early vintage designs.
81
World Energy Council
Performance of Generating Plant 2004 – Section 6
In another study completed just this year two fossil steam plants were analysed to find their statistically appropriate peer groups:
EFOR - PLANT A OLD CRITERIA (Coal; 100-199MW)
NEW CRITERIA
% difference
mean
6.47%
5.53%
-14%
median
4.78%
5.07%
+6%
best quartile
2.65%
3.26%
+23%
Here we can see that if we were to set a goal as the best quartile performers in our peer group, we would be setting unrealistically high expectations using the old criteria, compared to what we might set using the more appropriate peer group with new criteria. EFOR - PLANT B OLD CRITERIA (Coal; 800-1300MW)
NEW CRITERIA
% difference
mean
5.83%
7.63%
+31%
medial
4.55%
5.87%
+29%
best quartile
2.70%
3.97%
+47%
Clearly, these plants must have design and/or operational characteristics that create a more difficult challenge for their plant management to achieve the highest levels of reliability; a difficulty that might not be recognized through a benchmarking process that did not begin with a rigorous peer selection criteria analysis. Furthermore, the goals we might set using an inappropriate peer group may not be cost effective and we may end up spending more money than is justified to achieve these goals (for a more detailed discussion on optimising goals see our case study titled “Optimum Economic Availability”). Other plants studied have had just the opposite result. In those we may be setting our goals too low since the new peer group performs better than the one using the traditional criteria. These examples are not isolated cases; numerous other studies have demonstrated that while benchmarking is of unquestioned value in helping to achieve improved plant performance, we must first begin with as accurate a peer group as possible.
82
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 21: ESTIMATING A GENERATING PLANT’S FUTURE MAINTENANCE COST Robert R. Richwine Consultant BACKGROUND This month’s case study describes a technique that some international companies have applied in order to more accurately predict future maintenance cost for their generating plants without spending excessively. The need to accurately estimate these future costs has never been more important as the industry moves towards more market-based business environments. Even countries where there is no movement as yet toward more competition, there are still pressures to better anticipate future costs. APPROACH In addition to initial construction cost and fuel cost, there are three cost areas that need to be estimated: • • •
Recurring Operations and Maintenance (O&M) costs; Non-Recurring O&M costs; Retrofit Capital costs.
While we are usually confident that we can accurately predict recurring O&M costs, often we have little confidence in our ability to anticipate long term non-recurring O&M and retrofit capital cost. This study will focus on these last two areas and describe a low cost technique that has achieved reasonable results. APPROACH The process consists of a structured, detailed build-up of costs from the equipment level (bottom up view) as the basis of all analysis and comparisons. In addition some companies have been able to validate these bottom up values using cost benchmarking (top down view). When the top down and bottom up views yield comparable results, we will be confident that we can trust the estimates. However, when the two views are far apart, we must look deeper into the reasons to achieve the necessary confidence. BOTTOM UP VIEW A team formed of plant experts, equipment specialists and local plant staff first developed a list of the equipment that they would consider. Depending on the time available and the desired detail of analysis the equipment list could be extensive or relatively concise. The team then reviewed the equipment on the list and made estimates of the following items for each piece of equipment:
83
World Energy Council
• • • •
Performance of Generating Plant 2004 – Section 6
Probability of requiring major repair/replacement/upgrade over the time period being considered (typically 15-20 years); Probability of repair vs. replacement vs. upgrade (total probability=100%); Cost range for each in item 2 (above); Range of years in which the expenditure could occur and probability.
