Generating Plant

Performance of Generating Plant:
New Realities, New Needs

A Report of the World Energy Council

AUGUST 2004

ii

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
iii

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, action-
oriented 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 Committees 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

iv
Officers of the World Energy Council
Antonio del Rosario Norberto de Franco Medeiros
Chair World Energy Council Chair Programme Committee

Philip Aiken Shige-etsu Miyahara
Vice Chair Sydney 2004 Vice Chair Asia

Franois Ailleret Kieran OBrien
Chair Studies Committee Vice Chair Europe

Asger Bundgaard-Jensen Fred Phaswana
Vice Chair Finance Vice Chair Africa

John Derrick Carlos Pierro
Vice Chair North America Vice Chair Latin America/Caribbean

Alioune Fall Gerald Doucet
Vice Chair GEIS Initiative 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.)
Cte dIvoire
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

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

v
PGP Committee Membership

Chairman: K Theis VGB PowerTech e.V (Germany)

Members:
Brazil: S.R. Fernandes Eletronuclear
China: W. Guangyao State Power Corporation
Egypt: M. Sharaf /K. Yassin Egyptian Electricity Authority
France: B. Manoha Electricit 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 Velazquez Comision Federal de Electricidad
Poland: A. Jaworski Elektrociep!ownie 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

United Kingdom
B. Celik
M. Grasby
Eregil Iron & Steel Works Co.
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 Salomn
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)
vi
TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS iii
WEC OFFICERS AND MEMBER COMMITTEES iv
PGP COMMITTEE MEMBERHIP v
FOREWORD vii

INTERNATIONAL AVAILABILITY DATA EXCHANGE Section 1
FOR THERMAL GENERATING PLANT
G. S. STALLARD, Black & Veatch (US), R. RICHWINE Consultant (US)
V. MICALI ESKOM (South Africa)

THERMAL GENERATING PLANT UNAVAILABILITY Section 2
FACTORS AND AVAILABILITY STATISTICS
M. CURLEY, North American Electric Reliability Council (US)

NUCLEAR GENERATING PLANT UNAVAILABILITY Section 3
M. SZIKSZAINE TABORI, International Atomic Energy Agency (IAEA)

PERFORMANCE OF HYDRO AND PUMP STORAGE Section 4
PLANT
T. MOSS, Eskom Generation (South Africa)

PROPOSAL OF TECHNICAL, ENVIRONMENTAL AND Section 5
SOCIOLOGICAL PERFORMANCE INDICATORS
FOR RENEWABLE ENERGY SOURCES
B. MANOHA, Electricit de France (France)
M. HOPPE-KILPPER, Institut fr Solare Energieversorgungstechnik (Germany)
R. VIGOTTI, ERGA of ENEL (Italy)
E, HUGHES, Electric Power Research Institute (USA)

WORKSHOPS AND COMMUNICATIONS Section 6
CASE STUDIES OF THE MONTH
R. R. RICHWINE, Consultant (US)

vii
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 CO
2
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

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 Performance of Generating Plant 2004 - Section 1

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


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 Addressing Risk/Returns of Coal Sourcing on Plant Availability and
Production Costs......................................................................................................26
3.2 Evaluating Physical Positions of Underlying Assets within Power Trading
Framework...............................................................................................................27
3.3 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

2
REFERENCES................................................................................................................................... 36
CASE STUDIES ................................................................................................................................. 37
5.1 Use of Conditional Probability Methods to Evaluate Commercial Availability.....37
5.2 South Africas ESKOM Experience: Commercial Availability Application..........41
5.3 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 Governing Design Principles...............................................................................61
1.2 The First Phase....................................................................................................64
1.3 Peer Group Data Entry.....................................................................................66
1.4 Single Unit Data Entry........................................................................................67

3
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
todays 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 plants 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

4
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 Committees 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 Committees 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. WG6s 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 Councils Generating Availability Data System (NERC-
GADS), the International Atomic Energy Agencys Power Reactor Information
System (IAEA PRIS), Euroelectrics 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 Committees 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.

5
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 NERCs 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.

6

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 long-
term 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 RTOs. 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 FERCs
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

7
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 FERCs standard market design strategy. An RTO is a large-
scale (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).

RTOs 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 FERCs 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, co-
generators, etc. can co-exist within a viable market environment.


8
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 ECs 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

9
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 todays 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.
J uly 2004, all non-household customers (industrial, commercial and professional
customers) will be free to choose their supplier;
J uly 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. EURELECTRICs 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-the-
counter) 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:

10
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 fuel-
mix 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.

11
1.2 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.


12
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 Lderitz, 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.


13
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 generators 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

14
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 J inja, 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 J une 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 countrys 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 countrys
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 Africas
Eskom and Mozambiques HCB have recently imposed ultimatums on the utility to settle
its debts ZESA has been handicapped by debt as a result of Zimbabwes 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 countrys inflation rate.

Iran seems interested in investing in Zimbabwes power sector. It is reported that an
Iranian delegation has visited Zimbabwe with a view to becoming involved in
rehabilitation work at the countrys 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 Zimbabwes 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.

15
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 companys 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.

16

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

17
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 units 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.

18
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
the issues of EA as well as that of fuel
prices or marginal costs. The denominator
is the accumulated sum of the period in
question of the difference between Market
Price and generating costs times the
maximum capacity of the unit. This
difference between market price and
generating cost could be referred to as net
margin. This sum is accumulated on an
hourly basis over the time period in
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.

Method 3
Measured hourly when the unit is in the money
In the money occurs if market price cost of generation is >0
Cost of generation includes fuel, variable O&M, and emissions costs
Method 4
Measured hourly,
Denominator = Sum of In the money hourly generation x hourly market price less
marginal fuel cost.
Numerator = Hourl y i nstalled capacity x hourly market price less marginal fuel cost
Sum (Market Price - Generation Cost) x Max Capacity
Sum (Market Price - Generation Cost) x MW Available
Sum (Market Price - Generation Cost) x Max Capacity
Sum (Market Price - Generation Cost) x MW Available
Potential Margin
Actual Margin
Potential Margin
Actual Margin

19
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
planned outages and an expected number of
forced outages that occur during the
evaluation period which would have no
impact on the calculation. The planned
availability of the denominator is calculated
as a monthly average inclusive of planned
outage and an expected number/magnitude of
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.
Method 5
Uses the sum of hourly values
No penalty for scheduled outages
Expected forced outages converted to monthly average such that there
is no penalty for FOR=plan
Availability Ratio
Planned availability x (Market Price Unit Cost)
Actual Generation x (Market Price Unit Cost)
Planned availability x (Market Price Unit Cost)
Actual Generation x (Market Price Unit Cost)
1-EFOR
(target)
1-EFOR
(actual)
1-EFOR
(target)
1-EFOR
(actual)

20

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.

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).
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.
Revenue Costs Revenue < Costs
Dispatched
(Actual MW)
You added value
You destroyed
value
Not
Dispatched
You missed the
opportunity
You were not
competitive

Revenue > Costs Revenue < Costs
Dispatched
(Actual MW)
Adding Value Destroying Value
Not
Dispatched
Missing
Opportuniy Not Competitive

21

This family is illustrated below.

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.

Revenue Costs Revenue < Costs
Dispatched
(Actual MW)
Earnings x MW Earnings x MW
Not
Dispatched
Potl Earnings x
[Max{IC;D}
MW]
Potl Earnings x
[Max{IC;D}
MW]

22
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.

23

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

24
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.

25
OTHER ASPECTS OF MEASURING AND OPTIMISING
PLANT PERFORMANCE

For more that three decades, the main focus of the PGP Committees 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 decision-
making 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 plants 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.

26

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.
Profitabili ty
Opportunity
Based
" Planned"
Ops
Strategy
Outage
Plan
Plant
" Rol e"
Avai labil ity Efficiency Flexibi lity Prod Costs

27

Based on data collected, most often analysis is carried out within detailed techno-financial
models such as EPRIs 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 decision-
making 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.


28
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 GenCos/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

29
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

30
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.

31

As the above listings demonstrate, maximising performance is, indeed, a complex
subject.

32

33
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 todays 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.


34
At this time, the state of data collection and reporting systems are in a state of flux.
Historical tools such as UNIPEDEs 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.


35
4.1 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 Committees 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.


36
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
UnreliableIt'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 Co-
Generation 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


37
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 units 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 companys 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.


38
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 2 3 4 5 6 7 8 9 10
Hour Price
(Cost)
Var.
O&M
Gross
Margin
Unit
Size
Potential
Gross
Margin
Avail.
(Y/N)
Actual
Gross
Margin
Conditional
Probability
Goal
Gross
Margin
US$/
MWh
US$/
MWh
US$/
MWh
US$/
MWh
US$ US$/
MWh

1 20 10 300 3000 30 1 3000 .92 2760
2 20 - 300 - 15 1 - .90 -
3 20 5 300 1500 25 0 0 .92 1380
4 20 20 300 6000 40 0 0 .95 5700
5 20 40 300 12000 60 1 12000 .98 11760
6 20 80 300 24000 100 1 24000 .98 23520
7 20 60 300 18000 80 1 18000 .98 17640
8 20 30 300 9000 50 1 9000 .97 8730
9 20 30 300 - 10 0 - .90 -
10 20 - 300 - 20 0 - .90 -
TOTAL 73500 66000 71490

Traditional Avail ability = 6/10 X 100 = 60.00%
Forced Outage Rate (demand) = 2/7 X 100 = 28.57%
Commercial Avail ability (Actual) = 66000/73500 X 100 = 89.80%
Commercial Avail ability (Goal) = 71490/73500 X 100 = 97.27%
Gross Margin Achieved (above/below goal) = US$66000US$71490 = -US$5490

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:
Gross Margin Achieved (above/below goal) = US$72000-US$71490 = +US$510

39

EXAMPLE 3

If the unit in the example 1 above had been available in every hour except hour 6, then:

Gross Margin Achieved (above/below goal) = US$49500-US$71490 = -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.

40

41
5.2 South Africas 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.
(e.g. EAF)
Financial
Perf.
(Earnings)
Potential
(Inst. Cap.)
Actual
(Av. Cap.)
Expense
(Cost of Prod)
Value
(Sys. Price)

42

The Framework
There are two families of indicators: the Valuable Availability Balance (VAB) which is
one set generally used to measure the Traders 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.

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%).

Dispatched
(Actual (MW)
NOT
Dispatched
Revenue =CoS Revenue <CoS
(Added Value)
Earnings x MW
(Destroyed Value)
Earnings x MW
(Missed Opportunity)
Ptl Earnings x
[ Max{IC;Dcl} MW]
(Not Competitive)
Ptl Earnings x
[ Max{IC;Dcl } - MW]
Dispatched
(Actual (MW)
NOT
Dispatched
Revenue =CoS Revenue <CoS
(Added Value)
Earnings x MW
(Destroyed Value)
Earnings x MW
(Missed Opportunity)
Ptl Earnings x
[ Max{IC;Dcl} MW]
(Not Competitive)
Ptl Earnings x
[ Max{IC;Dcl } - MW]

43
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 GenCos
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.


44


45
5.3 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. GRTNs 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 ENELs 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 J uly 2007. This move complies with the framework of the Regulation EC
n. 1228/203 of J une 2003, on conditions for access to the network for cross-border
exchanges in electricity. In J une 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

46
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 J une 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 J uly 2004, with a full opening of the market by 1 J uly 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 co-
exist 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. Italys 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.

47
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 J une. 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)
Georthermal:
1.8%
Wind & PV:
0.6%
Coal & Other
Solids: 12.3%
Natural Gas:
36.8%
Hydro: 17.3%
Other Thermal:
4.8%
Oil: 26.4%

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.

48

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 J une 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

49
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 CO
2
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. Russias 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 J I 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 J I 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.

50

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 J I 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
CO
2
), 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
CO
2
.

Reference Scenario 2010: with measures already approved. A first reduction of 37,5 Mt,
down to 570,2 Mt CO
2
. 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 J I&CDM (12 Mt) and a further reduction of 11.2 Mt CO
2
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 J I& 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/CO
2
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 (J I and CMD): the same global target can be

51
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 Parliaments initiatives.

52


53
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 demand-
related 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).

54
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 = (Full) forced outage hours
EFDH= Equivalent forced derated hours
EFDHRS=Equivalent forced derated hours during reserve shutdown
SH= Service hours
f = Discount factor for FOH
= (1/r +1/T)/(1/r +1/T +1/D) Reference (4)
r = Average forced outage duration
= FOH/number of forced outages
T = Average reserve shutdown time
D = Average demand time (duty cycle time)
= SH/number of successful starts
T+D 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 forced-
outage 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 units events. This could be a costly effort and potentially subjective.


55

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. 3840.

[3] A.M. Adamson, Gas Turbine and Diesel Forced Outage Rates and Their
Application to Reliability Calculations, IEEE-ASME J oint Power Generation
Conference, September 2730, 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.


56
APPENDIX 2: NERC GADS

North Americas 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.


57

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 PJ M 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,

58
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
*100
(forced outage hours) +(service hours)+(service hours non-generating mode)
T FOR

=

And
FOH + equivalent forced derated hours
*100
SH+SH non-generating mode+FOH+ equivalent reserve shut down forced derated hours
T EFOR

=


59
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 energy-
based equations. Examples of the methodology are shown here for forced outage rates:

Time-based (unweighted)
1
1
*100
( )
n
i
i
n
i i
i
FOH
FOR
FOH SH
=
=

=

+

Energy-based (weighted)
1
1
( * )
*100
[( )* ]
n
i i
i
n
i i i
i
FOH NMC
WFOR
FOH SH NMC
=
=

=

+

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.

