Plenary Program
System Dynamics and the Sustainable Development
of the Electric Power Industry
Andrew Ford
Program in Environmental Science and Regional Planning
Washington State University
Pullman, WA 99164-4430
U.S.A.
ABSTRACT
This paper discusses the successful use of System Dynamics in the electric power industry.
It begins with a list of studies and articles documenting the extensive use of System Dynamics
to aid corporate and government planners around the world. I then speculate on why this
particular industry has been the focus of so many System Dynamics studies. The paper
provides a short summary of the long running System Dynamics project for the Bonneville
Power Administration in the Pacific Northwest region of the United States. Reflecting on the
Bonneville project, I observe two important trends. The first is a trend toward interactive
models that are easier for group operation. The second is a trend toward more attention on
the consumers of electric power. The paper concludes with results from a recent study of the
world aluminum industry, an industry that consumes a tremendous amount of electric power.
INTRODUCTION
System Dynamics has been used extensively in the electric power industry. A review of
articles from the System Dynamics Review, for example, shows the following applications:
1. simulating the controllability of feebates for electric vehicles
(Ford 1995),
2. description of the energy model used at the US Department of
Energy (Naill 1992),
3. analysis of the cost effectiveness of US energy policies to
mitigate global warming through (Naill, Belanger, Klinger and
Petersen 1992), and
4. an analysis of conservation policies in the Pacific Northwest
(Ford and Bull 1989).
And a review of papers published at the annual meetings of the System Dynamics Society
shows additional examples:
1. the prospects for American’s electric utilities as they prepare
for the new competitive environment (Lyneis, Bespolka and
Tucker 1994),
2. modeling the privatization of the electricity market in the
United Kingdom (Larsen and Bunn 1994),
3. electric power demand forecasting in India (Garga, Gupta, and
Thapar 1985), and
4. lessons from the use of System Dynamics in hearings before
the New Mexico Public Service Commission (Ford 1981).
60
System Dynamics ‘95 — Volume I
Additional examples may be found in the operations research and management science
literature (Bunn, Larsen and Vlahos 1993; Ford 1990), the energy and utility policy literature
(Ford, Bull and Naill 1987; Bunn and Larsen 1995), reports issued by consulting groups (SSI
1988; AES 1993), and in the early System Dynamics literature made available from the
University of Bradford’s publication of Dynamica (Zepeda 1975).
DISCUSSION
I believe that a combination of reasons explain the wide spread use of System Dynamics in
the electric power industry.. The most important reason is the energy crisis of the 1970s
which caused researchers to concentrate on the use and consumption of all forms of energy.
Another contributing factor is the requirement for long range planning due to the long lead
times to acquire large scale generating resources. And a third factor is the public regulation
of electric power which requires an open discussion of the utility’s operations and resource
plans. (Public regulation in the United States is performed at the state level by Public Utility
Commissions which are responsible for ensuring that electric rates are sufficient to recover
costs and that utility investments in new generating capacity are reasonable.)
I believe a fourth contributing factor is worth mentioning as well. Utility managers often
combine management experience with training in engineering. Their background leads to a
natural affinity for analytical efforts in general and modeling in particular. Indeed, many
electric power companies carry a wide range of computer models to aid in planning the many
functions of their business. But these models are usually designed around problems in each
function area (operations, finance, rate hearings, etc.) and are not connected together in an
integrated fashion. In my opinion, System Dynamicists’ main contribution in the electric
power business is their ability to help managers tie the separate pieces of the utility system
into an integrated system. To illustrate this contribution, I report on the extensive use of
System Dynamics to support conservation and electric resource planning in the Pacific
Northwest. But first, a historical summary.
HISTORY OF THE BONNEVILLE PROJECT
During the 1980s, electric utility companies in the USA became interested in programs to
encourage their customers to invest in conservation. The programs included general
information such as advertising, specific information such as audits, and direct financial
incentives such as zero interest loans. The programs were needed to help customers overcome
market obstacles that limited their investment:in measures which would improve the efficiency
of electricity use. Utility conservation programs were viewed as a better use of company
funds than investment in conventional coal or nuclear power plants.