These estimates were conditional on certain operational assumptions such as the unit remaining base-loaded or changing to two shift operations, fuel switching etc. Using the above data for all equipment included in the final list, a Monte Carlo simulation was run thousands of times in order to develop a probability distribution of future retrofit capital and non-recurring operations and maintenance cost (for a more detailed discussion of the Monte Carlo technique used in this process see Reference 1). This probability distribution, plotted against time, can then be added to the recurring (routine) O&M cost to get a total maintenance cost probability distribution. Example The following example is for one piece of equipment; boiler controls. The team member most familiar with this equipment made the following estimate (with advice from other knowledgeable team members or even from outside sources such as vendors, etc.): BOILER CONTROLS 1) Probability of requiring significant costs over next 20 years 2)
a) Probability of repair b) Probability of replacement in kind c) Probability of upgrade
3) Cost Range for 2a 2b 2c
- 80% - 20% - 20% - 60% total = 100%
US$200,000 - US$400,000 US$750,000 - US$1,000,000 US$2,000,000 - US$4,000,000
4) Years from now in which cost could occur-
84
16 years – 20 % chance 17 years – 20% chance 18 years – 20% chance 19 years – 20% chance 20 years – 20% chance Total = 100%
World Energy Council
Performance of Generating Plant 2004 – Section 6
This same type information would be compiled for all other equipment included on the list. Using the estimated probabilities the Monte Carlo simulation is then performed that randomly “draws” one possible future for every piece of equipment and sums the total costs resulting from that “draw”. DRAW #1 EQUIPMENT
MAJOR COST
TYPE PROJECT
COST
YEAR
1) Boiler Controls
YES
Repair
US$250,000 19 yrs
2) Turbine Shaft
….
….
….
….
3) …. . . . 50) ….
….
….
….
….
….
….
….
….
TOTAL
US$8,340,000
Therefore, for draw #1 (possible future #1) the cost we would see would be US$8,340,000. DRAW #2 EQUIPMENT 1) Boiler Controls
MAJOR COST
TYPE PROJECT
No!
….
COST
YEAR
….
….
2) 3) 50) TOTAL
US$12,320,000
For draw #2 the cost would be US$12,320,000
85
World Energy Council
Performance of Generating Plant 2004 – Section 6
DRAW #10,000 EQUIPMENT 1) Boiler Controls
MAJOR COST YES
TYPE PROJECT Upgrade
COST
YEAR
US$2,500,000 17 yrs
2) 3) 50) US$21,750,000
TOTAL
The total for draw #10,000 is US$21,750,000 Each time a new draw is conducted (luckily the computer does this for us) a different cost is calculated, based on the random selection of each of the individual probabilities. The resulting 10,000 possible futures can then be compiled into a final cost vs. time probability distribution which give us a range of possible costs in addition to the average or “expected” value. The figure below gives a typical result of the cumulative cost (at the expected, 5th and 95th percentiles) vs. the year obtained from this process.
86
World Energy Council
Performance of Generating Plant 2004 – Section 6
USES There are many ways in which this data has been used including: • • • • •
As a basis for financial terms in establishing Power Purchase Agreements; Evaluating economic viability of older generating plants; Evaluating alternative plans for existing plants (re-powering, mothballing, etc.); Acquisitions of plants; Long term budgeting etc.
The cost for compiling this data has averaged less than 100 man-days per plant, much of which can be supplied by internal sources. This is much less that what would be required to perform detailed inspections of the plant’s major equipment using various nondestructive examination (NDE) techniques and which have difficulty predicting beyond the next few years. This is not to belittle the value of these NDE techniques, only to question their cost-effectiveness when used for long-term maintenance forecasts. CONCLUSIONS The methodology described in this month’s case study and in the references (below) for collecting, collating and analysing generating plant capital and major non-recurring maintenance cost provides a comprehensive, logical and cost-effective approach to estimating future costs by: • • • • • • •
•
Ensuring as far as possible that all possible cost items are considered; Providing a structured approach to considering each item; Allowing uncertainties in timing and cost to be taken into account; Concentrating resources in considering each component from different perspectives, rather than having to make definite decisions about what exactly will be done in the future with exact timing and costs; Utilising the synergy available from the depth and breath of knowledge of team members. Providing a solid foundation for future business plans that can be easily updated during review of such plans; Allowing sensitivities to be carried out with the probabilistic data provided. For example, the financial viability of a generating unit can be examined using expected cost and then reappraised using a worse case such as the costs at the 95 percent confidence level. The latter has a firm basis rather than simply a set percentage of the expected costs; Providing a simple process that will become better and faster over time as local staff becomes more experienced with it.