60

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.


61
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 Committees 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

62
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 Councils 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.


63
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 Committees 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.


64
1.2 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
nekhaev@worldenergy.org.


65
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,

66
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.


67

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.

68
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.


Section 2

THERMAL GENERATING PLANT UNAVAILABILITY

MIKE CURLEY

Manager - GADS
North American Electric Reliability Council (US)

W. Guangyao (China)
D. Glorian (France)
K. Yassin (Egypt)
T. Tersztgansky (Hungary)
T. Uchida (J apan)
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
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

3
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 J oint UNIPEDE/WEC Committee, for many years
has been collecting statistical data on power plant availability using WECs 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.

4
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 J oint
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 fossil-
fuelled 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 J oint 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.

5
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.
6
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 J oint 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 Liquid &
fuels gaseous fuels
Coal Lignite Liquid Gaseous
(excl. lignite & others fuels fuels
& others)

Fig. 1: Steam Turbine - Fuel Types

7
Steam Turbine
>100 MW
100 to 200 to 400 to > 600
199 MW 399 MW 599 MW MW
200to 300 to 600 to 800 to > 1000
299 MW 399 MW 799 MW 999 MW 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.

8
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 EAF PUF UUF
Fossil, Al l Fossi l Fuels 84.0 8.3 7.7
Fossil, Solid Fossil Fuel s 84.4 8.1 7.5
Combined Cycle 87.0 8.2 4.8
Hydro 81.0 16.7 2.3
Al l US Base-loaded Units 83.8 7.0 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.
9
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 IPPs 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

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

4. Japan
J apan 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 J apan. The results are country based and presented in separate tables to
keep consistency with previous surveys.
The set of J apanese 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 J apan is in the same order of magnitude as the world sample (83.8 %), the
balance between PUF and UUF is again completely different. The J apanese 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 J apan 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.

10
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 users 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.

11
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 (%)

2000 2001 2002 Average
100 to 199 MW Count: 283 287 273 281
MW: 40330 40640 39117 40029
PUF: 10.5 9.7 10.3 10.2
EAF: 84.1 80.7 75.8 80.2
UUF: 6.0 8.3 11.7 8.7
200 to 299 MW Count: 132 119 118 123
MW: 31668 28206 28368 29414
PUF: 8.9 9.9 7.7 8.8
EAF: 82.5 83.2 80.7 82.1
UUF: 9.4 5.8 9.6 8.2
300 to 399 MW Count: 156 152 145 151
MW: 54601 53068 50536 52735
PUF: 11.5 11.7 11.7 11.6
EAF: 82.4 84.4 78.6 81.8
UUF: 4.8 4.2 6.8 5.3
400 to 599 MW Count: 230 196 201 209
MW: 115840 98502 101058 105133
PUF: 10.9 11.2 9.7 10.6
EAF: 82.2 84.4 84.3 83.6
UUF: 7.0 4.4 6.0 5.8
600 to 799 MW Count: 160 157 153 157
MW: 107355 105327 102466 105049
PUF: 15.5 12.2 10.4 12.7
EAF: 77.0 79.7 85.1 80.6
UUF: 9.3 6.6 4.5 6.8
800 to 999 MW Count: 38 35 38 37
MW: 32017 29545 32150 31237
PUF: 7.6 7.6 11.4 8.9
EAF: 88.5 86.4 82.3 85.7
UUF: 3.9 6.0 6.4 5.4
1000 MW & UP Count: 25 27 28 27
MW: 27828 29928 30928 29561
PUF: 15.2 14.9 13.0 14.4
EAF: 81.1 82.0 83.2 82.1
UUF: 3.8 3.1 3.8 3.5

List of countries: Brazil, Canada, Egypt, Japan, United States

12
TABLE 1-2 Solid Fossil Fuels, Fossil Steam Units, !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 239 234 234 236
MW: 34068 33512 33752 33777
PUF: 13.4 7.0 11.1 10.5
EAF: 78.8 78.1 76.9 77.9
UUF: 7.8 14.3 13.5 11.9
200 to 299 MW Count: 107 98 103 103
MW: 25795 23476 24793 24688
PUF: 8.8 8.1 7.1 8.0
EAF: 77.7 86.3 76.1 80.0
UUF: 13.5 5.6 16.9 12.0
300 to 399 MW Count: 69 68 70 69
MW: 24008 23622 24259 23963
PUF: 11.5 9.3 12.8 11.2
EAF: 84.1 88.1 70.2 80.8
UUF: 4.4 2.9 8.3 5.2
400 to 599 MW Count: 166 148 161 158
MW: 84341 75213 81697 80417
PUF: 8.5 9.5 8.1 8.7
EAF: 83.4 85.4 83.2 84.0
UUF: 8.1 5.1 8.6 7.3
600 to 799 MW Count: 106 106 106 106
MW: 72364 72574 72753 72564
PUF: 10.6 8.3 9.3 9.4
EAF: 84.3 83.7 84.2 84.1
UUF: 5.2 7.9 6.5 6.5
800 to 999 MW Count: 29 24 29 27
MW: 24508 20255 24543 23102
PUF: 10.4 9.2 9.6 9.8
EAF: 83.7 85.0 83.9 84.2
UUF: 5.9 5.7 6.6 6.1
1000 MW & UP Count: 16 18 19 18
MW: 18828 20928 21928 20561
PUF: 13.6 8.2 10.1 10.6
EAF: 80.6 86.4 83.9 83.6
UUF: 5.8 5.4 6.1 5.7

List of countries: Brazil, Canada, Japan, United States

13
TABLE 1-3 Liquid & Gaseous Fossil Fuels, Fossil Steam Units, !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 44 53 39 45
MW: 6262 7128 5365 6252
PUF: 7.6 11.0 9.6 9.4
EAF: 87.7 82.0 75.0 81.6
UUF: 4.8 5.4 10.3 6.8
200 to 299 MW Count: 25 21 15 20
MW: 5873 4730 3575 4726
PUF: 8.9 11.7 8.6 9.7
EAF: 87.3 81.7 84.2 84.4
UUF: 5.2 5.9 4.1 5.1
300 to 399 MW Count: 87 84 75 82
MW: 30593 29446 26277 28772
PUF: 11.4 12.9 11.2 11.8
EAF: 81.6 82.3 83.4 82.4
UUF: 4.9 5.0 5.9 5.3
400 to 599 MW Count: 64 48 40 51
MW: 31499 23289 19361 24716
PUF: 13.2 12.9 11.2 12.4
EAF: 80.9 83.4 85.4 83.2
UUF: 5.9 3.8 3.4 4.4
600 to 799 MW Count: 54 51 47 51
MW: 34991 32753 29713 32486
PUF: 19.2 15.2 11.1 15.1
EAF: 72.6 77.3 85.6 78.5
UUF: 11.8 5.9 3.2 7.0
800 to 999 MW Count: 9 11 9 10
MW: 7509 9290 7607 8135
PUF: 6.2 7.1 12.0 8.4
EAF: 91.0 86.8 81.7 86.5
UUF: 2.8 6.1 6.3 5.1
1000 MW & UP Count: 9 9 9 9
MW: 9000 9000 9000 9000
PUF: 16.8 21.6 16.0 18.1
EAF: 81.6 77.6 82.6 80.6
UUF: 1.7 0.8 1.5 1.3

List of countries: Brazil, Canada, Egypt, Japan. United States

14
TABLE 1-4 All Fossil Fuels, Fossil Steam Units, North America !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 270 265 251 262
MW: 38500 37625 36102 37409
PUF: 7.4 9.7 8.6 8.6
EAF: 86.6 83.4 84.4 84.8
UUF: 5.9 6.9 7.1 6.6
200 to 299 MW Count: 130 115 114 120
MW: 31248 27346 27508 28701
PUF: 7.9 9.9 6.3 8.0
EAF: 85.4 84.5 88.1 86.0
UUF: 6.7 5.7 5.6 6.0
300 to 399 MW Count: 94 89 82 88
MW: 33101 31205 28673 30993
PUF: 9.1 10.9 9.9 10.0
EAF: 82.9 82.7 83.1 82.9
UUF: 8.0 6.4 7.1 7.1
400 to 599 MW Count: 199 165 170 178
MW: 100856 83518 86074 90149
PUF: 7.8 8.7 8.2 8.2
EAF: 82.9 83.4 83.9 83.4
UUF: 9.3 7.9 7.9 8.4
600 to 799 MW Count: 108 103 99 103
MW: 74801 71373 68512 71562
PUF: 14.0 11.8 6.9 10.9
EAF: 76.4 77.3 85.5 79.8
UUF: 9.7 10.9 7.6 9.4
800 to 999 MW Count: 34 31 34 33
MW: 28601 26129 28734 27821
PUF: 7.2 5.2 11.1 7.8
EAF: 87.4 86.8 80.6 84.9
UUF: 5.4 8.1 8.3 7.3
1000 MW & UP Count: 11 11 11 11
MW: 13828 13828 13828 13828
PUF: 8.7 8.2 5.1 7.3
EAF: 81.7 83.2 84.9 83.3
UUF: 9.6 8.6 10.1 9.4

List of countries: Canada & United States

15
TABLE 1-5 Solid Fossil Fuels, Fossil Steam Units, North America, !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 239 230 230 233
MW: 34068 32930 33170 33389
PUF: 7.1 7.0 6.6 6.9
EAF: 85.8 85.9 86.0 85.9
UUF: 7.0 7.0 7.4 7.1
200 to 299 MW Count: 107 98 103 103
MW: 25795 23476 24793 24688
PUF: 6.9 8.1 4.1 6.4
EAF: 85.6 86.3 88.4 86.8
UUF: 7.5 5.6 7.5 6.9
300 to 399 MW Count: 67 65 67 66
MW: 23308 22559 23196 23021
PUF: 8.4 9.6 8.7 8.9
EAF: 85.4 86.4 86.8 86.2
UUF: 6.2 4.0 4.5 4.9
400 to 599 MW Count: 160 142 155 152
MW: 81341 72213 78697 77417
PUF: 5.5 7.5 6.3 6.4
EAF: 85.9 83.5 84.1 84.5
UUF: 8.6 9.0 9.6 9.1
600 to 799 MW Count: 96 94 94 95
MW: 65664 64474 64653 64930
PUF: 8.8 7.3 7.2 7.8
EAF: 83.9 82.3 83.5 83.2
UUF: 7.5 10.4 9.4 9.1
800 to 999 MW Count: 29 24 29 27
MW: 24508 20255 24543 23102
PUF: 10.4 9.2 9.6 9.8
EAF: 83.7 85.0 83.9 84.2
UUF: 5.9 5.7 6.6 6.1
1000 MW & UP Count: 11 11 11 11
MW: 13828 13828 13828 13828
PUF: 8.7 8.2 5.1 7.3
EAF: 81.7 83.2 84.9 83.3
UUF: 9.6 8.6 10.1 9.4

16
TABLE 1-6 Liquid & Gaseous Fossil Fuels, Steam Units, North America !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 31 35 21 29
MW: 4432 4695 2932 4020
PUF: 7.6 11.0 9.6 9.4
EAF: 87.0 82.1 83.5 84.2
UUF: 5.4 6.9 6.9 6.4
200 to 299 MW Count: 23 17 11 17
MW: 5453 3870 2715 4013
PUF: 8.9 11.7 8.6 9.7
EAF: 85.2 82.6 87.7 85.2
UUF: 5.9 5.7 3.7 5.1
300 to 399 MW Count: 27 24 15 22
MW: 9793 8646 5477 7972
PUF: 10.6 12.9 10.5 11.3
EAF: 80.2 78.3 79.1 79.2
UUF: 9.2 8.8 10.4 9.5
400 to 599 MW Count: 39 23 15 26
MW: 19515 11305 7377 12732
PUF: 10.1 9.8 10.1 10.0
EAF: 80.0 83.3 83.7 82.3
UUF: 9.9 6.9 6.2 7.7
600 to 799 MW Count: 12 9 5 9
MW: 9137 6899 3859 6632
PUF: 19.1 16.3 6.6 14.0
EAF: 69.0 72.3 87.6 76.3
UUF: 11.9 11.4 5.9 9.7
800 to 999 MW Count: 5 7 5 6
MW: 4093 5874 4191 4719
PUF: 3.9 3.1 11.9 6.3
EAF: 91.1 87.7 78.9 85.9
UUF: 5.0 9.2 9.2 7.8


17
TABLE 1-7 All Fossil Fuels, Fossil Steam Units, Western Europe !100 MW

ANNUAL AVAILABILITY
MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 171 90 75 112
EAF: 89.5 85.8 83.4 86.2
200 to 299 MW Count: 116 70 62 83
EAF: 86.4 90.7 80.4 85.8
300 to 399 MW Count: 167 91 81 113
EAF: 87.7 91.2 86.0 88.3
400 to 599 MW Count: 53 47 44 48
EAF: 88.9 85.0 76.6 83.5
600 to 799 MW Count: 36 24 22 27
EAF: 86.0 94.5 62.1 80.9
800 to 999 MW Count: 2 4 5 4
EAF: NR NR NR NR

