Westinghouse PWR Dominant design-MIT-Carlos Barrientos

WESTINGHOUSE PWR: THE RISE AND FALL OF A DOMINANT DESIGN IN THE
ELECTRIC POWER INDUSTRY
by
Carlos J. Barrientos
Master in Nuclear Engineering

Instituto Balseiro, Argentina, 1992
Submitted to the Engineering Systems Division

in Partial Fulfillment of the Requirement for the Degree of
Master of Science in Technology and Policy
at the
Massachusetts Institute of Technology
June 2002
@ 2002 Carlos J. Barrientos. All rights reserved
The author hereby grants to MIT permission to reproduce

and distribute publicly paper and electronic
copies of this thesis document in whole or in part.
Signature of Author: .
Technology and Policy Program, Engineering Systems Division
May 24,2002
Certified by;.r ~ .
John S. Carroll
Professor of Behavioral and Policy Sciences
MIT Sloan School of Management
Thesis Supervisor
Accepted by: .
Daniel Hastings
Professor of Aeronautics and Engineering Systems
Director, Technology and Policy Program
Chairman, Committee for Graduate Students
DEWEY
2
Westinghouse PWR: the rise and fall of a dominant design in the electric power industry
by
Submitted to the Engineering Systems Division on May 24, 2002
in Partial Fulfillment of the requirement for the Degree of Master of Science in
Technology and Policy
ABSTRACT
In the early 1950s the electric power industry was shaken by the introduction of a new type of energy
source that pledged to be the most cost-efficient way of producing electricity. Counting on the promise of
huge profits -attributed to large economies of scale- governments and utilities rushed to develop and
construct nuclear power plants. From 1952 to 1985, four hundred units were built all across the
industrialized world, from the United States to the Soviet Union and from Sweden to Taiwan. Nuclear
powt;r accounted for the most impressive capacity build-up in 100 years of electricity history.
As was the case with most of the strategic technologies developed in the 20th century, the first nuclear
prototypes were developed in the United States, with the sponsorship of the federal government. The
funding ended up mainly in R&D expenditures administered by the Atomic Energy Commission (AEC).
As a result of such a strong commitment to nuclear, the United States quickly became the technological
leader in the world. Such leadership meant not only building most of the nuclear power plants in its
territory -125 out of 480- but establishing industry standards in many areas: from design, licensing and
construction to commercial operation, and decommissioning.
This research explores one of the most substantial legacies of the U.S. nuclear power undertaking: the
PWR reactor developed by Westinghouse. The research examines the story of the PWR from its origins in
the drawing rooms of the Bettis Laboratories in the 1950s to its rapid adoption in the 1960s as the
dominant design in the industry. The main goal of the research is describing the dynamics of the process,
building at the same time a workable framework of analysis.
The fITst part of the research digs down into plants' data. Using a large database of reactors' records, the
dominant design hypothesis is tested thoroughly. The analysis confirms that the early design proposed by
Westinghouse quickly became the standard of the industry. Nearly two-thirds of all reactors built in the
world have their roots in the early Westinghouse design.
The reasons for the emergence of such a dominant design are numerous: (a) the influence of military
nuclear programs, such as the nuclear submarine; (b) the monopoly of the AEC regarding nuclear secrecy,
technology transfer and industry partnership; (c) the role of the cold war as a driving force in nuclear and
space policy; and (d) the obscure alliance between Westinghouse and GE regarding competition on
electrical components, notably the large steam-turbines used in nuclear power plants.
The emergence and consolidation of a dominant design in any industry has many consequences. In
nuclear power, some of the relevant issues are standardization, learning effects, economies of scale, and
regulation. All these issues are important to study not only in the United States context. The international
consequences are vast on nuclear power programs and policies of countries such as Japan and France.
Since most dominant designs have a rise and a fall, the research includes an analysis of why in the mid-
1980s the Westinghouse PWR collapsed, along with the entire nuclear power industry. The thesis is that
incumbent frrms that have successful dominant designs in the market, very often fail to be aware of subtle
-but disrupting- shifts in customer needs. Westinghouse was busy building large and complex units in
order to increase efficiency and profits, while customer needs were moving in a radically different
direction, towards less investment risk.
Thesis Supervisor: John S. Carroll

Title: Professor of Behavioral and Policy Sciences
3
In gratitude
When I graduated with the degree of nuclear engineer, I thought that in order to understand my profession
I needed to gain mastery in the intricacies of complex subjects such as nuclear physics, reactor theory,
fluid mechanics, thermal transfer, etc. After some years I realized that in order to fulfill my goal of
understanding nuclear engineering more profoundly I needed to move in other direction. I needed to
understand nuclear technology in its human dimension, as the outcome of the effort of a particular group
of people, who directed an enormous amount of work and capital pursuing an elusive and controversial
objective.
My quest for insight inspired me to visit Hiroshima, Japan in 1997 and more recently Los Alamos, New
Mexico. But more importantly, it motivated me to write this thesis which is, above all, a tenacious
attempt to understand the complex issues surrounding the birth and growth of nuclear power in the United
States and abroad. It also fulfills the requirement for the Degree of Master of Science in Technology and
Policy but, to be honest, that has been only a secondary objective.
I am particularly indebted to Professor John S. Carroll, who took the risk of supervising a young man's
way into MIT; to Sydney Miller for her support, encouragement and patience during a process that
became especially difficult in the last mile; to Dean Isaac Colbert and Sarah Gallop for their willingness
to help. In addition to these mentors, Professors Neil Todreas, Richard Lester, Ernest Moniz and Andrew
Kadak helped throughout the research with valuable comments and feedback. MIT has been for me the
fountain of knowledge and human generosity.
I am similarly indebted to Rotary Foundation for sponsoring my studies and for providing me the
opportunity to develop a lasting friendship with many Rotarians of the Boston area. Rotary's
ambassadorial scholarship has increased the value of my studies enormously and has helped me to go
through the turbulent waters of graduate school. In particular I want to mention Karen and EJ Swaim-
Babin, Chris and Dianne Fraser, and Alberto and Zareen Araoz. I also want to thank Yu-Ling Hsu for
helping me with the typing of several long and tortuous fragments of the text. Her support and her
unhesitating faith on my ability to finish the thesis have been an inspiration to me.
I owe the most to my family in Argentina -my parents Carlos and Martha- for my absence from home
during very difficult times. I dedicate my thesis to them.
Cambridge, May 24th 2002
4
Introduction
THE DAWN OF NUCLEAR ENERGY

Nuclear energy emerged after the strikes over Japan at the end of World War II. At 8:15 am of August 611\
1945, a bomb was dropped over Hiroshima from an aircraft flying at 31,600 feet. Fifty seconds later a
flash was seen by the pilot and immediately a shock wave hit the plane. Five minutes later, a dark gray
cloud some three miles in diameter hung over the center of Hiroshima. Because of this dense smoke
cloud, which ultimately rose to a height of 35,000 feet, it took several hours to have a clear view of the
devastated city. Two-thirds' of the total urban area was completely destroyed and, according to Japanese
official estimates, 71,000 people died and 68,000 were injured. Two days later, a second bomb was
dropped over Nagasaki from an altitude of 29,000 feet. This nuclear device seemed to be less lethal as
only 44% of the urban area was destroyed, and the fatal casualties, according to an US survey, amounted
to 35,000 people dead or missing (Del Sesto, 1979).
One can hardly imagine a more sensational and dramatic way to mark the dawn of a new technology. It
was by any account the most amazing display of energy ever unfolded by human civilization. As for the
planet harboring that civilization, it was one of the largest explosions registered in modem history, rivaled
th
only by the impact of a large meteorite in Siberia at the beginning of the 20 century.
Notwithstanding its spectacular and devastating introduction as a weapon, nuclear technology has made
its way in modem civilization primarily by means of peaceful applications. As many other post-war
branches of lmowledge, nuclear technology evolved during the second half of the 20th century from full
military monopoly to extensive civilian control. This technology has found a thorough utilization in fields
as diverse as medicine, materials science research, forensic analysis, archaeology, oil exploration, and
electricity production.
All across the world, scores of professionals use nuclear equipment in their routine activities. Physicians
use Cobalt irradiators for cancer therapy. Courts base their findings on forensic studies that involve the
use of sophisticated techniques such as neutron activation analysis. Archaeologists determine the
chronological date of items using the powerful Carbon-14 methodology ..
Even though the above-mentioned applications have been very important, no other civilian use of nuclear
technology has had a more profound impact in our society than the use of nuclear energy for producing
electricity. No other source of energy has evolved so rapidly over the last five decades and -without any
doubt- no other source of energy has inspired so much public controversy about its benefits and risks.
PURPOSE OF THE STUDY

This study is centered on the spectacular growth of nuclear power between the 1950s and the late 1980s.
Its main purpose is to determine the role of the PWR -a design developed in the u.s. by Westinghouse-
which quickly became the industry standard. In the same way that Douglas shaped the commercial
aircraft industry with the introduction of the DC-3 in the early 1930s, Westinghouse profoundly molded
the nuclear power industry with the inception of its PWR design.
The purpose of the study is thus twofold. On the one hand it has a historical focus on describing the
PWR's chronological progress in the two crucial dimensions of product development: technology and
Introduction
market. On the other hand the study has a policy goal which is to identify the public and private policies
that both favored the emergence of the PWR as an industry standard, and eventually caused its demise a
few decades later.
Studying the development of nuclear
power plants is an interesting exercise
for many reasons. First, nuclear power is 1989
a very complex technology that has an
enormous economic impact nowadays.
That by itself makes it an important
subject of study. But the analysis of
nuclear power reveals that the industry
has followed an observable technology
cycle characterized by a rapid
penetration of the market and in a like
manner, a rapid phase out. Analogously
to Kuhn's pioneering work on the
structure of scientific revolutions,
nuclear technology has described the
characteristic pace of technological
progress. It advanced slowly at first, Figure I. 1.Nuclear power development from 1952 to 1989
then accelerated, and then inevitably
declined (Kuhn, 1962).
In the particular case of nuclear power, the processes of acceleration and decline were embedded in the
designs that populated the marketplace. The thesis of this study is that those processes were shaped by the
fact that an industry standard -or dominant design- emerged at some point in the story (Utterback, 1994).
MAIN RESEARCH QUESTIONS

If the data presented in the study show sufficient indications to believe that the PWR achieved the status
of a dominant design, then a number of interesting questions arise:
1. Identifying and understanding the reasons that motivated the overwhelming adoption of this
particular design. In particular: (a) public policies that favored the PWR, (b) inherent competitive
advantages of this technology over other designs, (c) Westinghouse's value proposition to eventual
buyers, (d) the role of other participants of the value chain, including suppliers, and (e) the impact of
technology transfer policies to foreign countries.
2. Identifying and understanding the consequences of the emergence of a dominant design in the
nuclear power industry. In particular: (a) effects on the survival of firms, (b) impact on the
performance of the industry, (c) role in the industry's technology cycle, particularly in promoting -or
inhibiting- innovation, and (d) accountability for the ultimate collapse of the nuclear power industry.
3. Identifying and understanding the lessons of the Westinghouse PWR case for other rapid-growing
industries. In particular: (a) dominant designs and its role on technology cycles, (b) public policies
that both promote and regulate an industry, (c) early warnings regarding firms' destructive
competition, and (d) policies to stimulate innovation and industrial performance.
SCOPE OF THE STUDY

The scope of this study includes all the commercial nuclear power plants that were built in the world
before 1989. This large ensemble accounts for 481 units in more than 25 countries (Figure 1.2) and
consists of 100% of the market.
6
Introduction
Even though the scope of the study is over 481 units, the study has a particular focus on the 282 PWR and
BWR reactors, for reasons that will become clear in later chapters. The rest of the nuclear power plants -
. 199 units- were units of other designs: AGR, PHWR, LWGR, etc. In addition, there is an unintentional
focus on U.S. plants for two reasons: it is the nation where the PWR concept was originally developed
and where more PWR units were built.
USA 136 SWITZERLAND 6

FRANCE 60 TAIWAN 6
USSR 59 BULGARIA 5
UK 42 FINLAND 4
JAPAN 38 HUNGARY 4
GERMANY 30 ITALY 4
CANADA 21 ARGENTINA 2
SWEDEN 13 HOLLAND 2
SPAIN 10 MEXICO 2
CZECH REP. 9 S. AFRICA 2
KOREA 9 PAKISTAN 1
BELGIUM 8 YUGOSLAVIA 1
INDIA 7 TOTAL 481
Figure 1.2. NPPs that starled operation before 1989, divided by country and design.
hi order to draw valid conclusions, a specific database was built with data assembled from different
sources, especially from industry handbooks of the late 1980s (Cruickshank, 1989). This database was
developed in SPSS format and is composed of 481 cases and 51 variables. SPSS is a standard statistics
software that e~pedites the management of data and allows multiple operations.
It is important to note that the research scope is commercial nuclear power plants that started operation
before 1989, both units that are still functioning and those that have been shutdown. The following
reactors are excluded from the study:
1. Nuclear power plants that started operation after 1989, including units that are currently under
construction and those with works temporarily halted or cancelled.
2. Nuclear reactors that are not intended for electricity generation, namely (a) research reactors, (b)
radioisotope production reactors, and (c) plutonium production reactors.
3. Nuclear reactors developed for propulsion. In particular, (a) reactors for space propulsion, and (b)
reactors for naval propulsion such as the ones installed in submarines and aircraft carriers. As will be
shown in subsequent chapters, these reactors had a significant role in the emergence of the PWR as a
dominant design and that connection will be explored in detail. However these reactors are not the
primary focus of the research.
SYNOPSIS
Since the main thesis of this study is that a dominant design in the nuclear power industry emerged in the
1960s, the first part of the study is devoted to a historical recount of the PWR' s origins and a thorough
analysis of its progress in the technology and market dimensions. After establishing with reasonable
confidence that a dominant design effectively emerged in this industry, the final part of the study is aimed
to develop a feasible explanation of the PWR progress: from early development and market acceptance, to
its final demise. The goal is to develop a workable framework that serves to explain both the reasons and
the consequences of the emergence of a dominant design. In order to achieve these objectives
systematically, the material is distributed in the following chapters:
7
Introduction
1. The early origins of the PWR Design

The chapter starts explaining the foundation in 1946 of a civilian nuclear power program managed by the
ABC. Then it introduces the complex relationship between military and civilian objectives in the dawn of
the nuclear age, which was shaped by the advent of the cold war. The chapter also explains the early years
of the partnership between ABC and the industry for building nuclear reactors. In particular, the chapter
discerns the role of Westinghouse as main contractor of key military projects and ABC's decision of
selecting it for supplying the power system for the 15t nuclear submarine. The chapter includes a thorough
analysis of the environment at the time, namely: (a) the rush to finish the projects ahead of schedule and
-(b) the conflicting interests of stakeholders such as Congress, ABC, the military, and Westinghouse.
2. The rise of the PWR in the marketplace
Following the successful development of a nuclear submarine in 1954, Westinghouse started the
construction of the first PWR nuclear power plant at Shippingport using a design based on the submarine
reactor. This reactor and project manager -Admiral Hyman Rickover- shaped the nature of the
Westinghouse PWR and ultimately of the US nuclear power program. The chapter analyzes the PWR's
subsequent progress on the technology and market dimensions. It shows how the design quickly became
the most sold nuclear power plant in the world, capturing almost two-thirds of the market. The objective
of the chapter is to show that if a dominant design emerged in the industry that dominant design was the
Westinghouse PWR.
3. The dominant design
This chapter first describes the Westinghouse PWR design in detail, focusing on its distinctiveness from
other designs. A large amount of data is provided, with the proposition that all the main design parameters
of the reactor, such as pressures, temperatures, and mass flows, converged to the same numbers. The
dominant design quickly emerged as the industry standard not only in the U.S. but also in France, Japan,
Belgium and the rest of the western countries.
4. The BWR and the hidden link with the PWR
This chapter focuses on the development of the BWR design, i.e. the GE approach to the civilian nuclear
power plant and the main competitor to the Westinghouse PWR. The chapter starts with a thorough
description of the design and the differences and similarities with the PWR. The BWR also stabilized in
certain values and this will be shown following the same criteria explained in the previous chapter. Using
appropriate data, this chapter also shows that the main connection between the designs proposed by GE
and Westinghouse was through the use of similar suppliers and components. In particular both designs
have essentially the same steam turbine. This finding is used to argue that one of the main reasons why
the Westinghouse PWR design stabilized so quickly was because of the standardization of one of the
main component of nuclear power plants: the steam-electric turbine.
5. The relationship between Westinghouse and General Electric
This chapter is basically about antitrust. It starts with a summary of several cases of price fixing where
Westinghouse and GE were involved. The emphasis is on the notorious case in the early 1960s involving
large turbines, which led to the imprisonment of several corporate executives. Since the subject of the
thesis is not primarily antitrust, the stress is purely historical and within the context of nuclear power
plants. The question that the chapter raises is whether price-fixing strategies employed by Westinghouse
and GE in large electric components molded the rise of the PWR as a dominant design in the market.
6. The manufacturers game
Given the relevance of collusion and price fixing in the power industry, this chapter has an analysis of
manufacturers and their involvement in the design and construction of nuclear power plants. The analysis
8
Introduction
starts with a model of the order/construction process, which describes the role of utilities, architect-
engineers, civil constructors, reactor vendors and secondary suppliers. The way these firms interacted had
a profound significance on the consolidation of the Westinghouse PWR design. In addition, the chapter
focuses on the performance of the PWR design internationally. The data on France, Japan, Belgium,
Germany, Sweden, Spain, Taiwan and Korea seem to show that these countries overwhelmingly accepted
the Westinghouse and GE designs as the industry standards.
7. The demise of the dominant design
An analysis of the PWR's collapse in the mid-1980s suggests that incumbent fmns that have successful
dominant designs in the market, such as Westinghouse and its PWR design, very often fail to be aware of
subtle -but disrupting- shifts on the demand side. While Westinghouse was busy building large and
complex units in order to increase efficiency and profits the customer needs were moving in a radically
different direction, towards power units that had less inherent risk (as large capital investments). The
chapter shows that there was a fundamental change of mindset among industry participants and
consumers from the late 1950s to the early 1980s.
B. The 1990s and beyond: lessons from the Westinghouse PWR case
A rapid fast-forward through the 1990s explores the achievement of the long-awaited positive
externalities of a dominant design. It shows how the performance of nuclear power plants has increased
notably, through a combination of industry consolidation, specialization, cumulative learning effects, and
economies of scale. The chapter finishes with a summary of the Westinghouse PWR case and a
framework for understanding the issues from the point of view of public policy and management of
technology. Some of the issues discussed are: (a) dominant designs and their role on innovation and
industrial performance; (b) pros and cons of standardization; (c) factors that promote the emergence of
dominant designs; (d) dominant designs triggered by suppliers; and (e) dominant designs vs. customer
needs.
9
Chapter 1
THE EARLY ORIGINS OF THE PWR DESIGN
The chapter starts with the foundation in 1946 of a civilian nuclear power program managed by the AEC
and the complex relationship between military and civilian objectives in the dawn of the nuclear age.
Then it explains the early years of the partnership between ABC and the industry for building nuclear
reactors. In particular, the chapter discerns the role of Westinghouse as main contractor of key military
projects and ABC's decision of selecting it for supplying the power system for the 1st nuclear submarine.
The chapter includes a thorough analysis of the environment at the time, namely: (a) the rush to finish the
projects ahead of schedule and (b) the conflicting interests of stakeholders such as Congress, ABC, the
military, and Westinghouse
FROM 1934 TO 1946

Nuclear energy was one of the key technologies developed in the last century and one of the most
spectacular examples of rapid scientific advance. A giant technological leap occurred between the 1930s,
when the fission process was discovered, and the late 1950s, when commercial nuclear power broke into
the world's electricity market.
In 1934, Italian physicist Enrico Fermi conducted experiments that showed neutrons could split certain
atoms. Four years later in 1938, German scientists Otto Hahn and Fritz Strassman fired neutrons into
uranium. They were surprised to find lighter elements, such as barium, in the leftover materials. In the
next few months, they concluded that the barium and other light elements found during the experiment
resulted from the splitting of uranium. The fission process had been discovered.
During the war and under the auspices of the Manhattan Project a group of scientists led by Fermi were
recruited to put their theories into practice. By 1942, the team gathered at the University of Chicago was
ready to begin the construction of the world's frrst nuclear reactor, which was erected on the floor of a
squash court beneath the University's athletic stadium. On December 2, 1942, they obtained a self-
sustained nuclear reaction. Fermi and his group had successfully transformed scientific theory into reality
and the world had entered the nuclear age.
A number of other technical breakthroughs achieved under the scope of the Manhattan project lead to the
development of the first atomic bomb. By July 1945, the first prototype was tested in Alamogordo, New
Mexico and in August 1945, two full-scale bombs were dropped on Hiroshima and Nagasaki.
THE QUEST FOR GOVERNMENT CONTROL

The two atomic bombs dropped three days apart devastated two Japanese cities, killing some 106,000
people and injuring 136,000 others. In the aftermath of the explosions, U.S. policy makers promptly
realized that the new energy needed a strong national and international control. Within Congress, there
was consensus on the necessity of legislation ruling atomic energy. However, the complexities of the
technology and the difficult post-war challenges made the enacting of such legislation an extremely
difficult task.
The stakes were high for the United States. On the one hand there were a number of national security
issues that needed to be addressed properly. Among them: (a) the need to set up a sufficiently credible
1. The early origins of the PWR design
nuclear deterrent, (b) the need to avoid the proliferation of nuclear technology abroad, and (c) the need to
create an organizational structure capable of managing nuclear weapons.
On the other hand, there were non-military goals that concerned to average citizens. Among them the
need to (a) enhance scientific progress and disseminate it within the community, (b) benefit from the
peaceful applications of nuclear technology, and (c) capture the value created by the new technology.
Both types of objectives -military and civilian- were equally legitimate. The United States was emerging
from a long war and national security had a prime place in the domestic agenda, but it was also important
to reap the benefits of the new technology. The immense investment in the Manhattan project had
unveiled a new source of energy and it was a: legitimate claim to capture the economic benefits of the
innovation. It is important to note, however, that military and civilian objectives were generally at odds.
National security objectives would demand a tight control of the technology, while economic growth
objectives would demand opening the technology to free enterprise (Sweet, 1998).
THE COLD WAR AND ITS FRAMES OF THOUGHT

At the end of World War IT, the American society was extremely favorable to nuclear technology. There
was strong confidence in the ability of institutions to handle such a complex technology, which was seen
as the paragon of scientific progress (Duffy, 1997). This overconfidence contrasts with today's public
perception of nuclear energy -especially after Three Mile Island and Chernobyl.
By 1949, the Cold War had started, and the Soviet Union's rise was perceived as a threat to the United
States. On August 29, 1949 the Soviets exploded they first atomic bomb. The blast came as a total
surprise since experts believed that they would not possess that capability until 1952. This early success
was followed by the explosion ofa hydrogen bomb in 1953.
The ideological struggle of the Cold War had swiftly moved to the issue of technological superiority not
only regarding nuclear but also in space technology. To make it worse, the technical contest was
exacerbated by effective Soviet propaganda. The latter was dramatically shown in October 3, 1957, when
the Soviet Union placed a satellite -the Sputnik- in orbit at 900 Ian above the earth. At the same time the
announcement of the breakthrough went public, Moscow radio provided the frequencies to track the
spacecraft using a radio receiver. The wave of hysteria that this announcement generated in the United
States was monumental. Similar broadcasts proclaiming the existence of a Soviet Intercontinental
Ballistic Missile (ICBM) had a profound impact on the political environment within the United States
(McDougall, 1997).
The point is that in the mind of policy makers it was clear that the United States could not afford to fall
behind in space and nuclear technology. They had to do whatever was necessary in order to maintain U.S.
technological leadership in these strategic areas.
Besides the struggle to match Soviet technology achievements, there was an additional reason to speed up
the development of U.S. domestic nuclear industry: the potentially steep competition from the UK, France
and Canada. These nations had initiated their own nuclear power programs and threatened to reap the
profits of the new industry. That was something that the United States could not allow, as it was the
innovator that opened the market. In the mind set of policy makers; the nation that "created" the value had
to be the nation that "captured" that value.
MAIN STAKEHOLDERS
Figure 1.1 summarizes the main stakeholders involved in the development of nuclear technology. The
U.S. Congress was a fundamental stakeholder with the main objective of enacting sensible legislation to
both regulate and promote nuclear energy. In addition Congress had the intent to be a watchdog of the
11
Federal Government on nuclear issues, as it was perceived that the new technology had unprecedented
importance for the nation.
For the White House, nuclear technology was seen as an important tool for enhancing the myth of
presidential leadership and probably the single most important weapon of influence on foreign policy
issues. In addition, and especially during
the Eisenhower adnlinistration, nuclear was ...................................................................
. . .
: : :
: : :
seen as a burden to the resolute intent of the i : :
President to slash public spending Congress Senate ~ House ~ I'

(McDougall, 1997).
The military -represented by
Departments of War and Navy and later by
the 1I -.:
.1
............................................
. . _.II- "-.~l
1 Military
~ Labs ~
the Department of Defense (DoD)- were
stakeholders viewing the nuclear era in I Universities I : .. : .
terms of military applications. There was a j ~&D. ~ I:;.. White House " I::~ 1.::inePt•.Of, W,ar !":~
compelling interest in building a credible I organizations I Dept.~f Energy

nuclear deterrent. This was indeed an I Advocacy'l I Offt "rB
d' f I ~Dept.ofNavyj
urgent objective as the U.S. nuclear arsenal
was literally emptied by the war (Dawson,
L........~.~.~~~.~...........1 [::: ..~~~::~~::::::~ .... J
1976). Therefore there was a will - / i Private Utilities i Federal
promoted by the military- to build dual- Scientific
Comm.unity
I Defense. I Government
purpose reactors that would produce
weapons grade Plutonium and generate Contractors ::~:::::In dustlY
Manufacturers' of.
I:::::.'.
electricity at the same time. Additionally,

electrical assets
there was desire to have tight control of the : n :
technology in order to guarantee a high

Figure 1. 1. Stakeholders in nuclear technology
level of secrecy that would avoid eventual
nuclear proliferation abroad.
The issue of the control of the technology was therefore in the top place of the nuclear agenda. Whether to
hand full monopoly to the military or whether to create an autonomous civilian agency, were important
considerations in the legislators' minds.
The issue of secrecy was at odds with other important stakeholders: the scientific community. On one side
of the equation there was the principle of free dissemination of the new technology to promote scientific
progress and to maintain the traditional openness regarding scientific and technical literature. On the other
side, there were national security interests that needed to be protected. In the words of the prestigious von
Neumann, "It is now that physical science has become important in that painful and dangerous sense
which causes the state to intervene. The legislation on atomic energy represents the first attempt in history
to regulate science in this sense" (Del Sesto, 1979).
The private industry completed the club of main nuclear stakeholders. Electrical utilities viewed the
government-sponsored nuclear power program both as a threat and as an opportunity. In like manner,
former Manhattan contractors and a number of other potential entrants wanted to take a portion of the
future nuclear business.
Both the U.S. Congress and the Federal Government understood that the nuclear industry would take off
only via the participation of private industry, as had been the case with other technology-intensive
industries. But how to involve the private industry became a particularly difficult issue. Secrecy, product
liability, patents, and R&D costs, became formidable obstacles for the development of the technology.
12
CREATION OF THE AEC