87
World Energy Council
Performance of Generating Plant 2004 – Section 6
References: Burke, Frank; Duffy, Peter; Richwine, Robert; Newton, Thomas, Projecting Long Range Generating Plant Maintenance Costs. Richwine, R., Quinn, D., Utilizing Plant Examinations in Availability Improvement Programs. Lowe, Steve; et. Al., Economic Analysis to Determine Retirement Dates of Older Generating Units. Caffrey, G., Richwine, R., Prioritizing Expenditures for Availability Improvement. Richwine, R., Jenkins, K., Optimizing O&M Cost to Maximize Profitability.
88
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 22: AGING OR VINTAGE WHICH IS MOST RESPONSIBLE FOR DIFFERENCES IN BOILER TUBE LEAK RATES?: PART 1 Robert R. Richwine Consultant
BACKGROUND For many years there have been discussions concerning the problems that large power plants have experienced due to boiler tube leaks, the single largest contributor to power plant unavailability. Some believed that aging was primarily responsible for the fact that older units have had higher unavailability due to boiler tube leaks than that experienced by newer units. Others have contended that the newer plants have benefited from design improvements and would not degrade to the same level of unavailability of the older plants, even when they reached the same age. To investigate the issue a task force was formed under the direction of the Generating Availability Trend Evaluation (GATE) Working Group of the North American Electric Reliability Council (NERC). This month’s case study summarises the results of that investigation. Next month’s will continue to explore additional aspects of the study in this key issue. ANALYSIS The task force decided to consider coal-fired, 400 MW and larger units, divided into the following 4 sub-groups: 1. Subcritical units that entered commercial operation in or before 1974. (SUB 74) (54 units) (early vintage large subcritical units) 2. Subcritical units that entered commercial operation (SUB 75) (107 units) (recent vintage large subcritical units)
in
or
after
1975.
3. Supercritical units that entered commercial operation in or before 1974. (SUPER 74) (74 units) (early vintage supercritical units) 4. Supercritical units that entered commercial operation in or after 1975. (SUPER 75) (22 units) (recent vintage supercritical units)
89
World Energy Council
Performance of Generating Plant 2004 – Section 6
RESULTS The following is a summary of some of the more significant results (for detailed results see Table 1 in Reference Boiler Tube Failure Trends – available upon request): Subcritical Units •
There was a dramatic improvement (~70%) in the boiler tube failure rate from SUB 74 to SUB 75 units.
•
This improvement in failure rate occurred for all tube types: • • • •
Waterwalls Superheater Reheater Economiser
-
78% 47% 75% 73%
•
There was a slight increase in average repair time (15%).
•
Overall improvement was 71% in unavailability due to tube leaks – almost three full percentage points.
Supercritical Units •
There was a large improvement (~50%) in boiler tube failure rate from SUPER 74 to SUPER 75 units.
•
This improvement in failure rate occurred for all tube types: • • • •
•
-
53% 32% 38% 38%
There has been a substantial reduction in average repair time: • • • •
•
Waterwalls Superheater Reheater Economiser
Waterwalls Superheater Reheater Economiser
-
16.7 hours reduction 25.4 hours reduction 16.1 hours reduction 25.0 hours reduction
Overall there was a 60% reduction in boiler tube leak unavailability – almost three percentage points.
90
World Energy Council
Performance of Generating Plant 2004 – Section 6
SUBCRITICAL VS. SUPERCRITICAL SUB 74 vs. SUPER 74 • • •
SUB 74 units were having more leaks than SUPER 74 units (0.7 leaks per year more); SUB 74 units were having much shorter repair times that SUPER 74 units (26.4 hrs per leak less); SUB 74 units were having lower unavailability due to boiler tube leaks than SUPER 74 units (0.76% lower).