List of countries: Belgium, France, Germany, Italy, Portugal, Spain

18
TABLE 1-8 Solid Fossil Fuels, Fossil Steam Units, Western Europe !100 MW

ANNUAL AVAILABILITY
MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 103 65 65 78
EAF: 88.4 80.1 85.1 84.5
200 to 299 MW Count: 76 45 47 56
EAF: 83.6 89.5 84.9 86.0
300 to 399 MW Count: 90 82 72 81
EAF: 93.6 92.5 95.2 93.8
400 to 599 MW Count: 38 30 29 32
EAF: 92.3 89.4 84.8 88.8
600 to 799 MW Count: 16 20 20 19
EAF: NR NR NR NR
800 to 999 MW Count: 2 4 5 4
EAF: NR NR NR NR


19
TABLE 1-9 Liquid & Gaseous Fossil Fuels, Steam Units, Western Europe !100 MW

ANNUAL AVAILABILITY
MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 68 25 10 34
EAF: 90.2 88.6 81.8 86.9
200 to 299 MW Count: 40 25 15 27
EAF: 88.7 91.6 77.4 85.9
300 to 399 MW Count: 77 9 9 32
EAF: 83.7 88.6 67.8 80.0
400 to 599 MW Count: 15 17 15 15
EAF: 84.5 80.6 68.4 77.8
600 to 799 MW Count: 20 4 2 9
EAF: 86.0 94.5 62.1 80.9
800 to 999 MW Count: NR NR NR NR
EAF: NR NR NR NR

20
TABLE 1-10 All Fossil Fuels, Fossil Steam Units, .!100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 13 22 22 19
MW: 1830 3015 3015 2620
PUF: 7.3 12.2 15.3 11.6
EAF: 89.0 78.0 66.3 77.8
UUF: 3.7 9.8 18.4 10.6
200 to 299 MW Count: 2 4 4 3
MW: 420 860 860 713
PUF: 4.5 13.2 15.0 10.9
EAF: 91.6 80.7 80.7 84.3
UUF: 3.9 6.1 4.4 4.8
300 to 399 MW Count: 16 17 17 17
MW: 5000 5363 5363 5242
PUF: 15.2 11.2 19.9 15.4
EAF: 82.4 85.9 76.2 81.5
UUF: 2.4 2.9 3.9 3.1
600 to 799 MW Count: 2 2 2 2
MW: 1254 1254 1254 1254
PUF: 1.3 23.6 11.4 12.1
EAF: 65.0 72.2 87.8 75.0
UUF: 33.7 4.2 0.8 12.9

List of countries: Brazil, Egypt

TABLE 1-11 Solid Fossil Fuels, Fossil Steam Units, Other Countries !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 0 4 4 3
MW: 0 582 582 388
PUF: 0.0 8.1 6.5 4.9
EAF: 0.0 70.2 65.9 45.4
UUF: 0.0 21.6 27.7 16.4
300 to 399 MW Count: 0 1 1 1
MW: 0 363.0 363.0 242.0
PUF: 0.0 8.2 47.6 18.6
EAF: 0.0 90.1 48.6 46.2
UUF: 0.0 1.7 3.8 1.8


21
TABLE 1-12 Liquid & Gaseous Fossil Fuels, Fossil Steam Units, Other Countries
!100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 13 18 18 16
MW: 1830 2433 2433 2232
PUF: 7.3 14.3 19.8 13.8
EAF: 89.0 81.8 66.5 79.1
UUF: 3.7 3.9 13.7 7.1
200 to 299 MW Count: 2 4 4 3
MW: 420 860 860 713
PUF: 4.6 13.2 15.0 10.9
EAF: 91.6 80.7 80.7 84.3
UUF: 3.9 6.1 4.4 4.8
300 to 399 MW Count: 16 16 16 16
MW: 5000 5000 5000 5000.0
PUF: 15.2 12.7 6.1 11.3
EAF: 82.4 83.8 90.0 85.4
UUF: 2.4 3.5 3.9 3.3
600 to 799 MW Count: 2 2 2 2
MW: 1254 1254 1254 1254
PUF: 1.4 23.7 11.5 12.2
EAF: 65.0 72.2 87.8 75.0
UUF: 33.7 4.2 0.8 12.9


22
TABLE 1-13 All Fossil Fuels, Fossil Steam Units, Japan !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 0 0 0 0
MW: 0 0 0 0
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 0.0 0.0
UUF: 0.0 0.0 0.0 0.0
200 to 299 MW Count: 0 0 0 0
MW: 0 0 0 0
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 0.0 0.0
UUF: 0.0 0.0 0.0 0.0
300 to 399 MW Count: 46 46 46 46.0
MW: 16500 16500 16500 16500.0
PUF: 17.1 13.4 16.3 15.6
EAF: 82 85.8 81.6 83.1
UUF: 0.9 0.8 2.1 1.3
400 to 599 MW Count: 31 31 31 31
MW: 14984 14984 14984 14984
PUF: 14.5 16.5 11.0 14.0
EAF: 84.0 83.0 88.3 85.1
UUF: 1.5 0.5 0.7 0.9
600 to 799 MW Count: 50 52 52 51
MW: 31300 32700 32700 32233
PUF: 18.1 13.1 15.0 15.4
EAF: 81.1 85.3 83.5 83.3
UUF: 0.8 1.6 1.5 1.3
800 to 999 MW Count: 4 4 4 4
MW: 3416 3416 3416 3416
PUF: 8.5 15.0 12.3 11.9
EAF: 90.8 85.0 87.3 87.7
UUF: 0.7 0.0 0.4 0.4
1000 MW & UP Count: 14 16 17 16
MW: 14000 16100 17100 15733
PUF: 16.4 14.8 16.4 15.9
EAF: 82.1 83.5 81.8 82.5
UUF: 1.5 1.7 1.8 1.7

23
TABLE 1-14 Solid Fossil Fuels, Fossil Steam Units, Japan !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 0 0 0 0
MW: 0 0 0 0
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 0.0 0.0
UUF: 0.0 0.0 0.0 0.0
200 to 299 MW Count: 0 0 0 0
MW: 0 0 0 0
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 0.0 0.0
UUF: 0.0 0.0 0.0 0.0
300 to 399 MW Count: 2 2 2 2
MW: 700 700 700 700
PUF: 17.7 8.7 21.1 15.8
EAF: 81.4 89.3 58.5 76.4
UUF: 0.9 2.0 20.4 7.8
400 to 599 MW Count: 6 6 6 6
MW: 3000 3000 3000 3000
PUF: 14.6 16.6 13.5 14.9
EAF: 82.4 83.0 86.1 83.8
UUF: 3.0 0.4 0.4 1.3
600 to 799 MW Count: 10 12 12 11
MW: 6700 8100 8100 7633
PUF: 14.2 10.4 13.6 12.7
EAF: 85.2 86.6 85.6 85.8
UUF: 0.6 3.0 0.8 1.5
800 to 999 MW Count: 0 0 0 0.0
MW: 0 0 0 0.0
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 0.0 0.0
UUF: 0.0 0.0 0.0 0.0
1000 MW & UP Count: 5 7 8 7
MW: 5000 7100 8100 6733
PUF: 18.5 8.3 15.1 13.6
EAF: 79.5 89.6 82.9 84.4
UUF: 2.0 2.1 2.0 2.0

24
TABLE 1-15 Liquid & Gas Fossil Fuels, Fossil Steam Units, Japan !100 MW

MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
100 to 199 MW Count: 0 0 0 0
MW: 0 0 0 0
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 0.0 0.0
UUF: 0.0 0.0 0.0 0.0
200 to 299 MW Count: 0 0 0 0
MW: 0 0 0 0
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 0.0 0.0
UUF: 0.0 0.0 0.0 0.0
300 to 399 MW Count: 44 44 44 44
MW: 15800 15800 15800 15800
PUF: 17.1 13.6 15.9 15.5
EAF: 82 85.7 82.7 83.5
UUF: 0.9 0.7 1.4 1.0
400 to 599 MW Count: 25 25 25 25
MW: 11984 11984 11984 11984
PUF: 14.4 16.5 10.4 13.8
EAF: 84.5 83.0 88.8 85.4
UUF: 1.1 0.5 0.8 0.8
600 to 799 MW Count: 40 40 40 40
MW: 24600 24600 24600 24600
PUF: 19.3 13.9 15.6 16.3
EAF: 79.9 85.0 82.7 82.5
UUF: 0.8 1.1 1.7 1.2
800 to 999 MW Count: 4 4 4 4
MW: 3416 3416 3416 3416
PUF: 8.5 15.0 12.3 11.9
EAF: 90.8 85.0 87.3 87.7
UUF: 0.7 0.0 0.4 0.4
1000 MW & UP Count: 9 9 9 9
MW: 9000 9000 9000 9000
PUF: 15.2 20.0 17.5 17.6
EAF: 83.5 78.7 80.8 81.0
UUF: 1.3 1.2 1.7 1.4

List of countries: Japan
25
TABLE 1-16 Combustion Turbines, World, !30 MW

OF BASE-LOADED COMBUSTION TURBINES
MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
30 to 49 MW Count: 17 18 38 24
MW: 625 659 1561 948
PUF: 8.3 3.5 5.9
EAF: 92.9 94.9 89.4 92.4
UUF: 2.2 2.1 3.9 2.7
50 to 74 MW Count: 8 1 0 5
MW: 456 69 0 263
PUF: 3.7 18.9 0.0 11
EAF: 93.2 80.1 0.0 87
UUF: 3.1 1.0 0.0 2
75 to 99 MW Count: 0 0 1 1
MW: 0 0 81 81
PUF: 0.0 0.0 0.0 0.0
EAF: 0.0 0.0 95.9 95.9
UUF: 0.0 0.0 4.1 4.1

TABLE 1-17 Hydro Units, World, !50 MW

OF HYDRO UNITS (ALL OPERATING MODES)
MORE THAN
ONE YEAR OLD (%)

2000 2001 2002 Average
50 to 149 MW Count: 346 340 198 295
MW: 28831 28248 16425 24501
PUF: 25.2 10.7 9.6 15.2
EAF: 78.2 88.6 88.3 85.1
UUF: 1.7 1.2 1.2 1.4
150 and above MW Count: 111 109 96 105
MW: 25384 24821 27790 25998
PUF: 9.8 32.7 23.3 21.9
EAF: 87.6 73.4 80.1 80.4
UUF: 2.3 1.8 0.8 1.7

26
TABLE 1-18 Recent Combined Cycle Units, World, !100 MW

OF BASE-LOADED COMBINED CYCLE UNITS
MORE THAN
ONE YEAR OLD (%)
2000 2001 2002 Average
100 to 199 MW Unit-years: 73 71 75 73
MW: 10035 9867 10321 10074
Average MW/unit: 137 139 138 138
PUF: 10.39 9.77 9.63 9.9
EAF: 86.8 85.55 85.47 85.9
UUF: 2.28 4.35 3.1 3.2
200 to 299 MW Unit-years: 51 54 55 53
MW: 12262 13004 13308 12858
PUF: 12.19 8.48 7.24 9.3
EAF: 85.08 88.46 89.06 87.5
UUF: 2.73 4.94 3.3 3.7
300 MW & UP Unit-years: 24 30 34 29
MW: 11112 13636 15616 13455
PUF: 9.1 20 12.12 13.7
EAF: 90.66 83.2 87.45 87.1
UUF: 1.19 1.03 2.43 1.6

TABLE 1-19 Historical Combined Cycle Units, World, !100 MW

OF BASE-LOADED COMBINED CYCLE UNITS
MORE THAN
ONE YEAR OLD (%)
1991-1993 1994-1996
100 to 199 MW Unit-years: 148 220
Average MW/unit: 134 141
PUF: 14.4 12.5
EAF: 84.2 85.9
UUF: 1.4 1.6
200 to 299 MW Unit-years: 30 95
PUF: 9.5 10.8
EAF: 84.7 86.6
UUF: 5.8 2.6
300 MW & UP Unit-years: 12 56
PUF: 9.9 8.5
EAF: 78.0 88.3
UUF: 12.2 3.3

27

GRAPH 1-1 UNIT CAPABILITY FACTOR, WESTERN EUROPE

GRAPH 1-2 UNPLANNED CAPABILITY LOSS FACTOR, WESTERN
EUROPE


Section 3

NUCLEAR GENERATING PLANT UNAVAILABILITY

MARIANNA SZIKSZAINE TABORI
International Atomic Energy Agency (IAEA)


M. Szikszaine Tabori (IAEA)
S. R. Fernandes (Brazil)
W. Guangyao (China)
D. Glorian (France)
K. Yassin (Egypt)
A. R. Karbassi (Iran)
A. Torii (J apan)
T. Uchida (J apan)
K. Yoon Lee (Rep. of Korea)
A. A. Al Tuwaijri (Saudi Arabia)
J . Ruansup-anek (Thailand)

1

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 7
Worldwide
2. Conclusions 11
3. List of Countries Providing Information on 13
Nuclear Power to IAEA-PRIS

REFERENCES 13

LIST OF TABLES
Table 1: Nuclear Power Reactors in Operation and Under Construction 5
in the World (as of 31 December 2003)