Conservation was especially important in the Pacific Northwest region of the United
States. The region encompasses the states of Washington, Oregon, Idaho and part of Montana,
and it is famous for its great rivers and mountainous terrain which provide its huge hydro
electric potential. Development of this potential began in 1933 with the Bonneville and Grand
Coulee dams. The region now has one of the world’s largest hydro-electric systems and,
historically, some of the lowest electric rates. Because of the low rates, the region’s homes
and businesses have not made the same level of investment in conservation as in other parts
of the country, and the potential conservation savings is large.
61
Plenary Program
The long period of low electric rates ended in the 1970s when events combined to increase
utility spending on new generating capacity. Like many utilities across the country, those in
the Pacific Northwest looked to nuclear power plants to meet anticipated growth in demand.
And like utilities elsewhere, the Pacific Northwest companies were hit hard by double digit
escalation in construction costs and unanticipated reductions in the rate of growth in electric
load. The combination of problems led to soaring electric rates, cancellation of several partly
constructed plants, major defaults on bonds, and, finally, to the passage of the Pacific
Northwest Electric Planning and Conservation Act of 1980.
The Act created the Northwest Power Planning Council to take responsibility for setting
broad policies for the regional development of electricity resources. The Act also created
major new responsibilities for Bonneville, the arm of the U.S. Department of Energy with
authority to market power from federal resources in the region. The Act called for Bonneville
to act as a power broker and to take a central role in implementing the best plan for the
region. Three provisions of the act are especially relevant to conservation. First, Bonneville
was obligated to provide for any future need for electric power in the region. Second,
Bonneville now had the authority to acquire resources to meet loads received from other
utilities in the region. Third, and perhaps most important, Bonneville was required to consider
conservation (along with renewable and cogeneration) preferentially when determining the
resources it should acquire to meet future load.
Bonneville’s response to the new authority in conservation was swift and multi-faceted.
Within two years after the passage of the Act, utilities were operating five region wide
Bonneville conservation programs, and conservation planning was upgraded to "Office-Level"
status. In 1983, Bonneville’s Office of Conservation initiated a project to improve its ability
to model the effects of its conservation programs and consumer incentive designs for the
Pacific Northwest electric power system. A System Dynamics model was designed to build
from the results of existing models and databases and to facilitate rapid analysis of many
scenarios.
THE CONSERVATION POLICY ANALYSIS MODELS
Work on the new model began by adapting relevant structure from a simulation model which
had proven useful in studies for a major California utility (Ford and Harris 1984). The first
step was to design a regional model in which conservation programs, system operation,
capacity expansion, and electricity pricing were conducted by a single entity. The next step
was to construct a sub-regional model which would distinguish between the loads and
resources of the investor-owned utilities (IOUs), the publically owned utilities, and the federal
government. The models were known collectively as CPAM, or the Conservation Policy
Analysis Models (Ford and Naill 1985; Ford and Bull 1989).
Figure 1 portrays the overall design of the regional version of CPAM by depicting the five
important sectors in the utility system. The demand sector is the largest of the five sectors
shown in Figure 1; it was singled out for greatest detail to permit direct simulation of the
wide range of conservation policy options of interest to Bonneville. The remaining sectors
complete the representation of an integrated electric system. The price of electricity, for
example, is calculated in the regulatory sector which sets average retail rates based on
allowed expenses, allowed return on investment, and the effects of regulatory lag. This sector
also keeps track of construction financing and maintains an accounting of the balance sheet
and cash flow. The projected rates are passed on to the demand sector to ensure internally
consistent projections of customer investment in conservation throughout the simulation.
62
System Dynamics '95 — Volume I
The five sectors shown in Figure 1 work together to provide an integrated representation
of the region’s electric system. Information generated in each of the five sectors is available
during each time step of the simulation as needed in the remaining sectors. The utility cost
from conservation incentive programs, for example, is made available to the price regulation
sector where the expenditures are either capitalized or expensed in the rate-making
calculations. Electric rates, in turn, are recalculated after the appropriate regulatory lags, and
the rates for the next time step are used in the demand sector.