SUB 75 vs. SUPER 75 • • •
SUB 75 units were having fewer leaks than SUPER 75 units (1.2 leaks per year less); The repair times for SUB 75 and SUPER 75 were approximately equal; SUB 75 units were having lower unavailability due to boiler tube leaks than SUPER 75 units (0.76%).
AGING These results demonstrated conclusively that the newer units, both subcritical and supercritical, experienced much less unavailability due to boiler tube leaks than their older counterparts. If the reason was “aging” then we might expect that these newer units would degrade as they grow older, leading to potential significant reductions in the energy producing capability of the fleet of new large base-loaded coal-fired units. Therefore, the GATE task force expanded the study to include 10 years of data, including the early years of operation of the older groups of units. Figures 1 and 2 show these results for subcritical and supercritical units respectively. These clearly demonstrate that there has not been any deterioration in boiler tube unavailability for either technology, but rather a slight improvement.
91
World Energy Council
Performance of Generating Plant 2004 – Section 6
So if aging is not the primary cause, what is? VINTAGE A second possibility considered by the GATE task force was that the vintage of these units could explain the differences in their performance. In order to test the vintage hypothesis, Learning Curve Theory was applied to the data. The results of learning curve regressions on the data exhibited very strong logarithmic relationships between boiler tube unavailability and commercial date of operation; i.e., learning curve behaviour. Figures 3 & 4 show the results of those regressions (for more details about these analyses see reference Boiler Tube Failure Trends.
92
World Energy Council
Performance of Generating Plant 2004 – Section 6
CONCLUSIONS • • •
There was no significant change in large (>400 MW) coal-fired units boiler tube leak unavailability over the years included in this study; Units that were placed into service since 1975 have averaged almost three percentage points lower boiler tube unavailability compared to units in service prior to 1975; The poorer performance of the earlier vintage units does not appear to be due to aging (i.e. wear out) but does closely follow the performance predicted by learning curve theory.
Therefore, the most likely answer to our original question “Which is most responsible for differences in boiler tube leak rate – Aging or Vintage?” is Vintage.
References: Richwine, R.R., Lofe, J.J., Mills J.B., Boiler Tube Failure Trends. Lofe, J.J., Richwine, R.R., Prediction of Equivalent Forced Outage Rates for Future Supercritical and Subcritical Electric Generating Units Using Learning Curve Theory.
93
World Energy Council
Performance of Generating Plant 2004 – Section 6
94
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 23: BOILER TUBE LEAK STUDY: PART 2 Robert R. Richwine Consultant
In last month’s case study we summarised part of a study performed by NERC’s Generating Availability Trend Evaluation (GATE) Working Group in which we demonstrated that the reduced unavailability due to boiler tube leaks experienced in newer power plants was due to vintage effects rather than aging of the older plants and was strongly correlated to learning curve theory. This month we will continue summarising that study by considering possible areas where learning could have been incorporated into the design, construction plus the operations and maintenance programmes of the newer plants that could have contributed to their improved performance. In addition we will examine statistical evidence that suggests that although the older plants might have more difficult challenges, nevertheless some are meeting those challenges and are achieving performances that are close to the top performers in the recent vintage class. LEARNING If learning explains the superior reliability of recent vintage plants what changes in design, construction or Operations and Maintenance might be responsible? Discussions with knowledgeable industry experts identified several possible areas: DESIGN • • • • • • • • •
Newer units have had much larger volume boilers with lower heat release rates; Specifications for older units were made during times when utility emphasis was on lowest initial cost; Older units had operational problems when using lower quality coal that originally designed to burn; Design of the older, first generation large units was scaled up from versions of smaller units; High load growth during years when older vintage units were being designed meant short cuts were used during the design process; Old sub critical units had circulation problems that have largely been eliminated in more recent designs; Boiler tube metallurgy has been improved; Increased involvement and influence by operations and maintenance staff during the design and procurement stage; Improved steam condition control systems.