Two different proposed laws had been under discussion in the Congress at the end of 1945, where the
major point of disagreement concerned the role of the military: the May-Johnson Bill, which placed
heavy emphasis on the military control of the atom, and the McMahon Bill, which proposed the creation
of a civilian-controlled bureau. After a laborious eight-month debate on both legislative bodies, the
McMahon Bill prevailed, and was signed in August 1946 as the Atomic Energy Act.
In the declaration of the Atomic Energy Act, Congress aclmowledges the significance of the atomic bomb
for military purposes but stresses the need to promote and regulate the civilian applications of nuclear
energy. The Atomic Energy Act asserts that "the development and utilization of atomic energy shall be
directed toward improving the public welfare, increasing the standard of living, strengthening free
competition among private enterprise so far as practicable, and cementing world peace" (Duffy, 1997). It
was a clear commitment to strengthen the exploitation of nuclear technology for peaceful -and civilian-
purposes and an invitation for the involvement of private enterprise.
The Atomic Energy Commission (AEC), the agency that emerged from the 1946 Atomic Act, was a five-
man commission that had the mandate to both regulate and promote nuclear energy. The AEC was given
full control in the atomic energy development, both regarding military as well as civilian activities. In
what became an exceptional feature, the military aspects of the new technology were incorporated into the
civilian programs, without divesting the AEC of civilian control. There were dissimilar projects under the
same umbrella, such as (a) military and national security programs, (b) civilian power, (c) production of
fissionable materials, and (d) physical research programs.
The AEC was given a total monopoly on nuclear assets, including all fissionable materials, plants,
laboratories, facilities, and equipment. Even the technical information and related sources -patents,
drawings, experiment data, documents- became exclusive property of the new agency. This issue was
subjected to especially heavy attack during the debates, with opponents claiming that the compulsory
monopoly on nuclear patents and the restriction on the free dissemination of technical information would
transform the U.S. free-enterprise system into a socialist regime. Most lawmakers seemed to support the
idea of the freedom of scientific information but they were not in favor of removing all security
provisions. Thus the substance of the McMahon Bill finally prevailed untouched (Del Sesto,1979).
AEC's OPERATIVE ARMS

From an organizational point of view a five-man commISSIOn was considered necessary given the
complexity of nuclear issues. However, as several programs were already in progress -including activities
related with nuclear weapons- there was a need for a managerial approach. For this reason, the position of
General Manager was established who held enormous power and was directly appointed by the President
of the United States.
The Act also provided for the necessary links of the AEC with the main stakeholders in the nuclear field.
First, a Joint Committee on Atomic Energy (JCAE) was created in order to provide Congress with an
effective authority over the AEC. Congress' intent in this regard was not to leave regulatory power solely
in the hands of the AEC. Secondly, recognizing the complexities and extent of the military programs, a
Military Liaison Committee (MLC) was created. This was an advisory body to help the AEC deal with its
military duties. Finally, the General Advisory Committee (GAC) was created to provide the AEC with
scientific advice.
13
Figure 1.2 is a sketch that summarizes

..................................................................
...
the linkages between AEC and the ... .. ,
..
I ! !
nuclear stakeholders. The main point is j I ~ House of
r
that instead of creating an ABC that
Senate ~ I ! Representatives
would be an island within the Federal
Government, the 1946 Atomic Act
transferred the control of the new
technology to a rather larger subsystem
I JCAE Military
that included other participants, such as

GAC MLC
lawmakers, military experts, scientists,
and industry leaders.
AEC
JCAB was comprised of eighteen
members, representing the two bodies of
the U.S. Congress. Nine were from the '7 ;
II
House and nine from the Senate and no
....::::::::::::::::~Federal
more than five members from either Scientific
body were allowed to belong to the
same political party. JCAE's duties
Community I Govemmellt
were: (a) involvement on all bills,
resolutions and reports produced by
Congress, (b) effective watchdog power
over the ABC on nuclear energy and
nuclear weapons, and (c) authority on
L. __.~. __ ~ _ _J Illd~try
identifying, reviewing and analyzing

nuclear policy issues. Figure 1.2. The AEC and its linkage with stakeholders
MLC was an advisory arm on issues related to military programs. It was comprised of representatives of
the Departments of War and Navy (later Dept. of Defense). Finally, GAC was an advisory committee on
technical and scientific issues, comprised of nine members of relevant prestige appointed by the
President. Top-notch scientists held positions here, among them Fermi, Oppenheimer and Seaborg.
THE NUCLEAR 'SUBGOVERNMENT

Many political scientists have made the point that the organization invented by the Atomic Energy Act,
was the archetype of a sub-government, namely "a small, cohesive, and stable group of actors that
exercised considerable autonomy in policymaking". (Duffy, 1997). Most of the participants involved in
nuclear policy in the postwar had either an economic, political, or organizational stake. But unlike the
familiar example of subsystems -where government actors are lobbied by industry incumbents- in the
nuclear case there was no nuclear industry at the outset. Indeed, the creation of a nuclear power industry
was one of the objectives of the 1946 Act and, interestingly enough, the authority to conceive, develop,
and monitor these policies was within the ABC.
The emergence of a de-facto sub-government on nuclear issues was favored by a number of factors. First,
nuclear was perceived as an extremely complex technology and therefore only few people would be able
to understand policy issues. Because the United States was the only nation with the "atomic secret",
secrecy was considered to be an essential part of the effort to retain the atomic monopoly. In addition, the
onset of the Cold War with its prospects of a long-term conflict with the Soviet Union gave rise to strong
personal commitments toward the effective development of U.S. nuclear capabilities. People who were
involved in the nuclear effort during this time -lawmakers, military, scientists and industry leaders- had
little doubt on the need and importance of what they were doing. More importantly, the planning and
14
executing teams worked together through consecutive projects and objectives, with a very low attrition
rate and a high degree of stability and collective learning.
PROS AND CONS OF THE 1946 ACT

The 1946 Atomic Energy Act set up the grounds for the development of nuclear power in the United
States. Its positive elements were: (a) it set up the grounds for the participation of private enterprise in the
effort, and (b) it established an organization with high autonomy and the explicit mandate to invest
heavily in the development of commercial nuclear power plants.
The shortcomings were: (a) secrecy and monopoly of patents were restraints on entrepreneurial creativity,
(b) the taboo of product liability was an obstacle for industry involvement -particularly utilities, (c) the
opposing functions of promotion and regulation created conflict of interests, and (d) the de-facto sub-
government caused insufficient external scrutiny.
The shortcomings became evident and eventually forced the introduction of modifications to the
organization envisaged in 1946. In 1954 the Atomic Energy Act was amended to address the issues of
private patent rights and limited product liability. In 1974 the ABC was split into a regulatory body, the
National Regulatory Commission (NRC), leaving the promotion side in the hands of the Department of
Energy (U.S. Congress, HR 8631).
As Figure 1.3 points out, each of these changes can be seen as the transition from an organization with a
tight control of the technology towards an organization with a higher degree of entrepreneurial freedom.
Figure 1.3 shows that there was a trade-
off between exerting tight control of the
technology and favoring entrepreneurial
freedom. Clearly, it was impossible to 1974: NRG is created
manage a technology having at the same
time a tight control of the nuclear secrets- 1954: Atomic Energy
for national security reasons- and a full Act revised
"free-enterprise" legislation to promote
innovation, entrepreneurship and free
dissemination of knowledge. Surely
Congress understood the trade-off but
tilted the early legislation toward the side
of tight control, secrecy and national
security objectives. As the time passed the
direction was corrected in 1954 and 1974.
The question is how late these changes
were introduced and which were the Figure 1.3. AEG's beginnings as an organization with tight
consequences of this delay on the control of the technology but little free-enterprise
outcome of the nuclear power industry.
THE ROAD AHEAD

After the Atomic Energy Act set up in 1946 the grounds for a civilian nuclear program directed by the
ABC, a promising future for nuclear power came into sight. The conditions necessary for a rapid
development of the technology were reasonably outlined by the new legislation, namely: people,
resources, and goals. However, the strategy to meet these goals efficiently and in a timely manner was not
obvious.
15
The coexistence of two different nuclear programs -military and civilian- under the same umbrella
introduced a number of challenges, opportunities and difficulties. Given those boundary conditions, one
can picture the development goals of the first ABC's management team in two broad categories:
• Military projects: (a) Nuclear weapons, with the goal of creating a sufficiently credible nuclear
deterrent in terms of number of nuclear weapons and long distance strike capability. (b) Nuclear naval
propulsion, with the goal of developing operative nuclear-powered naval units with high fuel
autonomy and long-distance nuclear strike capability.
• Civilian projects: (a) Research reactors, with the goal of benefiting from nuclear applications, (b)
Nuclear power plants, with the goal of affordable and large scale commercial production of
electricity.
In summary, the four key product lines necessary to accomplish the goals pledged by the ABC were: (a)
nuclear weapons, (b) research reactors, (c) nuclear submarines and (d) nuclear power plants.
It does not mean that these were the only product-lines in ABC's plans but these four were the critical
trails. In addition, there were some key enabling technologies in the sense that without them it would not
be possible to carry out the projects. These enabling technologies were: (a) uranium enrichment, or
industrial-scale production of U235 for supplying power plants and weapons, (b) reprocessing, aimed at
producing weapons-grade Pu239, (c) front end technologies, such as uranium mining and manufacturing
of nuclear fuel, and finally (d) back en technologies such as the management of spent nuclear fuel and
radioactive materials.
Figure 1.4 summarizes ABC's four critical product lines in a sketch consisting of an S-curve. The latter is
a typical way to graph the relationship between effort put into developing a technology and the results one
gets back for the investment. For most products and processes, when the results are plotted, what usually
appears is a curve with an S-shape.
The reasoning behind this empirical
result can be explained as follows. In the Nuclear Power Plant "
very early stage of the technology, as
funds are poured into R&D, the
technical progress is very slow. When
the necessary cumulative resources and
learning are put in place the curve enters Research Reacto
in a second stage where huge advances
are obtained with relatively small effort ..
Finally, in the last stage of the curve, it -"""""""AtomicBomb
becomes more and more difficult to
make technical progress regardless of
how much money is spent in R&D. In
other words, there is always a limit on
Figure 1.4. Development roadmap in terms of products
the development of a product or process and cumulative R&D effort needed to develop them
using a technology (Foster, 1986).
S-curves are generally outlined for a single product or process. But in the particular case of early nuclear
technology one can extend the reasoning and think of a single and "broad" S-curve. This simplification is
possible because all the projects: (a) were managed under the same organizational umbrella, (b) using the
same people, and (c) building up collective learning from a pool of common shared innovations.
Unlike most of the conventional R&D effort nowadays, the key innovations in nuclear technology under
ABC sponsorship made extensive use of the state-of-the-art technology developed in prior phases of the
program. For instance, the theory of reactor physics was developed during the Manhattan project and was
useful when designing the first research reactors. The mow-how obtained in fuel manufacturing and the
16
operation of research reactors was then used in designing the first naval reactors. Such swift technology
transfer has few comparable cases in modem technology.
The resultant intricate path of civilian and military products and their multiple interrelations is shown
schematically in Figure 1.5. As it is shown, civilian and military product lines coexisted within the same
umbrella and no single product could have been developed without the cumulative know-how, R&D
investment, and learning effects due to previous products.
Figure 1.5. Conceptual roadmap of military and civilian nuclear projects pursued
by the AEC, and their complex interrelations .
MILITARY PROJECTS UNDER AEC

Figure 1.6 summarizes the military product lines that were on the drawing boards of the ABC in the late
1940s: nuclear weapons and nuclear submarines. The cornerstone of nuclear deterrence was having a
large inventory of nuclear warheads, but at the end of World War II, the United States possessed not a
single nuclear weapon. It is not surprising then that ABC's fITst major drive was towards developing an
extensive industrial capacity for producing enriched uranium and plutonium, which are the prime inputs
in nuclear weapons.
Fissile U235 uranium can be produced enriching natural uranium by gaseous diffusion, centrifugation or
electrostatic separation. Weapons-grade Pu239 plutonium can be produced only in reactors. Increasing the
energy of each device was an important objective and consequently, the H-bomb was developed. In
November 1st, 1952 a prototype of the H-bomb niclmamed Mike was exploded in the Marshall Islands.
Mike released an energy of 10,000 kilotons, roughly one thousand times the energy released by the A-
bomb dropped over Hiroshima in 1945. The mushroom-shaped cloud created by the explosion had a
radius of 8 miles and reached a height of 27 miles.
ABC's interest in developing the technologies for manufacturing nuclear fuel and building plutonium-
producing reactors was explicit. The construction of dual-purpose reactors that would produce weapons-
grade plutonium and electricity was pursued by the ABC from the outset. Considering that the US was
able to build a nuclear arsenal of nearly 20,000 warheads, one can have an idea of how many reactors
were needed in order to meet those needs. It is important to note that like the United States other countries
17
were also interested mainly in dual-purpose reactors. For instance, early British efforts focused on
developing dual-purpose reactors to meet the needs of the UK Atomic Energy Agency. In 1953, the UK-
ABA announced the future construction of a 60Mwe gas-cooled dual-purpose reactor in Calder Hall
(Patterson, 1985).
The point is that the first civilian projects of the ABC, namely constructing nuclear reactors, enrichment
facilities and nuclear fuel factories, had indeed a military objective: producing fissile materials for nuclear
warheads.
Having enough warheads was a necessary but not sufficient condition in order to have effective
deterrence capacity. There was also a need for having long distance strike capability. This was initially
achieved by long still-air range aircraft such as the B-52 but the ultimate devices in this segment were: (a)
intercontinental ballistic missiles -ICBM- and (b) nuclear-powered submarines equipped with short-range
missiles. Hence the interest of the ABC in cooperating actively with the development of the ICBM and
funding the nuclear submarine project (Suid, 1990).
Although some minor work had taken place previously, it was the detonation of the Soviet A-bomb in
August 1949 which precipitated the genesis of the submarine project. As a first step, a contract was
awarded to Westinghouse for constructing a prototype reactor, the Mark I, which became the laboratory
for testing the design concepts used in the submarine. It is interesting to point out that Westinghouse was
selected as a contractor in this crucial project not because of a competitive bid but because its main
competitors -GE, Allis-Chalmers, Gulf- were already involved in other projects. In particular GE was
working on a breeder reactor and a military reactor in Hanford, Washington (Beaver, 1990).
In summary, as it is shown in Figure 1.6, the development of military products had the ultimate goal of
attaining a credible long-range nuclear strike capability. Under this objeCtive, the nuclear submarine
project was a crucial step. Not only it condensed the state-of-the-art in nuclear reactor utilization, it also
set up the basis for future nuclear power plants. Westinghouse was awarded this important project
because its competitors had their hands full with other ABC contracts. As it will be shown afterwards, this
turned out to be a fortunate outcome for Westinghouse.
Figure 1.6. Military nuclear products and processes and the cumulative R&D effort
needed to develop these products
18
CIViliAN PROJECTS UNDER AEC

Figure 1.7 shows that the development of civilian projects under ABC auspices had the principal goal of
building a commercial nuclear power plant. The scheme also shows that the necessary technical expertise
to build such a plant was extracted from two predecessor projects: (a) the nuclear submarine project with
its prototype Mark I reactor, and (b) research reactors such as the Triga and the MTR.
Research reactors were built by smaller vendors like General Atomics and were marketed to Labs and
Universities for research, training and radioisotope production. Figure 1.7 also shows the fundamental
role of enabling technologies whose dominion was key for the development of reactors most notably (a)
the process of enriching uranium, and (b) the manufacturing of nuclear fuel.
Similarly, the role of the nuclear submarine in the development of a commercial nuclear power plant was
fundamental. As it will be shown in next chapter, the first power plant -a PWR reactor- was literally a
ground-based extension of the submarine reactor.
....... !\
Figure 1.7. Civilian nuclear products and processes and the cumulative R&D effort
needed to develop these products
THE AEC-INDUSTRY PARTNERSHIP

It is clear that in order to achieve its goals the ABC needed to develop a number of reactors, from research
reactors, to plutonium-producing units, to a workable prototype of a submarine power unit. In order to
carry out the development of such a variety of projects, the ABC eS,tablished programs aimed to provide
incentive to private industry for participating in nuclear development. These programs were (a) the
Reactor Development Program announced in 1948, and (b) the Power Reactor Demonstration Program
announced in 1955. Both programs had the intention to execute the different reactors planned by the ABC
and -as an expected extemality- to develop a capacity for building nuclear power plants within the private
sector.
19
Although in the long run the private nuclear industry indeed emerged in the United States, in the early
years 1945-1960, the so called ABC-industry partnership was basically a process to assign contracts to a
selected club of contractors. Moreover, since the technology was in a fluid phase and it was not clear
which reactor design would be the best, the AEC managed a portfolio of projects at the same time and
assigned arbitrarily these projects to different contractors.
ABC had a policy of maintaining product diversity. Some contractors would bid for a certain project that
looked interesting from a strategic point of view, but given the huge technical uncertainty of most of these
projects, one can argue that the process of assigning contracts was essentially a random process.
The nuclear submarine project illustrates the point for two top contractors: General Electric and
Westinghouse. General Electric had been awarded the project to develop a sodium-cooled reactor whereas
Westinghouse had been awarded the project for developing a water-cooled reactor. Both firms developed
land-based prototype reactors that started operation in the early 1950s. In the case of Westinghouse the
prototype reactor Mark I started operation in 1950 at the Idaho Testing Station. Four years later, in 1954,
a modified version of this reactor was installed in the Navy's first nuclear submarine, the Nautilus, which
was a highly successful vessel (Del Sesto, 1979).
General Electric on the other hand, developed a land-based prototype reactor, which started operation in
1955 at the Idaho site. Later, General Electric modified the design to fit the Navy's second nuclear
submarine, the Seawolf, which put to sea in 1957. But because of the danger of leaks and the violent
reaction of sodium with water, the Seawolf was removed from service the following year (Del Sesto,
1979).
With the Nautilus, Westinghouse not only developed the basic design concept that became the standard of
all nuclear submarines of the US Navy in the next four decades, it also set up the basis for the design that
would eventually become the first nuclear power plant. On the other hand, the Seawolf not only was a
faulty design concept for a submarine but also resulted in an inferior technological bid for future nuclear
power plants. Sodium-cooled reactors were abandoned a few decades later everywhere else in the world.
It is important to note that in 1949, when both prototypes were on the drawing boards, neither looked to
be superior and therefore the decision of which model was better was not really obvious. Indeed
Westinghouse was selected as a contractor in this particular project because General Electric was working
on a breeder reactor and a military reactor in Hanford, WA and ABC authorities thought that the firm
would not be capable to manage the development of yet another reactor (Beaver, 1990).
In addition to the above-mentioned reactors, the ABC developed a number of different projects under the
scope of the Reactor Development Program of 1948 and the Power Reactor Demonstration Program of
1955. For instance:
• Experimental breeder reactor developed by Argonne Laboratory. This reactor produced the world's
first electric power generated from a nuclear reaction in December 20, 1951.
• Material Testing Reactor (MTR) developed in 1950 Argonne and Oak Ridge Laboratories .
• Experimental Boiling Water Reactor (EBWR), developed at Argonne, and completed in 1957 .
• Sodium Reactor Experiment (SRE) developed in Santa Susana, Santa California in 1957 .
• Homogeneus Reactor Experiment (HRE 2) developed by the Oak Ridge National Laboratory, which
was abandoned in 1957.
20
Chapter 2.
THE RISE OF THE PWR IN THE MARKETPLACE
Following the successful development of a nuclear submarine, Westinghouse started the construction of the
first PWR nuclear power plant at Shippingport using a design based on the submarine reactor. This reactor
and project manager -Admiral Hyman Rickover- shaped the nature of the Westinghouse PWR and ultimately
of the US nuclear power program. The chapter continues with a recounting of the PWR's subsequent
progress on the technology and market dimensions. It shows how the design quickly became the most sold
nuclear power plant in the world, capturing almost two-thirds of the market. The objective of the chapter is to
show that if a dominant design effectively emerged in the industry the dominant design was the
Westinghouse PWR.
RICKOVER AND THE NUCLEAR SUBMARINE PROJECT

A nuclear-powered submarine had the potential to revolutionize naval warfare. First, its ability to remain
submerged for long periods made it the best vehicle for transporting nuclear devices to distant war scene.
Secondly, even without nuclear devices on-board, a nuclear submarine was fundamentally superior in
performance versus other comparable submerged vessels.
Westinghouse was granted the contract to build the nuclear submarine. But the effective control of the
project was in the hands of the Navy. The leadership of the nuclear submarine project was given to then
captain Hyman Rickover, largely because of his successful tenure as head of the Navy's electrical section
during World War II, and because he was one of the few individuals with knowledge of both naval and
nuclear issues.
By 1949, Rickover became chief of the ABC's Naval Reactors Branch. This was a dual agency serving both
the Navy and the ABC. Westinghouse was linked to Naval Reactors by the operation of Bettis laboratories in
Pittsburgh, PA. This was a laboratory built by Westinghouse under government contract and with the
exclusive objective of carrying out projects assigned to Westinghouse by the ABC and Naval Reactors. Thus
Rickover had to lead a three-hat organization made up by the ABC, the Navy and Westinghouse. ill
retrospective, working with three organizations under the same umbrella seems an impossible task. But
Rickover was no ordinary man. Rickover would become one of the most intriguing figures in nuclear history.
He was described as a genius, a man of vision and a master politician. Without any doubt he was a man of
strong leadership, able to motivate R&D engineers, naval officers and politicians towards his vision of
achieving high goals.
21
MARK-I AND NAUTILUS

The first achievement of Rickover and his staff was the design and construction of the Mark I, a prototype
reactor that would become the cornerstone of the PWR concept.
Despite the fact that many of the reactor's components had to be invented from scratch, the design evolved
swiftly. Many of the fundamental ideas that would eventually be used in the nuclear submarine and in the
PWR reactors were developed at Bettis during the design of the Mark I prototype. They developed a special
fuel (V02) to avoid damage during sailing conditions, and used zirconium alloy to protect the highly
enriched uranium from corrosion effects of water at 320 C.
In retrospective, the Mark I design is certainly the best design for a nuclear submarine. If today a committee
of nuclear engineers had to select a reactor design for a submarine, they would agree that a pressurized light
water reactor is the best alternative. It can't be a boiling water reactor because the latter requires taller
pressure vessels, spreads water contamination all across the submarine and it is noisy because of boiling
phenomena. Obviously it can't be a reactor using other coolant such as gas, sodium or heavy water. Ordinary
water is the best choice. Hence, in addition to his leadership skills, Rickover had the ability to make complex
technologies work and to make the right technical choice at the right time ..
Mark I went critical on March 1953. Subsequent test runs revealed that the reactor not only worked but it
worked very well. This reactor not only provided a basis for commercial PWR reactors in the United States,
it linked the name of Rickover and Bettis Labs as a synonymous of efficiency and unquestioned nuclear
expertise.
A modified version of the Mark I was installed on the first nuclear-powered submarine. By 1954, the
"Nautilus" was launched from the Electric Boat Yard in Croton, Connecticut. The Nautilus operated
successfully for over two years before having its reactor core removed for refueling for the first time. It
remained in the U.S. fleet for several years. Indeed, nearly one hundred nuclear submarines were built in the
U.S. afterwards, basing their design on the early Nautilus PWR type nuclear reactor.
SHIPPINGPORT
After the successful experience of Mark I and Nautilus, the ABC made the decision of extend the PWR
design to a commercial nuclear power plant. For a while the Eisenhower administration doubted building
government-sponsored nuclear civilian reactors. But two external events prompted the decision to move
ahead with the project. One was the start of a soviet demonstration nuclear plant at Obninsk in 1954. The
other was the advanced design status of a gas-cooled reactor in the UK -Calder Hall- which eventually went
critical in 1956.
The decision to support the PWR project demonstrated the urgent need to enhance American prestige and
counter Soviet and British advances in nuclear power. However, the ABC wanted to include an industrial
partner in order to avoid criticism from conservative foes. Luckily for ABC, Duquesne Light of Pittsburgh
agreed to participate in the project, Duquesne agreed to invest money building the electric generating section
of the plant, while the ABC funded the R&D and the capital costs of the reactor itself (Beaver, 1999).
The selection of a PWR design and Westinghouse as prime contractor of the power plant had two strategic
benefits. First, the ABC wanted to have a plant operating as soon as possible and Westinghouse and its Bettis
Laboratory already had a workable prototype (Mark 1) and a successful submarine (Nautilus) based on the
PWR concept. Second, the team that had worked under Rickover's direction had a record of achiev~ments.
Therefore it was a sound decision to hand a complex project over to one of the best nuclear engineering
teams available in the U.S.
The construction of the first nuclear power plant in the United States began in April 1955 in Shippingport,
PA. Since the ABC wanted the plant in operation by 1957, the construction process was virtually a "mission
22
in a hurry". The rush to complete the plant does reinforce the fact that Shippingport was more a political
project than a commercial endeavor (Beaver 1999).
Plant construction began under the rigid direction of Rickover and Naval Reactors. Westinghouse was the
prime contractor, while Stone and Webster acted as architect-engineer. Dravo was the construction company.
The turbine generator was provided by Westinghouse even though a number of technical difficulties arose
during the design, Rickover and his "three-hat" organization were able to finish the project in three years
which is a remarkable achievement. The plant was officially inaugurated in May 1958. It worked properly
for 25 years.
THE QUEST FOR TECHNOLOGY AND MARKET

AEC set up the grounds for a rapid development of nuclear power in the United States, particularly of
designs that received heavy government-sponsored R&D up front such as the PWR design. A giant
technological and commercial leap occurred between 1952, when the construction of the Mark-l prototype
started, and 1989 when nearly four hundred reactors were operative worldwide (half of them PWRs).
As it is shown in Figure 2.1, nuclear power plants evolved in the two classical dimensions of product
development, namely, technology and market. In the technology din1ension the advance was remarkable
given the complexity of this technology. Generating electricity from the atom is one of the highest scientific
and technological achievements of the 20th century, involving the mastery and integration of different
disciplines such as nuclear physics, material sciences, thermal-hydraulics, control systems, radiation
protection, etc.
Most. of these technologies had dual
applications for the development of
nuclear military devices as well as civilian
1989
commercial products ..In fact, most of the
contractors involved in the design of
nuclear power plants -including GE and
Westinghouse- were also involved as
defense contractors in projects ranging
from the development of a plutonium
producing facility to the design of a
nuclear submarine. The resulting
economies of scope were significant.
The second dimension of rapid product
development has been the market
dimension. Nuclear power has an
enormous economic impact nowadays.
The commercial nuclear power plants
currently operating in the world are the
3rd largest electricity supplier, behind Figure 2. 1. Nuclear power development from 1952 to 1989
only gas and coal, and ahead of hydro and
other renewable sources.
In the United States nuclear occupies second place with a strong 23 % of the total domestic electricity
generation (2000). With 103 operating reactors, the country also hosts the largest fleet in the world,
accounting for one-fourth of the world's 410 operating plants. In year 2000 U.S. nuclear plants produced 730
billion kilowatts hour (Figure 2.2). At about 6.7 cents per kilowatt-hour -the price an average utility charged
to consumers in year 2000- the nuclear power industry creates a value of about $50 billion per year.
In some countries -like France, Belgium and Japan- nuclear is the first source of electricity with an
involvement of between 40% to 75% of national generation. For these countries the importance of nuclear
23
power is given not only in terms of economic scale but also in view of other strategic issues such as energy
self-reliance, economic independence and sovereignty.
THE RAPID EXPANSION OF NUCLEAR POWER

Nuclear power is an interesting topic of study because of the speed at which this technology was developed
and introduced in the marketplace. In the U.S. the fIrst commercial nuclear power plant started its
construction in 1954 in Shippingport, PA with strong support of the government. It reached full power in
1957 and it was the first of a vast group of reactors subsequently identified as Pressurized Water Reactors or
PWRs. Two years later, in 1959, the first U.S. nuclear plant built entirely without government support
achieved a self-sustaining nuclear reaction.
After a strong government-supported introduction, the U.S. nuclear power industry grew rapidly by its own
wings. Private firms became more and more involved in developing reactors either as prime contractors,
architects-engineers, civil constructors, or components' suppliers. These firms fueled the already strong
market push driven by the government. They offered turnkey plants at very competitive prices and rapidly
secured early contracts. As a result, by the end of the 1960s, twenty plants were in full operation in the
United States.
Utilities saw this new form of electricity production as an opportunity to achieve economies of scale and
hence higher profits and rapidly jumped into the bandwagon. Consequently the decade of the 1970s
registered the largest increase in reactor orders. By 1973 -in the midst of a severe imported oil shortage- U.S.
utilities ordered the exceptional number of forty-one new units, widening the fleet to one hundred units
ordered in less than two decades. By 1986, the Perry power plant in Ohio became the lOOth U.S. nuclear
power plant under operation.
Such a rapid expansion of nuclear power
in the U.S. in the second half of the 20th 1,400 Coal
century introduced a drastic change in the
domestic electricity mix (Figure 2.2). 1,200
Only coal-fired power plants had a
comparable growth in the same period. It 1,000
is important to note that coal was already
a mature industry with large economies of 800
scale and extensive learning effects to
benefit from. It is clear that in the United
600
States nuclear power has been the de-
facto alternative to coal-fired plants. One
400
can go one step further to argue that
without the emergence of nuclear power,
200
the increase of coal-fired plants would
have been 1.6 times larger in the period
between from 1950 to 2000. This is an
1960 1965 1970 1975 1980 1985 1990 1995 2000
important environmental consideration
given the well-known connection between
Figure 2.2. Electricity Net Generation Addition at U.S.
carbon-dioxide emissions released by Electric Utilities, 1958-1999 (Billion Kilowatt-hours). Source
coal-fired plants and global warming DOE Energy Outlook '01
phenomena.
DOMESTIC AND INTERNATIONAL COMPETITION

After 1945, a number of reactor proposals quickly came out from different design teams working around-the-
clock in the United States, Soviet Union, United Kingdom and Canada.
24
As Figure 2.3 shows, these early designs lead to ten basic types of nuclear power plants that were eventually
constructed later. Obviously there were hundreds of other designs, but these ten were the only ones that made
it into the marketplace. Surprisingly these basic designs correspond to just three key decisions. The first is
the fuel itself, which can be an oxide compound of natural or enriched uranium (in which the concentration
of the fissile U235 isotope is increased from its natural 0.7% to about 3%). The second component is the
moderator, whose function is to slow down the neutrons from the fission process so that they are suitable to
sustain the chain reaction. In practice the options are graphite, ordinary water and heavy water. The third
major component is the coolant, where the options are helium, carbon dioxide, and ordinary/heavy water.
Moderator Coolant Fuel
Water
cooled
reactors
Gas
cooled
reactors
Sodiwn
cooled
reactors
Figure 2.3. Diagram of the designs that made into the market (1958-1989)
One can argue that given the state of the art

prevailing in the 1950s, any of these basic 250
Units under
designs could have become the commanding commercial
design. There were -a priori- no technical operation
200
reasons why any of these designs could be a
more judicious choice. Indeed, nations
committed their prestige .and public 150
resources to different designs. For instance,
the UK. committed to the Magnox and the
AGR, Canada supported the PHWR, while 100
the US ended up building PWRs and BWRs.
Figure 2.4 shows the outcome of the above-
50
presented designs by 1989, in terms of
number of units that started commercial
operation worldwide. It is clear that the PWR o
design was the winner as it was able to
capture 244 out of 481 power plants that
were put into operation by 1989. Its success
was matched only by the BWR -another U.S.
design- that was able to secure 101 units Figure 2.4. Summary of total number of NPPs that
built worldwide. started commercial operation before 1989 (World).
25
In terms of market share, the right dimension is

the installed electricity capacity, which means
the total MW e that each design managed to
connect to a commercial grid. Again the PWR
emerges as the industry leader, with a strong
61% (corresponding to 2x 105 MWe of
installed capacity). It is interesting to note that
there is a direct correlation between Figure 2.5
and Figure 2.4, due to the fact that most
designs had similar electric power per unit.
Therefore the leadership in number of units
sold gives rise to a similar leadership in
installed capacity.
It is also important to note that the summaries
presented in Figures 2.4 and 2.5 correspond to
worldwide data. Therefore, the dominance of
the PWR means the supremacy in various
Figure 2.5. Market share of NPPs that started
countries and with products manufactured by
commercial operation before 1989 in terms of
different sets of contractors. This is a strong installed electricity capacity (World).
indication that the PWR was indeed a
dominant design in the nuclear power industry.
WHAT IS A DOMINANT DESIGN?