The CPAM approach to model integration will be familiar to members of the System
Dynamics community. But this approach is NOT commonly used by utility planners to tie
models of different sectors together. The System Dynamics approach should be contrasted
with the more common utility approach in which several different models (usually constructed
in different departments within the same company) are designed to operate together as shown
in Figure 2. In this illustration, one begins with a set of electric rates needed as input for an
electricity demand model. The output of the demand model takes the form of electric load
projected for each of 20 years in the future, and the load projections are used as input for
a capacity expansion model. The output of the capacity planning model is a plan for new
power plant construction during the 20 years, and this plan is used to drive a costing model
which generates a set of electric rates needed to provide adequate revenues. The electric rates
emerging from this sequence of model projections are compared with the electric rates used
to start the calculations. If the two sets of rates are significantly different, the starting rates
are adjusted, and the sequence is repeated. Through artful manipulation of the starting rates,
one hopes to obtain a consistent set of projections within a reasonable number of iterations.
With the iterative approach shown in Figure 2, the output from an early model is not
provided to subsequent models until the early model is finished with a full 20 years worth of
results. In CPAM, output from the five sectors in Figure 1 are available to other sectors as
the model proceeds from one time step to another during the course of a 20 year simulation.
The main advantage of the iterative approach is the degree of detail that may be permitted
in each individual model in the sequence. Separate models may be coded in different
computer languages to allow analysts from different departments to find the best fit with their
topic area or to take advantage of existing models. With the CPAM approach, all five sectors
were designed from the outset to work together automatically over time. Each approach has
compelling advantages, and Bonneville used both approaches in the analysis of conservation
policies. The idea was to examine a broad range of conservation policies with CPAM before
initiating the iterative process with the more detailed models.
RECENT TRENDS
An important trend in the Bonneville project over the past few years is the emphasis on group
operation of the model in meetings designed to promote learning. This trend follows the lead
set by System Dynamicists around the world who have shifted their efforts to promote
Modeling for Learning Organizations (Morecroft and Sterman 1994). Bonneville’s model has
been redesigned to operate within a MicroWorlds cockpit (Diehl 1994). The MicroWorlds
version may be used in the traditional manner by the small group of analysts familiar with the
model details. But the important change in model use comes when wider groups of
Bonneville analysts and managers are turned loose to operate the model in group sessions.
Each participant is free to explore a variety of policies under different assumptions about the
future of the system. Each participant works at his or her pace. With this "learner-directed-
learning" approach, the role of the small circle of modeling experts is changed dramatically.
They no longer serve as the experts who must teach the wider team about the model insights.
63
Plenary Program
MEASURES.
DATA
PREPROCESSOR
CURVES
USERS.
USER |
SELECTION MENUS:
INTERFACE DOCUMENTED
LISTING
remeron { RESULTS
DOCUMENTOR
THE REGIONAL MODEL
CONSERVATION
>) INVESTMENTS REGULATION &
& ELECTRICITY ™~™ CONSTRUCTION |
DEMAND a FINANCING.
FEEDBACK EQUATIONS
CAPACITY |HYORO-THERMA\
EXPANSION SYSTEM
PLANNING OPERATION
CAPACITY
& ASSET:
ACCUMULATION
(BOOKKEEPING)
Figure 1. Design of an early version of the Conservation Policy Analysis Model.
START
DEMAND
—— ‘MODEL
™
REPEAT UNTIL.
STARTING PRICE
‘& ENDING PRICE
‘AGREE
a Nv
cost
MODEL
Figure 2. Iterative modeling approach often attempted by electric utility companies.
64
System Dynamics '95 — Volume I
Rather, the small group now serves the role of coach. Their new role is to allow members
of the wider group to explore the results of a variety of policies and arrive at their own
conclusions. In this setting, the key insights emerge as each individual conducts explorations
with the model. The insights are then confirmed and strengthened in group discussions.
A second trend is important as well. The new trend is toward the customer. The key
issues no longer seem to revolve around the choices that electric utilities must make about
generating technologies or finance. Rather, the key choices appear to be in the hands of the
electricity consumers. The System Dynamics model helps Bonneville planners move in this
new direction because the customers’ roles in the electric system are treated in considerable
detail. Of particular interest are large industrial customers whose electric loads are likely to
be sensitive to the price of electricity. Perhaps the most important of these customers are the
aluminum smelters.