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 8
Factor in 2002
Fig. 4: Average Availability and Unavailability Factors (1990-92, 9
1993-95, 1996-98, 1999-2001)
Fig. 5: Energy Availability Factors by Reactor Type (1999-2001) 10
2
Page

ANNEX 1

APPENDIX N1 - STATUS OF NUCLEAR POWER PLANTS A1-1
WORLDWIDE (AS OF 31 DECEMBER 2003)
APPENDIX N2 - NUMBER OF NUCLEAR POWER PLANTS BY A1-2
REACTOR TYPE AND REGION
(AS OF 31 DECEMBER 2003)
APPENDIX N3 - NUMBER OF NUCLEAR POWER PLANTS BY A1-3
AGE (AS OF 31 DECEMBER 2003)
APPENDIX N4 - DISTRIBUTION OF NUCLEAR POWER PLANTS A1-4
BY AGE AND TYPE (AS OF 31 DECEMBER 2003)
APPENDIX N5 - DISTRIBUTION OF NUCLEAR POWER PLANTS A1-5
BY AGE AND REGION (AS OF 31 DECEMBER 2003)
APPENDIX N6 - NUCLEAR SHARE OF ELECTRICITY GENERATION A1-6
IN 2003
APPENDIX N7 - ANNUAL WORLD ENERGY AVAILABILITY AND A1-7
UNAVAILABILITY FACTORS
APPENDIX N8 - ENERGY AVAILABILITY AND UNAVAILABILITY A1-8
FACTORS PWR
FACTORS BWR
FACTORS PHWR
FACTORS WWER
FACTORS LWGR
FACTORS BY AGE
APPENDIX N14 - ENERGY AVAILABILITY BY REGION A1-14
NORTH AMERICA
WESTERN EUROPE
EASTERN EUROPE
APPENDIX N17- ENERGY AVAILABILITY BY REGION A1-15
FAR EAST
MIDDLE EAST AND SOUTH ASIA
LATIN AMERICA
APPENDIX N20- ENERGY AVAILABILITY BY REGION A1-17
AFRICA

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 IAEAs 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.
4
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 IAEAs 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 re-
commissioning 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 Koreaand J apan.
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
5
a major component of national energy policy. The Industrys 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 industrys 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)

Reactors in Operation Reactors under
Construction
Nuclear Electricity Supplied
in 2000
Total Operating
Experience to 31 Dec.
2003
Country
No. of Units Total
MW (e)
No. of
Units
Total
MW (e)
TW (e).h % of Total Years Months

Argentina 2 935 1 692 7.03 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
Canada 16 11323 70.29 12.53 486 11
China 8 5977 3 2610 41.59 2.18 39 1
Czech R. 6 3548 25.87 31.09 74 10
DPR Korea 1 1040 0 0
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 11.01 32.69 74 2
India 14 2550 8 3622 16.37 3.30 223 5
Iran 2 2111 0 0
J apan 53 44139 3 3696 230.80 25.01 1123 7
Korea RP 19 15850 1 960 123.28 40.01 220 8
Lithuania 2 2370 14.30 79.89 36 6
Mexico 2 1310 10.51 5.23 23 11
Netherlands 1 449 3.80 4.48 59 0
Pakistan 2 425 1.81 2.37 35 10
Romania 1 655 1 655 4.54 9.33 7 6
Russian Fed. 30 20793 3 2825 138.39 16.54 761 4
S. Africa 2 1800 12.66 6.05 38 3
Slovak R 6 2442 2 776 17.86 57.35 100 6
Slovenia 1 656 4.96 40.45 22 3
Spain 9 7584 59.36 23.64 219 2
Sweden 11 9451 65.50 49.62 311 1
Switzerland 5 3220 25.93 39.73 143 10
UK 27 12052 85.31 23.70 1329 8
Ukraine 13 11207 4 3800 76.70 45.93 279 10
USA 104 98298 763.74 19.86 2871 8

Total 439 361094 31 25387 2524.03 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.
6
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 (1999-
2001) 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: all units in the world
North America: Canada and the USA
Latin America: Argentina, Brazil and Mexico
Western Europe: 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 J apan
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).
7
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 non-
prototype 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.

73.1
74.4
75.4
76.7
78.1
77.1
79.1
80.8
82.1
83.3
83.7
50
55
60
65
70
75
80
85
90
95
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002

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).
8
0
10
20
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
70-79% 80-89% >90%

Fig. 2: Distribution of Reactors with High Availability Factor

84
98
62
10
70
49
19
15
1 3 2
0 0
2
0
20
40
60
80
100
120
<
3
0
%
3
0
-
3
5
%
3
5
-
4
0
%
4
0
-
4
5
%
4
5
-
5
0
%
5
0
-
5
5
%
5
5
-
6
0
%
6
0
-
6
5
%
6
5
-
7
0
%
7
0
-
7
5
%
7
5
-
8
0
%
8
0
-
8
5
%
8
5
-
9
0
%
9
0
-
9
5
%
(
%
)

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.
9
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 1999-
2001. 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.

5
10
15
20
25
30
0 3 6 9 12
Unplanned EUF (%)
P
l
a
n
n
e
d

E
U
F

(
%
)
70%
75%
90%
80%
85%
65%
55%
1990-1992
1993-1995
1996-1998
2002
1999-2001
60%
10
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 1999-
2001. 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.
84
76
85
62
81
85
70
0
10
20
30
40
50
60
70
80
90
ABWR AGR BWR LWGR PHWR PWR WWER
E
A
F

(
%
)

Fig. 5: Energy Availability Factors by Reactor Type (1999-2001)

11
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 5
th
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.
12
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.
13
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 - J apan -
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, 16
th
World Energy Council Congress, Tokyo, J apan, 1995.

[6] Unavailability Factors of Thermal Generating plant and Availability Statistics
1998, D. Glorian, EdF, France and R. Spiegelberg-Planer, International Atomic
Energy Agency, 17
th
World Energy Council Congress, Houston, USA, 1998.

[7] Exchange of Availability/Performance Data on Gas Turbine and Combined Cycle
Generating Plant, 15
th
World Energy Congress, Madrid, Spain, 1992.
World Energy Council Performance of Generating Plant 2004 Section 3, Annex 1

A1-1
APPENDIX N1 - STATUS OF NUCLEAR POWER PLANTS WORLDWIDE

Nuclear Electricity Total Operating
Experience
Reactors in Operation Reactors under Construction
Supplied in 2003 to 31 Dec. 2003
No of Total No of Total % of
Country
Units MW(e) Units MW(e)
TW(e).h
Total
Years Months

ARGENTINA 2 935 1 692 7.03 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
CANADA 16 11323 70.29 12.53 486 11
CHINA 8 5977 3 2610 41.59 2.18 39 1
CZECH R. 6 3548 25.87 31.09 74 10

DPR KOREA 1 1040 0 0
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 11.01 32.69 74 2
INDIA 14 2550 8 3622 16.37 3.30 223 5
IRAN 2 2111 0 0
J APAN 53 44139 3 3696 230.08 25.01 1123 7

KOREA RP 19 15850 1 960 123.28 40.01 220 8
LITHUANIA 2 2370 14.30 79.89 36 6
MEXICO 2 1360 10.51 5.23 23 11
NETHERLANDS 1 449 3.80 4.48 59 0

PAKISTAN 2 425 1.81 2.37 35 10
ROMANIA 1 655 1 655 4.54 9.33 7 6
RUSSIA 30 20793 3 2825 138.39 16.54 761 4
S. AFRICA 2 1800 12.66 6.05 38 3

SLOVAK R 6 2442 2 776 17.86 57.35 100 6
SLOVENIA 1 656 4.96 40.45 22 3
SPAIN 9 7584 59.36 23.64 219 2
SWEDEN 11 9451 65.50 49.62 311 1

SWITZERLAND 5 3220 25.93 39.73 143 10
UK 27 12052 85.31 23.70 1329 8
UKRAINE 13 11207 4 3800 76.70 45.93 279 10
USA 104 98298 763.74 19.86 2871 8

Total 439 361094 31 28387 2524.03 11143 5

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.

A1-2

APPENDIX N2 - NUMBER OF NUCLEAR POWER PLANTS BY REACTOR
TYPE AND REGION (AS OF 31 DECEMBER 2003)

0
20
40
60
80
100
120
140
160
WWER 48 2
PWR 2 2 1 92 69 46 1
PHWR 2 1 16 6 13
LWGR 17
GCR 12
FBR 1 1 1
BWR 2 20 35 31 2
AGR 14
ABWR 2
Africa Latin America
Eastern
Europe
Western
Europe
North
America
Far East
Middle East
and South
Asia

A1-3
APPENDIX N3 - NUMBER OF NUCLEAR POWER PLANTS BY AGE

2
6
3
6
4
4
3
6
4
6
9
6
4
10
11
14
22
24
32
33
22
19
22
21
7
15
14
16
11
23
15
10
11
5
7
3
1
2
2
1
3
0 5 10 15 20 25 30 35
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
34
35
36
37
38
43
44
A
g
e

(
y
e
a
r
s
)
No. of Reactors

A1-4
APPENDIX N4 - DISTRIBUTION OF NUCLEAR POWER PLANTS BY AGE AND
TYPE (AS OF 31 DECEMBER 2003)

0
20
40
60
80
100
120
140
N
u
m
b
e
r

o
f

R
e
a
c
t
o
r
s
FBR
GCR
AGR
LWGR
WWER
PWR
PHWR
BWR
ABWR
FBR 1 1 1
GCR 12
AGR 4 6 4
LWGR 1 5 4 7
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
0 - 5 6 - 10 11 - 15 16 -20 21 - 25 26 - 30 > 30

A1-5
APPENDIX N5 - DISTRIBUTION OF NUCLEAR POWER PLANTS
BY AGE AND REGION (AS OF 31 DECEMBER 2003)

0
20
40
60
80
100
120
140
160
N
u
m
b
e
r

o
f

R
e
a
c
t
o
r
s
> 30 1 3 21
26 - 30 8 6 1 4 32 14
21 - 25 8 15 19 22
16 -20 2 21 14 2 2 26 51
11 - 15 23 15 1 2 30 31
6 - 10 3 16 1 2 7 6
0 - 5 5 9 1 6 1 5
Af r ica
East er n
Eur ope
Far East
Lat in
Amer ica
Middle
East and
Sout h
Nor t h
Amer ica
West er n
Eur ope


A1-6
APPENDIX N6 - NUCLEAR SHARE OF ELECTRICITY GENERATION IN 2003(%)

80
78
57
56
50
46
40
40
40
38
36
33
31
28
27
25
24
24
20
17
13
9
9
6
5
5
4
3
2
2
0 10 20 30 40 50 60 70 80 90
LITHUANIA
FRANCE
SLOVAKIA
BELGIUM
SWEDEN
UKRAINE
SLOVENIA
KOREA, REP.
SWITZERLAND
BULGARIA
ARMENIA
HUNGARY
CZECH REP.
GERMANY
FINLAND
JAPAN
UNITED KINGDOM
SPAIN
USA
RUSSIA
CANADA
ROMANIA
ARGENTINA
SOUTH AFRICA
MEXICO
NETHERLANDS
BRAZIL
INDIA
PAKISTAN
CHINA
(
%
)


A1-7
APPENDIX N7 - ANNUAL WORLD ENERGY AVAILABILITY AND
UNAVAILABILITY FACTORS

Ener gy Availabilit y Fact or
76
79
8 1
76
50
55
6 0
6 5
70
75
8 0
8 5
9 0
9 5
10 0
1
9
8
2
1
9
8
3
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)
Y ear l y P = 1 P = 2 P = 3 Li f et i me

Planned Ener gy Unvailabilit y Fact or
16
12
16 14
0
5
10
15
20
25
30
1
9
8
2
1
9
8
3
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)
Year ly P = 1 P = 2 P = 3 Lif et ime

Unplanned Ener gy Unavailabilit y Fact or
6
4
8
6
0
2
4
6
8
10
12
14
1
9
8
2
1
9
8
3
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

A1-8
APPENDIX N8 - ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS
FOR PWR

PWR - Energy Availability Factors
81
85
77
50
55
60
65
70
75
80
85
90
95
100
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)
Yearly P = 1 P = 2 P = 3 Lifetime

PWR - Planned Ener gy Unavailabilit y Fact or s
14
11
10
14
0
5
10
15
20
25
30
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

PWR - Unplanned Ener gy Unavailabilit y Fact or s
6 6
4
7
0
2
4
6
8
10
12
14
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

A1-9
BWR

BWR - Ener gy Availabilit y Fact or s
75
80
85
74
50
55
60
65
70
75
80
85
90
95
100
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

BWR - Planned Ener gy Unavailabilit y Fact or s
17
15
11
18
0
5
10
15
20
25
30
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

BWR - Unplanned Ener gy Unavailabilit y Fact or s
7
4
3
7
0
2
4
6
8
10
12
14
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)


A1-10
PHWR

PHWR - Ener gy Availabilit y Fact or s
70 70
81
73
50
55
60
65
70
75
80
85
90
95
100
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

PHWR - Unplanned Ener gy Unavailabilit y Fact or s
13
16
6
12
0
2
4
6
8
10
12
14
16
18
20
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)
PHWR - Planned Ener gy Unavailabilit y Fact or s
13
12 12
11
0
5
10
15
20
25
30
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)