Perhaps the best way to define a dominant design is to provide examples of dominant designs available in the
marketplace. The IBM PC format is a dominant design in the market of small personal computers. In the
early 1980s it came to define the concept of the personal computer for at least 80 percent of the market. The
Palm III is probably also a dominant design in the handheld Personal Digital Assistants (PDAs) market -now
challenged by the pocket PC, both seems to capture most of the features that users expect from a PDA. A
Corona typewriting machine with a QWERTY keyboard, visible type, tab and shift keys, and carriage
cylinder is also an example of a more aged dominant design. Finally, the classic example of a DC-3 aircraft
with two wings containing the fuel inventory, two engines positioned in each wing, a retractile landing gear,
single tail, and movable flaps. The DC-3 is an illustration of a dominant design that emerged in the 1930s but
still influences modem aircraft designs such as the Boeing B767-200 or the Airbus A320.
In other words, a dominant design is the product that wins the enthusiasm of the marketplace. It is the design
that all competitors and new entrants must adhere to if they hope to command significant market share. Thus,
the IBM PC format is still manufactured nowadays not only by IBM but by other firms such as Compaq, HP
and Dell. Since most of the users are interested in performance of the product, one can affirm that the
dominant design is the one that summarizes all the explicit performance requirements of the users and makes
them implicit to the design (Utterback, 1996).
Thus an airline pilot will no longer demand a retractile landing gear on a plane; he/she will take for granted
that any "reasonable" large commercial aircraft will include such a feature. The dominant design that
emerged with the DC-3 made the retractile landing gear an implicit performance requirement of a large
commercial plane. It is so implicit that nobody questions it anymore.
26
WAS THE PWR A DOMINANT DESIGN?

This is the fundamental question of this research and will be answered in three steps. First, the previous data
shows the long run leadership of the PWR design and its adoption by a large percentage of the market, which
is a strong indication. Second, this section will show its market share progression over time. One might
expect that a dominant design would gain the market progressively over time, since by hypothesis a
dominant design is the product that wins the enthusiasm of the marketplace. Finally, in the next chapter it
will be shown how the architecture of the PWR design achieved a stable status.
Table 2.1 summarizes the number of
nuclear power plants sold as a function of Market Share
time, as well as the resulting new -PWR
100%
electricity capacity resulting from these \ -BWR
new units. The data show what happened \ - -LWGR
with the PWR design in the period \ - PHWR
between 1955 and 1989. The market \ - - Magnox
reception to the PWR increased rapidly \ .... 'FBR
both in terms of units sold and in terms of

market share, which supports the notion
\
\
- Other
/'
PWR
\
of dominant design. 50% LWGR
From the data of Table 2.1, the share in

sales can be calculated and plotted in a
user-friendly way. The latter is an
incremental value, defined as the share of
all new operating nuclear electricity
capacity that is added to the marketplace
within a given period. If a design has
increasing sales share over the time, one
can conclude that the buyers are
1955 1965 1975 1985
increasingly enthusiastic with the product.
This supports the thesis that the PWR was Figure 2.6. Progress of PWR's sales' share -in terms of
indeed the dominant design in the market new nuclear electricity capacity. Data corresponds to
of nuclear power plants. units starting operation in the world in the same period.
Figure 2.6 is a"plot of sales share data -in percentage points corresponding to the portion of new nuclear
electricity generated using this tecbnology- as a function of time. It shows the progress of PWR sales share
over time, versus comparable designs. It is very clear from the data presented in Figure 2.6 that the P.WR
increased its sales acceptance over time. While in 1965 only 19% of all new NPPs were a PWR, by the end
of the 1980s the sales share had increased to 68%. The latter is a strong indicator that the PWR emerged as a
dominant design in the industry.
Figure 2.6 also shows that comparable designs did not have the same fortune in the same period, as PWR
grew at the expense of the other designs. The BWR for instance could not emerge consistently and after a
peak of 27% in 1970, its sales share decreased to 17% in 1990. The PHWR remained marginal at less than
10% of the market. The most dramatic changes occurred with two graphite-moderated reactors: the LWGR
and the Magnox. Before 1965, these designs were the most successful in the marketplace but later their sales
plunged to zero. The LWGR is a graphite-cooled light-water reactor primarily constructed by the Soviet
Union, and its most notorious example is Chernobyl Unit II. The Magnox on the other hand is a reactor
developed in the UK. See Figure 2.7 for schemes of some of these reactors.
27
Concre'e Pressurised water reactor IPWR )
sh~
Basic gas-cooled reactor IMAGNOX)
Pressurised heavy WIIter reactor (CANDU)
Pnmairy Fuel ""~f"rM'O&II1 t

........ r .... nwnt5 he.t
ncheng.'
Sodium-cooled fast reactor
Figure 2.7. Schemes of some of the most common world nuclear reactors
28
Table 2. 1. Progress of NPPs that started commercial operation between 1950 and 1989 in number of
units and additional total electricity output (in MWe). Worldwide data.
B1:F::::T~~2';2;Y~;21::,Hf:~:;:2~~::~:E::Year.rof:Sta'rEof:CoDllrierCiarOperatloW~:~::":~:=::E::::E:~EEE~:?:::I::l::2::
E:f~~9751ts:;:Ei9'5~?~9}~~ :1:96"57692: :d9jol74":::~:":19JS?:t9,:: :T980784::~:""i985hf9:::
~:~i9.6'O/64~~t
,,,~,,,L~Q~~Efs:;i~:;~i~ ...............;..,..;~'..""'N .. "'A.;
................... .. -
1 4 2 6
:JI~:~~;=E: 2 5 11 20 24 17 22
~»~~~~+
-- ... -~"--~' -. - -.
2 3 1 3 3 1 2
~~~0 2 2
~~Rf;f~I~~f 2 2 3 3
Biiq;[~i~E 1 2 1 1
r$W~RtrS~~ 1 2 5 2 2 8 7 3
2~~~~![#f 6 9 11 2
~iB!iI£~E
'- .-'."-', ...... -', ~'- - "', ,
2 2 8 4 9 6
~;£~~:~'~.;~::~ 2 4 8 35 44 65 86
16 32 42 75 88 101 126
~~:~~~:,~~~: ..:..~~~,~~:~~.;,:.~",:~,,~~':;~J:.~.'.,~~.v;.J';:~~:~~~~.~~~~~~~:~.'~l.~~'~~~:-,.~.~~~'~"~'.~,:,:,:~~.~.",.M,,~~' ....

"'...: ....,:. I••••• ,...
"",;.,.....', ..........
f; Start:o f~C()nt merci al~Op eta ti on :;::::;:;:;:~~

~~:.;;'~::;;~~:;:;;::;:;:;;,'::.~.:':.:::;::;~:;:,~::::;~year:o ..:~:.:::;"~=;;;.::;. ~,::::'~":::,:.:.
I._~~'''v_,~",.""
~i95.oT54::; £19"55159;: ~I9'6oi64£;;1":965169.= £I[7"o/?;4fl:;]Q~75/'J9,:; ;1980/84F2 ~1985/89'::
. ... . • - ..... . , ... -
-
~3TotaTE;:~£::TorarEE ::~z:TQt3I:t;tS~otal::'~:;;;~Total:-=: ::~;,Totnr~::.:'" ~,;:;~'T:otal:::::
--..,.
...
.................
, ......
<.I, ..... :, ....................
:'~.:Total~::::
' .. ,..,0,11.
....... ........ __ .......u ouo__ .. .., ..... __ .. '
;';':.;.><;',
:AGR::'::*~£~":?;r(MWe}~~Ei 36 2640 1320 3960
""",.:~;,;
ill:WREE::~:;,iMWe)E&
..
10 513 1902 13256 18625 16660 22264
t:~:~ ,- ............
;.;.::;.~;.
..;. .. ;.;.,: ....
~:~,:
::FBRE2Fa
..... " " "., ~'(MWe}qE 23 115 65 412 390 600 1255
;.~_ .. ~;.
f~~~1.!~~~~(M;We):~;.:
~J';';~:';'
;U'("':'J~: .......... , .... - • ~
....., .. ...,.:."'...,00.: ... "''''',.,. .•

,"Y
84 249
rGCif~ss.:::
;'oc'c;::~;' ;,1.
:~.~.~~
;,.;~;<;:"..:..
....
;';';';""';'
~(~:W~)rtf;
,,,,.,,,.,.,,,.,.,,,, -,,,,"'- ......... ,.,
44 125 990 151 6
f.i!tG~ii;ili: :fM}Y~tti 1 57 342 308

~'o;:;'
~'
:"'~'';''~
..
'.;';'"
:t:: j;WGit~;~ ~lo'Yl" ........ "
.:.;.,;.;.;.;..
~~~': ........ _.., ....... ,. ... 'K'
ifMw~t:t 5 200 508 1054 1012 5036 7500 3500
~~-~~~~
f(M~~):::
~~tt!9x:q ................................. 364 1298 3156 1310
i~i::
~; ..~
..................
.. .......
v .....
~.. __.......................
"
37 2 76 3148 3280 5117 4010

:.;...;.;. ... ,3HI~:;;;'2
~.;.;.,.;.
:~~jj~
~;...~~;.,;.
~'';'';';';''
-- -~--
":p.WR~;~~~:~:~:ii;;:~.~~:~E
:.; ;.;....;. ._._.~..... .....
.. _. .,.".....~
-
.... ~u ..... _
-
.. ___
..................... ... ;~-. .... l<,;
64 461 2876 223 70 35128 541'68 88571
"'"----_
~,:..~~:~~:..:~~~:..;.~~:~;~~~~~'!~':~J.~~ ~
.................. ... "'''~ ...... --......''' ... _. ~ ..... -". ..... _,
~(i. ~~ii P.~1Eo~iil~f.::i~:~:;,~:;;:i; 5 5 705 3094 10460 432 73 65441 85365
....... ..
...... _' .. ~"" ..
" .......... w; ... ~_ .. .,~ .. ~
................................ ... ...,. .. ~ " ....
29
Chapter 3.
THE DOMINANT DESIGN
This chapter first describes the Westinghouse PWR design in detail, focusing on its distinctiveness from
other designs. A large amount of data is provided, with the proposition that all the main design parameters of
the reactor, such as pressures, temperatures, and mass flows, converged to the same numbers. The dominant
design quickly emerged as the industry standard not only in the U.S. but also in France, Japan, Belgium and
the rest of the western countries.
THE PWR AS A SUBGROUP OF POWER PLANTS

A nuclear power plant is just an electricity-generating machine with a nuclear heat source to produce heat
and a conventional turbine/generator group to transform heat into work and into electricity. The unique
characteristic of a nuclear plant is the heat source. While coal, oil, or gas is burned in conventional plants, in
a nuclear reactor the heat is obtained from nuclear fission.
In a nuclear power plant the heat source
is referred as the
system while the
often known as the
nuclear heat supply
rest of the plant is
conventional system.
As Figure 3.1 shows, a fluid -steam or
gas - flows between these two systems
-;
·
II- -
HotFluid ~
Conventional
System 11"-
and transports the energy from one
~ Cold Fluid ~
system to the other. It captures the heat
produced by the nuclear reaction and
through a complex energy conversion Figure 3. 1 Conceptual division between nuclear and
process it converts heat into work and conventional systems in a nuclear power plant
into electricity. Table 3. 1. Characteristics of the PWR and other designs
The main component of the Fuel Coolant Moderator
conventional system of a nuclear power
U02 in zircaloy Light water H20 at Light water
plant is a turbine/generator, but there are PWR clad, enriched at 160 bar (boiling not H2O
other minor conventional equipment 3.0% U235 allowed)
such as feed pumps, heat exchangers, U02 in zircaloy Light water H20 at Light water
and condensers. BWR clad, enriched at 75 bar (boiling is H2O
3.0% U235 allowed)
Since the steam or gas flowing into the U02 in zircaloy Heavy water 020 Heavy water
conventional system is not significantly PHWR clad, not enriched at 90 bar (boiling 020
(0.7% U235) not allowed)
contaminated with radiation, nuclear
U metal in a Carbon Dioxide Graphite
power plants use commercial off-the- magnesium alloy CO2 at '20 bar
shelf parts manufactured by standard Magnox
clad, not enriched
vendors. These vendors market their (0.7% U235)
products to both nuclear and fossil U02 in stainless Carbon Dioxide Graphite
fueled power plants. AGR steel clad, enriched CO2 at 40 bar
at 2.5% U235
Unlike the conventional system, the nuclear heat supply system is unique to each type of reactor design. The
main differences arise on the selection of the materials used to produce the nuclear fission reaction inside the
re8:ctor. As Table 3.1 shows, there are three major reactor components where the choice of a material has to
be made. The first is the fuel itself, which can be an oxide compound of natural or enriched uranium (in
30
which the concentration of the fissile U235 isotope is increased from its natural 0.7% to about 3%). The
second component is the moderator, whose function is to slow down the neutrons from the fission process so
that they are suitable to sustain the chain reaction. In practice the options are graphite, ordinary water and
heavy water. The third major component is the coolant, where the options are helium, carbon dioxide, and
ordinary/heavy water.
As Table 3.1 shows, it is the permutation of the available options that gives rise to the variety of nuclear
designs that populate the nuclear power fleets of countries like the US, France, Japan, Canada and the UK.
The PWR uses a unique selection of these materials.
BRIEF DESCRIPTION OF THE PWR DESIGN

Westinghouse and its engineering team at Bettis Laboratories originally developed this design for submarine
propulsion. In addition to this firm, Babcock & Wilcox and Combustion Engineering have built this type of
reactor in the United States while Mitsubishi, Framatome, and Siemens/KWU have done it in the rest of the
western world. As it is shown in Figure 3.2, the PWR reactor has a single pressure vessel-where the reactor
core is located- and two separate cooling subsystems.
The primary cooling subsystem is the
only subsystem that passes through the
reactor core. The fission process heats
up the water that passes upward past the
reactor core, from a temperature of
about 290°C to a temperature of about
320°C. Outside the reactor pressure Sleam line
vessel, the primary system is divided in

two to four cooling "loops" connected to
a central pressurizer. The latter TurDlne
genet.lor
....
maintains the primary pressure constant
at 160 bar. At this pressure water
remains always liquid since its
Control rod. COnll .....
saturation temperature is well above cooling
wat.,
r
320°C and therefore no boiling can

occur. Each loop contains a primary
pump and a steam generator. The water
coming from the reactor is pumped to
the primary side of the steam generators
where it transfers heat to the secondary Figure 3.2. Simplified scheme of a PWR nuclear power plant
coolant subsystem.
The secondary cooling subsystem consists of cold water, which passes on the secondary side of the steam
generator, is heated and converted to steam. The later is possible because the water in the secondary
subsystem is at a pressure of about 70 bar and its saturation temperature is 280°C. Therefore boiling can
occur, taking into account the high temperature of the primary subsystem. The steam then passes to the
turbine and turn shafts that are mechanically connected to the electricity generator. The redundant steam
from the turbine condenses the water, which is then pumped again to the steam generators. The condenser is
the ultimate heat sink of the reactor. Here the heat that was not converted into work is cooled by cold water
pumped through the condenser. As a cold sink the condenser uses water extracted from the sea, river, or a
water reservoir. Another choice of cold source is the atmosphere and on that occasion the device for heat
rejection is a cooling tower.
One can go one step further, defining the PWR architecture in terms of system, subsystem, components and
main design variables. Following the systemic approach traditional in engineering settings, the following
definitions are adopted: (a) system is an independent part that can be treated as a whole, (b) subsystem is an
31
element of a system, which performs a particular function, (c) key components are the most important parts
that comprise a subsystem, and (d) main variables are the critical physical properties of a given subsystem
and/or component.
For example, one can define the conventional system as the part of the nuclear power plant whose function is
to take steam coming from the nuclear heat supply system in order to extract electricity. The steam
subsystem can be defined as the part of the conventional system that has the function of processing the steam
coming from the reactor and converting part of its energy into work. In this case the key components are the
turbine, the condenser and the secondary pumps, while the main subsystem variables are the steam pressure
and the coolant mass flow.
Table 3.2 and Figure 3.3 summarize the PWR architecture taking these considerations into account. The
selection of key components and main variables is not arbitrary. It is based on the premise that the selected
items and variables are the crucial elements of the design concept. Less important components -valves, tanks,
process subsystems- are not taken into account, as they are irrelevant for the analysis. The same holds for
physical variables which are intrinsically dependent on others (e.g. the saturation temperature is an intrinsic
function of fluid pressure).
It is important to note that a particular subsystem can be exposed to either architectural or component
changes. An architectural modification would be, for instance, if instead of using only one turbine the
designer decides to use four smaller turbines. On the other hand, the changes can be at the component level if,
for example, instead of using traditional rotating pumps the designer introduces piston-type pumps. Finally
any subsystem as a whole can undergo incremental innovations without changing the basic architecture or
the core components leading to the enhancement of a certain performance variable, for instance higher
thermal cycle efficiency.
Table 3.2. System description of the PWR architecture
System Subsystem Key Components Design variables
Primary coolingPrimary Pumps

Nuclear ubsystem team generators
Heat Pressurizer
Supply
System Reactor Vessel
Reactor Internals
Reactor Core
urbine
econdary pumps
Conventional
Condenser
system
Generator Gross Power (Mwe)
Figure 3.3. Conceptual scheme of the main components of a PWR design
32
CORE DESIGN CONCEPT

In the previous section, the PWR design concept was described as it was originally conceived in ~he mid-
1950s. However, there are several architectures which can meet this rather basic design concept and, in
principle, a variety of different components and materials that can be utilized. As a matter of fact, the PWR
design concept as it was introduced can be considered as vague and undefined as it can be an SUV concept in
a newspaper's automobile section.
What transforms a set of engineering choices into a design concept are not only the components and concepts
that are utilized. It is the use of an architecture that links the components and core concepts in a certain
manner that not only is consistent but also stable over time. In the SUV example above, it is the constant use
of an architecture defined by a high power gasoline engine, above 4.5 liters, a 4-wheel drive automatic
transmission, and a light external structure with a sport appearance mounted on a frame rather than on the
axles. Furthermore, it is the way these components are linked in order to make the driver feeling like driving
a 2002 Saab car rather than a 1965 Chevy truck.
In the particular case of the PWR, the core concepts, components and linkages between them can be
summarized as follows:
1. The reactor core is contained by a cylindrical pressure vessel made of stainless steel (it could have been
prestressed concrete as in the Magnox).
2. The fuel is uranium dioxide enriched at 3% with a can-shaped zircaloy clad (it could have been a lower
enrichment and a clad of stainless steel as in the AGR).
3. The primary coolant is ordinary water maintained at a fixed pressure of about 160 bar by a single
pressurizer connected to the primary cooling subsystem and whose job is to avoid bulk boiling within the
primary (the coolant could have been allowed to boil as in the BWR).
4. The heat is transferred from the primary coolant to the secondary system through tube/shell steam
generators (this intermediate process could have been avoided as in the BWR).
5. The mass flow rate is provided by rotating pumps, which are located outside the reactor (it could have
been piston-driven pumps or natural convection).
6. The reactor control drive mechanisms are on top of the pressure vessel (they could have been located in
the bottom of the pressure vessel as in the BWR).
7. The containment of the nuclear heat supply system is a "dry" spherical stainless steel structure designed
to withstand up to 5 bar of internal overpressure (it could have been a "wet" pressure-suppression
containment with a cylindrical shape as in most BWR reactors).
TESTING THE DOMINANT DESIGN HYPOTHESIS

Up to this point, the main proposition has been that a complex system like a nuclear power plant can be
accurately described in terms of a handful of systems, subsystems, components and variables; which will
account for the most important physical processes involved in the system. Namely, that a PWR system can
been reduced to the simplicity of Figure 3.3.
The next step in the analysis is proposing a definition as to when a design concept fulfills the conditions
necessary to become a dominant design. A reasonable criterion is that the emergence of a dominant design
should be manifested at least by the following features: First, two consecutive products must maintain
exactly the same architecture in terms of systems and subsystems, and must have exactly the same number
and type of key components. Second, at some time the main physical variables of the system should attain a
convergence or consensus value.
For example, if the secondary cooling subsystem consists of a single turbine, a single condenser and two feed
water pumps, then the number and type of these components must be maintained constant in two consecutive
33
products. Moreover, variables like the gross power of the turbine or the steam pressure must also stay
invariable.
It is important to note that architectural stability is a necessary but not a sufficient reason for a particular
product to become a dominant design because it does not take into account the market, which mayor may
not adopt the product. It is just a strong indication of the emergence of a dominant design. A firm will freeze
the design and focus its energy in achieving economies of scale, only when the firm senses that it has a
product sufficiently attractive in the marketplace. Otherwise it will continue to experiment with alternative
concepts.
EVOLUTION OF THE PWR DESIGN

Figures 3.4 to 3.10 contain a graphic representation of the main information concerning the PWR design
evolution.
After an initial turmoil, which occurred in the
1950s and early 1960s, a consensus design 170
clearly emerged. N
{3 .1.1970 a consensus emerges
en 160 \V
The design was refined and improved in the ~ D do&J":W1:::''1:;rF't::i'WD~~~-a.JQ.llf'~-:::z,
following decades but basically the PWR ;. 150 D D 0
architecture remained unchanged. The type ~ D D
and number of components in all the

Q)
:5
~
140 tD DDD D 1976
Fifty PWR One-
1986
subsystems were maintained between two 'iij units on- hundred
.~ 130
consecutive products and, similarly, the most ~ line PWR units
significant variables attained a consensus ~ on-line
~ 120
value. 5-
01
.~
For example, one can explore the history of co 110 D
CD
the most critical thermal variables in the c.
a 100
plant, that affect the process of extracting o 50 100 150 200
heat from the reactor and converting it into
Cumulative number of PWR reactors
work. These variables are pressures,
temperatures and heat rates. Figure 3.4. Evolution of primary pressure in PWR units
that started commercial operation in the western world
Since the first PWR started commercial
operation in 1957, the primary pressure
160
varied significantly in the early designs, until ~
it attained a convergence pressure of around C-E
158 Kg/cm2• The latter is arguably the most
critical design parameter in the PWR design
and achieved a remarkable consensus as can
be seen in Figure 3.4. The major PWR
vendors converged on a figure that has been .. / ,/
maintained in PWR reactors even today . .,
, '. WEST
/
It is important to note that the data presented /
,,
/
in Figure 3.4 correspond to worldwide data -
except the Soviet Union. Therefore it can be 1/1
concluded that 158 kg/cm2 was the gold
standard of all PWR reactors built either in 100
52 54 56 58 60 62 64 66 68 70
the United States or abroad.
Construction permission (year)
Figure 3.5 provides a closer look at the same Figure 3.5. Same as 3.4 but for early PWR units as a
data but as a function of the year that the function of the year construction permit was issued
construction permit was issued. In the late
34
1960s there was a delay of about five years

between construction permit issuance and 340
commercial operation. As shown, the major o
rJ CJ a R [JLl
PWR vendors in the United States tt\ a~?-.rfjl, 0 ~ (Ill lIIJ Y ~A~¥.fi~ 0
Westinghouse, Combustion Engineering and 320. CJ
Si'
o~
(l
CI
rJ.\1 lilt]
OJ,
r "~~ll11l:JCrliiiimul
[60
a r:f1
l 5~Pr6.ll!1l 0
u B
~
0,!! l"J 00 CllJ
Babcock & Wilcox- seemed to agree on a 00 nn ~~dlb 0 rl 1

consensus figure. o 0 a5b j Core Outlet! UJ
Given the fact that the three mam PWR 300. : " I Temperature I
manufacturers were independent finns 1 o~oco
competing for market share, their agreement is a 00 j orJil0 0 ttJ ~o~
olIlm5l %~~~ oDD 0 00 0 ~ [51'-""

a 0
strong indication that a dominant design J Jj
CD 00
t9b:D0
.
0 ~
llll.
"tD
mIII)iI!Ib.
0 u 0 1m
00'
0 ~ c
00
effectively emerged in the 1960s. The data also 280. 0 lJl r5b j j
o DO' : CD
show that these three finns competed evenly in - ~ Core Inlet l

the beginning of nuclear power in the United ~
Q,)
J
o
~
..
Temperature ~
J.. 0 1976 1986
States, building similar numbers of units. Later, :J E
CiS 260.
this equal market penetration finished in a strong 10-
Q,)
leadership of Westinghouse. However, the three Co
EQ,) :J
firms provided essentially the same design and
utilized the same supplier base. For instance, ...240
Babcock & Wilcox and Combustion Engineering o 50 100 150 200
did not manufacture turbine generators and
therefore were forced to buy these key Cumulative number of PWR reactors
components from either Westinghouse or
General Electric. Figure 3.6 Evolution of primary temperatures in PWR
units that started operation in the western world
As a result of the convergence in the primary
pressure, the core outlet temperature, or the 340
temperature at which the coolant exits the reactor
vessel, also followed a convergence trend. The
final consensus temperature was about 320°C.
The same holds for the core inlet temperature,
which stabilized at 290°C.
The convergence of the two core temperatures to
their consensus values is shown in Figure 3.6, 300
I.'
which summarizes worldwide data -excepting The three
the Soviet Union. As it is shown, the WEST .,~~. major PWR
A- .' • :".-;. ',; " constructors
convergence evidence is overwhelming.
Figure 3.7 provides a closer look at the same
data presented in the previous figure but as a
.' B&W
~ agree in the
same figure
function of the year that the construction permit

was issued. Note that there were five years 260
between construction permit issuance and 52 54 56 58 60 62 64 66 68 70
Construction permission (year)
commercial operation.
Again, as shown in Figure 3.7, the major PWR Figure 3.7. Same as 3.6 but for early PWR units, as a
vendors in the United States -Westinghouse, function of the year construction permit was issued
Combustion Engineering and Babcock &
Wilcox- seemed to agree on a consensus figure.
35
A very interesting feature in the PWR design

evolution appears when comparing the
number of primary loops and pumps as a
function of time. Initially PWR reactors
started with one or two loops and one or two
1 loop
primary pumps. When the power was
increased the number of loops was also
increased. At a final stage of the evolution,
the PWR dominant design consisted of four
loops and four pumps. Figure 3.8 is a
scheme of this process.
Notwithstanding the fact that the number of
pumps changed, the fundamental parameter Figure 3.8 Scale up in the PWR design: how higher ratings
associated with the PWR design -which is were achieved using more cooling loops
the mass flow rate- was maintained constant.
As can be seen in Figure 3.9, the mass flow
30000
rate per pump attained a consistent
convergent trend. By the mid-1960s all the - ~
.s::
C
PWR designers were using pumps rated at ~
about 19,000 ton/hr. a.
E
:J
The convergence in mass flow rate suggests a.
~
that all the. designers were using the same C1l
a.
15000
.
....IV 'C-E
. .:
type of pumps, perhaps provided by the C1l
same manufacturer. However this hypothesis ~