95
World Energy Council
Performance of Generating Plant 2004 – Section 6
CONSTRUCTION Newer units have benefited from: • • • •
Better construction management; Increased quality assurance/control; Reduced site welds and total welds; Greater involvement and control by owners/operators on boiler construction techniques.
OPERATIONS AND MAINTENANCE* Newer units have also benefited from: • • • • • • • •
Improved operator training and procedures; Increased monitoring of boiler tube metal temperatures and adherence to design limits; Earlier detection of potential areas for failures; Better water chemistry control; Clearer understanding of root causes of failures; Reduction in operating boiler in overpressure condition to increase output; Improvement in welding techniques and weld verification; Better communication throughout industry on problems and cost-effective solutions.
*Although most Operations and Maintenance improvement techniques can be applied to older units, damage already incurred through poor practices in their early years may have already reduced the long-term potential reliabilities of these units. CONCLUSIONS In last month’s case study we reached the conclusion that vintage effects were more likely to explain the superior reliability of newer plants than aging effects of the older plants. Does this result mean that we must be satisfied with poor performance of older plants? Absolutely not!!! Some interesting statistics support this answer. There is a very large variation in the performance of the older units such that even though the “average” older unit is much worse that the “average” newer unit there are still a large percentage of older units with very good performance. Figures 1 and 2 display the cumulative probability distribution of boiler tube unavailability for the older and newer units.
96
World Energy Council
Performance of Generating Plant 2004 – Section 6
One striking statistic stands out: more than 1 out of 5 (21%) of the older units are performing better than the average value for newer units! This indicates that older units can be good performers if the necessary steps are understood and taken. A survey of companies with older units in the “good performer” category revealed that their unit’s good performance was no accident. Rather it was only through commitment and consistent application of resources (money, time, manpower) that these results were achieved (see our case studies for Jan – April 2003 for a detailed discussion of this successful process).
97
World Energy Council
Performance of Generating Plant 2004 – Section 6
Another important factor contributing to superior performance of top units is application of the work that the Electric Power Research Institute (EPRI) has done to identify the root cause and possible solutions for almost all boiler tube failure modes. Case studies of some of the applications of this work have shown that a company can be successful if they are persistent in their pursuit of good performance and refuse to accept less.
References: Richwine, R.R., Lofe, J.J., Mills J.B., Boiler Tube Failure Trends.
98
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 24: CHANGING GAS TURBINE DESIGN REQUIREMENTS Robert R. Richwine Consultant BACKGROUND The worldwide movement towards more competitive, market-based systems of power pricing is leading electric utilities and independent power developers to change their methods of making technology and manufacturer decisions to more fully integrate these new financial influences into those choices. This, in turn, should cause generation suppliers to rethink the processes they use in designing, developing and marketing their products so as to incorporate their client’s value system into their design trade-off options evaluations and to be flexible enough to more effectively tailor their products to match particular user applications. GAS TURBINE APPLICATIONS EXAMPLES As an illustration of the new relationships emerging between gas turbine designers and owners, we can look at three different gas turbine powered generating plants being developed. The owners and managers of each of these plants face different challenges in trying to meet their goals of improving their company’s profitability. Plant A consists of simple-cycle gas turbine generator sets, used almost exclusively for meeting the company’s peak load requirements. Each GT is expected to operate fewer than 200 hours per year at 4 to 5 hours per start. Plant B will be made up of gas turbine generator sets connected to heat recovery steam generators, which power steam turbines and will be used for daily cycling duty. The plant will therefore start up each day and run on average for 12 hours before being shut down each night. Plant C will be an integrated coal gasification combined-cycle power plant that, because of its thermal efficiency and low fuel cost, will be continuously dispatched at full load except when it is out of service for maintenance or because of a forced outage. PAST In the past most electric generating companies have been highly regulated and the price a company charged its captive customers was determined by taking its “prudently incurred cost” and adding a legally mandated return on investment (in the United States that might be 12-14 %) so that PRICE = COST (prudent) + PROFIT (mandated 12-14%).