THE PIVOTAL ROLE OF THE ALUMINUM INDUSTRY
The aluminum industry is a major consumer of electric power around the world. Aluminum
smelters locate where power rates are low, often immediately adjacent to the generators. In
1895, for example, a smelting facility was opened next to Niagara Falls in the State of New
York. (The smelter was the first customer of the Niagara Falls Power Company.) Aluminum
smelters use electrolytic reduction in which the raw material (alumina) is dissolved in a stable
solvent. Smelters consume around 7-8 kilowatt-hours of electric energy to produce a single
pound of aluminum metal. For large smelters (which can produce 200 thousand metric tons
of aluminum per year), the electricity demands are enormous. For example, the ten smelters
operating in the Bonneville service area can create an annual average electric load of almost
3,000 megawatts, accounting for around one third of Bonneville’s electricity sales in the
region.
The price of electricity is perhaps the single most important factor in determining the
competitive position of individual smelters in the world aluminum market. Consequently, the
electric load from the smelting industry can be quite sensitive to changes in the electric rate.
Were the rate to increase, for example, some smelters operating on the edge of profitability
might close down. Their closure would cause a loss of revenues to the electric utility. But
the electric utility usually faces interest obligations and other fixed costs which must be
satisfied. The historical approach has been to raise the electric rate to cover these fixed costs.
But this additional rate increase could trigger the closure of additional aluminum smelters,
leading, in turn, to further loss of revenues and further rate increases. The prospects of higher
and higher rates coupled with lower and lower sales to price sensitive customers is sometimes
called the
death spiral
or the "spiral of impossibility (Ford and Youngblood 1983).
Planners disagree whether Bonneville is locked in a "death spiral" with its large aluminum
customers. But regardless of one’s view of this downward spiral, the large consumers of
electric power play a pivotal role in almost any analysis of the future of the Pacific Northwest
power system. Even when one is studying some other aspect of the system (ie, the proper
incentives for customer conservation in the residential sector), the aluminum companies end
up playing an important role because of the information feedback loops that work their way
through their corner of the system. Several recent studies have been completed at Bonneville
65
Plenary Program
focusing on options to maintain a stable aluminum load. These include incentives for
investment in more efficient smelters, rate incentives that respond to changes in the world
aluminum price, and rate incentives that incorporate an advance notice requirement if a smelter
plans to permanently close operations.
THE MODEL OF THE WORLD ALUMINUM INDUSTRY
The world aluminum industry is of special importance to developing countries contemplating
the financial challenges of starting the development of large scale hydro-electric generation.
Large aluminum smelters offer a steady customer whose payments can be counted on to
provide a large share of cash flow needed to cover interest expenses. When demand growth
in the residential and commercial sectors is highly uncertain, a large, predictable aluminum
smelter may appear as an ideal partner for initiating the development of a region’s hydro
electric potential. Because of the importance of this industry to developing countries, I close
this paper with a short summary of a System Dynamics model of the world aluminum
industry.
The world aluminum model was developed with support from the Idaho National
Engineering Laboratory (INEL), one of the major energy laboratories in the Pacific Northwest.
INEL operates a variety of research programs on behalf of the US Department of Energy on
the aluminum industry, and their managers were interested in the potential for System
Dynamics to shed light on the "sunset problems" faced by smelters in the Pacific Northwest.
The INEL commissioned the development of a model that would include all the world’s
smelters (excluding smelters in the CIS, the Confederated Independent States formerly known
as the Soviet Union). INEL envisioned a model which would allow planners in the USA to
simulate research initiatives that could help USA smelters with "sunset problems" within an
industry where the "rising sun" has moved to countries like Australia, Brazil and Canada.
Another objective was to understand the great volatility in the world industry that is evident
in Figure 3.
The initial version of the world model was quite simple. We started with the stocks and
flows shown in Figure 4. We then added the key converters shown in Figure 5. We included
the recycling of "old scrap" as shown in Figure 6. The model breaks the demand for
aluminum down into nine end use categories (ie, automobiles, buildings, containers, etc.), and
allows the recycling parameters to vary from one end use to another. Recycling of "new
scrap" was also captured by the variables shown in Figure 7. Figure 8 concludes this brief
view of the model structure by portraying the feedback of market price information to govern
the aluminum ingot production at each of the 146 smelters in the wester world. An
overview of the model’s system boundaries is portrayed in Figure 9.