::0 ....
..
could not be verified. It should be pointed
u:
out that the primary pumps are large II)
II)
components subjected to stringent IV
:i: .~.:- B&W
requirements both in terms of temperature
WEST ""JI"
and pressure. 0
N ~LO <0 co 0 N ~<0 <0 co 0
The mass flow rate was maintained LO LO LO <0 <0 <0 <0 r--
constant -at a consensus value- even though Construction permission (year)
there was a drive to increase the power
ratings of nuclear power plants in order to Figure 3.9 Mass flow rate per pump as a function of the date
the construction permit was issued
achieve economies of scale. The scale up
process -which lead to 1,300 MWe units-
was achieved without changing the core 100 1200
concepts of the design. The scale up process
-i
"0
occurred just by adding cooling loops. This 0
'i:
80
IV 1000
was a way to reduce the need for further a. 800 ~
IV
.s:: 60 IV
R&D costs. .... 600
::
0
.S a.
In summary, it can be affirmed that a ~ 40 IV
':; 400 t»
dominant PWR design appeared in the late .c ~caIV
1960s. It had four primary pumps rated at ~ 20
II)
200 ~
c:
19,000 ton/hr, operated at a pressure of 158 ::J
0 0
kg/cm2, with coolant temperatures of 320°C
55-59 60-64 65-69 70-74 75-79 80-84 85-89
and 290°C.
Period (years)
Regarding power rate -thermal power and
electricity output- the units advanced Figure 3.10 Average electricpower rating (MWe) of PWR
constantly to a target rating that changed plants (worldwide) that went on line prior to 1989
36
over time. As it is shown in Figure 3.10, in the

16
late 1980s, the average PWR had an electrical H
rating of 1,000 MWe, although some units
14
attained 1,300 MWe.
Consequently, one can argue that the designers
were able to maintain a stable dominant design
even though there was an extensive scale-up
operation going on. This is also an important
indication of the robust architecture of the PWR.
It allowed power increases without a significant
-L-
a>
a>
E
co
:c
change in the design concept itself. Other "ffi
s:::
designs, such as the Magnox, faced significant
difficulties when designers tried to scale up
(Patterson, 1985).
-
'-
Q)
'C
-C
co
To strengthen the point, it is interesting to note .c
en :(:1
that notwithstanding the increase in power, large '(ii

components of the primary system, such as the :I: o
reactor pressure vessel, did not change
o 150 200
significantly. As an illustration, Figure 3.11
Cumulative number of PWR reactors
summarizes the evolution of the two main
Figure 3. 11 Evolution of PWR pressure vessels
dimensions of the pressure vessel, showing that
it did not change at all.
A PATTERN OF INCREMENTAL INNOVATION

The previous sections defined a "general" description of a PWR design. Some units had features slightly
different and some minor innovations at the component level. What is important to note however is that all
the PWR units developed in the US or abroad had the same design from an architectural point of view.
Namely, the way the core concepts and components were integrated and linked together has not
fundamentally changed even though there were minor innovations at the component level. The above-
described pattern is consistent with the incremental innovation pattern defined in the framework of
incremental, architectural, modular and radical innovation (Henderson and Clark, 1990).
Henderson and Clark's framework classifies
innovations along two dimensions: (a) the Core Concepts
innovation's impact in components and (b) the Reinforced Overturned
innovation's impact in the linkages between
different components that compose the system. Unchanged Modular
This idea is summarized in the matrix presented innovation
in Figure 3.12.
Based on the data of all the PWR units Changed Architectural Radical
constructed before 1989, one can argue that the innovation innovation
PWR design had only incremental changes after
the concept was introduced in 1957. Its core
design concepts were reinforced and the
Figure 3. 12 Henderson and Clark framework for
linkages between its components ~emained defining innovation
unchanged.
A comparative example of a product that reached an era of pure incremental innovation is the diesel engine.
Even though diesel engines still improve continuously, there are no big innovations that shake the core
concepts or the way these concepts and the components are linked. Engineers do not reevaluate the decision
to use a particular core concept of the diesel engine every time they develop a new product.
37
Architectural stability is a necessary condition for a particular set of engineering choices to become a
dominant design. Once a dominant design is established, progress takes the shape of improvements in the
components within the boundaries of a stable architecture. The core engineering choices of a dominant
design are not revisited every time a new product comes out from the factory because firms cannot waste
valuable resources in re-inventing the wheel. In fact the dominant design often emerges in response to the
opportunity given by stable architectures to obtain economies of scale and/or to take advantages from
externalities (Abernathy, 1978).
The same holds for the PWR. From 1957 to 1989 the core concepts of the design have not changed and the
design engineers focused on improving the performance of stable architectures. The parameters that
summarize the balance of the plant, such as pressures, temperatures, mass flows, all attained a consensus
value. Some innovations have occurred in the component side, for instance, design improvements in pumps,
control rods, nuclear instrumentation and steam generators. However the basic PWR architecture and its core
concepts did not change, which is a finding that strongly supports the dominant design hypothesis.
FROM PRODUCT INNOVATION TO PROCESS INNOVATION

The previous sections described the evolution of the PWR as a product. The design concept evolved from an
initial stage of high variety and uncertainty to a dominant design, where there was only incremental
innovation on a largely standardized product. The industry itself changed from an era where it was
fragmented, unstable, and where many products were making their way to the marketplace to a more mature
era in the late 1980s with standardized products.
The above-presented pattern is consistent with the framework of product and process innovation developed
by Abernathy and Utterback. The framework hypothesizes that the rate of major innovation for both products
and processes follows a general pattern over time and that product and process innovations are related as
shown in Figure 3.13 (Abernathy, 1978).
The rate of product innovation is highest
during the initial formative "fluid" phase.
This is a period where product Product Innovation
experimentation occurs both in terms of
testing different options and gaining
experience. During this fluid period of high
product innovation, much less attention is
given to the processes by which products
are made so the rate of process innovation
is less rapid.
There is a second period, however, where
things are reversed. This is the 'transitional Fluid Phase Transitional Phase Specific Phase
phase" where the rate of product
innovation slows down and the rate of Figure 3.13 The dynamics of innovation (Utterback, 1994)
process innovation takes the lead.
The reason for this is the fact that product variety gives way to standard designs that have proved to be well
accepted in the marketplace. At this point, a dominant design emerges which is in general the product that
best satisfies the buyers' needs (Utterback, 1994). Finally, there is a third period, where the rate of both
product and process innovation declines. Here is where the industry becomes more focused on cost, volume
and capacity.
Based on the data available, one can argue that the PWR design emerged as a dominant design in the late
1960s, after a flurry of product innovations such as the Magnox, the FBR, the AGR, etc. Since then process
innovation became the preferred way to introduce changes and these changes were incremental innovations
on the dominant PWR design. The manufacturing progressed from heavy reliance on R&D to more
38
systematic and repetitive engineering tasks. The organizations in charge of building reactors changed from
small, cohesive and creative groups- such as Bettis Labs - toward a more hierarchical organization with
defmed tasks and procedures. Engineers became more focused on improving the performance of the design
or at least what was considered performance of a nuclear power plant at that time, namely: (a) increasing the
efficiency of the thermal cycle, (b) building larger units and (c) increasing the number of redundant safety
features.
A relevant point is whether the change from product innovation to process innovation was made at the right
time. fu other words, could a better design have emerged if the industry spent more time in a fluid phase? But
as was explained in chapter 1, there was a government-sponsored "rush" to move forward in the learning
curve and attain a truly "operational" U.S. nuclear power capacity.
Abernathy and Utterback's framework also predicts substantial changes in the industry arising from the
emergence of a dominant design, but these changes are discussed in Chapter 6. Some of these issues are: (a)
competition from new entrants building PWR reactors, (b) organizational control through proj ect and task
groups, and (c) vulnerability of industry leaders.
39
Chapter 4.
THE BWR AND THE HIDDEN LINK WITH THE PWR
This chapter focuses on the development of the BWR design, i.e. the GE approach to the civilian nuclear
power plant and the main competitor to the Westinghouse PWR. The chapter starts with a thorough
description of the design and the differences and similarities with the PWR. The BWR also stabilized in
certain values and this will be shown following the same criteria explained in the previous chapter. Using
appropriate data, this chapter also shows that the main connection between the designs proposed by GE
and Westinghouse was through the use of similar suppliers and components. In particular both designs
have essentially the same steam turbine. This finding is used to argue that one of the main reasons why
the Westinghouse PWR design stabilized so quickly was because of the standardization of one of the
main component of nuclear power plants: the steam-electric turbine ..
THE BWR AS A SUBGROUP OF POWER PLANTS

Allis-Chalmers and General Electric had introduced commercially the BWR design in the United States
in the early 1950s. Although Allis-Chalmers abandoned the business in 1962, GE remained in the market
and was able to sell nearly fifty units within the U.S and abroad. In addition to U.S. vendors, Hitachi,
Toshiba, Asea and Siemens/KWU built BWR plants abroad generally as licencees of GE. As a result, one
hundred nuclear power plants were built worldwide before 1989, making the BWR the second best sold
design in the world. To illustrate this point, Figure 4.1 compares the market evolution of the BWR with
respect to the PWR.
The first BWR was built at Argonne
National Laboratory partially with
involvement of Allis-Chalmers. It
481
received the name of Experimental lUfA..
Boiling Water Reactor (EBWR) because
it was originally conceived as a pilot ~
Q)
plant to test the concept of a reactor .c

E
where boiling was allowed to occur. In ::s
c:
addition to its nominal thermal rating of Q)
20MWth it was expected to produce >

:;:;
5MWe of electricity using a turbine ~
::s
generator manufactured by Allis- E
::s 1
Chalmers. The facility went into U
operation in late 1956 and was operative
until mid-1959 when it was shut down
in order to be modified for a number of
experiments, including an upgrade to
produce 100MWth, five times the Figure 4.1. Comparison of PWR & BWR units that started
original heat power rating. Thereafter, operation before 1989 (worldwide data)
EBWR operated till 1967.
After EBWR, the participation of Allis-Chalmers as designer and constructor of BWRs had increased
throttle in 1957, when the firm was granted a contract to build the Pathfinder reactor in Sioux Falls, SD.
This was a 59 MWe unit operated by Northern States Power from July 1966 until late 1967.
40
The BWR made a commercial leap when it became the first reactor sold on a purely competitive basis
and without government support. This reactor, Dresden-I, achieved a self-sustaining nuclear reaction at
Morris, IL, in October 1959. Under the ownership of Commonwealth Edison, Dresden 1 operated
commercially at a rating of 200MWe until its programmed shutdown in 1978. The reactor heat supply
was designed and constructed by General Electric as well as its turbine generator. Bechtel was the other
participant in the project in the double role of constructor and architect-engineer.
44 out of 47 BWR nuclear plants built in Table 4.1. Main characteristics of the PWR and BWR
the United States before 1989 followed
the same formula of the Dresden 1 Fuel Coolant Moderator
reactor. Namely, GE as manufacturer of UOz in zirca/oy Ught water HzO at .Ught water
the reactor heat supply system and the PVVR dad, enriched at 160 bar (boiling not HzO
turbine generator and a major engineering 3.Cf/o U235 allowed)
firm, such as Bechtel, Stone & Webster, UOz in zircaloy Ught water HzO at
or Sargent & Lundy, taking the central BVVR dad, enriched at 75 bar (boiling is
3.Cf/o U235 aI/owed)
role of architect-engineer and constructor.
BRIEF DESCRIPTION OF
THE BWR DESIGN
In the same way as a PWR, the heat
source of BWR reactor is referred as the
nuclear heat supply system while the rest
of the plant is often lrnown as the
conventional system. As Table 4.1 shows, St.am IInl
both designs make use of essentially the

same fuel -with minor differences- and
the same coolant. The main difference is Turbin.
g.n .... 'ar
....
that boiling is allowed to occur in the
BWR and the steam produced in the
reactor pressure vessel directly drives the
Conttol rod. COnd." ....
turbine generator with no intermediate cOOling
.a'.,
steam generators in the pathway.
Since the steam flowing into the
conventional system comes directly from
the reactor in a once-through cycle, the
BWR turbine generator is slightly more
complex in order to avoid contamination
with radiation. Excluding these minor
differences, turbine generators of BWR
and PWR reactors are fundamentally the
same. Both operate at a pressure of about Tu,OInl
_"10'
....
70-75 bar.
Unlike the PWR design, the BWR reactor
has only one cooling subsystem (see Concl.n ..
cOOling
Figure 4.2 for comparison). First, in the .attt
reactor section of the cooling subsystem,

water circulates through the reactor core
removing heat as the water moves past the
fuel assemblies. The water eventually is Figure 4.2. Schemes of the PWR and BWR designs
41
heated enough to convert to steam. The later is possible because the water inside the reactor is at a
pressure of about 70-75 bar, its saturation temperature is 280°C and therefore boiling can occur.
Steam separators in the upper part of the reactor remove water from the steam. The steam then passes
directly to the turbine and turns shafts that are mechanically connected to the electricity generator. The
redundant steam from the turbine condenses into water, which is then pumped again into the reactor.
The condenser is the ultimate heat sink. Here the heat that was not converted into work is cooled by heat
exchange with cold water pumped through the condenser. As a cold sink the condenser uses water
extracted from the sea, river, or a water reservoir. Another choice is the atmosphere and the device for
heat rejection is a cooling tower.
One can go one step further, defining the BWR architecture in terms of system, subsystem, components
and main design variables. Using the following definitions: (a) system is an independent part that can be
treated as a whole, (b) subsystem is an element of a system, which performs a particular function, (c) key
components are the most important parts that comprise a subsystem, and (d) main variables are the critical
physical properties of a given subsystem.
Table 4.2 summarizes the BWR architecture taking these considerations into account. The selection of
key components and main variables is based on the idea that the selected components and variables are
the crucial elements of the design concept.
Table 4.2. System description of the BWR architecture
System Subsystem Key Components Design variables

Nuclear
Reactor Vessel hermal Power (MWth)
Heat
Heat source Reactor Internals essel diameter (m)
Supply
Reactor Core essel height (m)
S stem
Inlet Temperature (OC)
Feedwater pum ps Outlet Temperature CC)
Cooling
urbine Steam Pressure (Kg/cm2)
Conventional ubsystem
Condenser Coolant Mass flow (tn/hr)
system
Electricity
enerator Gross Power (Mwe)
ubsystem
Figure 4.3. Conceptual scheme of the main components of a BWR design

It is important to note that only key components and subsystems are taken into account. Less important
components are not taken into account, as they are irrelevant for the analysis. For instance primary
42
circulation pumps and the subsystem for water purification are less significant items and do not determine
the boundaries of the BWR design. The same holds for physical variables which are intrinsically
dependent on others.
CORE DESIGN CONCEPT

In the particular case of the BWR, the core concepts, components and linkages between them can be
summarized as follows:
1. The reactor core is contained by a cylindrical pressure vessel made of stainless steel (the same as the
PWR but taller in order to allow natural convection to occur).
2. The fuel is uranium dioxide enriched at 3% with a can-shaped zircaloy clad (same as the PWR except
that rods are slightly thicker).
3. The primary coolant is ordinary water but unlike the PWR, bulk boiling is allowed to occur within the
primary, at a saturation pressure of 70 bar.
4. The heat is directly transported by steam from the primary coolant to the turbine secondary system
(there are no intermediate tube/shell steam generators as in the PWR design).
5. The mass flow rate is part natural convection and part forced convection provided by jet pumps (in
the PWR design there are circulating pumps and natural convection effects are negligible).
6. The reactor control drive mechanisms are located on the lowest part of the pressure vessel whereas in
the PWR control rods are on the highest part of the pressure vessel.
7. The containment of the nuclear heat supply system is a "wet" containment, namely a cylindrical
stainless steel structure designed to withstand up to 5 bar of internal pressure. Unlike the PWR "dry"
containment, the BWR utilizes steam condensation as main mechanism of heat rejection.
EVOLUTION OF THE BWR DESIGN

Figures 4.4 to 4.10 summarize the main information concerning the BWR reactor. The evolution of this
design follows essentially the same pattern as in the PWR case.
After an initial turmoil in the early
100
1950s, a stable BWR architecture N
E
emerged in the 1960s. Since then, the u
en "
90 "
BWR design was refined and improved ~
I
but basically its architecture remained >

a..
unchanged. Similarly to what happened ex:: 80
{1970 a consensus emerges
·"\.~.-~'""."1'
Q) a "
""t
Q
with the PWR, the number

components in all the subsystems were
of =Q)
"0
'iij 70 'b
a
",,'II': ~w':m"".~..: W
'\.
,S ""a
maintained between two consecutive Q)
products and the most significant 5 60

C/l
C/l 1976 1986
variables attained a consensus value. For Q)
a. Forty BWR Chernobyl,

instance, the key thermal variables in the Ol
c:: 50 units on- eighty BWR
:;:;
plant that affect the process of extracting co line units on-line
G3
heat from the reactor and converting it a. "
0 40
.
into work. 0 20 40 60 80 100
As it is shown in Figure 4.4, in the years Cumulative number of BWR reactors

when the first BWR reactors were
Figure 4.4. Evolution of reactor pressure in BWR reactors
ordered, the primary pressure varied that started commercial oceration in the western world
significantly. But afterwards the
43
pressure attained a consensus value of

about 72 Kg/cm2• The data correspond to 100
all reactors that started operation
A-C
worldwide. The 72 kg/cm2 was therefore
the gold standard of all BWR reactors N l~1962
built either in the United States or abroad E / Allis-Chalmers
0
(the Soviet Union did not develop the en GE " exits the
~ / business
BWR concept to an operational level) Q)
'-
::J I~
Figure 4.5 provides a closer look at the
same data but as a function of the year the
en
en
Q)
'-
a.
70 ~
, I
I
construction permit was issued. In the late ~ ,

I
,,
ra
1960s there was a delay of about five E
'i:
years between construction permit a. I
I
issuance and commercial operation. As it I
is shown, the major BWR vendor in the

United States -General Electric- set up
..
I
54 56 58 60 62 64 66 68 70
the consensus, because the other BWR
builder -Allis Chalmers- abandoned the Construction permission (year)
business after building only three BWR

Figure 4.5. Same as 4.4 but for early BWR reactors as a
reactors. Some years later, other
function of the year construction permit was issued
international manufacturers -notably
Toshiba, Hitachi and Asea - initiated the
production of BWRs and eventually built 320
as many reactors as GE did. But these
firms did not change GE' s standards. One
reason is that there was appreciable
Q:
operating experience accumulated by a.l 01970 a consensus emerges
1 1
300
General Electric to discourage changes.
The other was that GE secured license
agreements with the Japanese 0 0 00 0: 0 0 0 0 0
i
(5 OaJrIUJ
manufacturers to build BWRs and market aa a 000 aCUJ rDDcD en c an::c 0 am a D~a Da a:E!J t~hDDaDaJ
them in Japan. In fact some of the early

units built there were jointly built by GE
2.0: 0 t~7~OO 00 tS6
and one of the two Japanese builders. ~ Forty BWR Chernobyl,
As a result of the convergence in the E units on- eighty BWR
primary pressure, the core outlet tt 260 line units on-line
o ~ ~ w ~ 100
temperature also followed a convergence
trend. As it is shown in Figure 4.6, the Cumulative number of BWR reactors
final temperature was about 285°C. The
latter is an obvious result since the BWR Figure 4.6 Evolution of primary temperatures in BWR
is a boiling reactor and hence there is a reactors that started operation in the western world before
relationship between pressure and
saturation temperature.
The same trend prevailed in the rest of the main variables associated to the BWR design; such as
dimensions and geometry of the reactor vessel, reactor internals, core inlet temperature, etc.
As an illustration, Figure 4.7 summarizes the evolution of the two main dimensions of the pressure vessel,
showing that they did not change at all. As shown,t the BWR vessel is taller and wider than the PWR
vessel, in order to allow coolant circulation and -to some extent- natural convection
44
4. The B WR and the hidden link with the PWR
It is important to note that although the

BWR attained a "stable" architecture it 30
did not become the dominant design in
the sense that it did not become the
preferred product in the marketplace. H
o~
00 cUb a 0 dloo~~oO 00 ,jP tqp
a
The dominant design was the PWR. One 0 a WO [D000 qsno~o a w Cb 0
20 OOCo D a
can argue that GE decided to pursue Irn wB9o~o
0
a
standardization or architectural stability 0
Eli
c:
0
in order to achieve economies of scale 'Ci)
c:
and learning effects. Namely the same Q)
E 0
reasons that Westinghouse had when it :0 OJ
decided to freeze the PWR concept. Q) 10

00
0
.-0+
rn
rn 0
0
Q)
General Electric expected its strategy > 0
n
Ul 0q).
II W
J'
WUll
III 0 IIUX
:l' tJ
WI"",
CJ LlWOUJ1tV',fIDUJ
't?cp
0
un .
L1IbIUO
'b tl 0
Q)
11 (J 0 0
would eventually match the :srn u 0 [)
[1
0
J
competencies developed by rn
Q)
~~IJ
Westinghouse with its PWR. But 0: 0 -.- -. .

Westinghouse had an early lead in 0 20 40 60 80 100
nuclear power with the design of the

Cumulative nu mber of BWR reactors
nuclear submarine "Nautilus" and the
Shippingport PWR plant. General
electric was a market follower. Figure 4.7 Evolution of BWR pressure vessels in reactors
that started oDeration before 1989
THE LINK BETWEEN PWR & BWR DESIGNS

An unexpected feature that arises from
the survey conducted over the BWR 1400
reactors and the PWR reactors is that
A A A OOO~~
although these designs are different in 1200 0
architecture, they surprisingly share A A AA Y A ~ Ai AA; ~
dh 0 ~4:J 0 Jtp ~ OoCb D
0r5';rJ:>.
some common features. 1000 °A 0 ~ A- A-
A A~O 0 0
A A 1i&.Lt-~ Al:/).
First,in both types of reactors,the total
power rating attained a similar value. 800 o{}~ffhtfJ
tt'l4.'1;!f c: • ~ ••• 0 0
If••
0
Q) A
This agreement is quite interesting ~ cO AnA 0
because there were two differentreactor
architectures, involving two sets of
e.
1-
600 60
~B!r
o
00
6 OA
A 0
o 00
At. A A
A
0
Q)
AIIlC 0
competitors supposedly not interrelated. ~ 400 A A
0
In theory both designs should have a.. fn. 60
followed different paths. But as Figure
co
()
6
'C
200 ~ A
4.8 shows, this was not the case. Both t5
Q)
design concepts agreed on similarvalues 0 A P
[jJ
for reactor power ratings.It is important a3 0
Z -200 B
to note that the reactor power ratings 0 100 200 300
increased as a function of time, as
nuclear power plants increased size in Cumulative number of PWR and BWR units
order to achieve economies of scale.But
at any given time, the BWR units and Figure 4.8 Power ratings of PWR (P) and BWR (B)
the PWR units that started commercial reactor units that started commercial operation before
operation had similar ratings. 1989 in the western world.
Particularly in the 1970s, the consensus
45
was 900 MWe, but this consensus shifted to

1300 MWe in the 1980s as shown in Figure 120
4.8. This change was pursued basically by
Japanese and French vendors: Mitsubishi and
Framatome.
100 1
The second unexpected agreement appears N

E A
when plotting the gross power of turbine

generators as a function of the cumulating -::.
CJ
C)
A
number of reactors in operation at a given Q)

J..
:]
time. As it is shown in Figure 4.9, the turbine en
en
gross power rapidly attained a consensus Q)
J..
Co
value. It is important to note that the turbine C)
gross power is defined as the maximum c:
:;::
power that the turbine/generator can provide m
J..
Q)
to the electrical grid. This is different from Co

o
the net electrical power actually produced by Q)
c:
OD
P
the plant because turbine/generators most of :c
J..
11
U
D
the time work below full power capability. It :]

I- 20 '" A B
is common, for example, that a 1200 MW e o 100 200 300
turbine will be operating at 1150 MWe. The
data show that most of the BWR and PWR Cumulative number of PWR and BWR units
reactors used essentially the same type of
turbine/generator, with the turbine Figure 4.9 Gross power generated by turbine of all
manufactured by GE in the lead. PWR (P) and BWR (B) units that started
commercial operation before 1989
Taking this agreement into account, it is not
unreasonable to conclude that the key factor 100
in limiting the power of the nuclear power
A
plants was the fact that, at that time, the A
standard large-scale turbine/generator in the
United States market was 1200 MWe. All the N 80
players in nuclear system design agreed to E
CJ n
A
use this particular type of unit. Either (a) 0, n
~ 6. ~
DA
larger units were not available at the time or Q)

J.. ~~a~~&ft~~~~6.~~O~ A A
BcaDD 0
::J
(b) the cost of developing and optimizing a II) 60 6. nO ~ IJ
tJ
II) 0"16° CIa 8
larger unit was far beyond the expectations to Q)
J..
nt
a a
0
a. []
obtain profit from economies of scale. This C) (J
a~ffiCi
f'j .. ~ u
Cl
.. agRnaAS
u8 . i:I []
conclusion is also supported by the fact that .:: a

10 A
(J
not only the turbine gross power achieved a J..

Q) 40 A
convergence value. Results presented in a. n
0 0
Exhibit 4.10 show that, in addition, the main Q)

C
variables involved in the process followed a :c
J.. (]
(J
0
p
::J
similar pattern. For instance, the figure l- 20 6. B
50 55
n
.
shows that the steam temperature converged 60 65 70 75 80 85
to a nominal value of about 280°C. Since the
steam temperature corresponds to a Order (year)
saturation condition, the same plot also
illustrates the fact that the operating pressure Figure 4.10. Pressure of turbines of reactors
inside the turbine was kept constant: around PWR(P) and BWR(B) in the world as a function of
60 Kg/cm2• the year of initial commercial operation
46
CONSEQUENCES FOR THE PWR DESIGN