A1-11
WWER

WWER - Ener gy Availabilit y Fact or s
64
67.5
72
70
50
55
60
65
70
75
80
85
90
95
100
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

WWER - Planned Ener gy Unavailabilit y Fact or s
25 25
21
22
0
5
10
15
20
25
30
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

WWER - Unplanned Ener gy Unavailabilit y Fact or s
8
4
2
5.5
0
2
4
6
8
10
12
14
16
18
20
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)


A1-12
APPENDIX N12 - ENERGY AVAILABILITY AND UNAVAILABILITY FACTORS -
LWGR
LWGR - Ener gy Availabilit y Fact or s
58
60
63
67
50
55
60
65
70
75
80
85
90
95
100
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

LWGR - Planned Ener gy Unavailabilit y Fact or s
29
36
29
25
0
5
10
15
20
25
30
35
40
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)

LWGR - Unplanned Ener gy Unavailabilit y Fact or s
3.6
2.7
2.4
4.1
0
2
4
6
8
10
12
14
16
18
20
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
(
%
)


A1-13

BY AGE
Ener gy Availabilit y Fact or by Age
50
55
60
65
70
75
80
85
90
95
100
0 1 2 3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
9
2
0
2
1
2
2
2
3
2
4
2
5
2
6
2
7
2
8
2
9
3
0
3
1
3
2
3
3
E
A
F

(
%
)

Planned Energy Unavailability Factor by Age
0
5
10
15
20
25
30
0 1 2 3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
9
2
0
2
1
2
2
2
3
2
4
2
5
2
6
2
7
2
8
2
9
3
0
3
1
3
2
3
3
E
A
F

(
%
)

Unplanned Energy Unavailability Factor by Age
0
5
10
15
20
25
30
0 1 2 3 4 5 6 7 8 9
1
0
1
1
1
2
1
3
1
4
1
5
1
6
1
7
1
8
1
9
2
0
2
1
2
2
2
3
2
4
2
5
2
6
2
7
2
8
2
9
3
0
3
1
3
2
3
3
E
A
F

(
%
)


A1-14
APPENDIX N14 - ENERGY AVAILABILITY BY REGION
NORTH AMERICA

North Ameri ca
74 73
77
78
77
72
80
86
88
89
90
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
(
%
)

WESTERN EUROPE

Western Europe
74
79
80 80
83
84
83 82 83
84 84
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
(
%
)

A1-15
EASTERN EUROPE

Eastern Europe
66
64
61 61
65
67
64
65
68
70
74
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
(
%
)

APPENDIX N17- ENERGY AVAILABILITY BY REGION
FAR EAST

Far East
76
78
76
80
81
83 83
82
83 83
81
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
(
%
)


A1-16
MIDDLE EAST AND SOUTH ASIA

Mi ddl e East and South Asi a
56
46
39
52
50
60
68
76
78
75
78
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
(
%
)

LATIN AMERICA

Lati n Ameri ca
68 69
63
71
72
83
78
80
67
82
80
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
(
%
)


A1-17
APPENDIX N20- ENERGY AVAILABILITY BY REGION
AFRICA

Afri ca
58
45
62
70
73
79
85
79
81
69
76
30
40
50
60
70
80
90
100
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
(
%
)


Section 4

PERFORMANCE OF HYDRO AND PUMP STORAGE
PLANT

TERRY MOSS
Eskom Generation (South Africa)


T. Moss (South Africa), Chair
R. S. Chadha (India)
R. N. Misra (India)
C. Cakmak (Turkey)
G. Varol (Turkey)
1
PERFORMANCE OF HYDRO AND PUMP STORAGE PLANT

The Largest Source of Renewable Energy
Hydropower is the worlds largest source of renewable energy used for power generation;
it accounts for 17% of the worlds 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
hydropowers 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 three-
year 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.
2
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 energy-
based 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 J une 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-ids and passwords for individual users
can be issued on request. As the PGP Committee moves towards collection of unit-
specific 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.


Section 5

PROPOSAL OF TECHNICAL, ENVIRONMENTAL
AND SOCIOLOGICAL PERFORMANCE
INDICATORS
FOR RENEWABLE ENERGY SOURCES

BRUNO MANOHA
Electricit de France (France)

MARTIN HOPPE-KILPPER
Institut fr 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)

1
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 18
th
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, inter-
comparisons, 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.

RESUM
Le prsent travail a t ralis dans le cadre du Conseil Mondial de lEnergie (CME) et de son
Comit sur les performances des centrales de production lectrique (PGP). Son objectif est de
dfinir des indicateurs de performance pour les filires utilisant des sources dnergie
renouvelable (ENR).
Une premire phase (1999-2001) a consist proposer des indicateurs techniques, comme cela a
t fait par le pass pour les centrales nuclaires et thermiques classiques par le CME et
lUNIPEDE. Ces indicateurs techniques ont t proposs pour lolien, le photovoltaque, la
biomasse et la gothermie, et le travail a t consign dans un rapport prsent au 18
me
Congrs
du Conseil Mondial de lEnergie Buenos Aires en Octobre 2001.
La seconde phase prsente ici (2002-2004) a pour principal objectif dtendre ltude des
indicateurs environnementaux et sociologiques, les ENR ayant pour caractristique de participer
fortement ce quon a coutume dappeler le dveloppement durable. Comme pour la premire
phase, dminents spcialistes ont travaill la dfinition de ces indicateurs, en liaison avec divers
organismes internationaux reconnus en la matire (IEA, Eurelectric, IGA, CME, etc.), afin que les
indicateurs proposs puissent tre considrs comme aussi standard que possible. Des exemples
sont prsents de manire fournir de premiers ordres de grandeur de ces indicateurs de
performances.
Nous esprons que ces indicateurs savreront trs utiles pour toutes les personnes,
organismes et pays souhaitant dvelopper les nergies renouvelables. A lavenir, ils
devraient aider, travers des travaux dinter comparaisons, de benchmarkings, voire la
mise en place ventuelle de bases de donnes, comparer et positionner les diffrentes
techniques utilises, permettant ainsi de mieux identifier leurs faiblesses ventuelles,
damliorer leurs performances et daider, de faon gnrale, au dveloppement des
nergies renouvelables.


2
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.


3
TABLE OF CONTENTS

Page

SUMMARY / RESUM 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. Overview of the Present Development and Future Expansion 8
of Renewable Energy
4. Review of Technical Performance Indicators Proposed in Phase 1 9
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

LIST OF FIGURES

Fig. 1: Measured Power Curve of Turbine with Pitch Control and 12
Variable Speed (a) and Stall Control and Constant Speed (b).
Fig. 2: Wind Speed Frequency and Weibull Distribution 13
Fig. 3: Distribution of Wind Energy Based on Energetically Weighted 13
Frequency Distribution of the Wind Direction in 12 Sectors
Fig. 4: Gross Wind Energy Supply by Years (Germany 1993-1999) 14
Fig. 5: Monthly Capacity Factor of the Wind Turbines 14
in the Scientific Measurement and Evaluation Program
(WMEP) of ISET, Germany, 1990-1999


4
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 51
of the Various RES Technologies
ANNEX 4: List of Proposed RES Indicators 52
ANNEX 5: Commercial Availability 54


5
1. 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 17
th
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 18
th
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.


6
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.

7
2.3 Scope of the Work

The following conditions were agreed for the first phase:
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.

2.4 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.

8
3. 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,
J apan 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.

9
4. 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

Type of RES
Ease of measurement /
estimate
Comparability
between sites
Plant size (MW) All Easy OK
Expected life time All Rather difficult OK
Capacity Factor (%) Wind, Biomass,
Geothermal
Easy OK
Load Factor (%) Geothermal Easy OK
Specific Energy
Production (kWh/m2)
Wind Easy OK
Equivalent Full Load
Hours (hours)
All Easy OK
Availability Factor (%) Wind, Biomass,
Geothermal
Easy OK
Reference Yield
(hours/day)
PV Rather difficult OK
Array Yield
(hours/day)
PV Rather difficult OK
Final Yield (hours/day) PV Rather difficult OK
Performance Ratio (%) PV Rather difficult OK
Efficiency (higher heat
value)
Biomass Easy More or less
Fuel Moisture Biomass Easy More or less
Annual energy sent out
(MWh/year)
Biomass Easy 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:

10
- Age;
- Vintage;
- Type of technology (type of PV cells, of wind turbine wings, of biomass fuel,
of geothermal technique);
- Cycle of operation.


11
4.1 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/m
2
]. 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/m
2
];
Equivalent full load hours i.e. the annual energy production in relation to the rated
power of the turbine, in hours;
Capacity factor
1
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'

12

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 -

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
0,2
0,4
0,6
0,8
1
1,2
0 5 10 15 20 25
wind speed [m/s]
p
o
w
e
r

o
u
t
p
u
t

[
P
/
P
r
]
0
0,2
0,4
0,6
0,8
1
1,2
0 5 10 15 20 25
wind speed [m/s]
p
o
w
e
r

o
u
t
p
u
t

[
P
/
P
r
]
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 (BIN-
method) 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).


13
0%
5%
10%
15%
20%
25%
wind speed in m/s
f
r
e
q
u
e
n
c
y
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
rel. frequency
Weibull distribution
Weibull parameter: A =8,2 m/s ; k =2,2 measured in 30 m height

Fig. 2: Wind Speed Frequency and Weibull Distribution

0
15
30
45
60
75
90
105
120
135
150
165
180
195
210
225
240
255
270
285
300
315
330
345

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.

14

159
107
81
69
147
182
173
179
209
183
119
92
76
113
125
111
87
84
98
102
84
70 70
61
71
74
52
50
0
50
100
150
200
250
1993 1994 1995 1996 1997 1998 1999
coastline, islands (1993-'99: 176 W/m)
Highlands (1993-'99: 106 W/m)
Northern lowlands (1993-'99: 87 W/m)
Northern lowlands, wooded (1993-'99: 64 W/m)
W/m

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 J uly, to 32% in J anuary.
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 J anuary 1993.
0
10
20
30
40
50
60
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
C
a
p
a
c
i
t
y

F
a
c
t
o
r
%
Maxi mum Val ues 1990-99
Mi ni mum Val ues 1990-99
Mean Val ues 1990-99

Fig. 5: Monthly Capacity Factor of the Wind Turbines
in the Scientific Measurement and Evaluation Program (WMEP) of ISET, Germany,
1990-1999


15
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)
F
d
(%) =

_____________________________________________________________________________________________
(F
w
, F
m
, F
y
) kWh (or money) forecasted one day (one week, one month, one year)
before


16

4.2 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, Y
R

The Reference yield Y
R
is the daily (monthly or annual) in-plane irradiation E
S,A

divided by the STC reference in-plane irradiance G
STC
( =1 kW/m)
] [kW/m G
d)] [kWh/(m E
Y
2
STC
2
A S,
R

=
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, Y
A

The array yield Y
A
is the daily (monthly or annual) array energy output E
A,d
per kWp
of installed PV array power Po:
[kWp] Po
[kWh/d] E
Y
d A,
A =
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, Y
f

The final yield Y
f
is the daily (monthly or annual) plant useful energy output E
use
per
kWp of installed PV array power Po:

[kWp] Po
[kWh/d] E
Y
use
f =

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

17
The Performance Ratio, PR indicates the overall effect of losses on the arrays rated
output due to array temperature, incomplete utilisation of the irradiation, and system
component inefficiencies or failures. It is given by the ratio:

R
f
Y
Y
= PR

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, Y
R

[kWh/m]
[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

PV roof top plant Bovisa (3 kWp)
1999
Cesi (3 kWp) 1999 Cagliari (3 kWp)
1999
Reference Yield, Y
R

[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

b) 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):


18

Year 2000
Reference yield Y
R
h/d
3.84
Array Yield Y
A
h/d
3.56
Final Yield Y
f
h/d
3.32
Performance Ratio PR
%
86.00


19

4.3 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).

20

Technical Indicator Units Definition
Typical
Value
Low
Value
High
Value
1 Size
(1)
MWe Plant size in net power output 25.00 8.00 65.00
2
Generation (based on
# 1, 3, 4)
GWh/year
Total electricity (net) generation
per year
153.30 35.04 455.52
3
Capacity factor
(annual)
(2)

Total net electricity generation
per year divided by plant size in
net MW times the 8760 hrs in a
year
0.700 0.500 0.800
4
Hours per year
equivalent of the
capacity factor (if run
at full power)
(3)

hours
Number of hours that would
give the net annual generation if
all operation were at full net size
of plant
6132.00 4380.00 7008.00
5
Efficiency (higher heat
value)

Net output in heat unit
equivalent of the electricity
divided by the fuel heat input
using the higher heating value
of the fuel
0.201 0.155 0.263
6 Heat Rate (SI units) MJ/kWh
Fuel heat input per net
electricity output
18.00 13.50 23.00
7
HHV Higher heating
value (at moisture &
ash free)
(4)

MJ/kg
Heat content (HHV basis) if
there were no moisture and no
ash
19.70 18.60 20.90
8
Fuel ash content on
dry basis
%
Fraction of the total fuel input
by weight (dry, not as-received,
basis) that is ash
2% 1% 10%
9
Dry Higher heating
value (at ash given, but
no moisture)
MJ/kg
Heat content (HHV basis) if
there were no moisture, at ash
given
19.3
(2% ash)
17.7
(10% sh)
19.5
(1% ash)
10 Fuel moisture (5) %
Moisture (H2O) as a fraction of
the total weight of fuel (wet, not
dry basis)
30% 10% 60%
11
Wet Higher heating
value (at moisture
given, typical 2% ash)
MJ/kg
HHV at typical 2% ash content,
at moisture given
13.5
(30%
moisture)
7.7
(60%
moisture)
17.3
(10%
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


21

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.