99
World Energy Council
Performance of Generating Plant 2004 – Section 6
Therefore, in this business structure a generating company’s primary concern was whether or not an investment decision could be defended as “prudent”, regardless of the actual effect the decision had on the cost of electricity. With no real competition among generating companies, the rates one company charged had little effect on another’s decision making, leading to a strong risk-avoidance mindset (risk in this context refers to the outcome of an investment decision that in not absolutely certain). From a stockholder’s perspective risk-avoidance was the most appropriate criteria to use since a very successful “risky” decision (one that might give a 10 to 1 return but is not guaranteed) could only yield a profit legally allowed (12-14% in the US) while an unsuccessful investment might be entirely disallowed in a prudence review. Therefore, the maximum return a series of “risky” investment decisions could yield would be no greater than the return from a series of “safe” decisions. And because not all “risky” investment would be successful (otherwise they would not be risky) the return is very likely to be less. When considering a new plant in this environment, a generating company would assess its needs and the various technologies and manufacturers available in terms of the initial cost while specifying certain performance expectations to be demonstrated during relatively short-term tests. Little, if any, weight was given to differences in the long-term impact on the company’s cost due to such items as maintenance requirements, reliability, efficiency degradation over time, operator training and other factors. Gas Turbine designers, when evaluating design trade-offs, had to consider the evaluation process its clients would use and develop their products accordingly so as to make sales. Designs thus emphasised low initial cost (achieved in part by developing a standard design for all user applications) and high efficiency (achieved with high inlet temperatures, tight tip clearances on rotating blades and sophisticated controls and instrumentation systems, etc.). This did not necessarily lead to investment decisions that would lower the company’s cost of generating electricity, but they could be defended as “prudent” because that was the industry standard method of evaluating bids. FUTURE In the new evolving competitive business environment the price of electricity will be determined by the marketplace and the basic equation governing a generating company’s decision-making changes to: PROFIT = PRICE (market) – COST Now profit is the dependent variable (instead of price) and will force a basic change in the decision maker’s attitude toward risky investments since the higher risk options can lead to higher returns (profitability). Instead of avoiding risks, the decision maker must now become adapt at identifying, quantifying and managing risks in order to economically evaluate their cumulative impact on his company’s financial performance. Since the generator decision maker will be evaluating his purchasing options on the basis of their risk/reward ratio, the gas turbine designer must now identify all of the elements that will enter the generator’s decision process, transform them into useable cost figures and incorporate them into his design trade-off decisions. These elements will include all cost and revenue streams over the life of the project including performance incentives and penalties. In addition the designer, instead of designing one generic product for all
100
World Energy Council
Performance of Generating Plant 2004 – Section 6
applications (and requiring his marketing force to try and convince the customer that it’s best for his specific application), should now develop a series of products to meet the needs of particular power segments and customers. GAS TURBINE APPLICATIONS RECONSIDERED We can now reconsider our three plants: Plant A, a simple-cycle gas turbine plant needs a simple, low cost design with high starting reliability. There is less concern with efficiency since the plant will not generate many MW-HRS per year, but it needs a very low Operations and Maintenance Cost. Plant B, a combined-cycle daily cycling plant with its almost constant transient loading conditions, requires a more sophisticated instrumentation and controls system with increased emphasis on efficiency. This is balanced against the risk of increased forced outages due to its more difficult operating regime. Because this plant values Availability more than Plant A, additional expenditures can be justified for increased operator and maintenance personnel training, spare parts inventories, condition monitoring systems, outage planning and other needs. Plant C, the integrated coal gasification combined-cycle plant, will run almost continuously at full load. Ideally, it would start up; run for 8000+ hours at full load, if possible; then shut down for annual maintenance. The plant staff must be careful during start-ups to avoid blade tip rubs, etc. but does not have to worry about transient conditions until it is time to shut down. The primary concerns are efficiency and running reliability, and additional money can be justified for more elaborate on-line performance monitoring systems and other diagnostic tools to maximise the plant’s reliability and efficiency during steady state operating conditions. The following generic table indicates the relative* importance of several key factors influencing a power producer’s gas turbine equipment purchase decision according to its anticipated future duty cycle.