Turbine/generators utilized in the nuclear industry were not specifically developed for this purpose. In
fact, these products were a standard component of other power plants. By 1950, both General Electric and
Westinghouse were the leaders in the manufacturing and design of turbine/generators used in coal and gas
electric-generating stations. This helps us understand why both companies offered the same standard
product with a gross power of nearly 1200 MWe. A feasible explanation is that since they were both
involved in the larger market of conventional electric-generation plants, their strategy was to provide a
standard product. In this way, they increased the chances to compete against outsiders.
As mentioned previously, the driving force in the development of nuclear reactors was the desire of
obtaining economies of scale by means of increasing power. The criterion of "high power output implies
more efficiency" was consistently pursued by reactor vendors. However, in order to meet the strict goal of
efficiency, designers were faced with the dilemma of increasing power but at the same time utilizing
standard components. As was accurately pointed out by a nuclear researcher who worked in the industry
during those days:
The pressure of the marketplace and the desire to make a profit in this business-after many
years of losing. money- was the main motivation to limit the size of nuclear reactors and
standardize the designs. (Lahey, 2000)
In the early sixties the designers were confronted with the fact that they were not able to increase the
power of nuclear plants indefinitely. The "ceiling" in the size of plants was defined by the availability of
large-scale-unit turbine/generators. And this ceiling became visible at 1200 MWe.
In simple words, it was like trying to design "the biggest four-wheel-drive car utilizing a 3 liter engine".
Under such severe constraints there were not too many design alternatives and obviously a dominant
design will appear rapidly.
It is notable that the same firms that monopolized the turbine market were also involved in the reactor
design. However, it is important to note that these were different divisions of the same corporation and, to
the extent of this investigation, they should be understood as different business units. For that reasons,
there are no grounds to speculate that the nuclear designers had any influence in reshaping the design of
the turbine generators. Designers were simply compelled to use the state-of-the-art standard product
provided by their corporation.
47
Chapter 5
THE RELATIONSHIP BETWEEN WESTINGHOUSE AND GENERAL

ELECTRIC
This chapter is basically about antitrust. It starts with a summary of several cases of price fixing where
Westinghouse and GE were involved. The emphasis is on the notorious case in the early 1960s involving
large turbines, which led to the imprisonment of several corporate executives. Since the subject of the
thesis is not primarily antitrust, the stress is purely historical and within the context of nuclear power
plants. The question that the chapter raises is whether price-fixing strategies employed by Westinghouse
and GE in large electric components molded the rise of the PWR as a dominant design in the market.
THE MARKET FOR LARGE TURBINE GENERATORS

Historically, the market of large turbine generators has been dominated by only a handful of players. The
reason for such concentration of market power within a small number of manufacturers is the existence of
high barriers to entry. These barriers to entry are due -basically- to the following factors: (a) to
manufacture the product a heavy capital investment is required up front, (b) R&D costs are huge, and (c)
there are" strong learning effects arising from cumulative output. All these issues constitute an effective
deterren t against newcomers
When nuclear power emerged in the United
States in the mid 1950s, only three firms -
General Electric, Westinghouse and Allis-
Chalmers - were in condition to build the
large steam turbine generators required by
nuclear power plants. Given the fact that
turbine generators produced over 80% of the
U.S. power supply, it is clear that firms
competed over the dominance of a big
market.
In the late 1950s, General Electric was
probably as diversified as it is nowadays.
General Allis-
The manufacturing of large turbine Westinghouse
Electric Chambers
generators was part of its key division of
heavy-capital goods, which accounted for • Total Sales ($MM) 2340 1260 340
nearly one-fourth of the corporation's total o Total R&D ($MM) 254 163 8
sales. As Figure 5.1 shows, General Electric II Shipments (GW) 102.7 53.1 11.3
was the clear leader in the market for large
turbine generators, both in terms of Figure 5. 1. U.S. manufacturers, aggregated financial
accumulated experience and total sales. results from turbine generators from 1948 to 1962
Nearly 60% of all the new capacity installed in the United States in the period between 1948 and 1962
was captured by General Electric. Its sales almost doubled the sales of its immediate competitor,
Westinghouse. Thanks to its favorable position in the market, General Electric had also been the price
leader, with its two main competitors usually matching GE's book prices.
General Electric was also the leading R&D investor in the segment, reinvesting nearly 15% of its
revenues on developing product and process innovations. As a result, General Electric had not only been
the market leader in terms of sales but also in terms of technology. It had historically been the pioneer in
the introduction of innovations that led to the remarkable increase of energy conversion efficiencies of
large turbine generators.
Westinghouse was also a diversified manufacturer of capital goods -among them large turbine generators-
and was a corporation as big and as profitable as General Electric. By the late 1950s, Westinghouse held a
solid second place in sales with an average market share of 300/0.But in the particular market of large
turbine generators Westinghouse was widely perceived as a market follower, with most of its marketing
and technology strategies designed to match GE's leadership. The main focus of Westinghouse was in
standardizing units, in order to benefit from economies of scale, reduce costs and -in the long run- being
able to challenge the price leadership of General Electric.
Allis-Chalmers was the third competitor on large turbine generators in the United States, although its
market share was quite small. It could not match the prices of the two incumbents and had not been able
to sell enough units to benefit from economies of scale as its competitors did. In addition, Allis-Chalmers
had not invested enough in R&D and therefore its products were becoming obsolete. Unable to
differentiate itself in terms of quality or price, the firm finally decided to exit this market in 1962 after
decades of marginal results.
Finally, two European firms, Brown Boveri and Parsons, had attempted to break into the U.S. market but
were quickly neutralized by sharp price cuts from both GE and Westinghouse. In addition, they lobbied in
Washington for tariffs, arguing that the entry of these firms in the domestic market was considered a
threat to energy independence. The two incumbents effectively barred potential foreign competition by
raising entry barriers, using pricing and lobbying strategies.
CHARACTERISTICS OF THE PRODUCT

As it is shown in Figure 5.2, a large turbine generator is rather complex and expensive equipment whose
main function is to convert heat rate into work rate and ultimately into electricity.
Figure 5.2. A large turbine generator (1,000 MWe unit). (1) the turbine converts heat into mechanical work,
(2) the generator converts work into electricity, and (3) the generator alterix maintains the electric field.
Although manufacturers have made a strong effort to standardize the product, most of the times engineers
need to customize each unit to a particular specification. As a result, the engineering and marketing costs
to adapt a product to the buyer are relevant (on the order of 10% of the total cost of an order).
49
As a heavy capital-intensive good, most of the costs are direct costs due to parts and labor. The overhead
costs are small compared to other manufactured capital goods, accounting for less than 25% of total cost
(USGPO, 1981).
The main performance parameter of a large turbine generator is the thermal efficiency. The latter is
defined as the ratio of electricity output to total energy input. In the 1930s the standard thermal efficiency
of turbine generators was around 20%, but a number of innovations introduced in the 1940s boosted this
parameter to 35%.
However, improving the efficiency had a technological limit -which turned out to be 35%- and by the
turn of the decade manufacturers had to pursue another way to increase value to their customers and to
differentiate themselves from competition. This was by offering larger and more complex units, which
reduced operating costs and also decreased the investment cost (in dollars per installed electric kilowatt).
Utilities grasped the opportunity for
increasing profits and began to buy larger Electricity Simple design
and larger units. In the early 1950s the Cost
largest turbine generator available was ($/kWh)
200 MW e while in the early 1960s units
had been introduced with ratings well Complex design
~
above 500 MWe. By the end of the 1960s,
the standard large turbine generator
offered to utilities was around 1000 MW e,
I
a trend that was magnified by the massive
introduction of nuclear power plants. The
latter is a case of strong economies of Size (MWe)
scale. As Figure 5.3 shows, when
economies of scale are manifested, larger, Figure 5.3. A typical price-size curve in the large turbine
more complex designs exhibit a different generator market
price-size curve from smaller, simpler
designs (Shy, 1998).
Turbine generators' buyers -generally large private-owned utilities- benefited from economies of scale
because, by buying and operating a large turbine generator, the total cost per unit of output was reduced,
even though the product became more complex. This makes sense for any large utility, as smaller and
simpler units are competitive only in the niche market of small grids (e.g. remote isolated areas).
In addition to economies of scale -which
allow the buyers to obtain more profits- a Cost per
large turbine generator is a product with unit sold
intrinsic learning effects. This means that ($/MWe)
as the manufacturer produces more of a
product, the unit cost of production
reduces at a decreasing rate. This
phenomenon is frequently described with
a learning curve -as shown in Figure 5.4-
namely a plot of the cost of producing a
unit versus the organizational experience. Cumulative number 0
The price per unit drops by virtue of units produced

increasing the cumulative number of units Figure 5.4. A typical learning curve in the large turbine
produced. generator market
Learning curves have been documented in a number of organizations in both the manufacturing and
service sectors, and the yardstick for comparison has been the progress ratio. According to its definition,
50
each doubling of cumulative output leads to a reduction in unit cost to a percentage of its fonner value.
This percentage is the progress ratio. Thus, an 80% progress ratio means that each doubling of cumulative
output leads to a 20% reduction in unit cost (Argote~ 1990).
According to the literature, the learning curve of large turbine generators exhibited a progress ratio of
90% (Porter, 1986). The latter means that each doubling of cumulative output led to a 10% reduction in
unit cost. The result is comparable to learning curves in other manufacturing sectors of large capital goods,
for instance, commercial aircraft and merchant ships (Ghemawat, 1985).
ill summary, the buyers- utilities- had an incentive to buy large complex turbine generators instead of
small units, and the manufacturers had a strong incentive to increase sales volume in order to benefit from
learning effects.
THE BUYING PROCESS

ill 1962, the U.S. market for large turbine generators was segmented in three categories: (a) 143 private-
owned utilities such as Duke Power or Edison Electric, (b) 62 Federal Government utilities such as the
Tennessee Valley Authority, and (c) over 1,500 Municipal utilities. Figure 5.6 summarizes the market
share of the three segments.
ill the 1960s, the sale of a large turbine generator Municipal utilities Miscellaneous
.«~~
was a process that took around two and a half years. 11% \ :/ 1%
The long delivery times were due first to a six-
month negotiation time. Then it was the engineering
and construction time of a year to eighteen months.
The rest of the delivery time reflected the
manufacturer's order backlog. With these long
Federal
~l
government~. ::::::::
utilities .
delays, the turbine generator was the bottleneck in
constructing a new power plant and accordingly
utilities' managers spent a lot of effort in trying to
speed up the entire buying process. The buying
process itself differed according to the market Figure 5.6. Market segmentation in the early
segment. 1960s
For Government-owned utilities as well as Municipal authorities, the process generally took the form of a
standard public procurement process, namely: specifications, sealed bids, and purchase decisions based
on a competitive basis. Private-owned utilities had a rather more complicated and dubious process.
Whenever a utility had an interest in buying a new asset, it would ask the manufacturers to submit price
quotes. Then a long negotiation process followed which generally took place under incomplete
information on both sides. Manufacturers did not know each other bids, and utilities did not know the real
manufacturer's costs. Even after an order was placed, the mystery about the final price was not revealed.
Given the high price tags of large turbine generators, final negotiations were generally conducted by top-
executives and involved a long process of proposals and counterproposals where the main negotiation
point was price. Degrees of quality, post-sale service, and delivery times were similar among competitors
and played a minor role in the negotiations. The clear differentiating factor was price (Gilbert, 1996).
THE PRICE-FIXING SCANDAL

Between 1961 and 1965 there were a number of judiciary prosecutions involving executives of both
General Electric and Westinghouse. Prosecutors had initiated investigations after a complaint from the
government-owned Tennessee Valley Authority concerning identical bids they were getting from
manufacturers of large turbine generators even though the bids were submitted in sealed envelopes (Geis,
1996).
51
The investigation lead to one of the most notable antitrust cases in the United States. Prosecutors argued
that Westinghouse, General Electric and other minor electrical equipment manufacturers were engaged in
price-fixing practices. They claimed that top executives were engaged in a willful price-fIXing conspiracy,
which is a crime contrary to the spirit of the Sherman Antitrust Act of 1890, which forbade price-fixing
arrangements as a horizontal restraint upon free trade.
As a result of the investigation, four grand juries were ultimately convened and subpoenaed a large
number of persons, some of whom cooperated with the prosecutors revealing the modus operandi of the
association. As a result, 20 indictments were issued involving at least 45 executives. Given the weight of
the evidence against these individuals, most of them pled guilty, avoiding in this way the social cost of a
public trial.
Millions of dollars in fines and treble damages were charged against the individuals and corporations
involved in the scandal but, most importantly, several executives were sentenced to brief jail terms. Seven
executives from General Electric and Westinghouse, four of them vice-presidents of these two companies,
ended the day "handcuffed in pairs" and conducted to the Montgomery County Jail in Norristown,
Pennsylvania. As an illustration of the magnitude of the scandal surrounding the case, during their term in
prison, none of the seven men "had visitors during the Wednesday and Saturday periods reserved for
visiting; all indicated a desire not to be seen by their families or friends" (Geis, 1996)
WHEN IS PRICE-FIXING A CRIME?

The antitrust statutory body is given by the following acts: (a) the Sherman Act of 1890, (b) the Clayton
Act of 1914 later amended in 1936 by the Robinson-Patman Act and in 1950 by the Celler-Kefauver
Anti-merger Act, and (c) the Federal Trade Commission Act of 1914.
The Sherman Act remains as the cornerstone in antitrust statutes, prohibiting contracts, combinations, and
conspiracies that restrain free trade. The infringement of this law triggers criminal penalties when
enforced by the government. Violation can also result in fines and, for individual transgressors, prison
terms (Kaye, 1990).
The most common violations of the
Sherman Act, and the most harshly treated,
are the so-called horizontal restraints of l~
trade. These are concerted actions among

firms in actual or potential competition with 120
one another. The reason for the zero

110
tolerance to these kinds of practices is that
antitrust laws postulate a competitive
marketplace where rival firms compete in
terms of price, product and services. Any
arrangement reducing the competitive
environment is against the essence of
antitrust laws (Kaye, 1990)
7Il
Price-fixing is the capital crime in
horizontal restraint of trade. Any agreement
among competitors with respect to prices is
illegal regardless of who is involved and
what the surrounding circumstances are.
First, competitors can not agree on the
actual prices they will charge or pay for a Figure 5.7. Index of turbine generator book prices and
product or service. Secondly, competitors actual order prices (Civil Action, Ohio Elec. vs. GE 1965)
52
can not agree on a price-range within which they will compete, or on a common list or book price.
Thirdly, competitors can not divide markets by territory or by customers.
Prosecutors found that Westinghouse and General Electric had effectively set a book price for large
turbine generators. Figure 5.7 shows the time behavior of book prices and actual order prices for large
turbine generators (Civil Action, Ohio Elec. vs. GE 1965). Salesmen from both firms would visit utility
executives and offer prices according to a jointly accorded price. Whenever it was necessary, both firms
would modify the book price to cooperatively drive prices up or down.
Having a book-price however is not the only condition to trigger antitrust penalties. There are three
necessary conditions that need to be put in place: agreement, monitoring and enforcement. The agreement
was largely demonstrated in court by the existence of a price-book, and was not the main issue.
The main issue in court was whether there were monitoring and enforcement provisions in order to make
the price-fixing scheme work. The latter was the core of the conspiracy charges. In other words, the
question was if there was an established organization, which made price-fixing arrangements, monitored
these arrangements and enforced penalties to cheaters.
The outcome of the process, with executives pleading guilty and serving terms in jail and corporations
paying million-dollar settlements, overwhelmingly demonstrates that indeed there was a conspiracy.
WHITE-COLLAR CRIME
When most of General Electric and Westinghouse executives were sentenced to jail, a newspaper reporter
described them with a certain degree of sarcasm as "middle-class men in Ivy League suits- typical
business men in appearance, men who would never be taken for lawbreakers" (Geis, 1996).
On the same line of thought, Edwin H. Sutherland coined the phrase "white collar crime" for defining a
crime committed by a person of respectability and high social status in the course of his occupation.
Consequently the definition excludes many crimes of the upper class -murders, abuse cases- that are not
committed in the workplace, and also excludes wealthy members of the underworld, since they are not
persons of respectability and high social status. Hence, the sophisticated offenses that white collar crime
encompasses are things like restraint of trade, deceptive advertising, and patent infringement. (Sutherland,
1996).
Sutherland points out two important characteristics of white-collar crime. First, most of the offenders are
not really aware of the seriousness of their acts. For most of them, the borderline between duty and crime
is blurred by a corporate culture where incentives are tilted towards productivity, sales performance, and
profit margins. In the case of large turbine generators, a high-ranking executive categorically denied the
illegality of his behavior. "We did not fix prices" he said, "all we did was recover costs" (Geis, 1996).
The second characteristic of white-collar crime is the recurrence of crimes. In the period between 1930
and 1955, General Electric and Westinghouse had prompted a number of decisions in courts regarding
restraint of trade -13 decisions for GE and 10 for Westinghouse. Despite this record of indictments both
firms continued to practice price-fixing, and have probably continued to do so after 1965 (Sutherland,
1996 and Fuller, 1962). Westinghouse in particular had a similar experience regarding the uranium
business (Joskow, 1976).
REASONS FOR COOPERATIVE ARRANGEMENTS

The price conspiracy that unfolded in the late 1950s and early 1960s involved large turbine generators
and other heavy electrical equipment. The reasons why a cooperative -though illegal- arrangement
between GE and Westinghouse emerged are numerous and difficult to confirm. The following are some
of the most likely reasons:
53
1. Response to destructive competition. Very often cooperative arrangements arise after a prolonged
destructive competition between firms on the grounds of price. This generally occur with
homogeneous products such as oil (OPEC), copper (CIPEC) or diamonds (DeBeers). In the case of
turbine generators the product is not homogeneous but the buyers -utilities- were remarkably engaged
in price bargaining. Product performance, post-sale service and delivery times were similar across all
competitors.
2. Inelastic demand. Price-fixing schemes can only be applied in markets with inelastic demand,
namely, where the product demand is not significantly reduced by price increases. Since large turbine
generators were key components of power plants and utilities had to necessarily buy these assets, they
ended up paying the asked price.
3. High entry barriers. Price-fixing arrangements work perfectly when there are barriers to entry. As it
has been shown in previous sections, the large turbine generator market had huge barriers to entry.
These were due to the heavy capital investment that is required up front in order to being able to
manufacture any unit, along with the R&D costs necessary to be technologically competitive. In
addition, there are strong learning effects arising from cumulative output that favor the incumbents
versus the newcomers.
4. Weak legal impediments. Historically antitrust laws have been difficult to enforce because (a)
conspiracies are difficult to prove and (b) the all-embracing statutory language of the laws is not
applied. For instance, in 1911 by the Supreme Court introduced the "rule of reason" ruling that
despite the all-embracing statutory language, the Sherman Act reached only those trade restraints
which are unreasonable. In other words, courts weigh the anti-competitive consequences of a
challenged practice against the business justification, leaving most offenders unaccountable.
CONSEQUENCES FOR THE INDUSTRY

In the particular area of large turbine generators, the consequences of the price-fixing arrangement
between General Electric and Westinghouse were enormous. First this might have been the main reason
for the exit of Allis-Chalmers from the business, and also the main entry deterrent to other competitors -
like Parsons from the UK and Brown Boveri from Switzerland. Using these practices, the two incumbents,
General Electric and Westinghouse, were able to seize a larger portion of the turbine generator market
and reap more profits per unit sold. In addition they benefited from the externalities arising from larger
cumulative manufacturing experience in such a critical area.
In the utility sector, the increasing costs -depicted in Figure 5.7- were directly charged to end users,
because this was the common practice during those years. Large overnight capital costs meant higher
electricity prices to end-users. Since the electricity demand is inelastic -at least in the short run- the
electricity demand was not noticeably affected by this practice (Lean, 1982).
CONSEQUENCES FOR THE PWR DESIGN