22

4.4 Geothermal Energy

The following technical performance indicators have been proposed for geothermal
energy:

4.4.1 Definitions of the Proposed Performance Indicators

Capacity Factor = Total MWh generated in period x 100
Installed Capacity (MWe) x period (hours)

Load Factor = 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).


23

4.4.2 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 J apanese 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 J apanese
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 %


24


25

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 (km
2
).
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:

AvCO
2
(t/MW/y) = Avoided CO
2
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 Solar / PV Biomass Geothermal
*Wind Power
Denmark: avoided
CO
2
=
2000 t/MW/year

*US AWEA: 750kW
wind turbine avoids
1500t CO
2
/year

Must take into account CO
2

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 CO
2

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 CO
2
in 20 years
(i.e.
930t/MW/year)


26
Total European wind
farms (16300 MW)
avoid 30 million
tonnes CO
2
/year
(source ADEME
France)
In Paris, France, 106 m2 of
roof-PV cells avoid emissions
of 22t CO
2
per year
Roughly speaking, 1
ton of wood avoids 1
ton of fossil fuel CO
2

emission (see
comments below)

Italy: in 2010,
2500MW wind
should avoid
2MtCO
2
(i.e.
~800t/MW/yr)
(based on emissions
of modern CCGT).
New wind capacity
5.5 TWh.
A 60 years old US oak
and maple forest
stocks 2 tonnes C / ha
/ year

Comments:
The reference technology displaced must be quoted (coal, oil, gas). The avoided CO
2

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 CO
2
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 CO
2
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 CO
2
where compared to nuclear
generation. As shown in this case, the avoided CO
2
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 CO
2

abatement depends on the wind conditions and hence on the wind plant factor over the
year. The best way to express the CO
2
abatement is to put it in terms of tonnes of CO
2
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 CO
2
(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);

27
The specific costs of avoiding CO
2
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 CO
2
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:

Q
CO2
, Q
SOx,
Q
NOx
(g/kWh) =Quantities of CO
2
, SOx and NOx emitted per kWh during
the life cycle of the plant

Sample values:

Type of pollutant Wind Solar / PV Biomass
(energy crops
current practice)
Geothermal
CO
2
( g/kWh) 7 9 98 - 167 17 - 27 79 (source IEA)
122 (source IGA)
- Newest flash for
steam plants: 1 lb
CO
2
/MWhe (source
ORMAT)
SO
2
( 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 CO
2
emissions in geothermal plants, it should be noticed that CO
2
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 CO
2
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

28
does not create new CO
2
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 CO
2
, 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 (km
2
):

Definition:

A (km
2
) =Necessary area for exploitation of a plant producing 1 GW of electric power .

Sample values:

Land area for a
1 GW plant
Wind Solar / PV Biomass
Geothermal

100 km
2
30 km
2
5 000 km
2
200 km
2

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 km
2
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%;

29
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 km
2
for 50 MW. The proposed value is 200 km
2
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 1/10-
1/20 of the quoted number (source IGA).


30

5.1.4 External (environmental) cost (%):

Definition:

External unit cost
Ext (%) =
_______________________________
( 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 Wind Solar / PV Biomass Geothermal
External cost per MWh* 0.5 to 2.6
95

PV connected to
the grid ~ 1.4 to
3.3
95
(only one
study)
~ 2.0 to 50
95

(highly
dependant on
the type of
biomass).
0.2 0.5 ***
Average production cost
per MWh **
~ 45 ~ 650 ~ 128 30 100 ***
Corresponding Ext (%) 1.1 to 5.8 % 0.2 to 0.5 %

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:


31

Type of plant

* Average
investment cost
( / kW)

* Production
cost
(including
exploitation,
investment
amortization, fuel
and R&D)
( / MWh)

** External
costs
(
95
/ MWh)
(bracket
depending on
the countries)

% external
costs /
production
costs

Nuclear

1730

29
0.05 4.8
(0% discount
rate)
2.3 18.8
(3% discount
rate)
0.2 17 %
(0% discount
rate)
8 65 %
(3% discount
rate)
Coal (CFB) 1273 33 19 99 58 300 %
Gas (CCGT) 488 35 7 - 31 20 89 %
Hydro large scale 40 0.04 6.03 0.1 15 %
Hydro small scale 1250
Wind (on shore) 1000 45 0.5 2.6 1.1 5.8 %
Solar PV grid-
connected
8000 650
1.4 3.3 (only
one study)
0.2 0.5 %
Biomass 1200 128 2.0 50 1.6 39 %
Geothermal 2400/1800/1400*** 55/45/37 *** 0.2 0.5 *** 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;

32
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
d
min
(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
- 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 Systmes
Solaires)

33

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

34
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.

Q
CH4
(g/kWh) =Quantities of CH
4
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

H
2
S emissions during the life cycle of the plant (g / kWh)

Definition Comments and sample values *

Q
H2S
(g/kWh) =Emissions of H
2
S during the life of
the plant, in g per kWh

- 1 g H
2
S/kWh

- H
2
S is 1% in volume of the CO
2

released: this is a rough world-wide
average

* Source: International Geothermal Association (IGA)


35

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 jobyears 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 +J apan);
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.


36
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.

37

6.2 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 utilitys staff;

38
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.

39

7. 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.


40
8. 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 - 17
th
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 18
th
WEC Congress,
Buenos Aires, Oct. 2001
(3) - La production dlectricit dorigine renouvelable dans le monde Cinquime
inventaire - Edition 2003 - ObservER - EDF - Systmes Solaires Paris Dc 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 J aneiro, April 2000

General Web sites:

World Energy Council: www.worldenergy.org
UNIPEDE / EURELECTRIC: www.eurelectric.org
U.S. Department of Energy / Energy Efficiency
and Renewable Energy Network(EREN): www.eren.doe.gov
International Energy Agency (IEA) / Renewable
Energy Working Party: www.iea.org/techno
National Renewable Energy Laboratory (US): www.nrel.gov
Systmes Solaires / ObservER : www.systemes-solaires.com
CADDET Renewable Energy Newsletter : www.caddett-ee.org
WREN - World Renewable Energy Network: www.wrenuk.co.uk
Sustainable Energy and Development: http://solstice.crest.org
Renewable Energy DataBase: www.osti.gov/html/eren/eren.html
www.eeca.govt.nz
Sustainable Energy and Development: http://solstice.crest.org

Annex 2 provides additional web sites specifically for wind, solar, biomass and
geothermal energies.


41
ANNEX 1

Contributors to WEC PGP Work Group
On RES Performances Phase 2

Surname First Name Position Organisation Town Country
WEC WG Members:
* = Sub-group leaders **
=Chairman

Bertani * Ruggero Executive
Director
International Geothermal
Association
Pisa ITALY
Hoppe-Kilpper * Martin Head of Division
Information and
Energy
Economy
Institut fr Solare
Energieversorgungstechnik e.V.
(ISET)
Kassel GERMANY
Hughes * Evan EPRI Palo Alto USA
Manoha ** Bruno Environment
Managerial
Adviser
Electricit de France, Division
R&D
Chatou France
Vigotti * Roberto Renewable
Energy Unit
Manager
ERGA of ENEL
Business Development
Milan ITALY

WEC members:
Glorian Daniel Senior Advisor
to the Director
EDF - Division Recherche et
Dveloppement,
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
Electricit 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 Project Manager Mitsubishi Research Institute
Inc.,
Tokyo J APAN
Li Albert Senior engineer China Light & Power
International
Kowloon HONG-KONG
Millborrow David Consultant Wind Stats Lewes ENGLAND
Pineau Dominique Research
engineer
Electricit de France Chatou FRANCE
Rosillo-Calle Frank Research Fellow King's College London London ENGLAND
Salvaderi Luigi Consultant on
power systems
Rome ITALY

42
ANNEX 2

Renewable Energy in the World

Electricity in the World:

The following figures show the electricity production in the world in 2002 (ObservER
Systmes 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
Electricity
production 2002
(TWh)

Share % 2002
Average
annual growth
1993 2002
(%)
Annual
growth
2001 2002
(%)
Fossil 10 484.0 65.0 % 3.4 % 3.3 %
Hydro (incl. Pump storage) 2 643.0 16.4 % 1.0 % 1.8 %
Nuclear 2 720.0 16.9 % 2.4 % 2.1 %
Biomass 175.1 1.1 % 4.9 % 3.2 %
Geothermal 49.3 0.31 % 2.5 % 1.2 %
Wind 53.6 0.33 % 29.2 % 35.1 %
Solar 1.95 0.01 % 13.5 % 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.

43
Renewable Energies in the World by the end of 2002:

Installed
power
Forecast 2010 Installed in
2002
Annual
growth
rate (%)
Average
cost /
kWh
Top countries Observations
WIND 31 412 MW

(39 294 MW
in 2003)
90 000 MW (65
000 MW for
Europe)
6 868 MW

(7 948 MW in
2003)
35.1 %

(25,1 % in
2003)
4 to 7 US
cents
Germany 19.4 TWh
Spain 8.4 TWh
USA 8.1 TWh
Denmark 4.9 TWh India
3.0 TWh
Italy 1.4 TWh
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)
SOLAR / PV 1 246 MWp 11 300 MWp

(5 GWp J apan +3
GWp Europe
+2.14 GWp USA)
535 MWp
cells &
modules
produced
19.8 % up to
~1 US $
J apan 433 MWp
Germany 189 MWp
USA 171 MWp
Australia 34 MWp Italy
24 MWp
Netherlands 21 MWp
Exampl es : Carrisa Plains (USA) =5
200 kWp / Napoli (Italy) =3 300 kWp /
Saijo (J apan) =1 000 kWp / Munich
(Germany) =1 000 kWp / Toledo
(Spain) =1 000 kWp
BIOMASS 175.1 TWh
produced in
2002
2.6 Gtep average
scenario 2025
3.2 %
(4.9 %
mean
annual
1993-
2002)
~3 to 7
US cents
USA 69.2 TWh
J apan 14.1 TWh
Germany 11.7 TWh
Finland 10.4 TWh
Brazil 9.5 TWh
Canada 7.0 TWh
Various types of biomass :
Wood energy / Industrial wastes /
Domestic wastes / Landfill biogas
Size : 2 to 100 MW (average ~20MW)
GEOTHERMAL 8 356 MW
(49.3 TWh
produced in
2002)
between 21
GW and 32 GW
(depending on
scenarios)
382 Mwe
between
2000 and
2002
1.2 %
(2.5 %
mean
annual
1993-
2002)
3 to 5.5
US cents
USA 2 018 Mwe
Philippines 1 834 Mwe
Mexico 953 Mwe Italy
862 Mwe Indonesia 797
Mwe
Various types of technologies :
- Aquifers (180 to 350 C)
- Binary (90 to 150 C)
- Fractured deep rocks

W
o
r
l
d

E
n
e
r
g
y

C
o
u
n
c
i
l

P
e
r
f
o
r
m
a
n
c
e

o
f

G
e
n
e
r
a
t
i
n
g

P
l
a
n
t

2
0
0
4

S
e
c
t
i
o
n

5

4
3


44
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)
0
10
20
30
40
50
60
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
T
W
h
Western Europe
USA
Total World

Solar Electricity Production (TWh)
(Heliothermodynamics + PV)
0
0.5
1
1.5
2
2.5
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
T
W
h
Western Europe
USA
Total World


45
Biomass Electricity Production (TWh)
0
20
40
60
80
100
120
140
160
180
200
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
T
W
h
Western Europe
USA
Total World

Geothermal Electricity Production (TWh)
0
10
20
30
40
50
60
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
T
W
h
Western Europe
USA
Total World


46
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)
0
2
4
6
8
10
12
14
16
18
20
G
e
r
m
a
n
y
S
p
a
i
n
U
S
A
D
e
n
m
a
r
k
I
n
d
i
a
I
t
a
l
y
U
n
i
t
e
d

K
i
n
g
d
o
m
N
e
t
h
e
r
l
a
n
d
s
R
e
s
t

o
f

t
h
e

W
o
r
l
d

Solar Production in 2002 (TWh)
(Heliothermodynamics + Photovoltaic)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
U
n
i
t
e
d
S
t
a
t
e
s
J
a
p
a
n
G
e
r
m
a
n
y
I
n
d
i
a
A
u
s
t
r
a
l
i
a
I
t
a
l
y
M
e
x
i
c
o
N
e
t
h
e
r
l
a
n
d
s
S
p
a
i
n
R
e
s
t

o
f

t
h
e
w
o
r
l
d


47
Biomass Production in 2002 (TWh)
0
10
20
30
40
50
60
70
80
U
S
A
J
a
p
a
n
G
e
r
m
a
n
y
F
i
n
l
a
n
d
B
r
a
z
i
l
C
a
n
a
d
a
U
n
i
t
e
d
K
i
n
g
d
o
m
N
e
t
h
e
r
l
a
n
d
s
R
e
s
t

o
f

t
h
e
W
o
r
l
d

Geothermal Production in 2002 (TWh)
0
2
4
6
8
10
12
14
16
U
S
A
P
h
i
l
i
p
p
i
n
e
s
M
e
x
i
c
o
I
t
a
l
y
J
a
p
a
n
N
e
w