To facilitate interactive operation by groups of analysts, the model is embedded in a Micro
World’s cockpit which permits the user to control key policy and scenario variables without
a knowledge of the DYNAMO equations. The MicroWorld’s version allows for a wide variety
of graphs, tables, and custom tailored reports like the reports shown in Figure 10. The main
applications to date have been to explain the underlying causes of the great volatility in the
world aluminum market and to test the industry’s response to major research initiatives. These
applications are portrayed in Figure 11 which shows (1) high price volatility in a base case
exploration and (2) the price changes made possible by the assumption that a highly efficient
inert anode, stable cathode technology is available starting in the year 2000. The model may
also be used to simulate how the impacts of a new technology (like the inert anode, stable
cathode technology) play out differently from one country to another in the world aluminum
system.
66
System Dynamics '95 — Volume I
120
3
8
8
a
,————}
=
=
ol NEN UTE TAT TTT
AL Prica In CentsAb in 1987$
3
f-——-t
—
ii
a
I
40,360 1965 1970 1975 1980 1985 1990
Figure 3. World aluminum price in cents/pound, constant 1987$.
ingot inventory fabricators’ inventory Products in use
ingot production ingot shipments product shipments products discarded
Figure 4. Key stocks and flows in the world aluminum model.
time in inventory time to fabricate average life of products
ingot inventory fabricators’ inventory products in use
“oO”
ingot production
ingot shipments product shipments products discarded
Figure 5. Adding the converters to control key flows in the model.
67
Plenary Program
fraction recycled
recycled old scrap -
time in inventory
time to fabricate
ingot inventory,
oo
ingot production
fabricators’ inventory Products in use recycle potential
ingot shipments product shipments
discarded aluminum
average life of products
Figure 6. Simulating the recycling of "old scrap" in the world model.
new scrap metalic recovery fraction
. new scrap collection fraction
New scrap effective recovery fraction
Product discards
ingot production
ingot inventory products in use
demand for ingots C)
indicated demand for aluminum products
Figure 7. Simulating the recycling of "new scrap" in the model.
68
System Dynamics '95 — Volume I
old scrap
as product discards
O fab discards
new scrap
product shipments
ingot production
ingot shipments
coverage in
ingot inventory
fabricators' inventory products in use
days
demand
production from the
146 existing smelters
is based on a
smelter-by-smelter
calculation of
profitability
transactions price of aluminum
Figure 8. Simulating the
monetary
exchange rates
air pollution
regulations
indicated demand in
Westem world
‘nat import from
the CIS
car production
loop controlling ingot production from existing smelters.
EXCLUDED
price of
steel and copper
EXOGENOUS,
electricity
requirement
hydro-electric
potential
ENDOGENOUS
production from,
old smetters
demand for
aluminum
cost of
alumina
price of
aluminum
production from
new smetters
recycling collection
fractions
govemment
stock piles
Figure 9. Bull’s eye diagram (system boundary diagram) for the world model.
69
Plenary Program
Aluminum Flows: Diagram World Aluminum 8/15
old scrap 5.3
6.3
4.3 product discards
Year 2005
14.5
primary production
ie)
products in use
166.4
ingot shipments —_ product shipments
23.3 17.3
Note: flows in million metric tons/year; stocks in millions of metric tons
Aluminum Flows: Picture World Aluminum 8/15
secondary
production
9.6
old scrap
new scrap
4.3
product
shipments
17.3
fabricators’
discards YZ
3
for the year: 2005
Figure 10. Two Reports from the Micro Worlds version of the
world aluminum model mid-way through a base case exploration.
70
System Dynamics '95 — Volume I
90 q ria
80 |
ia WW
70 L ys i
e et
5 go & A NWA ian
§ 60 ; !
c]
8 \ WY
3 WV
= 50
.-3
7
= 40
2
3
i=
8 30
20
10
0
mwonrwoeoorarmnrwvrnMnor Doorn nwvr
OO? ooodceodececceoo0odececcrrrrere:r
Ooo ooccscdcdccdcececscooccscdccs
rrr Tr TH NNNNNHNH NN NNN NNN
Figure 11. Illustrative simulation results from the world model.
The open squares show the results of a base case exploration to learn if the
model can exhibit the price volatility shown previously in Figure 3. The cycles
in price shown here are typical of commodity production cycles (Meadows
1970). This volatility emerges from "inside the system." That is, there are
only a few, major external changes to disturb the system.
The black squares show a simulation in which the inert anode, stable cathode
smelting technology is available in the year 2000. Investment in the new
technology is calculated inside the model based on a payback target of 2 years.