The story of price-fixing in large turbine generators became public when the first nuclear power plants
were beginning to operate. In 1957 the first nuclear power plant in Shippingport went on line. In October
1959, Dresden-I, the first U.S. nuclear plant built entirely without government funding, achieved a self-
sustaining nuclear reaction. In August 1960, the third U.S. Nuclear Power Plant, Yankee Rowe Nuclear
Power Station, achieved a self-sustaining nuclear reaction. All these plants utilized large turbine
generators. All in all, although their operation started a few years later, sixteen nuclear powered units
were ordered in the period between 1952 and 1963.
54
BWR units:
1955 Dresden 1, IL 200MWe Commonwealth Edison Opel'. Jut-60
1957 Pathfinder, SD 59MWe Northern States Power Opel'. Jul-66
1958 Humboldt Bay 3, CA 63MWe Pacific Gas and Electric Opel'. Aug-63
1958 Elk River, MN 18MWe River Coop. Power Opel'. Jul-64
1959 Big Rock Point, MI 69MWe Consumers Power Co. Opel'. Nov-65
1960 Puerto Rico Bonus 72MWe P.Rico Water Authority Opel'. Aug-64
1962 La Crosse, WI 51 MWe Dailyland Power Opel'. Nov-69
1963 Nine Mile Point I,NY 615MWe Niagara Mohawk Power Opel'. Dec-69
1963 Oyster Creek,NJ 650MWe General Public Utilities Opel'. Dec-69
PWR units:
1953 Shippingport, PA 60MWe Duquesne Opel'. Dec-57
1955 Indian Point 1, NY 257 MWe Consolidated Edison Opel'. Jan-63
1956 Yankee-Rowe, MA 175MWe Yanquee Atomic Co. ' Opel'. Jul-61
1957 Saxton, PA 4MWe Saxton Nuclear Exp. Opel'. Jut-59
1959 Carolinas, SC 17 MWe Carolinas Nucl. Power Opel'. Dec-63
1963 Haddam Neck, CT 582 MTiVe Connect. Yankee Power Opel'. Jan-68
1963 San Onofrel,CA 436MWe Southern Cal. Edison Opel'. Jan-68
All these plants were acquired by privately-owned utilities with some support from the AEC. Most were
turnkey projects where the reactor system and the turbine generator were offered together as a bundle. To
show the manufacturers that were involved, Table 5.1 provides data on vendors of turbine generators
versus vendors of the reactor systems, in these 16 power plants.
Table 5.1. PWR (3) and BWR (#) units ordered from 1952 to 1963
Turbine generator manufacturers versus reactor manufacturers
~~~i>.in,ejge~~~:ator~~;';fa~(~~eI:
~::7.:'~:
•.. ~....~.~:
$E;Ai..~=3:m~~_¥-=~:::6 ~~t~:: ~Ii';4ii;~;.?£

~If~~ctritj;;:d
t~EG:~;~'S ### 3
~.;~~ ###### H ~
~:~#~r~~:~:~~
~':'-""""""":'-~~"":-~.~';"~"~'.'~"''''~~,,~''''j.'':':'
;););););) 5
:*QjIiI:-:iE::l::::'2::ii::::; 3 6 7 16
As Table 5.1 shows, GE and Westinghouse benefited evenly in terms of sales during this period, with
Allis-Chalmers falling behind. Allis-Chalmers not only built turbine generators; it also built three of the
first nuclear reactor systems in the United States. The self-evident question is therefore whether the
duopoly exerted by GE and Westinghouse indirectly prompted the exit of a potentially harmful
competitor like Allis-Chalmers.
The evidence is just circumstantial. There is evidence that GE and Westinghouse were involved in price-
fixing in the large turbine generator sector in general, but there is no direct evidence that these practices
were used in the particular case of nuclear power plants -regarding reactor systems, turbines or a bundle
of both products. Chances are however that many of these deals were indeed non-competitive.
The modality of turnkey projects that prevailed during these early contracts offered an excellent
opportunity for tying together the reactor system and the turbine. The same manufacturer would build
55
both the reactor system and the turbine generator. As Table 5.1 shows, this was the favored option for
Westinghouse (5 cases) and General Electric (6 cases). The danger of this practice is when incumbents
use horizontal and vertical restraints of trade to keep a competitor out of the game. Although the evidence
is not conclusive, this looks to be the case with regards to Allis-Chalmers.
The cooperative arrangement between GE and Westinghouse seemed to be a win-win deal. Both were
able to exploit their technical competencies in manufacturing large turbine reactors and at the same time
gaining cumulative experience in building reactor systems. Economies of scale, economies of scope and
learning effects all seemed to be in favor of the two incumbents.
It is true that both firms had different designs, General Electric developed BWR reactors while
Westinghouse developed PWR reactors. But as it was shown in previous chapters, both designs are light
water reactors and their designs are very similar and complementary. It is clear then that any non-
competitive practice that existed in the sector helped both the BWR and the PWR designs, at least in the
short run -obviously to the detriment of other designs such as gas cooled reactors. In the long run one of
the two "rewarded" designs would eventually prevail as the favorite of the buyers. It turned out to be the
PWR but it could have been the BWR. The market made its choice in this matter, although with only two
options on the table.
The question is why the buyers preferred the PWR over the BWR if both products had similar
performance and similar economic benefits? The most probable answers come from two areas. First,
operators were not comfortable by the fact that in the BWR design water contamination spreads all over
the plant, including the turbine generator. For them the PWR had the advantage of clearly separating the
plant in two different systems -conventional and nuclear. The second reason is related with the supplier
base. The PWR design had a broader suppliers base in the U.S. market -Westinghouse, Babcock &
Wilcox, Combustion Engineering- while the BWR was manufactured -with high vertical integration- by
General Electric only. In the minds of the operator managers, a broader supplier base meant less
investment risk (Kadak, 2002).
56
Chapter 6
THE MANUFACTURERS GAME
Given the relevance of collusion and price fixing in the power industry, this chapter has an analysis of
manufacturers and their involvement in the design and construction of nuclear power plants. The analysis
starts with a model of the order/construction process, which describes the role of utilities, architect-
engineers, civil constructors, reactor vendors and secondary suppliers. The way these firms interacted had
a profound significance on the consolidation of the Westinghouse PWR design. In addition, the chapter
focuses on the performance of the PWR design internationally. The data on France, Japan, Belgium,
Germany, Sweden, Spain, Taiwan and Korea seem to show that these countries overwhelmingly accepted
the Westinghouse and GE designs as the industry standards.
THE PRODUCT
Nuclear Power Plants are complex electricity-generating assets that bring together a nuclear heat source -
the reactor- with a conventional system -turbine generator- that converts heat into work and into
electricity .
The main component of a nuclear power plant is the reactor itself. This is a very complex machine and its
design requires the competence of a number of sophisticated technologies such as neutron physics,
thermal-hydraulics, radiation shielding, reactor dynamics, radioprotection, etc. Most of these technologies
were still under development in the late 1950s and therefore only a handful of firms with extensive R&D
capabilities could handle the development of nuclear reactors.
The conventional system, on the other hand, is just an extension of an existing product line, the large
steam-driven turbine generators. This electricity equipment was part of most of the fossil-fueled power
plants in the United States (about 80% of the U.S. power supply in the early 1960s). Thus, there were
\huge economies of scope and scale, which the incumbent vendors of large turbine generators could
benefit from. For a manufacturer of large turbine generators, nuclear power plants were only a new
market segment with similar needs. There was little need for additional R&D effort as adapting the
turbine generator utilized in a coal-fired plant to a nuclear power plant was relatively straightforward.
In summary, in the early 1960s nuclear power plants were simple product substitutes addressing an
existing market for electricity-generating assets. The new product was half based on a revolutionary
innovation -the nuclear reactor- and half was a conventional machine that enabled the conversion of
thermal energy to electricity.
Nuclear power plants broke into the electricity-generating market when electricity operators were looking
for units of larger power ratings. Unable to boost profit margins by means of increasing efficiencies' of the
steam thermal cycle, as they had done in the 1940s and 1950s, electricity operators envisioned a future
where power plant ratings would be in the thousands of megawatts rather than in the hundreds. The vision
of higher profits by means of economies of scale was shared by U.S. utilities and by other electricity
operators in the rest of the world, like Electricite de France or Tokyo Electric Power. Rapidly, reactor
vendors focused on enhancing the power output of nuclear power plants.
As it is shown in Figure 6.1 the

average size of a nuclear power plant 14
shifted from 200 MW e in the early

12
1970s to 900 MWe in 1980s. Nuclear
power plants surfaced as a 10
commercial product just at the 19605

perfect time. The new technology
could provide the means of building
large power plants with very low fuel
cost.
Starting in the late 1960s a number Std. Dev" 188.53
Mean" 210.4
of environmentalist groups pushed N = 75.00
for increasing levels of safety in 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0
nuclear power plants and advocated 50.0 150.0 250.0 350.0 450.0 550.0 650.0 750.0 850.0
for the application of tougher 60
environmental standards. For

instance, there was concern over the 50
thermal pollution of nuclear power 19805
plants, i.e. the fact that water from 40
the condenser would be discharged

directly into the sea, lakes, and 30
rivers; increasing ambient

temperatures. 20
In 1966 the dispute had emerged 10 Std. Dev '" 264.54

publicly during a disagreement over Mean" 928.4
an application for a construction N = 213.00

0.0 200.0 400.0 600.0 800.0 1000.0 1200.0 1400.0
permit for the Millstone Nuclear 100.0 300.0 500.0 700.0 900.0 1100.0 1300.0 1500.0
Power Station in Connecticut.

Unit's Electric Power (MWe)
Shortly later, Vermont Yankee
Power had a similar problem during
the licensing of a nuclear power
plant, and rapidly the furor over Figure 6. 1. Histogram of power ratings (in MWe) of PWR and
BWR reactors in the decades 1960-69 and 1980-89
thermal pollution extended to the
licensing of any new power unit in
the United States.
The thermal pollution controversy Environmental
finally died down in 1970 after the and Safety
introduction of the Water Quality Regulation in
Nuclear Field
Improvement Act, which requested 1952: Reactor
that federal agencies, including the Development
ABC, would consider thermal Program
pollution issues in the course of their
licensing reviews. This legislation
prompted the mandatory installation Promotion of
of cooling towers, which cool the nuclear power in
water from the condenser -via natural the U.S.
convection - through the atmosphere
(Duffy, 1997). Figure 6.2. Sketch of the AEC evolution from tight control of the
technology and little free-enterprise to openness
The change of safety culture is
58
schematically shown in Figure 6.2 as a trade-off between almost exclusive promotion in 1952, to almost
exclusive nuclear regulation in 1974. There is a trade-off because it is obviously impossible to have high
regulation and high promotion at the same time and under the same government agency. In part due to
public concerns, the ABC, which had the double role of promoting and regulating the nuclear industry,
tilted towards the side of safety. This change of policy was accentuated in 1974 when Congress dissolved
the ABC and created the Nuclear Regulatory Commission (NRC) with the sole responsibility of enforcing
safety standards.
Similar safety and environmental controversies prompted the introduction of changes to existing safety
systems. For instance, in the early 1970s, there was a huge controversy over the emergency core cooling
system (ECCS) that ended up in tougher regulatory standards.
One of the peculiarities of power plants is that reactors were built with an unusual degree of variability
and diversity regarding civil works and piping. From a "macro" point of view units of the same design
concept were equal. But at a more detailed level they were very different. Essentially every reactor was
custom-designed and custom-built. The degree of standardization was minimal (USGPO, 1981).
The required safety level of a nuclear plant is
35.00
orders of magnitude higher than the safety
required in conventional power plant designs. !ill Other
30.00
This fact caused a significant increase in the _JAPAN
complexity of the licensing process. For 25.00 o FRANCE
instance, after Three Mile Island, nuclear
power plants were the only electricity-
generating assets that required a probabilistic
safety analysis (PSA) in order to be granted a 15.00
construction permit. The PSA is a
sophisticated methodology that allows the 10.00
identification of all possible initiating events

5.00
and the calculation of the consequences of
these events in terms of collective dose.
52 56 59 62 66 69 72 75 78 81
Although it is a nice standard, a PSA
demands an extremely time-consuming Year
process that involves thousands of
engineering hours in order to be performed Figure 6.3. Reactors orders in the U.S. France, Japan
adequately. The fact that almost every reactor and Rest of the World (except Soviet Union)
was one-of-a-kind caused difficulties in
identifying initiating events and verifying the
safety of individual plants.
-- --- ---
The complexity of safety systems and the
environmental concern experienced in the
late 1960s and 1970s led to longer licensing
Number
process times. The ABC (and later the NRC)
of units
ended up allocating more human resources ordered
per reactor application. It is important to note
that this is an abnormal feature because
normally the licensing process becomes more
efficient relative to the cumulative number of
operating units.
To make it more difficult, the early 1970s
Figure 6.4.Effect of compounding on licensing delays
were the years when the orders of new power
59
plants peaked. As shown in Figure 6.3, in 1973, thirty-five units were in operation in the United States
and utilities ordered the exceptional number of 41 new nuclear power plants.
Figure 6.4 illustrates schematically the compounded processes of (a) increasing the complexity of safety
features and licensing applications and (b) the number of applications per year. The result is obviously an
increase in the average length of the licensing process.
BUYING PROCESS
The natural buyers for nuclear power plants were private and government owned electricity operators.
As Figure 5.6 shows, in the United States 75% of the electricity capacity was owned by private utilities,
13% by the federal government and 11% by municipal authorities. In addition to the above-mentioned
natural buyers, there were new entrants willing to participate in the nuclear undertaking. An example was
Yankee Atomic Power, which was an organization created explicitly for the construction and operation of
a nuclear power plant built in 1960 in Rowe, MA. Yankee Power was owned by a consortium of New
England utilities.
Figure 6.5 has a geographic distribution of nuclear power sites. As can be seen, there was little
consolidation in the industry and several utilities -all across the country - embarked on nuclear power
projects. Some of them were not aware of the difficulties and complexities inherent in building and
operating a nuclear power plant.
As a former chief executive of Yankee
Atomic Power recalls, "lured by the
prospect of higher profits, operators
underestimated the complexity of
undertaking a project with a myriad of
participants and managing a highly
complex plant" (Kadak, 2002).
In the first years of the AEC, from 1946 to
1954, it was very difficult to get private
utilities committed to participate in nuclear
power. Some of the issues that discouraged
these firms were (a) the compulsory
government ownership of any nuclear
reactor, (b) the risk of unlimited product ... Ucenstd 10 Dperlle
liability in the event of a reactor accident, Note: There Ire
DO commerclalruclors In Alaska or lIawllI.
(c) the mandatory government ownership of Figure 6.5. Nuclear power sites in the U.S. by 1989
the fuel elements utilized by any reactor,
and (d) the uncertainty inherent to a new technology.
Most of these barriers were lifted in the period between 1954 and 1963. The Amendment of the Atomic
Act passed in 1954 paved the way for a new relationship between the AEC and the industry. It allowed
private ownership of nuclear reactors under AEC licensing and provided for near-normal patent rights.
The government maintained ownership of special nuclear materials -such as Plutonium- but was allowed
to lease these materials to the industry.
Regarding the question of liability in the case of an accident, in 1955 the AEC established a group from
the insurance industry to study the problem of insurance coverage for peaceful uses of atomic power. This
and other studies were the fundamental roots of the Price-Anderson Act passed in 1957. The Act required
licensees of nuclear power plants to furnish financial protection to cover public liability claims up to an
amount specified by the AEC, and provided government indemnity in the amount of $500 million for
60
each nuclear incident over and above the amount of financial protection required. The Act further
specified that the maximum coverage available from private sources was $60 million (Dawson, 1976).
In addition to removing all barriers for commercial participation, the ABC developed a program of
subsidies for nuclear power plants. The program designed to foster industry participation in nuclear power
as named as Reactor Development Program (RDP). The pioneer deal in this respect was Shippingport, the
first PWR developed in the United States and which started commercial operation in late 1957. The
participant utility -Duquesne Light- offered to furnish the site in Pennsylvania, build and maintain the
electrical generating section of the plant, contribute $5 million to the construction of the reactor and
purchase the steam produced by the reactor for the equivalent of 8 mills per kilowatt hour. For its part, the
ABC agreed to finance 90 percent of the reactor costs, build the reactor plant, and assume legal liability
for the plant. The ABC would own the reactor and the nuclear fuel while Duquesne owned the
conventional part of the plant.
A contract like the one signed by Duquesne, was really a sweet deal. The Shippingport plant ended up
costing $55 million and produced electricity for 25 years until its programmed shutdown in 1982.
Duquesne invested only $5 million in the nuclear heat source and about $10 million in the conventional
electricity-generating unit. In the case of Yankee Atomic Power, the contract signed in 1956 included a
commitment from the ABC to fund up to $5 million in R&D in support of the project. The ABC also
agreed to waive its normal charge for use of special nuclear materials to fuel the reactor, a figure
approximately $3 million. The total cost of the plant was $34 million (Dawson 1976).
Similar cases of AEC subsidies paved the way for private participation of the industry. With time these
subsidies vanished and by early 1970s, the full capital cost of nuclear power plants was absorbed by
private utilities. But by this time the industry had sufficient R&D investment and collective learning to fly
by itself.
It is important to note that the subsidies provided by the ABC were the "carrot" to attract private utilities
into the nuclear power business. There was a "stick" that would be present in the minds of utility
managers which was the concern about government-owned-and-operated nuclear power plants. Much of
the concern stemmed from the Atomic Act of 1946, which allowed only the government to own nuclear
materials. Given the fact that the government-owned Tennessee Valley Authority had nearly 13% of the
power supply in the United States in the late 1950s, it is clearly that the US government could become a
formidable competitor if utilities did not rush to build nuclear plants themselves.
CONSTRUCTION PROCESS
The process of designing and constructing a nuclear power plant was not performed by only one firm.
Generally the process involved the participation of a number of different companies. As it is shown
conceptually in Figure 6.6, each of the major participants in the process of designing and constructing a
nuclear power had a particular role within the project organization:
1. The nuclear heat supply system vendor designed and built the NHSS.
2. The architect-engineer firm was in charge of the balance of plant.
3. A main construction company was in charge of civil works.
4. A supplier of the turbine generator.
5. A supplier of large nuclear components (reactor pressure vessel and reactor internals)
61
The total number of companies

involved may well have been in
the hundreds, but the above-
Reactor
mentioned effectively controlled vendor
the major decisions during the
Turbine
project.
manufacturer
Whenever an electric utility
decided to build a nuclear power Turbine
plant, the first step was hiring an Generator
architect-engineer to help
estimate costs, select the other
industry participants and execute
the actual management of the
project.
As a prime contractor, the
architect-engineer's main function
was to design the balance of the
plant. Figure 6.6. Components of the design & construction process
The balance of the plant consisted of all the equipment, in addition to the nuclear heat supply system and
turbine generator, which is necessary to produce electricity from a nuclear power plant. In other words,
the Architect-Engineer had the responsibility to match the requirements of the NHSS with the
requirements of the turbine generator and assure that the nuclear power plant would work as an integrated
system rather than a mere collection of disconnected pieces.
The architect-engineer had the most critical
role in the design and construction of a
nuclear power plant for two reasons. First it Contingency
was economically relevant as almost 80
6%
percent of the budget of the plant
corresponded to the balance of plant. The Professional
second reason was a safety-related
preeminence, as 90 percent of the necessary 13%
paperwork required by the NRC to license
the plant was the responsibility of the
architect-engineer. Given the complexity of
the licensing process it is clear that a
smooth passage through NRC scrutiny
depended overwhelmingly on the technical Construction
competency of the architect-engineer. Materials &
Architect-engineer
equipment
One of the top functions of an architect- responsibility
engineer was to recruit the suppliers of the 27%
two key components of the plant: the NHSS
and the turbine/generator. In general there Figure 6.7. 1,000 MWe overnight construction cost
were only a handful of options in this bhreakuP. (flnot. including finan?ial costs, escalation
.... c a/iges In atlOn and tax deductIons)
respect, as It WIll be explamed In subsequent '
sections. Very often it occurred that the electric utility had already an agreement with some of the
suppliers. This weakened the position of the Architect-Engineer firm as chief project contractor.
In particular, in the United States the incumbent suppliers of the NHSS and the turbine/generator were the
same companies -General Electric and Westinghouse- and became the controlling part of the project. The
62
construction company was generally hired by the Architect-Engineer and its function was relatively minor
regarding the control of the project. In summary, the firms with higher power within the project -
additionally to the electric utilities- were the architect-engineer and the suppliers of the NHSS and the
turbine generator. Their relative share in the total costs of a typical project is summarized in Figure 6.7.
As it is shown in figure 6.7, the relative economic weight of the nuclear system vendor was small
compared with the magnitude of responsibility on th~ side of the architect-engineer.
ARCHITECT-ENGINEERS AND CONSTRUCTORS

In the United States, 22 firms participated as architect-engineers and 25 as civil construction companies.
With only a few exceptions, the companies that build nuclear power plants were the same architect-
engineer firms that designed them. The role of the construction firm is to build the plant according to the
specifications given by the architect-engineer and the reactor system manufacturer.
It is important to note that nearly 80% of the design of the plant is in the hands of the architect-engineer.
The most relevant architect-engineers in the U.S. power industry have been Bechtel (41 units built); Stone
&Webster (17 units), and Sargent &Lundy (15 units)
Civil Constructor
UECH CWE DUKE Dravo EBSO S&L S&\\' TVA lJE&C
34 2
3
4
6
2
8 4
2 13
6
37 8 7 2 8 14 6 8
Figure 6.8 Cross tabulation of civil constructors and architect-engineers
REACTOR SYSTEM AND TURBINE VENDORS

Five firms participated in the business of supplying nuclear systems in the United States: General Electric,
Westinghouse, Combustion Engineering, Combustion Engineering and Allis-Chalmers. Regarding turbine
generators only three participated: General Electric, Westinghouse and Allis-Chalmers.
63
Firms participating in the design and construction of nuclear power plants selected different strategies for
capturing value in the business. These strategies were based on several tradeoffs, which can be
summarized in the following categories: (a) economies of scope, (b) diversification, (c) market power, (d)
vertical integration, and (e) cumulative learning.
PRESSURE VESSEL MANUFACTURERS

A-C B&\V C-E CB&I GE RDl\'1 WEST
2
11
15
7 11 16 8 2
9 32 3 4 18
2 27 58 19 8 6 19
1970-79
1960-69
1950-59 0C-E -B&W
0 20 40 60 80 100
Market Share (0/0)
Figure 6.9 shows how C-E and B& W dominated the market of reactor
pressure vessels in terms of (a) units sold and (b) market share increase
The first dimension, economies of scope, refers to benefits arising from cumulative output in parallel
industries. For example, when Honda introduced its first car in the early 1970s, the company benefited
from the economies of scope arising from the previous cumulative output of motorcycles. Although cars
and motorcycles are different products, there have similar manufacturing and development processes.
The ability to expand horizontally to create new products leads to an increasing diversification of the
firm. Sometimes the diversification of product portfolios goes too far and the organization becomes
unmanageable, the brand unrecognizable and the profits shrink. Therefore there is an optimum balance
between these two competing forces. Honda seems to be doing fine with motorcycles, cars, and SUVs and
does not look very eager to jump into the market for large trucks.
Babcock & Wilcox and Combustion Engineering benefited from economies of scope because these two
firms were the incumbent producers of large equipment for applications in fossil-fueled power plants,
such as heat exchangers, pressure vessels and high temperature circuits. It is not a surprise, therefore, that
these two firms became the incumbent manufacturers of reactor pressure vessels, steam generators, and
reactor internals. They used the experience accumulated in conventional power plants to enhance its
performance in the nuclear market.
64
Figure 6.9 shows that the market share of these two firms in producing these items increased over time.
Although successful in using economies of scope to build key reactor components, B&W and C-E had
limited experience in designing complete nuclear power plants. Consequently and after an early
successful introduction, these companies were unable to erode Westinghouse's dominance as PWR
vendors. Previously, in the market of turbine generators, these two firms had tried to compete with GE
and Westinghouse. But it was clear -even before the 1950s- that it was an attempt doomed to failure.
TURBINE GENERATORS MANUFACTURERS

A-C BBC GE Hitachi KWU MHI S-Laval WEST
2
5 6
9 7
40 6
2 2 17 2 3 49
4 2 71 1 2 3 68
1970-79 31.7 48.1
1960-69 33.9 41.1

DGE -West
1950-59 33.3
'0 20 40 60 80 100
Market Share
Figure 6.10 shows how GE and Westinghouse dominated the market of

reactor systems and reactor turbines in terms
In summary, B&W and C-E entered into the nuclear business benefiting from economies of scale, but this
effect was insufficient to make them solid NHSS vendors. As large diversified firms General Electric and
Westinghouse also benefited from economies of scale when they became NHSS manufacturers. Both had
already marketed a huge production of large turbine generators and had developed entire conventional
power plants. Their experience in R&D of complex power assets was already significant. Hence, both
firms were able to manage the process of building nuclear power plants. As a result, their market share in
the segment of nuclear systems increased over time as it is shown in Figure 6.10, as the same time as they
increased their predominance in turbine-generators.
The third dimension, market power, refers to the total market share that a firm can have in a given market
or segment of the market. For instance Microsoft has enormous market power in operating systems, while
Intel excels in computing processors.
Both of these firms have high market power due to their large market shares. They can obtain higher
profits by commanding output and price in the segment. However, Microsoft and Intel have been able to
obtain high market power, in part because they are focused firms and address a market of little vertical
integration. Neither of them tried to approach the computer business with high vertical integration as
Digital did in the 1980s.
65
6. The manufacturers game .
When firms compete for market power, the tradeoffs are (a) whether to concentrate efforts in a particular
segment in order to gain market share or to diversify, and (b) whether to attack a particular segment of the
industry or try to be an integrator. In the nuclear power business, GE approached the nuclear power
business with high degree if vertical integration. In a given project GE developed the reactor system, the
turbine generator, the pressure vessel, etc. Whereas Westinghouse focused its efforts only in the critical
areas of reactor systems and turbine generators. Regarding pressure vessels, for instance, it gave up the
market to Combustion Engineering and Babcock and Wilcox.
To summarize, the following are the market share ratios in the different segments of reactor construction:
Reactor Pressure Reactor Turbine

System Vessel Intern als Generator
# Firms in the business 5 10 9 6
Main two players GE, WEST C-E, B&W GE, WEST GE, WEST
2-firm market share 76% 64% 71% 95%
The market power in this industry was heavily concentrated in the hands of two players, General Electric
and Westinghouse. Only in the small segment of pressure vessels was their predominance overshadowed
by other participants. The most severe case of concentration was in the turbine generator segment. With
such concentration,
66
67
68
Chapter 7
THE DEMISE OF THE DOMINANT DESIGN
This chapter explores the reasons why the Westinghouse PWR collapsed in the mid-1980s. Data suggests
that incumbent firms that have successful dominant designs in the market, such as Westinghouse and its
PWR design, very often fail to be aware of subtle -but disrupting- shifts on the demand side. While
Westinghouse was busy building large and complex units in order to increase efficiency and profits the
customer needs were moving in a radically different direction, towards power units that had less inherent risk
(as large capital investments). The chapter shows that there was a fundamental change of mindset among
industry participants and consumers from the late 1950s to the early 1980s.
THE DEMAND CRASH

After a remarkable surge in orders between 1965 and 1974, the U.S. demand for nuclear power plants
unexpectedly collapsed and since 1980 there have been no new orders. Moreover, in the 1980s a number of
reactors under construction were cancelled or temporarily shutdown. Elsewhere in the world, reactors
continued to be built but inevitably the demand was reduced even in countries that strongly supported
nuclear energy such as France and Japan. Current worldwide demand is less than one-tenth of the U.S.
demand typical of the late 1960s. Such impressive decline is unusual in the energy sector, which is
characterized by long product cycles and stable technologies. The demise of a well-accepted product such as
the PWR is therefore a extraordinary episode in nuclear history. Several reasons have been considered in
order to account for such a decline. The most commonly mentioned are the following:
Three Mile Island and Chernobyl: The huge negative impact of these accidents on public opinion is
probably the single most mentioned reason for the demise of nuclear power. Namely, that electrical utilities
could not afford to buy new nuclear power plants after the wide public opposition against the construction of
these kinds of electricity-generating assets because of underlying safety concerns. A referendum in Sweden
banning nuclear power additions and similar government-lead efforts in Holland, Germany and Spain are
often mentioned as examples of a global anti nuclear endeavor that gained momentum after Three Mile
Island (1979) and Chemobyl (1986).
Notwithstanding the importance of these
160 180000
accidents in nuclear power demand, Figure
7.1 provides a contradictory view regarding 150000
this point. It shows that the demand of 120
nuclear power collapsed even before the 120000
!!
'c:::I CII
accident of Three Mile Island in 1979.
~ 80 90000
~
Actually, in the five years between 1975-79 z
CD
demand was for just 13 new reactors while 60000

40
in the previous five years demand had been 30000
for more than 140 new reactors. Following
a simple cause and effect reasoning, it looks 0 0
en -.;t 0> -.;t en
like the major nuclear accidents had a LO (0 to I'- I'-
ab 6 ab 0
I
rh
secondary role in the collapse of nuclear LO
en
(0
en
to
0>
I'-
en
I'-
0>
T"" T"" T"" T"" T""
power. Presumably they made things

worse, making a demand recovery less
likely to occur. But it cannot be said that
the demand slump was caused by the 1979
and 1986 events.
Regulatory complexity. Many people argue that the increasing intricacy of the regulatory process,
particularly in the United States, discouraged electrical utilities from ordering new nuclear plants. The
licensing process had become so cumbersome that most private firms had decided to dodge these difficulties
by relying completely on other electricity-generating assets, such as coal-fired plants and gas combined
cycles, which have a very good economic performance and less regulatory paperwork. It is important to note,
however, that after the collapse of demand in the 1980s, the NRC made a substantial effort to speed up the
regulatory process and conducted a profound turnover of its licensing standards. One can argue that
nowadays there would be few regulatory obstacles to the licensing of a new nuclear power plant but
nevertheless there is no demand upsurge. This might be showing that regulatory complexity is not by itself
the sole reason for nuclear power decline.
Competition from other energy sources. It might be that nuclear power was abandoned because it could
not compete with other fossil-fueled electricity sources such as coal and gas. Although competition seems to
have been a compelling reason for the nuclear power demise, there are a number of facts that show that this
by itself was not the only reason. First, although it is true that nuclear power has not been competitive during
some periods of time, its competitiveness has been highly dependent on the prevailing prices of fossil fuels,
which have been highly unstable oyer time. According to DOE releases, the 2001 nuclear power bus-bar
costs were lower than the costs of average gas-fueled plants, but this by itself has not prompted a bandwagon
of orders from the electrical utility sector. The truth is that the decision to buy large electricity-generating
assets obeys an elaborate tradeoff that is more profound than simple short-term energy prices. This tradeoff
involves: (a) volatility of energy costs in the long run, (b) reliance on imported energy sources, and (c) the
underlying financial risk of electricity-generating assets.
It is true that real overnight costs of nuclear
power plants increased notably in the period
when demand soared -between 1965 and
1975- as is shown in Figure 8.2. An ~ 4000
::>
overnight cost is the cost that manufacturers N
co
ask to an electrical utility for delivering the 0>
~
generating-asset. It is called "overnight" cost I 3000
because it is settled as if the manufacturers CD
could deliver the product overnight -not in
five years. As shown, the unit cost of nuclear
-
~
~
..... 2000
en
..
..
installed capacity in the United States
o
increased from about $600/kWe in 1966 to o
$3,000/kWe in 1980. A prevailing higher !:::
z
overnight cost had definitely an impact in ::>
--!
1000
..
•
U.S. nuclear power competitiveness but its «
w
role seems to be coupled with other factors. 0:: o
For instance, it could be the case that the 66 68 70 72 74 76 78 80
increase on the overnight price tag of nuclear
power assets was simply a demand vs. supply Year of start of construction
response to market pressure. The only
Figure 7.2 Real overnight costs of nuclear power
conclusion at hand from Figure 7.2 is that plants built in the United States (1982 U.S. dollars).
there were diseconomies of scale in the
construction cost.
In summary, there is empirical evidence that the profound decline of reactor orders was somehow influenced
by safety concerns, adverse public opinion, regulatory intricacy, and lower costs of alternative energy
sources. But none of these reasons is sufficiently strong to explain thoroughly the demise of nuclear power. It
seems that these issues are small parts of a broader picture that has not yet been fully understood. The next
sections are an attempt to do so.
70
THE UNFULFILLED EXPECTATIONS THESIS

The fundamental proposal is that the demise of nuclear power was caused essentially by unfulfilled
expectations on the part of buyers -the electrical utilities- and an incorrect reading of user needs on the part
of the manufacturers of reactor systems.
In the early 1960s, the main expectation from
electrical utilities was to benefit from economies Electricity
of scale by buying large nuclear power plants. It Cost Vie desig/:(coal & g:::nPlex design
was thought that the unit cost of producing ($/kWh)
electricity -in dollars per KWh- was going to be ~ (large nuclear
lower for larger nuclear units, even if the design plants?)
of these units was much more complex than the
design of simpler, smaller plants. An illustration
of this reasoning is provided in figure 7.3. It
shows that scale economies would provide for Size (MWe
large profits if there were sufficiently large units
Figure 7.3. Typical price-size curves for large
in the market. The estimated breakeven size of electricitY-Qeneration assets
price-competitive units was around 1,000 MWe.
Reactor system vendors and electrical utilities both adopted the same postulate. Reactor system vendors
developed standard designs that could easily scale up. For instance, the PWR can be seen as an effort from
Westinghouse to provide a product platform that could easily work for power ratings between 100 MW e and
1,000 MWe, with no fundamental changes of the design concept. In fact, the PWR was a reactor concept
derived from a nuclear submarine power unit (-1 OMW) and therefore it started and evolved through
successive adaptations.
As for electrical utilities, the hope of economies of scale triggered a different course of actions. First, they
were inclined to buy the reactor design that was more suitable for achieving scale economies, namely, a
design that could easily adapt to larger power ratings. The market data seem to indicate that the PWR was the
preferred concept for utility managers because it is arguably easier to scale up than its main competitor, the
BWR. To illustrate the point, the introduction of jet pumps in the BWR design was a major change of the
original concept prompted by the need to achieve higher power ratings (early BWRs relied only on natural
convection). The PWR did not have this problem.
Under these assumptions, it seems that the inexorable leadership of the PWR in the marketplace was a self-
fulfilling prophecy. It was a design concept suitable for producing higher power ratings without significant
design changes and it addressed a market where a premium was paid to firms that could come up with such a
design.
At the outset of the bandwagon of reactor orders -circa 1964- an electrical utility would base its buying
decision on two key issues: (a) the hope of attaining economies of scale, and (b) the promotional benefits
stemming from ABC's poliCies. Ten years later the expectation of actual economies of scale simply vanished,
and without such promise the demand collapsed:
c:
.S
C ,
El
Q ,
s..
Co I>
U
~
1964 1974 (expected) 1974 (real)
71
LISTENING TO THE WRONG SIGNAL

Christensen cites a number of cases where well-managed firms fail in the long run because of listening
excessively to their mainstream customers. These firms invest aggressively in product improvements that
would provide their customers the kind of benefits that they want. After a meticulous analysis of customer
needs, they allocate resources towards the goal of satisfying the needs explicitly stated by the buyers, but
paradoxically, many of these firms fail in the long run and lose their leadership positions (Christensen, 1997).
The point is that many times mainstream buyers demand the wrong performance dimension and induce the
vendors to excel in this dimension while the real "unexpressed" needs are moving towards a radically
different dimension.
In the particular case of nuclear power plants, the single most important performance parameter for electrical
utilities was scale. The .larger the unit the better, as it was assumed that larger sizes implied lower electricity
costs (in dollars per kilowatt-hour).
Reactor vendors worked hard to
achieve the goal of higher
power ratings and were quite
Ci) 1400
- --•..•.•......---.
successful in the effort. To
-
~ 1200 ..•..•; ~
..it
illustrate this point, Figure 7.4
g> 1000 1975
./•
summarizes the achievement of ;;
higher ratings as a function of ca
'- 800
'-
./
cumulative R&D effort. The Q)
latter is a typical way to graph ~ 600

o
the relationship between effort
put into developing
technology and the results one
a
c.
E
::s
E
400 1955
~..~. ,J ~1965
200
gets back for the investment.
'xca
~ ............ ~. --
.. .. ..-
..•.
/
F or most products and o

processes, when the results are o 20000 40000 60000 80000
plotted, what usually appears is
Cummutative R&D investment ($millions)
a curve with an S-shape (Foster,
1986). In the particular case of Figure 7.4. US" CUNe of maximum nuclear power rating (in MWe) as
nuclear power plants, the a function of cumulative R&D effort. Data valid for U.S. nuclear
picture is also an "S" shaped power plants and prototypes that started operation before 1975.
curve.
The reasoning behind this empirical result can be explained as follows. In the very early stage of nuclear
technology, funds coming from the ABC poured into R&D on nuclear power plants, but the technical
progress was very slow, By 1955 there were no commercial power plants in operation and early prototypes
had ratings no higher than 50 MWe. When the necessary cumulative resources and learning were put in place
the curve entered a second stage where huge advances were obtained with relatively small additional R&D
effort (circa 1965). Finally, in the last stage of the curve, it became more and more difficult to make technical
progress regardless of how much money was spent in R&D. In other words, a physical limit appeared on the
development of larger products using the same technology. This limit seems to have been 1300-1400 MWe
and became visible in 1975.
The cumulative R&D effort that was used to obtain Figure 7.4. was extracted from ABC costs (summarized
in Table 7.1) and therefore it relies on two critical assumptions: (a) the ABC was the single largest R&D
investor in nuclear power, and (b) all the investment made by the ABC was beneficial for developing nuclear
power plants. The first assumption is certainly true as no other organization -private or public- invested in
nuclear power as much as the AEC did. The second assumption is a little bit more questionable as the AEC
invested in other product lines -such as nuclear weapons- that are less likely to have a role in nuclear power
development.
72
Given the expectations regarding the benefits of economies of scale inherent in large power generating
assets, one can imagine that in 1974 most directors at Westinghouse and the ABC would have been very
pleased looking at Figures 7.1 and Figure 7.4. What these two figures were telling was that the demand of
new power plants was robust -especially PWRs- and that the prevailing power ratings of units was exactly
what electrical utilities were expecting. In other words, there was a greedy market for PWRs and
Westinghouse had achieved the long awaited goal of power plants in the 1,300 MW e range. Westinghouse
was clearly listening to its customers and had a strong product to address their needs.
However, listening to market
.
demand and customer needs
14
was exactly the wrong thing
en
to do in 1974. The right
signals to listen were the
~
Q)
12
•
," ~ ... -J?197
increasingly long times

-
.~ 10
.+.,."
A~.tual
.~./
required to construct plants . c
Originally in the early 1960s, ~o 8
when utilities jumped into ~...
.. .
1965 ~./
the nuclear bandwagon, they enc 6
o •• __ """1:r:- - - - - -Expected
expected construction times ~-fF' • -- ../
in the order of 5 years with ~ 4 ~~ ... +
en "yr •
ns ~'
slight increases due to ; 2
increasing complexity. In
/+
~
other words, using the same o
x-axis of Figure 7.4, namely o 20000 40000 60000 80000
cumulative R&D effort, a
flat curve was expected. But Cummulative R&D investment ($millions)
as Figure 7.5 clearly shows
this was not the case. Figure 7.5. Curve of average construction times as a function of
Construction times actually cumulative R&D effort (same as Fig 7.4). Data valid for U.S. nuclear
increased and peaked in power plants and prototypes that started operation before 1975.
1975 at around 12 years.
The first nuclear power plant in the
20
United States, Shippingport, was built in .... Countries
.
....
three years, and in days to come it was USA
predicted that normal construction times
D SWI
of larger units would take on average 15
..
five years. But this prediction totally
-
0
.. . .......................... D
SWE
underestimated the complexity of

nuclear plant construction in the United
~CIl
Q)
>. 10
. .. .............. .~
,
"
0 ..
D
SP
ROC
States. Just to provide an example, it e. .. •
. .. ~
.&1'" ...... o•
took 17 years to build the Diablo

Canyon nuclear power plant in
£
Ol
c:
~ ...
. a a A.()
. ."
f1
0 ~.D
......,
()IO. ......
D
DO
a
4Cll
II 0 0*11,
0+0
- KOR
o JAP
.. ..
c: 1tI 0- oe.o _
California (1968-85) . 5 ..• &(lIiI "'000 00
.Q DI D 1lIP- II ..... .. cxm 0 GER
U:J ;"
.... ~ 0
Figure 7.6 contains a plot
construction time versus net electrical
of
u;
c:
0
. * FR
power for all nuclear power plants built

l> 0
0 200
. D BEL
400 600 800 1000 1200 1400
in the western world, between 1955-89.
As is shown, it seems that the larger the Net Electrical Power (MWe)
nuclear power plant, the larger the

Figure 7.6. Scatter plot of construction time versus net
construction process. One has to be very
electrical power for all nuclear power plants built in the
cautious however regarding this world (exceot Soviet Union) between 1955 and 1989.
73
seemingly logical statement. First, the same graph shows that there are many countries -like Germany and
France- that were able to build complex plants of 1,300 MWe in five years. Second, beneath the role of
higher power ratings there are other issues that are highly correlated, like year of construction. Whereas all
small plants were built in the early 1960s, most of the large complex units were constructed in the 1980s
when the regulatory and industrial environment was different. Hence it is difficult to separate the effects of
size from other issues.
In summary, whereas the manufacturers of nuclear reactor systems were gratified by the fact that they had a
seemingly strong market demand and value proposition for buyers, construction times were skyrocketing and
that was very dangerous for the future of the industry.
WHY CONSTRUCTION TIMES ARE IMPORTANT

The longer the time required to construct a nuclear power plant, the more expensive is the capital cost of the
project and consequently the lifetime-Ievelized busbar cost of electrical energy. To understand the previous
statement it is necessary to note that the cost of any power plant is made up by three terms: (a) capital-related
costs, (b) operation & maintenance costs, and (c) fuel costs.
Each of the three components of electricity cost depend on financial charges such as the annual rate of
inflation, the cost of capital, mortgage rates, tax charges, etc. But the capital cost is the most sensitive
component because of the long times required to construct power plants. A large lump sum is required up
front as an investment and the revenue stream is unleashed only when the plant starts commercial operation.
To understand the tradeoffs involved in nuclear electricity economics, a simple model of the lifetime-
levelized busbar cost of electricity eb is provided below:
For plants fueled by coal and natural gas:
e
b
= 100Cl>Io
8760Cp
[1+ x +2 y]n + 100MI
8760C,..
[1+ y(L 2-1)] + 10 -6 ~[
HuF; 1 +
y{L -1)]
2
\. J \. I \. I
v
(:af'i/~/('/J.IiI Fixec/O&'" COS/.f Fill'! cos /_ jo.ui!
For a nuclear power plant (only the fuel term changes):
eh _- 100cDlo [1+ -- + y]n

x 100M,
+ ----'-- [1+ ---'----'-
y(L -1)] + ---
100F 2 [ y(L
1 + --- -1)]
8760CF
\.
2 I
8760CF
\.
2 I
2411Bu
\.
2 J
Caf'iI~/Co.'it FixeclO&M COS/.f Fuel co;/- jO,f.fil
The description of the various input parameters and their typical values are given as following:
Natural Gas Coal Nuclear
Overnight cost of the plant 10 ($/kW) 1,000 2,000 2,500
Capacity factor during operation CF 0.92 0.90 0.90
After tax cost of capital x (%/yr) 9 9 9
Inflation/escalation rate y (%/yr) 4 4 4
Plant construction time B (yrs) 3 4 6
Plant operation lifetime L (yrs) 30 30 30
O&M cost in 1st year of operation M1 ($/kW) 30 60 120
Plant thermodynamic efficiency 11 0.33
Plant heat rate HR (BTU/kWh) 7,000 9,800
Fossil fuel cost F1 (cents/million BTU) 600 150
Nuclear fuel cost of 1st core F2 ($/kgU) 1,800
Nuclear fuel burnup Bu (kWd/kgU) 40,000
74
Substituting the input data in the above-presented formulas one can obtain the total cost of electricity (in
cents/kWh) of the three alternatives, and one can compare with a lifetime-Ievelized bus bar cost. This is the
cost that any electrical utility would need to pay over the entire lifetime of the plant in order to break even.
Numerical results are presented in
Figure 7.7. They show that the main cost
• Capital cost III Fixed O&M cost 0 Fuel cost
component of electricity generated by
nuclear means is the capital cost, which
accounts for 70% of the total cost. Point Nuclear
number one is therefore that the capital
cost is the fundamental parameter in the Coal
competitiveness of nuclear power plants.
Point number two is the relative
Natural gas . .......
importance of the construction time B in
the cost of capital. In other words, the
cost of capital changes significantly o 5 10 15
from B=6 yrs and B=12 yrs. To test this Cost component (cents/kWh)
dependence, one can readily use the
model presented above changing the
parameter B while maintaining the rest
of the variables constant. The results are 18 18
shown in Figure 7.8. By duplicating the 16
construction time from 6 to 12 years the
total cost increases 33%.
14
12 ....... - -- total cost -- -- 16
14
12
In summary, from 1965 to 1975, 10 capital cost 10
construction times of nuclear power 8 8
plants increased abnormally, from an 6 6
average of 6 years to an average of 12 4 O&M cost 4
years. This rise in construction times
o
2 :-:=~. -=:-:=~.
-=- -=- -= fuel cost . =. -=- 2-=~.:~.. -=~.=~.
o
increased the capital costs of nuclear
power plants by nearly one-third of the
...
en
o 6 8 10 12
(J
original "overnight" capital cost. Since
Construction time (yrs)
70% of the cost of electricity generated
by a nuclear power plant is due to the up
front capital cost, these delays increased Figure 7.8 Cost increase due to longer construction times
in a nuclear power plant
the underlying cost of these assets by at
least 30%. Such an abrupt increase in
cost simply could not be afforded by electrical utilities.
The original promise of economies of scale could not be fulfilled. Instead of having lower unit costs (in
$/kWh) the electricity producers were having higher unit costs. Utilities virtually had to manage large
complex and challenging generating assets to have input costs that were .above market value. It was not a
sound investment at all, i.e. economies of scale affected O&M and fuel costs which were the smaller portion
of total costs.
WHAT WENT WRONG?

Certainly the participants did not want or encourage long construction times. It was most likely an
unexpected outcome of the construction process. The question is what caused this sharp increase in
construction times.
75
In order to answer this question it is

14.1
useful first to see whether this was a
11.6
general pattern of the nuclear industry or
it was a manifestation of the U.S. industry 8.9
only. Figure 7.9 summarizes average

5.91
i ij I
construction times of nuclear power
3.0
plants in the United States, Germany,
France, Japan and Sweden. These I
countries account for 60% of the 1952.55 1956-59 1960-64 1964-67 .1968-71 1972-75 1976-79 1980-83
worldwide production of nuclear USA

electricity. In particular, France with 60
7.5
i iiI I ii Ii i
reactors and Japan with 45 reactors are
two of the strongest advocates of nuclear
energy in the world.
As Figure 7.9 shows, the average times 1964-67 1968-71 1972-75 1976-79 1980-83
1964-67 1968-71 1972.75 1976-79 1980-83
required for construction skyrocketed
JAPAN FRANCE
only in the United States, for projects
starting around 1965. The rest of the 8.5 8.5 8.5
countries had minor increases or managed
iiIi I iill
to reduce their construction times. In 4.0
Japan, reactors built in 1980-83 were built
in 5 years. In France construction times
were slightly higher, between 6 and 7 1968-71 1972-75 1976-79 1980-83 1964-67 1968-71 1972-75 1976-79 1980-83
years.
SWEDEN GERMANY
The conclusion is therefore that whatever
caused the slowdown of construction
Figure 7.9 A verage construction times in selected
times was inherent to the U.S. nuclear
countries versus the year when the project was half-done
industry only.
There are a number of hypothesis as to why nuclear power plant construction in the United States took such
long times. The fITst is the intricacy of the regulatory process, which introduced innumerable delays to
projects. As it was mentioned in previous chapters, the U.S. nuclear regulatory body had a hard time
processing the burst of reactor orders that occurred between 1965 and 1974 (see Figure 7.1). The second
hypothesis is that the long construction times were caused by the large number of participants in the
construction process and their relative inexperience in building power plants. As it was mentioned in chapter
6, the main participants on the design & construction process were: (a) the electrical utility, (b) the architect-
engineer, (c) the reactor system vendor, (d) the turbine generator manufacturer and (e) the construction
company.
In'the United States, there were 60 utilities operating 115 nuclear power plants, 15 architect- engineers and a
similar number of construction firms, 4 reactor system vendors and 2 turbine generator manufacturers. The
possible permutations of that many players are utterly immense. It is similarly huge the complexity of the
construction process and the organizational challenge of undertaking an extremely technically difficult
project, such as a nuclear power plant, with so many participants on board. Notice that the process is
progressive since as more reactors are built, more participants are attracted, there are more firms involved
and ultimately construction times increase. That would explain why in the beginning (before 1965) reactors
were built in the United States in less than 5 years. At that time there were few firms involved.
Compare this situation with France, which has only 1 utility operating 60 reactors, Electricite de France,
which is also the architect-engineer of plants. There is 1 reactor vendor, Framatome, and 1 turbine generator
manufacturer, Alshom. Obviously the degree of organizational complexity is much lower in France than it is
in the United States.
76
It is important to note that the first hypothesis

is somehow included in the second hypothesis.
Utility
Namely, if the construction process in the U.S.
was chaotic because of the large number of RPV Architect- / \
"
participants then one can expect that the whole Intemals engineer --- Constructor
licensing process would be difficult and long. vendor
After all it would be in the NRC' s best interest
RPV
~ /
Reactor Turbine
to detect potential sources of problems arising vendor -- Vendor
vendor
from messy organizational structures.
However, the first hypothesis is really difficult
to prove. It is not easy to show that reactors
r' core
vendor
/
with longer construction times faced more
regulatory difficulties than reactors that were
built faster. For these reasons, we shall focus Figure 7.10 Critical interactions during the design
only on testing the second hypothesis, i.e. firm and construction of a nuclear power plant
diversity.
To test the influence of the number of firms on total construction times, using a statistical analysis tool
(SPSS), some "diversity" parameters have to be defined. These parameters evaluate diversity by pairs, and
refer to the most critical interactions that are expected to occur during the design & construction process.
These interactions are summarized in Figure 7.10.
AExCIV: Architect-engineer and civil constructor (=0 if are the same, .=2 otherwise)
RSxTUR: Reactor system vendor and turbine vendor (=0 if are the same, =2 otherwise)
UTlxAE: Utility and architect-engineer (=0 if are the same, =2 otherwise)
UTlxCIV: Utility and civil constructor (=0 if are the same, =2 otherwise)
RSxRPV: Reactor system vendor and RPV vendor (=0 if are the same, =2 otherwise)
TURxSTR: Turbine vendor and steam riser vendor (=0 if are the same, =2 otherwise)
TURxAE: Turbine vendor and architect engineer (=0 if are the same, =2 otherwise)
RSxAE: Reactor system vendor and architect-engineer (=0 if are the same, =2 otherwise)
st
RSxCORE: Reactor system vendor and 1 core vendor (=0 if are the same, =2 otherwise)
RSxINTER: Reactor system vendor and RPV internals (=0 if are the same, =2 otherwise)
DIVERSE: Average of all previous variables (=0 lowest diversity, =2 highest)
The correlation matrix between these parameters is shown in Table 7.2 for worldwide data corresponding to
279 reactors built before 1989 (189 of them PWRs). For clarity, the main results involving the correlation
between diversity parameters and construction length are summarized below, organized by country:
0:: W 0::
0:: > I- 0:: w w
> ::> a.. en w o I- en
C3 0:: () Z 0::
x ~ X ~ ~ X X w
w en en ::> en en en >
« 0:: 0:: I- 0:: 0:: 0:: is
USA .081 ..029 -.295* -.229* .080 .131 .052 .8 -.077 -.154* -.031*
JAPAN -.548* -.165* .8 8 .128 -.206* -.242* -.223* -.056 -.205* -.337*
SWEDEN -.123 -.070 -.031 -.186* .119 .026 .093 .078 -.258* -.163* -.666*
BELGIUM -.117 -.083 -.038 -.170* .093 .011 .087 .072 -.248* -.155* -.842*
ALL -.119 -.086 -.040 -.171* .098 .007 .085 .071 -.236* -.147* -.163*
* Correlation is significant at the O.05 level (1-tailed). a cannot compute a result because one of the two parameters is constant
77
First, let us focus on the United States. The results show that there is no statistically significant correlation
between DIVERSE (which accounts for overall diversity) and the construction length of the projects. This
might be indicating that the number of firms by itself does not make a project messy. It is the way the
participants orchestrate themselves that makes a project a big success or a failure. One can argue that Japan
had a very organized and well-managed system for delivering nuclear power plants on schedule, while the
United States had a less efficient system; not because there were many players on board but because they
were badly organized.
However, the results show some parameters that are statistically significant and are negatively correlated
with construction times. The first, UTIXCIV, accounts for the relationship between the construction firm and
the electrical utility. Projects where the constructor was the utility had substantially shorter construction
times. This finding might be indicating that one of the bottlenecks of the construction process was in the
construction firm and how it interacted with its two immediate partners: the architect-engineer and the
electrical utility. The other parameters that are negatively correlated are RSXCORE and RSXINTER. These
parameters account for the relation between the reactor system vendor and two of its main suppliers: the 1st
core supplier and the core intemals manufacturer. Since the correlation is negative with construction length it
means that projects where the reactor vendor was also the supplier had substantially shorter construction
times. This finding might be showing that in the U.S. most delays in the construction process were
introduced by late deliveries of these components, which are in the critical path of the construction of the
reactor. For instance, without a timely delivery of the reactor internals, a number of extremely time-
consuming tasks can not be done on time.
Let us focus now Japan, Sweden and Belgium. Note that, as it was mentioned before, these nations managed
to have lower construction costs and -arguably- a more organized construction and delivery process. The
matrix correlation shows a statistically significant correlation between construction time and most of the
diversity variables -including DIVERSE. Again, since the correlation is negative with construction length it
means that projects where there were many participants involved had substantially longer construction times.
In the particular case of Japan, for instance, there is a strongly negative correlation between construction
length and AEXCN, which means that projects where the architect-engineer and the civil constructor were
the same company, had shorter construction times. The same conclusion holds for RSXTUR, TURXAE and
TURXSTR, namely, when the turbine generator manufacturer was also the reactor vendor, the architect-
engineer and the steam-rising supplier, then the project was more likely to finish on time.
France is not included in the correlation matrix because most of the parameters were constant. In other
words, all the French reactors were built using exactly the same formula -as was pointed out- and therefore it
is impossible to make a correlation. However it is important to note that France has a good record of timely
construction as it is shown below:
Number Average Number Average
of BWRs Construction of PWRs Construction
built time (yrs) built time (yrs)
JAPAN 19 5.9 16 4.7
SWITZ. 2 7.0 3 4.7
FRANCE 48 6.5
GERMANY 7 8.1 13 6.5
BELGIUM 7 6.7
KOREA 8 6.9
SWEDEN 9 5.4 3 7.3
SPAIN 2 7.5 7 8.6
USA 47 7.7 82 8.7
It is important to note that there is no significant difference between the PWR and other competitive designs
regarding construction delays. The problems associated with mismanagement of nuclear power projects in
the United States affected both BWR and PWR projects.
78
Table 7.1. Summary of AEG's Research and Development costs from 1948 to 1974
79
Table 7.2 Correlation matrix between diversity parameters. Data valid for 279 reactors, 189 of them
PWR, built worldwide before 1989
Chrrelations
'~' •••. "''''-'''~"' ,~,";.": •.',.,;,.;.;,, •• .,;,..;.~ .. .:..." ~ .. ~u,,~~ ",.~,;.;;;~::;~ ~;.:.;..:;~;;.;'; ";::::~~;;::::~"Z,
:::.;~~::;~-::::;:;;:.
:::::'~;;';:;:;;.:;~"~lriictioit~
:~ ~~~xr.rn:~] :tiri~1\7;f~i :~:~ ;~ ~~1.~;Wi1 ~% g~}~~
'Alxcw:.':". ;:::.:.- J.OOO .J17 -.141' .144 .005 .056 ...1ij. ...123' .199'l .274 .174~ ...119
~::~~~~~
.. .. . " - -'
-.m6
':
-.. ........... 01':><'
.117 1.000 -.161 ...004 ...381' •778 .259' .289' .209' .127 .641Ji
~~~:;~:::.:
..
........... ... -
':'':' ':''''"'':'-~.,;,-~'''''.':-'':'.;..
...16J J.OOO .251 .469' ...141' ...J86 ...J80' ...444 ...