Z
e
l
a
n
d
I
n
d
o
n
e
s
i
a
R
e
s
t

o
f

t
h
e
w
o
r
l
d


48
Among the national and international standards and directives proposed for the
development of RES, we can note the following:
In theUSA, 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:

1995 2010
Biomass 44.8 Mtoe 135 Mtoe
Small hydro 9.3 GW 13.8 GW
Wind 3.5 GW 40 GW
1

Solar (thermal) 6.5 Mm2 100 Mm2
Photovoltaic 0.075 GW 3 GW
Geothermal electric 0.5 GWe 1 GWe
thermal 0.75 GWth 2.5 GWth
Others 0 1 GW

1
Considering the present trend, this target could be increased up to 65 GW in Europe in
2010

49
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): www.ewea.org
American Wind Energy Association (AWEA): www.igc.apc.org/awea/
IEA programme on Wind turbines: www.afm.dtu.dk/wind/iea/
Windenergie Report Deutschland 2003: www.iset.uni-kassel.de
Budesverband WindEnergie 2002: http://wind-energie.de
Wind Power monthly: www.windpower-monthly.com
Wind Stats Newsletter: www.gridwise.com/windstats/

Specific web sites for solar / PV:
European Solar Energy Association: www.eurosolar.org
Solar Energy Industry Association: www.seia.org
Utility Photovoltaic Group: www.upvg.org
IEA programme on Solar Power and
Chemical Energy Systems (Solar PACES): www.demon.co.uk/tfc/SolarPACES.html
IEA programme on Solar Heating and Cooling: 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: www.oit.edu/~geoheat/
Geothermal Database and Publications: www.smu.edu/~geothermal.htm
Geothermal Education Office: http://geothermal.marin.org/
Geothermal Energy Technology: www.doe.gov/get/getright.html
Heat Pumps: 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

50

51

ANNEX 3

Main Advantages and Drawbacks of RES Technologies

WIND SOLAR PV BIOMASS * GEOTHERMAL HYDRO

MAIN
ADVANTAGES

- Inexhaustible
- No emissions
- Local energy

- Inexhaustible

- Ideal for many remote
regions

- No polluting emissions

- Local energy

- Low maintenance

- Cheap energy

- Large variety of
fuel types

- Local energy

- Wastes elimination

- Agricultural
reconversion

- Clean, local and cheap
energy

- Independent on climate
or season

- High capacity factor

- Variety of direct uses

- Clean and cheap energy

- Instantaneous energy

- Storage of energy

MAIN
DRAWBACKS

- Discontinuous

- Visual impacts

- Noise

- Danger to birds

- Discontinuous

- Recycling of toxic
materials

- Grid connection

- Still very expensive

- Collection and
transport

- Air pollution

- Season or climate
dependant

- Competition with
other uses

- Limited zones

- Displacement of population

- Impact on river ecosystems

- Discontinuous

- Risks of conflicts for share
between countries
* Very much dependant on the type of biomass
W
o
r
l
d

E
n
e
r
g
y

C
o
u
n
c
i
l

P
e
r
f
o
r
m
a
n
c
e

o
f

G
e
n
e
r
a
t
i
n
g

P
l
a
n
t

2
0
0
4

S
e
c
t
i
o
n

5

5
1


52

ANNEX 4

Complete List of Proposed RES Indicators

Performance Indicator Type of RES Easiness to
measure /
estimate
Comparability
from one site
to another
Technical indicators

Plant size (MW)
All Easy OK
Expected life time All Rather difficult OK
Equivalent Full Load Hours (hours) All Easy OK
Capacity Factor (%) Wind,
Biomass,
Geothermal
Easy OK
Load Factor (%) Geothermal Easy OK
Specific Energy Production (kWh/m
2
) Wind Easy OK
Availability Factor (%) Wind,
Biomass,
Geothermal
Easy OK
Weather Variability Control Wind, PV Rather difficult OK
Reference Yield (hours/day) PV Rather difficult OK
Array Yield (hours/day) PV Rather difficult OK
Final Yield (hours/day) PV Rather difficult OK
Performance Ratio (%) PV Rather difficult OK
Efficiency (higher heat value) Biomass Easy Rather difficult
Fuel Moisture Biomass Easy Rather difficult
Annual energy sent out (MWh/year) Biomass Easy Rather difficult

Environmental indicators
Duration of operation to produce the
energy consumed for its manufacturing
(years)
All Rather easy Rather difficult
Man-hours to return to original state
(hours/kW)
All Rather easy OK
Avoided CO
2
(t/MW/y) All Easy Rather difficult
Q emissions CO
2
during production and
during the whole life cycle (g/kWh)
All Rather easy OK
Q emissions SO
2
during production and
All Rather easy OK
Q emissions NOx during production and
All Rather easy OK
Land area required for 1 GW (km
2
) All Easy OK
External cost (in /kWh, or in % of
production cost)
All Very difficult Difficult

53
Q
tox
toxic materials contained in the cells
& batteries
PV Difficult OK
Q
ash
quantities of ashes emitted (g/kWh) Biomass Rather difficult OK
QCH
4
quandities of CH
4
emitted
(g/kWh)
Biomass Easy OK
QH
2
S quantities of H
2
S emitted (g/kWh) Geothermal Easy OK
Dmin Visual / landscape protection
distance (m)
Wind Rather difficult Rather difficult
Sf Sound at foot of turbine (dB) Wind Easy OK
St Sound of turbine (dB) at standard
distance H+D/2 (according to norm IEC
61400-11)
Wind Easy OK
S500 Sound 500m away from turbine
(dB)
Wind Rather difficult Rather difficult
Birds killed (no/year) Wind Difficult OK
Shadow casting (hours/year) Wind Difficult Rather difficult

Sociological indicators
No J obs created direct (full time
employees) and indirect (n/MW)
All Rather difficult Rather difficult
Na People access to electricity (n/MW) All Easy Rather difficult
SAR number of accidents / 1,000,000
man-hours
All Easy 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.


54
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 Avail ability Value (BAV)
(Nor malised in %)
67.4%
90.4%
-0.9%
-2.7%
-7.7%
-28.9%
-1.9%
-40.0% -20.0% 0.0% 20.0% 40.0% 60.0% 80.0% 100.0%
Stati on A
Stati on B
Adding Value Not Competitive Missing Opportunity Destroying Value


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)


ii

R. Richwine (US)
C. W. Kim (Korea)
G. S. Stallard (US)
E. Nekhaev (WEC)

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 ESKOMS 90:7:3
PROGRAMME
Vincent Micali - Eskom, Generation, Energy Management Department.. 7

CASOM 4: DESIGN OR MANAGEMENT - WHICH INFLUENCES YOUR
PLANTS RELIABILITY MOST?
Robert R. Richwine - Consultant... 15

CASOM 5: GENERATING UNIT AVAILABILITY FOLLOWING PLANNED
OUTAGES

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

CASOM 9: AVAILABILITY IMPACT OF FLUE GAS DESULPHURISATION
SYSTEMS

CASOM 10: ESTIMATING NEW TECHNOLOGY RELIABILITY

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

iv
STEP 2 IDENTIFICATION

STEP 3 EVALUATION

STEP 4 IMPLEMENTATION

CASOM 16: ARE RELIABILITY MEASURES UNRELIABLE?
PART 1

PART 2 USING COMMERCIAL AVAILABILITY

CASOM 18: THE RELATIONSHIP BETWEEN SCHEDULED MAINTENANCE
AND FORCED OUTAGES AND ITS ECONOMIC IMPACT ON SELECTING
AVAILABILITY GOALS

CASOM 19: STATISTICS OR UNDERSTANDING: WHICH ONE DO YOU
BELIEVE?

CASOM 20: PEER SELECTION FOR BENCHMARKING - DOES IT MAKE A
DIFFERENCE?

CASOM 21: ESTIMATING A GENERATING PLANTS FUTURE
MAINTENANCE COST

CASOM 22: AGING OR VINTAGE
WHICH IS MOST RESPONSIBLE FOR DIFFERENCES IN BOILER TUBE
LEAK RATES?: PART 1


v
CASOM 23: BOILER TUBE LEAK STUDY: PART 2
Robert R. Richwine - Consultant...95

CASOM 24: CHANGING GAS TURBINE DESIGN REQUIREMENTS

CASOM 25: THE FUTURE IS NOT WHAT IT USED TO BE
Robert R. Richwine - Consultant..... 103


vi
INTRODUCTION
The main objective of the PGP Committees Work Group 7 is to develop the Committees
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 WECs 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 months 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: PGP Workshop Reliability Measurement,
Analysis & Benchmarking
9-11 October, Berlin: VGB Annual Conference
2003:
24-25 March, Amsterdam: Operational Outages for Power Generation
Conference
16-19 J une, Atlanta: ASME Conference
9 September, Kiev: WEC Executive Assembly, PGP Workshop
15-17 September, Copenhagen: VGB Congress Power Plants 2003
2004:
15 April, Rome: Seminar & Round Table Improving Reliability
of Generating Plant: The Real Value

1
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.

2

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

3
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
J ersey, USA, Minimising High Impact Low-Probability Forced Outages, published PWR-
Vol. 28 1995, J oint Generation Conference, Volume 3 ASME 1995.


4

5
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 cost-
effective 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.


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.


7
CASOM 3: AN APPLICATION OF BENCHMARKING TO ESKOMS
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 Africas 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

8
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 STATHAMs 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: Slow incremental allows resources accumulation
B: Hybrid incremental, as in A with rapid accumulation
C: Constant rate
D: Hybrid, medium resources & breather period
E: Step Change, resource intensive

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.
T*
E

9

Figure 2: KPI Gaussian Distribution

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).

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).

20%
Target
P
r
o
b
a
b
i
l
i
t
y

D
e
n
s
i
t
y

F
u
n
c
t
i
o
n

KPI
20%
Target
P
r
o
b
a
b
i
l
i
t
y

D
e
n
s
i
t
y

F
u
n
c
t
i
o
n

KPI (Profits)
Normal Weibull

10
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


11
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 Kick-
in: 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% Target of 90% UCF
1997: 91,5% had 20% chance of being achieved

Figure 5: Original UCF Strategy and Actual Performance


12
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 high-
performing 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.


13

Figure 7: Components of a 3,2% UCLF

IMPORTANT POINTS

Eskoms 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.

Generator Stator
Boiler Tube
Leaks
Milling Plant
Outage Slip
Combustion
Unit Electrical
Coal Plant

14
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. 17
th
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., J acobs N.B., Development of Production and Supply Agreements and
Performance Contracts, Generation Group Procedure GGP0422, Eskom, J ohannesburg,
February 1996.

Campbell I.G., Sustainability Index Managers Manual, Doc No. 7741 A 162 S, Eskom,
J ohannesburg, February 1998.


15
CASOM 4: DESIGN OR MANAGEMENT - WHICH INFLUENCES
YOUR PLANTS RELIABILITY MOST?

Robert R. Richwine
Consultant

BACKGROUND

There are many occasions when a plants 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
Councils (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 groups probability distribution of their units
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.


16
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 plants 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 plants future performance will need to
devote much more time and energy to evaluating the quality of the plants 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., J enkins, A.K., Establishing Realistic
Availability Goals Using Statistical Benchmarking Techniques.
Richwine, R.R., J enkins, A.K., Optimizing O&M Costs to Maximize Profitability.


17
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 units 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 Councils (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 doesnt, 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.


18
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 plants 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.


19
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 plants Optimum Economic
Availability (OEA). The first step in estimating a plants OEA is to determine the plants
Cost versus Availability curve (Figure 1).
4
$
AVAILABILITY 100%
Cost

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).

8
Cost
$
AVAILABILITY 100%

Figure 2


20
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.
9
Cost
$
AVAILABILITY 100%

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).

5
Worth
$
AVAILABILITY 100%

Figure 4

21
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 staffs understanding the
value of their plants performance.

By subtracting the minimum Cost curve from the Worth curve, we get the Net Value
curve (Figure 5).

6
Net Value
(Worth Minus Cost)
$
AVAILABILITY 100%
X
X
OEA

Figure 5

The point where that curve is at its maximum value is the plants 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 plants 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 plants OEA is a dynamic
goal, changing to meet the changing technical, operational and economic environment
that we are constantly facing.

Todays 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 plants 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 companys long-term
financial goals.


22
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., J enkins, K., Establishing Realistic Availability
Goals Using Statistical Benchmarking Techniques, 1995 International J oint Power
Generation Conference.

Richwine, R., Niebo, R., Peer Unit Benchmarking: Assessing Factors Affecting
Availability, 1992 International J oint Power Generation Conference.


23
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 Councils (NERC) Generating Availability Data Systems (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
1960s 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 units 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

24
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
MODE OF OPERATION
VINTAGE
Sub Super
CRITICALITY
Cycling
<1972
Baseload
Size
Draft Type
Fuel
Boiler Mfr.
Draft Type
Size
<1972

25
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, IJ PGC Conference,
1995.
Peer Unit Benchmarking, IJ PGC Conference, 1992.
Unit Benchmarking Demonstrates compatibility of European and North American Data
Collection Systems, WEC Congress, 1995.