101
World Energy Council
Performance of Generating Plant 2004 – Section 6
RELATIVE* IMPORTANCE OF CHARACTERISTICS BY APPLICATION Gas Turbine Characteristic
Gas Turbine Plant Application Peaking
Cycling High (day)-Low (night)
Baseload
Availability
Low
High
Start-up Reliability
High
High
Low
Running Reliability
Low
Medium
Very High
Efficiency
Low
Medium
Very High
Initial Cost
High
Medium
Low
O & M Cost
High
Medium
Low
*The actual relationships among the factors listed in the table (plus other factors not mentioned) will depend on their value in the specific application. For example, a plant sited where fuel is relatively cheap would value efficiency less than a plant located where fuel is expensive. Another example might be for an off-shore oil platform where any interruption in generation could have overwhelmingly large financial consequences due to reduced oil production. Its value of availability would be much greater than for a similar facility that had a reliable source of back-up power. The important principle here is that these factors must be quantified in monetary terms according to the customer’s specific value system in order to provide the optimal design. SUMMARY Changes in the basic economic framework of power generation in many countries around the world, fostered by increasing competitive pressures, are altering the decision making criteria traditionally used by power generators. This has created incentives for increased cooperation between power producers and equipment manufacturers. The most successful manufacturers, in this author’s opinion, instead of using a one-design-fits-all approach, will take aggressive steps to understand and quantify their customer’s priorities and values and incorporate them into a range of designs, each one optimal from the owner’s perspective, in order to help meet the requirements of a changing business environment.
Reference: Richwine, R.R., Newton, T.U., Gas Turbine Design Requirements Changing for U.S. Utilities, published in the Global Gas Turbine News, Jan. 1996.
102
World Energy Council
Performance of Generating Plant 2004 – Section 6
CASOM 25: THE FUTURE IS NOT WHAT IT USED TO BE Robert R. Richwine Consultant INTRODUCTION We are often called upon to make estimates of future plant reliabilities for use by in a variety of ways from expansion planning and design optimisation studies to economic dispatch analysis to goal setting using benchmarking techniques and to Operations and Maintenance improvement of existing plants. As has been noted in previous case studies, historical reliability data of plants with similar design and mode of operations is a key element in forecasting these reliabilities. However, using only this data can introduce significant inaccuracies in that these data are statistical measures based on observed events during some previous time period, whereas what is needed is a prediction of what is going to happen in the future. This probability forecast is not necessarily the same as yesterday’s statistics. EXAMPLES As discussed in previous WEC case studies there are many examples that demonstrate the difference between historical and future power plant performance (for complete case studies visit http://www.worldenergy.org/forward.asp?page=pgpand select “See Previous Studies at the bottom of the banner page): Other examples that have not yet been summarised as WEC case studies include the effect on plant reliability of aging equipment or the impact due to changes in fuel quality. Also, many analysts agree that there is likely to be a significant impact on “traditional” reliability indices as countries move into a more de-regulated and competitive business environment (for more details see the monographs published by other working groups of the WEC Performance of Generating Committee on the WEC website www.worldenergy.org. Another aspect that may be very important is related to the impact of management on reliability in which the relationship between proactive and reactive maintenance cost is quantified and related to reliability. SUMMARY In summary in order to more accurately predict a unit’s future reliability we should start with a good understanding of the historic reliabilities of similar generating units as well as a comprehensive understanding of the prevailing conditions that influenced those reliabilities. But from there we will have to estimate what the future prevailing conditions will be and use our understanding of the past to develop better estimates of the future. Since we cannot be sure about what the actual future conditions will be, we might use a range of possible futures, weighted by their probabilities, and utilise statistical techniques such as Monte Carlo simulations in order to arrive at the most appropriate set of predictions.
103