491 .143 ...010
.144' ...004 .251 J.OOO ...091 .006 ...103' .175 .J9(1' .268 -.171
.005 .469' ...091 1.000 ...27/f' ...124' ...

243' -.247 ...111' .131 JJ98
.056 .778' -.341' .006 ...27/f' J.OOO .115' .070 .19~ .1Jl' .573 JJ07
...18.1' .259' ...186 ...101' ....124' .115' 1.000 .855 -.331 ...198' .146 JJ85
...123' .289' ...J80' ...

243' .070 .855 1.0()() ...J94' JJ71
............ -
RSXCffiE:::::::::: ~
..""'_ .._-" .. .199' .2fJ9' ...
444' •J75 ...247' .J95' -.311 ...311' JJJOO .653' ...058 -.236
'~:::.:.~-::':..:,":':::::';.:;
-'_\..I,Oo~' _"_"'I_'<JoUoo.-'"
'.
......
......... ,-_ .. _~.....-;.o ... ......, .......... ~
.274' ...491 .J9(1' ...Jl1' .1J2' ...198' ...194' .65P 1.000 .049
~XlNrER ~=:..::_
......;.~-~""':
DIVERiEr::,::'::::'.~ -
-, ;..,: -..... .174 .14~ .168' .131 •571' .346 .198' ...058 .019 1.000 ...019
.~~"': " .... _''''''':''':''~ '':' ~ .~~
~~jr)~==i=;
~!1i:t.~~(~!!.)~~):;..: _
...119 ...m6 ...010 ...171 .098 .007 .m5 .071 ...147 ...049 1.000
*. CmeiatiOl is sig,1i6cat at theO.05IC\e1 (l-lailed) .

... Carelabcn is sigi 6cmt ;l tIe O.01IC\e1( l-tliled)
80
ChapterB
THE 1990S AND BEYOND: LESSONS FROM THE PWR CASE
This chapter is a rapid fast-forward through the 1990s. It explores the achievement of the long-awaited
positive externalities of a dominant design. It shows how the performance of nuclear power plants has
increased notably, through a combination of industry consolidation, specialization, cumulative learning
effects, and economies of scale. The chapter finishes with a summary of the Westinghouse PWR case and
a framework for understanding the issues from the point of view of public policy and management of
technology .
CONSOLIDATION OF NUCLEAR OPERATORS

In 1990 there were 112 nuclear power plants under commercial operation in the United States, which
were operated by 54 electrical utilities. This means that a typical electrical utility owning nuclear
electricity-generating assets would on average own only two nuclear reactor units. The extreme
fragmentation among nuclear power operators before 1990, was a consequence of the overconfidence that
prevailed among utility managers at the time when most plants were constructed. For them, a nuclear
power plant was just a large electricity-generating asset not very different than standard 'plants fueled by
coal and natural gas. For the most part, the unique organizational and technical skills required to operate
nuclear plants efficiently were greatly underestimated at that time (Kadak, 2002).
During the 1990s, things started to change.
Under a new environment marked by Top 4 firms "-
Average number of reactor ~
deregulation and competition, electrical units per electrical utility
utilities reevaluated the value of nuclear
assets. Some firms decided to get rid of
nuclear power plants while others actively
pursued the acquisition of these power units.
As a result, the structure of nuclear power
operators changed significantly.
As it is shown in Figure 8.1, the number of
firms was reduced from 54 in 1990 to 24 in
1,990 1,995 2,000
the year 2000. The average ownership raised
from 2 reactors per nuclear operator to 4
reactors. In 2000, the top four firms owned
on average 11 reactors, with some owning as
many as 17 units.
8. The 1990s and beyond: lessons from the Westinghouse PWR case
The process of consolidation within the nuclear power industry occurred primarily through two parallel
processes: (a) mergers and acquisitions between electrical utilities, and (b) purchases of nuclear power
plants. The first process -mergers and acquisitions between different firms- increased momentum
between 1998 and 2001 and the outcome is summarized below (including Figure 8.2):
AJ/iant Energy (1)

Northern States (3) }NUMCO(7J Dominion
Wisconsin Electric (3) 6%
Virginia Power (4)
Northeast Utilities (2) } Dominion (6)
Duquesne (2)
First Energy (2) } First Energy (4)
Florida P&L (5)
Entergy (6) Entergy/FPL (6)
New York Power (2)
GPU (2)
CommEd-Unicom (11) Exelon (17) 1998 2001
PECD (4)
Figure 8.2. Consolidation of U.S. electrical utilities owning
nuclear power plants between 1998 and 2001~
Note: Number of reactors between • In 1998, 4-Firm ratio was 30 %
parenthesis • In 2001, 4-Firm ratio was 45 %
The process of mergers and acquisitions between nuclear power operators is not an isolated story. It has
occurred at the same time energy markets in general have been restructured. Vertically integrated utilities
have been broken up. A new-generation industry has sprung up with new players and business dynamics.
In the nuclear arena, mergers and acquisitions led to multiple plants among fewer owners. Most
companies that own nuclear plants have nuclear as their core business. The emerging nuclear operating
companies are focused on the management and oversight of their operating plants with increased
economies of scales and resource sharing among a larger plant base. In brief, nuclear consolidation seems
to occur in conjunction with specialization.
In particular, Exelon Corporation is
building a case for nuclear power. Table 8.1. Between 1998 and 2001, fifteen U.S. nuclear power
Exelon currently represents nearly plants were acquired by other electrical utilities.
one-fifth of the U.S. nuclear power • In 1998-99, 4 reactors sold for $404 million.
capacity, with 17 reactors at 10 sites. • In 2000101, 9 reactors sold for $3. 7.billion
Exelon's strategy seemingly is to be
ready with a plan when nuclear
projects become viable, and in the
13x819 $100m
meantime enj oy the benefits of
market power and economies of 8x650 $10m
scale. For instance, since Exelon 26x933 $182m
spends about $600 million a year on
Pilgrim 1 11x655 $112m
uranium fuel its buying power is
significantly higher than other Indian Point 3 15x965 $967m
FitzPatrick 14x816
smaller buyers. As a result, better
Nine Mile Point 1 2J
8x613:27x1143$824m
deals can be brokered and fuel cost
reductions can be achieved. Savings 10x986;12x98E $602m
are also realized by spreading O&M 14x870: $1305m
and overhead costs across the 25x1154
company, and using the same
contractors during refueling outages
82
and maintenance activities. Thus while the industry average down time for refueling is about 32 days,
Exelon units were down an average of 22 during 2000. The same year, Exelon's nuclear plants had a
capacity factor of 93.8 percent compared to an industry average of 89.6.
Regarding the second process leading to nuclear power consolidation, i.e. the purchasing of electricity-
generating assets, it is important to note that a few years ago nuclear power plants were seen as expensive
liabilities. But today their perceived value has increased. As shown in Table 8.1, facilities designed for
shutdown have been recently sold to the highest bidder. Three Mile Island 1, the sister unit to Three Mile
Island 2, the reactor that had an accident in 1979, started the path when it was sold in 1998 for $100
million. The sale of Clinton 1 followed course when the reactor was purchased form Illinois Power Co.
for $182 million. The latter cost illinois Power $4.2 billion to build.
The early purchases of nuclear plants -between 1998 and 1999- were cheap compared to more recent
prices of nuclear electricity-generating assets. The reason why these early sales were less expensive is
because some companies were looking to lean their assets in order to be competitive -in a deregulated
environment- and nuclear power was not their core business. Companies that were actively looking to be
in the nuclear business benefited from this situation and bought assets at bargain prices, i.e. Exelon,
Entergy, NumCo. Once market forces started to act, competition increased and assets began to be sold for
higher prices. Hence, when in 2000 Entergy bought Indian Point 3 and Fitz-Patrick nuclear plants from
New York Power Authority, it had to pay $967 million. Again, the surge of sales in the northeast reflects
the relatively pace of deregulation in states like New York, Massachusetts, Connecticut and Pennsylvania.
To illustrate the point of
increasing bidding prices,
1,400
Figure 8.3 provides a summary
of purchases versus the -
~
~
underlying value of the
~ 1,000
electricity-generating asset. The Q)
later is plotted in the x-axis and o
"i:
accounts for the product of C. 600 1998/99
useful life times the electricity c:
o
gener~ting capacity. This in E
theory is the maximum amount .~ 200
:J
of energy (in kilowatt-hour) that c-
a nuclear operator can extract o
from an asset. One can consider
<
10,000 20,000 30,000 40,000
this number as something "in
theory" because the NRC is Useful Life (rv1W x year)
granting operational extensions
that can extend the useful life 10 Figure 8.3. Between 1998 and 2001, the acquisition price of nuclear
years and even 20 years in some power plants increased significantly.
particular cases. However these • In 1998/99, $400million for 10,000 megawatt-year.
extensions are analyzed on a • In 2000/01, $150million for 10,000 megawatt-year.
case-by-case basis and cannot
be taken for granted.
Additional nuclear plants are expected to be up for sale in the next years as more utilities divest their
electricity-generating assets in response to restructuring initiatives in several states that remain heavily
regulated.
83
B. The 1990s and beyond: lessons from the Westinghouse PWR case
IMPROVED OPERATING PERFORMANCE