26

27
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 didnt know. Therefore we decided to investigate by
comparing the actual EFORs 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.

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 systems 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 companys 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
6
EFOR
ACTUAL
- EFOR
PREDICTED

28
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 units 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 units 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 years reliability (no surprise this is what we had been doing);
Previous years duty cycle;
Current year Operations and Maintenance activities including outage time and
spending;
Previous years 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 EFORs for that year. The results are shown in the dashed line
frequency distribution in Figure 2 below (the solid line is the original distribution).

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.
1
0
EFOR
ACTUAL
- EFOR
PREDICTED

29
References:

Predicting Unit Availability: A Top-Down Perspective, published by the North American
Electric Reliability Councils (NERC) Generating Availability Trend Evaluation (GATE)
Working Group, 1991.
Predicting Unit Reliability, published by NERCs GATE Working Group, 1995.


30

31
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 units availability and
reliability. Therefore the North American Electric Reliability Councils (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 NERCs
Generating Availability Data System (GADS) as well as information available from
federal agencies, industry organizations and internal data from the task forces
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
NERCs 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 systems 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

32
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 units capacity;
Station Service (house load or internal station electricity requirements) annual energy
requirements for the FGD systems averaged 1.67% of the units capacity (median =
1.28%). J ust 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 J oint Power Generation Conference, Atlanta, Georgia, October 18-
22, 1992.


33
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 plants 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 plants 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, J une 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 Plants 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).


34
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 Generators Performance Past,
Present and Future.

Figure 1

1
2
Ou t ag e Rat es v er s u s Year
o f In i t i al Op er at i o n

35
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 plants
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 plants reliability (or
Operations and Maintenance costs either, but thats another study). My Reliability
Engineering Department decided to undertake an analysis comparing generating plants
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 companys 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 plants 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 CFs 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.

36
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 Councils
(NERC) Generating Availability Data System (GADS) we plotted the regression line of
FOR vs. OF as shown in the figure below.

RELIABILITY vs. DEMAND
N NE ER RC C
G GA AD DS S

37
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 operators 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
OFs. 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 operators
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 OFs, 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 plants 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, IJ PGC Conference,
1995.
Peer Unit Benchmarking: Assessing Factors Affecting Availability, IJ PGC Conference,
1992.
The Importance of Data Recording and Analysis in Managing Availability Improvement,
WEC Performance of Generating Plant Conference, 1985.

38


39
CASOM 12: PERFORMANCE DATA TO PERFORMANCE
IMPROVEMENT:

STEP 1 AWARENESS

Robert R. Richwine
Consultant

US$80 Billion per year PLUS 1 Billion tonnes of CO
2
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 Plants Reliability Most?). This months 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 Councils (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 months 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 companys 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:


40
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
todays 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 todays 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 companys available (but
limited) resources. Next months 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
projects 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.

41
AWARENESS

The first step in successful Performance Improvement Programmes is to create awareness
in the companys generation executives and staff of the opportunity for improvement in
their plants 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 plants 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 months 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 plants 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).


42
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
units 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 units 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 plants limited resources.

Value of improved performance of the companys 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 companys
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 companys 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.


43
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 staffs
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 plants 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.


44

45
IMPROVEMENT:


Robert R. Richwine
Consultant

This months 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 months 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.

PROGRAMMES

Step 1 - Awareness

exists at each of the companys 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:


limited) resources. This months case study will describe in more detail some of these
identification techniques being used by top performing companies.

46
Step 3 Evaluation

After all viable options have been identified (Step 2) their cost and technical impacts on
quickly and cost-effectively. A future months case study will describe this process in
more depth.



STEP 1 - AWARENESS


There are many ways in which historical plant data can be of invaluable assistance in
identifying trouble spots and viable options for improving a plants 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.

47
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
plants 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 Turbo- Feed Fuel Boiler Other Total
Tubes Generator Water Burn

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.

48
Unplanned Unavailability (%) by plant System
(With peer plants results added)

Plant Boiler Turbo- Feed Fuel Boiler Other Total
Tubes Generator Water Burn

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 3.92 0.55 1.15 0.44 1.56 1.88 9.50
(Average)

Peers 2.56 0.32 0.81 0.33 1.01 1.55 6.58
(Best quartile)

Peers 1.18 0.12 0.23 0.11 0.50 0.77 2.91
(Best decile)

By examining this data we can clearly see that the area with the most potential for cost-
effective improvement is plant D in the feed water system (no surprise since its 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 plants equipment configuration (e.g. a reliability
block diagram) showing the amount of equipment redundancy and the percentage that
the plants capacity is reduced when one or more components are out of service;
Failure rate and repair time for each of the components;


49
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 plants 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 plants 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 plants 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.


50

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 companys
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.

51
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
plants 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 companys 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 plants reliability most? This study
demonstrated that between 75%-80% of a plants 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 plants 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 technologys 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
months 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.


52
References:

Richwine, R.R. The Importance of Data Recording and Analysis in Managing
Availability Improvement, presented at the World Energy Council Conference.


53
IMPROVEMENT:

STEP 3 EVALUATION

Robert R. Richwine
Consultant

This months 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.

PROGRAMMES

Step 1- Awareness

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:


limited) resources.

Step 3 Evaluation

quickly and cost-effectively. This months case study will describe this process in more
depth.

54

This post installation comparison of expected versus actual results will then be fed back

STEP 1 - AWARENESS
(For details of Step 1 AWARENESS see CASOM 12)

(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:
J ustification;
Prioritisation.


55
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 dont
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.


56
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.


57
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 J ustification we asked:

Is this a good project Yes or No?

Now we must answer a more difficult question:

If the company doesnt 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-years delay this would be the difference between the companys cost of
capital and the forecast inflation rate of the projects 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.


58
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 plants
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 near-
term 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 plants staff and corporate

59
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.

60

61
IMPROVEMENT:


Robert R. Richwine
Consultant

This months case study will conclude a four part series that has demonstrated how top-
performing 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.

PROGRAMMES

Step 1 - Awareness

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:


available channel to identify those viable options, from original equipment manufacturers
limited) resources.

Step 3 Evaluation


62
quickly and cost-effectively. This months case study will describe this process in more
depth.

Step 4) Implementation


STEP 1 - AWARENESS
(For details of Step 1 AWARENESS see CASOM 12)

(For details of Step 2 IDENTIFICATION see CASOM 13)

STEP 3 EVALUATION
(For details of Step 3 EVALUATION see CASOM 14)


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.

63
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 decision-
maker must select from among many good projects that group of projects, which will
make best use of the companys 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 plants Optimum Economic Performance (OEA). This process will also
estimate the money required to achieve and maintain the plants OEA. If there is
insufficient funding provided to achieve this OEA goal (each plants unique point of
diminishing returns) then with the actual funding provided, the plants 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 plants 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

64
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:

Common Elements
In Successful Performance
Improvement Programs
A
w
a
r
e
n
e
s
s
I
m
p
l
e
m
e
n
t
a
t
i
o
n
I
d
e
n
t
i
f
i
c
a
t
i
o
n
P
r
i
o
r
i
t
i
z
a
t
i
o
n
Performance
Improvement


65
CONCLUSIONS

This months 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 plants 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., J enkins, 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.


66

67

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 companys bottom line from the failure to generate during
those hours.

This months 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%

68
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 dont come close to describing the units 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 units 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 units reliability so that

Reliability =100% - 25% =75%

So it appears that FOR (and EFOR when forced deratings are present) are good measures
of a units 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 units 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.

69
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 units 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 plants 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 units 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 units 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 non-
peak 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
units 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.


70
This choice also had the effect of minimising the forced outage hours resulting from this
event.

However, when the exact event occurred during the companys 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 plants availability beyond that of Option 1.

As we can see, often the objective of maximising a units technical performance (in this
case availability) brings it into direct conflict with the companys 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 plants goals and the companys 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
UnreliableIts Time for a Change; Published in the IGTIs 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 17
th
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 18
th
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.


71

PART 2 USING COMMERCIAL AVAILABILITY

Robert R. Richwine
Consultant

INTRODUCTION

Last Month we examined why traditional measures of a power plants
availability/reliability (EAF, FOF, UCF, UCLF and EFOR) were inadequate in todays
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 plants availability has on the companys 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 1990s
following the deregulation of its power industry into a market system. Since a plants
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 plants
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 plants impact
on the companys profitability. (E.g. during hours when the gap is US$20/MW-HR the
plants 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 plants 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).


72
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 plants economic
incentive to generate at different times. Since the greater the economic incentive to
generate is, the better the plants 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
units 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
(thats the probability part) that the unit can deliver the requested amount of energy
during a specified time period corresponding to that units demand profile (thats 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 units unique
economics we can then arrive at a goal Commercial Availability objectively

73
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 plants 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 companys 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?

74
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 Councils 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
UnreliableIts Time for a Change, Published in the IGTIs 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 17
th
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 18
th
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.

75
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 units Optimum Economic Availability (OEA)
have focused on comparing the units 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.


76
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 units 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 plants
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 plants outages on the companys financial performance, no
single definition has been adopted by the industry as yet and there are other implications
that have not been fully resolved.

77
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 counter-
productive 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.


78

79
CASOM 19: STATISTICS OR UNDERSTANDING:
WHICH ONE DO YOU BELIEVE?

Robert R. Richwine
Consultant

This months 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 companys power plant performance data undertook a study that
changed the maintenance philosophy of his company, eventually helping to increase the
plants 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 youve been showing us? Ive been placed in charge of
collecting our plants 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 companys 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 dont 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).

80
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 companys 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 months 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 dont conform
with the understanding expressed by your plant staff, Id 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.


81
CASOM 20: PEER SELECTION FOR BENCHMARKING-
DOES 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 RECENT VINTAGE

EFOR (mean) 15.60% 9.68%
EFOR (median) 12.17% 8.08%
EFOR (best quartile) 8.14% 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.


82
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 NEW CRITERIA % difference
(Coal; 100-199MW)

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 NEW CRITERIA % difference
(Coal; 800-1300MW)

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.

83
CASOM 21: ESTIMATING A GENERATING PLANTS FUTURE
MAINTENANCE COST

Robert R. Richwine
Consultant

BACKGROUND

This months 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:

84
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 - 80%

2) a) Probability of repair - 20%
b) Probability of replacement in kind - 20%
c) Probability of upgrade - 60%
total = 100%

3) Cost Range for 2a US$200,000 - US$400,000
2b US$750,000 - US$1,000,000
2c US$2,000,000 - US$4,000,000

4) Years from now in which cost could occur- 16 years 20 % chance
17 years 20% chance
18 years 20% chance
19 years 20% chance
20 years 20% chance
Total = 100%


85
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


1) Boiler Controls No! . . .

2)

3)

50)

TOTAL US$12,320,000

For draw #2 the cost would be US$12,320,000


86
DRAW #10,000


1) Boiler Controls YES Upgrade US$2,500,000 17 yrs

2)

3)

50)

TOTAL US$21,750,000

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.


87
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 plants major equipment using various non-
destructive 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 months 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.


88
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., J enkins, K., Optimizing O&M Cost to Maximize Profitability.


89
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
months case study summarises the results of that investigation. Next months 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 in or after 1975.
(SUB 75) (107 units) (recent vintage large subcritical units)

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)


90
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 - 78%
Superheater - 47%
Reheater - 75%
Economiser - 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:

Waterwalls - 53%
Superheater - 32%
Reheater - 38%
Economiser - 38%

There has been a substantial reduction in average repair time:

Waterwalls - 16.7 hours reduction
Superheater - 25.4 hours reduction
Reheater - 16.1 hours reduction
Economiser - 25.0 hours reduction

Overall there was a 60% reduction in boiler tube leak unavailability almost three
percentage points.


91
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.


92

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.


93

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.


94

95
CASOM 23: BOILER TUBE LEAK STUDY: PART 2

Robert R. Richwine
Consultant

In last months case study we summarised part of a study performed by NERCs
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.

96
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 months 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.


97

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 units
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 J an April 2003 for a detailed discussion of this
successful process).


98
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.


99
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 clients 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 companys profitability.

Plant A consists of simple-cycle gas turbine generator sets, used almost exclusively for
meeting the companys 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%).

100
Therefore, in this business structure a generating companys 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 anothers
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
stockholders 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 companys 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
companys 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 companys
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 makers 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 companys 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 generators 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

101
applications (and requiring his marketing force to try and convince the customer that its
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 plants reliability and efficiency
during steady state operating conditions.

The following generic table indicates the relative* importance of several key factors
influencing a power producers gas turbine equipment purchase decision according to its
anticipated future duty cycle.


102
RELATIVE* IMPORTANCE OF CHARACTERISTICS BY APPLICATION

Gas Turbine Gas Turbine Plant Application

Characteristic Peaking Cycling Baseload

Availability Low High (day)-Low (night) 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 customers 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 authors opinion, instead of using a one-design-fits-all approach,
will take aggressive steps to understand and quantify their customers priorities and
values and incorporate them into a range of designs, each one optimal from the owners
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, J an. 1996.

103
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
yesterdays 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 units 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.

Generating Plant

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Generating Plant

Uploaded by

Copyright:

Available Formats

Performance of Generating Plant:

New Realities, New Needs

You might also like