There are a number of ways to measure
the operating performance of a nuclear 90
Reactors in operation
power plant. In what follows, some of the
All reactors
most common performance factors are ~
~ 80
presented, in order to show that the 1990s
were indeed years of uninterrupted '-
o
improvements. t) 70
cu
LL
Figure 8.5 shows an increase of the
capacity factor of nuclear power plants ~(,) 60
from around 60% in the early 1980s to cu
Co ,
90% in the late 1990s. Capacity factor is cu o
(
I
defined as the yearly ratio of electricity u 50
that a nuclear power plant actually
80 82 84 86 88 90 92 94 96 98
supplies to the electrical grid over Year
installed capacity. Note that according to
this definition a plant with a capacity Figure B.5. Between 19BO and 2000 there has been a
factor of 100% is operating at full power dramatic increase of the capacity factor among U. S.
for 365 days a year. Most of the increase nuclear power plants
in capacity factors is certainly due to
100
simple learning effects arising from the -r-- -
more experienced operating crews at U.S. 80 r-- - f--
power plants, but a significant portion is - I-- f-- f--
due to consolidation and specialization of fJ) 60 - t-- I-- ~ -

~ ......
nuclear power operators cu - - ....,..
C
40 t--
I--
t--
I--
I--
I--
I-- -
f--
I--
0_ -
-:--r Average
....:...
Figure 8.6 shows another plant t-- t-- I-- t-- ..:... I-- .:.. - - - > Median
performance parameter, the refueling 20 l- I- l- I- - t-- - .....:. ~
outage time, which is intimately related to t-- t-- I-- I-- ....:., I- ~ - -
capacity factor. Every 18 months, more or o I I I I I
less, nuclear power plants need to 90 92 94 96 98

shutdown for refueling. In the meantime a Year
number of critical maintenance activities Figure B.6. Between 1990 and 2000 there has been a
are conducted. Since the plant does not reduction of refueling outage times
produce electricity during refueling, it is
critical to speed up this process. Figure
INDUSTRIAL SAFETY ACCIDENT RATE
8.6 shows that this goal was achieved and 1/1 2.1
there has been a significant reduction ~
~
~ 0
during the 1990s. Cl)J:.
a.~ 1.5
The achievement of' higher operative ...c: ~~
1/1 Cl) 1.34
performance within U.S. nuclear power Cl) 0
1.03
i
plants was not constrained to parameters :2 ::
uo 77
that are intimately related with higher Uo
<s:o
profits. Figure 8.7 shows that the o
o IJO' 0.46
industrial safety accident rate also
decreased substantially during the last two
decades. This indicates better
N
80 84 88 90 92
~J 94 96 98 99 2000
management and more experienced
crews. Figure B.7. Reduction of the rate of accidents due to
industrialreasons (not necessarily nuclear-related incidents)
84
SUMMARY OF THE 19905

The dramatic improvements in performance in the 1990s can be explained in the following terms. First,
consolidation of nuclear operators and increasing specialization necessarily lead to higher efficiency and
profits. This statement is schematically illustrated in Figure 8.4. Consolidation and specialization lead to
higher efficiency for a number of reasons:
1. Economies of scale stemming from the shared use of infrastructure and O&M resources among a
larger base of nuclear power plants.
2. More efficient use of human resources, including having an in-house critical mass in technical areas
and a management team more specialized in managing nuclear power plants.
3. Learning effects arising from cumulative experience operating a broader base of reactors.
4. More buying power regarding suppliers such as uranium producers, service organizations, etc.
5. More lobbying power over critical agencies in the nuclear energy sector such as NRC, DOE and EPA.
Figure 8.4 shows a loop that is positively reinforced. Once higher efficiency is achieved, profits also
increase and the management becomes more committed to pursue even more specialization and
consolidation. Empirical data shows that in the 1990s the reinforcing loop has been functioning very well.
It is important to note that the reinforced loop
explained above did not appear in the 1970s and
1980s. Indeed, some scholars found no evidence
that a learning process was going on in nuclear
power plants. The question is therefore what
changed in the 1990s that led to substantial
learning and scale effects. One can argue that the
answer is embedded in the process of
consolidation and specialization of nuclear
operators. The fact that nuclear operators
became more focused on managing several
nuclear plants with more specialized
management and operation teams, certainly had
a role on the overall performance of the industry.
Figure 8.4. Schematic representation of the effect of
ill summary, The 1990s finally brought the consolidation among nuclear operators and their
promises of economies of scale and large profit increasing nuclear specialization
margins that the industry analysts had foreseen
forty years before. As it is shown in Figure 8.8,
nuclear power actually increased in the 1990s by
30% while the number of nuclear power plants
was reduced from 112 to 103 reactors. Had the
efficiency of existing plants not increased,
almost 30 new reactors would have been needed
to produce the same amount of electricity. The
question is why the nuclear industry did not
achieve the same results in the mid-1970s when
the demand of power plants was in its peak?
1990 1994 1998 1999 2000
Why did Westinghouse, who had the successful
PWR in the marketplace, lose the grip on the
technology and the market? These questions are Figure 8.8. In the 1990s electricity produced by
U. S. nuclear power plants increased roughly 30%
answered in the following section.
85
LESSONS FROM THE WESTINGHOUSE PWR CASE

There are two possible ways to analyze the PWR case. One is from the point of view of public policy,
addressing the issue of the use of law and public funds to promote the nuclear power industry in general
and the PWR in particular. The other is the point of view of private policy, addressing the issue of how
Westinghouse, benefited from a particular regulatory and competitive environment to develop and market
a product. The conclusions stated subsequently address the latter case, namely, the analysis of strategies
that Westinghouse carried out to sell its successful PWR product in the marketplace.
In order to address the issue in order, the PWR story is divided in two phases. The first phase is the
process of value creation and goes from 1945 to 1960, when Westinghouse was able to develop a
promising product from scratch. The second phase is the process of value capture and goes from 1960 to
1975, when Westinghouse was able to sell a large number of units in the United States.
First phase
In the late 1940s the U.S. government wanted to develop nuclear power plants as soon as possible. As it
was explained in Chapter 1, the main reasons for the rush for building power plants were the externalities
associated with military programs, such as the massive production of weapons grade plutonium, and the
need to capture the value created by this innovation ahead of other nations. In addition, getting ahead of
the Soviet Union in terms of nuclear technology was a national objective.
As was shown in Chapter 2, Westinghouse benefited from the business opportunity and -using
government funding- developed an outstanding product, the PWR. A number of other products -Magnox,
PHWR, BWR, LWGR, FBR- were developed at the same time but none of them became more successful
than the PWR.
The PWR evolved in the right direction because it was intended for the propulsion system of a nuclear
submarine. This turned out to be a fortunate factor for three reasons: (a) the nuclear sub was a challenging
product that requested a reliable design, (b) the project had the highest government priority, and (c) the
rush allowed Westinghouse engineers to have a rapid learning process.
The nuclear submarine was a small-scale power unit with all the complexities of a nuclear power plant
plus the severe requirements of underwater operation. For instance, engineers had to design a new type of
fuel, which had higher strength and turned out to be better than fuels designed previously. On can argue
that Westinghouse made the right decision in committing to this project because the project turned out to
have a higher potential than other projects sponsored by the ABC. The difficulties of designing a
challenging reactor created an environment favorable for product innovation.
Thanks to the nuclear submarine project, a remarkable group of engineers and scientists was formed at
the Bettis Laboratories. This group worked under the leadership of Rickover and the joint umbrella of
ABC, Naval Reactors and Westinghouse. The same people worked on a platform of products using the
same technology, such as the Mark-l prototype, the Nautilus and the eventually the first PWR plant -
Shippingport. The continuity of the Bettis group provided an excellent opportunity for collective learning
and rapid technical improvement.
The U.S. government recognized the value of the Bettis group and selected the PWR design to be the first
U.S. plant. The government wanted to have a plant operating as soon as possible and Westinghouse and
its Bettis Laboratory already had a workable prototype and a successful submarine based on the PWR
concept. Additionally, the Bettis team had a record of achievements and it seemed a sound decision to
hand a complex project over to one of the best nuclear engineering teams available in the U.S.
In summary, Westinghouse made a number of correct decisions during the first phase of the development
of nuclear power plants, when product innovation rate was much higher than process innovation. This
early phase is summarized in Figure 8.9, using the framework of innovation patterns introduced in
Chapter 3 (Utterback, 1994).
86
The rate of product innovation is

highest during the "fluid" first phase,
which for nuclear plants occurred c:
o
approximately between 1945 and ~ Product Innovation
nJ
1960. This is a period where product >
o
c: k: Process Innovation
experimentation occurs both in terms c:
of testing different options and gaining L.
experience. Westinghouse took the .~o

most challenging product -the nuclear :E
submarine- and invested time and '0
effort to create a first class engineering ...ns
G,)
staff -the Bettis group. By 1960 ~

Westinghouse was in a great 1 45 1960
competitive position: it had created a
1st Phase 2nd Phase 3rd Phase
dominant design that served for
submarines and power plants, and it
had gathered substantial construction Figure 8.9 The dynamics of innovation (Utterback, 1994)
expenence ..
Second phase
As shown in Figure 8.9, during the first "fluid" period of high product innovation, little attention is given
to the processes by which products are made so the rate of process innovation is less rapid. There is a
second period, however, where things are reversed. This is the "transitional" second phase where the rate
of product innovation slows down and the rate of process innovation takes the lead. That occurred with
the PWR design after 1960. Westinghouse became more focused on cost, volume and capacity. Process
innovation became the preferred way to introduce changes and these were just incremental innovations on
the dominant PWR design. The manufacturing progressed from heavy reliance on R&D to more
systematic and repetitive engineering tasks. The organizations in charge of building reactors changed
from the small, cohesive and creative group at Bettis Lab toward a more hierarchical organization with
defined tasks and procedures. Engineers became more focused on improving the performance of the
design or at least what was considered performance of a nuclear power plant at that time.
After securing the first deals in the first half of the 1960s -by means of turnkey projects and the use of
government subsidies- Westinghouse was ready for a second phase where the main goal was to capture
value from the innovation. In other words, to make profits out of the PWR product.
As shown in the scheme of Figure 8.10,
Westinghouse chose to focus on its role Westinghouse
A
of supplier of two of the main
components of a nuclear power plant: "'
the nuclear system and the turbine
generator. Westinghouse did not seek a
high level of vertical integration by A rchitect-
entering in the areas of balance of plant engineer
and civil construction. On the one hand
these areas were not part of
Westinghouse's core competencies. On
the other, most managers seemed to act
as if the demand for reactor systems and
turbines was infinite and left the rest of
the proj ect in the hands of Figure 8.10. Components of the design & construction
utilities, architect-engineers and
87
construction firms. Westinghouse adopted this strategy probably because it was the familiar way. As was
shown in Chapter 5, Westinghouse had a history of selling large turbine generators for conventional
power plants. In this segment of the market Westinghouse had been a pure player, with no involvement in
other areas of the plant. Therefore, when the first nuclear power plants started to be built, the extension of
this role of pure supplier was natural. However, to dodge the complexities of the construction process and
to act as if the demand was infinite had a downside: long construction times that jeopardized the
competitiveness of nuclear power plants.
The PWR is probably one of the few dominant designs that emerged because of the role of suppliers, in
this case the supplier of the reactor system - Westinghouse- and the supplier of the turbine generator -also
Westinghouse. As it was shown in Chapter 5, tying these two components together is probably on the
borderline of antitrust infringement. Indeed there was an antitrust case against GE and Westinghouse
regarding large turbine generators. One can argue that this non-competitive strategy was a win-win deal
for both firms that hurt other designs, such as gas-cooled reactors, but by itself it does not explain the
early success of the PWR.
General Electric took the same strategic approach as Westinghouse but it was less successful probably for
three reasons: (a) utility managers did not like to have water coming from the reactor inside the turbine
generator and therefore preferred the PWR over the BWR, (b) the supplier base that Westinghouse was
able to assemble was broader, and (c) from the outset Westinghouse issued licenses to firms developing
the PWR, thus there were a number of manufacturers of PWR technology such as Framatome, Mitsubishi,
Babcock& Wilcox and Combustion-Engineering.
Westinghouse also made the mistake of listening excessively to its customers, the electrical utilities,
which were asking for larger power plants. As it was shown in Chapter 7, in the 1960s and 1970s the
single most important performance parameter for electrical utilities was scale. The larger the unit the
better, as it was assumed that larger sizes implied lower electricity costs (in dollars per kilowatt-hour).
Reactor vendors worked hard to achieve the goal of higher power ratings and were quite successful in the
effort, building power plants with power ratings on the order of 1,300 MWe. However, listening to market
demand and customer needs was exactly the wrong thing to do. The right signals to listen were the
increasingly long times required to construct plants.
Originally in the early 1960s, when utilities jumped into the nuclear bandwagon, they expected
construction times in the order of 5 years with slight increases due to increasing complexity. The first
nuclear power plant in the United States, Shippingport, was built in three years, and in days to come it
was predicted that normal construction times of larger units would take on average five years. But this
prediction totally underestimated the complexity of nuclear plant construction in the United States. Just to
provide an example, it took 17 years to build the Diablo Canyon nuclear power plant in California.
Average construction times increased and peaked in 1975 at around 12 years.
Chapter 7 argues that the long construction times were caused by the large number of participants in the
construction process and their relative inexperience in building power plants. In the United States, there
were 60 utilities operating 115 nuclear power plants, 15 architect- engineers and a similar number of
construction firms, 4 reactor system vendors and 2 turbine generator manufacturers. The possible
permutations of that many players are utterly immense. It is similarly huge the complexity of the
construction process and the organizational challenge of undertaking an extremely technically difficult
project, such as a nuclear power plant, with so many participants on board.
Compare this situation with France, which has only 1 utility operating 60 reactors, Electricite de France,
which is also the architect-engineer of plants. There is 1 reactor vendor, Framatome, and 1 turbine
generator manufacturer, Alshom. Obviously the degree of organizational complexity is much lower in
France than it is in the United States.
88
The rise in construction times increased the capital costs of nuclear power plants by nearly one-third of
the original "overnight" capital cost. Since 70% of the cost of electricity generated by a nuclear power
plant is due to the up front capital cost, these delays increased the underlying cost of these assets by at
least 30%. Such an abrupt increase in cost simply could not be afforded by electrical utilities. Hence, the
original promise of economies of scale could not be fulfilled. Instead of having lower unit costs (in
$/kWh) the electricity producers were having higher unit costs. Utilities virtually had to manage large
complex and challenging generating assets to have input costs that were above market value. It was not a
sound investment at all, i.e. economies of scale affected O&M and fuel costs which were the smaller
portion of total costs (see Figure 8.11).
The fact of the matter is the first priority
of an electrical utility is to have a low • Capital cost Il Fixed O&M cost 0 Fuel cost
degree of financial risk and that goal

was not achieved. Nuclear power plants Nuclear
had a huge financial risk because many
plants were not completed on time and Coal
more than one was cancelled prior to the
opening. Had Westinghouse put 6.64 .
Natural gas
investment security as first priority then
its strategy would have been different,
probably with more control over the o 5 10 15
management of proj ects. Instead of Cost component (cents/kWh)
focusing on the role of supplier
Figure 8. 11 Cost components 0 power plants
Westinghouse should had a more
proactive role in project management.
Westinghouse made the decision to standardize the PWR in order to speed up the delivery of products and
benefit from learning effects. Arguably, that was the right decision at the right time. However, the rest of
the plants were largely not standardized, and were built on a one-of-a-kind basis. This lack of overall
plant standardization jeopardized the PWR because units could not be delivered on time. Construction
times skyrocketed and licensing applications became a mess.
The roles of architect-engineers and construction firms are fundamental in the timely delivery of a nuclear
power plant. Also important is how these firms interact with the utility that will operate the plant.
Nonetheless, the careful analysis and of these interactions were absolutely overlooked by the ABC and
reactor system vendors. In the rush for building power plants, these critical participants did not think
about what turned out to be the Achilles' heel of the U.S. nuclear power industry.
Westinghouse should have been more proactive in having control over project management since it was
the number one stakeholder in the PWR technology. An architect-engineer such as Bechtel can be the
prime contractor of coal-frred plants, combined cycle plants, etc; and therefore not be committed to a
particular design such as the PWR reactor. As a lesson for Westinghouse, sometimes suppliers of critical
components need to look for more control of the project, especially when there are many participants and
none of them is particularly engaged with the technology.
Electrical utilities underestimated the difficulties of building and operating a nuclear power plant. They
rushed to buy these fancy electricity-generating assets but could not manage the process efficiently. Three
decades later, a large consolidated operator like Exelon has learned to benefit from the economies of scale
possible in large nuclear power plants. Regulated electricity markets did not provide the opportunity for
the emergence of a nuclear operator such as Exelon in the 1960s. Only the deregulation environment of
the 1990s allowed this to occur.
Policymakers were naYve regarding the partnership of electrical utilities with the ABC to build nuclear
power plants. Given the heavily regulated environment, the creation of large government-owned and
89
largely nuclear-specialized utility would have been the best choice. The example of EdF in France is
worthwhile to mention here. EdF managed to build 60 nuclear power plants with construction times on
the order of 6 years (half the time most U.S. utilities spent building plants back in the 1970s).
90
Bibliography
NUCLEAR HISTORY
Beaver W. 1990. Nuclear power goes on-line: a historyl of Shippingport. New York: Greenwood Press
Cruickshank A, ed. 1989. World Nuclear IndustlY Handbook. Surry, UK: Reed Business Publishing
Dawson FG. 1976. Nuclear power: development and management of a technology. Seattle: University of
Washington Press.
Del Sesto 8L. 1979. Science, politics, and controversy: civilian nuclear power in the United States,
1946-1974. Boulder, Colo.: Westview Press
Duffy RJ. 1997. Nuclear politics in America : A history and theory ofgovernment regulation. Lawrence,
Kan.: University Press of Kansas
Hatch MT. 1986. Politics and nuclear power: Energy policy in Western Europe. Lexington, Kentucky: The
University Press of Kentucky
Hecht G. 1998. The radiance of France : Nuclear power and national identity after World War II
Cambridge, Mass.: MIT Press
Kadak, A, 2002 Personal communication (March 2002). Massachusetts Institute of Technology
Lahey RT, 2000. Personal communication (June 2000). Rensselaer Politechnic Institute
McCaffrey DP. 1991. The politics of nuclear power: a history of the Shoreham Nuclear Power Plant.
Dordrecht ; Boston: Kluwer Academic Publishers
McDougall WA. 1997. The heavens and the earth: Apolitical histmy of the space age. Baltimore and
London: The Johns Hopkins University Press
Moreland J. 1968. Current status and future technical and economic potential of light water reactors.
Washington: U. S. Govt. Print. Off.
Patterson WC. 1985. Going critical. London: Paladin
Pringle LP. 1979. Nuclear power: from physics to politics. New York: Macmillan
8uid LH. 1990. The Army's nuclear power program: the evolution of a support agency. New York:
Greenwood Press,
Bibliography
Sweet W. 1988. The nuclear age: atomic energy, proliferation, and the arms race. Washington, D.C.:
Congressional Quarterly
Termuehlen H. 2001. 100 years of power plant development: focus on steam and gas turbines as prime
movers. New York.: ASME Press
U.S. Congress 1975. H.R. 8631, to amend and extend the Price-Anderson act: hearings before the Joint
Committee on Atomic Energy,Congress of the United States, Ninety-fourth Congress, fITst
session, on ... September 23 and 24, 1975. In United States. Congress. Joint Committee on
Atomic Energy. Washington: U.S. Govt. Print. Off.
USGPO 1981. Nuclear Power Plant standardization: light water reactors., Congress of the U.S., Office of
Technology Assessment. For sale by the Supt. of Docs., U.S. G.P.O." Washington, D.C.
Walker ML. 1983. Nuclear Power Stnlggles: Industrial Competition and Proliferation Control. London,
Boston, Sydney: George Allen & Unwin
PRICE-FIXING
Fuller JG. 1962. The gentlemen conspirators; the story of the price-fixers in the electrical industry. New
York: Grove Press
Gilbert G. 1996. The Heavy Electrical Equipment Antitnlst cases: price-fixing techniques and
rationalizations. In Corporate and governmental deviance: Problems of organizational
behavior in contemporary society, ed. S Edwin. New York: Oxford University Press
Kaye. 1990. Sholer's Antitnlst Deskbook
Lean DE 1982. Competition and collusion in electrical equipment markets :An economic assessment.
Washington, D.C.: Bureau of Economics, Federal Trade Commission
Ohio Electric vs. General Electric 1965. District Court of New York. In Defendants Pre-trial Brief and other
parts: United States District Court of New York Ohio Valley Electric vs. General Electric
Civil Action 62 Civ 965 Second US
Porter M.E. 1986. General electric v.s. Westinghouse in large turbine generators (A). In Porter, M.E. &
Ghemawat, P.: Harvard Business School
Sultan RGM. 1975. Pricing in the Electrical Oligopoly, Volumes I and II,. Cambridge, Mass: Division of
Research, Harvard Graduate School of Business Administration,
92
Bibliography
Sutherland EH. 1996. Corporate and governmental deviance: problems of organizational behavior in
contemporary society / The heavy electrical equipment antitrust cases: price-fixing
techniques and rationalizations. New York: Oxford University Press
USGPO 1980. Steam-electric plant construction cost and annual production expenses /, The Supt. of Docs.,
U.S. G.P.O. The Administrator: The Supt. of Docs., U.S. G.P.O, Washington,D.C.
TECHNOLOGY AND INNOVATION
Abernathy WJ, & Utterback, l M. 1978. Patterns of industrial innovation. Technology Review, 80: 40-7
Argote LE, D. 1990 Learning curves in manufacturing. Science, 247: 1990
Christensen CM. 1997. The innovator's dilemma: When new technologies cause great firms to fail. Boston,
MA: Harvard Business School Press
Christensen CM, Suarez, F. F., & Utterback, lM. 1998. Strategies for survival in fast-changing industries.
Management Science 44: s207 -s20
Cooper AS, C. 1992. How establishedfirms respond to threatening technologies. The Academy of
Management Executive 6
Foster R. 1986. The s-curve: A new forecasting tool. In Innovation: The Attacker's Advantage, ed. R Foster,
pp. P.88-111. NY: Summit Books, Simon and Schuster,
Ghemawat, P. 1985. Building strategy on the experience curve. Harvard Business Review March-April,:
143-9.
Gould Sl 1987. The Panda's Thumb ofTee/mology. Natural History
Henderson RM, & Clark, K.B. 1990. ArchiteCtural innovation: The reconfiguration of existing product
technologies and thefailure of established firms. Administrative Science Quarterly, 35.:
P.9-P30
Joskow PL. 1976. Commercial impossibility, the uranium market, and the Westinghouse case. Cambridge,
Mass.: M.I.T. Dept. of Economics: M.LT. Dept. of Economics
Kuhn T. 1962. The Structure of Scientific Revolutions. Chicago: The University of Chicago Press
Porter M.E. 1991. Towards a dynamic theory of strategy. Strategic Management Joumal12: P.95-117.
Tushman L. PAM. 1991. Managing through cycles of technological change. Research/Technology

Management, May/June: 26-31
93
Bibliography
Rosenbloom R, &'R. Burgelman. 1989. Technology strategy: An evolutionary process perspective.

Research on Technological Innocation, Management and Policy 4
Shaw V, & Shaw, C.T. 1998. Conflict between engineers and marketers. Industrial Marketing Management
27: p.279-91.
Shy O. 1998. Industrial organization: theory and applications. Cambridge, MA.: The MIT Press
Teece DJ. 1986. Profiting from technological innovation: Implications for integration, collaboration,
licensing and public policy. Research Policy, 15,: P.285-305.
Teece DJ. 1987. Capturing valuefrom technological innovation: Integration, strategic partnering, and
licensing decisions. In Technology and Global Industry, ed. NAo Sciences. Washington,
D.C.: Courtesy of the National Academy Press
Tushman MLR, L. 1992. Organizational determinants of technological change: Toward a sociology of

technological evolution. Research in Organizational Behavior, 14: P. 311-47.
Utterback 1M. 1994. Mastering the dynamics of innovation. Boston, MA: Harvard Business School Press
94
~;x:-.,:'" ")
~o ,./..)

Westinghouse PWR Dominant design-MIT-Carlos Barrientos

Uploaded by

Copyright:

Available Formats

Westinghouse PWR Dominant design-MIT-Carlos Barrientos

Uploaded by

Document Information

Original Description:

Original Title

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

Westinghouse PWR Dominant design-MIT-Carlos Barrientos

Uploaded by

Copyright:

Available Formats

WESTINGHOUSE PWR: THE RISE AND FALL OF A DOMINANT DESIGN IN THE

ELECTRIC POWER INDUSTRY

Master in Nuclear Engineering

Submitted to the Engineering Systems Division

Master of Science in Technology and Policy

Massachusetts Institute of Technology

@ 2002 Carlos J. Barrientos. All rights reserved

The author hereby grants to MIT permission to reproduce

Thesis Supervisor: John S. Carroll

THE DAWN OF NUCLEAR ENERGY

PURPOSE OF THE STUDY

MAIN RESEARCH QUESTIONS

SCOPE OF THE STUDY

USA 136 SWITZERLAND 6

1. The early origins of the PWR Design

THE EARLY ORIGINS OF THE PWR DESIGN

FROM 1934 TO 1946

THE QUEST FOR GOVERNMENT CONTROL

THE COLD WAR AND ITS FRAMES OF THOUGHT

President to slash public spending Congress Senate ~ House ~ I'

compelling interest in building a credible I organizations I Dept.~f Energy

electricity at the same time. Additionally,

technology in order to guarantee a high

CREATION OF THE AEC

AEC's OPERATIVE ARMS

Figure 1.2 is a sketch that summarizes

that included other participants, such as

identifying, reviewing and analyzing

THE NUCLEAR 'SUBGOVERNMENT

PROS AND CONS OF THE 1946 ACT

THE ROAD AHEAD

MILITARY PROJECTS UNDER AEC

CIViliAN PROJECTS UNDER AEC

THE AEC-INDUSTRY PARTNERSHIP

THE RISE OF THE PWR IN THE MARKETPLACE

RICKOVER AND THE NUCLEAR SUBMARINE PROJECT

MARK-I AND NAUTILUS

THE QUEST FOR TECHNOLOGY AND MARKET

THE RAPID EXPANSION OF NUCLEAR POWER

DOMESTIC AND INTERNATIONAL COMPETITION

Moderator Coolant Fuel

One can argue that given the state of the art

In terms of market share, the right dimension is

WHAT IS A DOMINANT DESIGN?

WAS THE PWR A DOMINANT DESIGN?

reception to the PWR increased rapidly \ .... 'FBR

both in terms of units sold and in terms of

From the data of Table 2.1, the share in

Basic gas-cooled reactor IMAGNOX)

Pressurised heavy WIIter reactor (CANDU)

Pnmairy Fuel ""~f"rM'O&II1 t

Sodium-cooled fast reactor

~~:~~~:,~~~: ..:..~~~,~~:~~.;,:.~",:~,,~~':;~J:.~.'.,~~.v;.J';:~~:~~~~.~~~~~~~:~.'~l.~~'~~~:-,.~.~~~'~"~'.~,:,:,:~~.~.",.M,,~~' ....

f; Start:o f~C()nt merci al~Op eta ti on :;::::;:;:;:~~

....., .. ...,.:."'...,00.: ... "''''',.,. .•

f.i!tG~ii;ili: :fM}Y~tti 1 57 342 308

:t:: j;WGit~;~ ~lo'Yl" ........ "

37 2 76 3148 3280 5117 4010

:~:,~: ..:..~,:.;,:.~",:~,,~~':;~J:.~.'.,.v;.J';::.~~~:~.'~l.'~:-,.~.~~~'~"~'.~,:,:,:~~.~.",.M,,~~' ....