As smelters around the world adopt the new technology, the market clearing
price is around 10 cents/pound lower than in the base case exploration. The
magnitude and periodicity of the cyclical behavior is not changed by the
introduction of the new smelting technology. Because of the lower market
clearing price, the simulated benefit of the new technology for smelter
operations in the Pacific Northwest is lower than what one might have expected.
71
Plenary Program
CONCLUSION
System Dynamics has been and continues to be an important approach to aid corporate and
government planners in the electric power industry. System Dynamicists have demonstrated
an ability to contribute in a "model intensive" industry because of their ability to tie the
separate pieces of the utility system together in a holistic manner. In reflecting on the use of
System Dynamics for the Bonneville Power Administration, I notice two important and useful
trends. The first is increased attention to making models easier to use in group sessions.
The second is greater attention to the customers’ side of the system. The stronger focus on
the customer is especially important for large industrial customers that may be highly
sensitive to the price of electricity.
A recently developed model of the world aluminum industry combines these two important
trends. The world model is especially important for planners studying the sustainable
development of their nation’s electric resources. The new aluminum model provides a
foundation for building a greater understanding of the complex "partnership" between the
electric power industry and the aluminum industry.
REFERENCES
AES 1993 The AES Corporation, "An Overview of the Ideas Model," 1001 North 19th St,
Arlington, VA 22209, October 1993
Bunn and Derek Bunn and Erik Larsen, "Assessment of the Uncertainty in Future UK
Larsen 1995 Electricity Investment Using an Industry Simulation Model," Utilities Policy,
forthcoming.
Bunn, Larsen Derek Bunn, Erik Larsen and Kiriakos Vlahos "Complementary Modelling and
Vlahos Approaches for Analysing Several Effects of Privatization on Electricity
1993 Investment," Journal of the Operational Research Society, Volume 44, No.10,
1993.
Diehl 1994 Emst Diehl, "Managerial Microworlds as Learning Support Tools,” in Modeling
for Learning Organizations, John Morecroft and John Sterman, editors,
Productivity Press, 1994.
Ford 1981 Andrew Ford, “Lessons from Case 1454," Proceedings of the 1981 System
Dynamics Research Conference," Rensselaerville, NY, October, 1981.
Ford 1990 Andrew Ford, "Estimating the Impact of Efficiency Standards on the Uncertainty
of the Northwest Electric System," Operations Research, July-August, 1990.
Ford 1995 Andrew Ford, "Simulating the Controllability of Feebates," System Dynamics
Review, Spring 1995.
Ford and Andrew Ford and Michael Bull, "Using System Dynamics for Conservation
Bull 1989 Policy Analysis in the Pacific Northwest, System Dynamics Review, Winter
1989.
72
System Dynamics '95 — Volume I
Ford, Bull Andrew Ford, Michael Bull and Roger Naill, "Bonneville’s Conservation Policy
and Naill Analysis Models," Energy Policy, April 1987.
1987
Ford and Andrew Ford and Steve Harris, "A Simpler Method for Calculating the Cost of
Harris 1984 Conservation Subsidies for an Electric Utility," Energy Policy, Sept. 1984.
Ford and Andrew Ford and Annette Youngblood, "Simulating the Spiral of Impossibility
Youngblood in the Electric Utility Industry," Energy Policy, March 1983.
Garga, Gupta Krishan Garga, N.K. Gupta and B. Thapar, "Dynamics of Power Supply and
and Thapar Demand," Proceedings of the 1985 International Conference of the System
1985 Dynamics Society, Keystone, Colorado, July 1985.
Larsen and Erik Larsen and Derek Bunn, "Modelling Electricity Privitisation in the UK,"
Bunn 1994 Proceedings of the 1994 International System Dynamics Conference, Stirling,
Scotland, July 1994.
Lyneis, James Lyneis, Carl Bespolka and Ben Tucker, “Preparing for a Competitive
Bespolka and Environment: The Prospects for American’s Electricity Utilities,"
Tucker 1994 Proceedings of the 1994 International System Dynamics Conference, Stirling,
Scotland, July 1994.
Meadows Dennis Meadows, Dynamics of Commodity Production Cycles, Wright Allen
1970 Press, 1970
Morecroft John Morecroft and John Sterman, editors, Modeling for Learning Organizations,
and Sterman Productivity Press, Portland, Oregon 1994.
Naill 1992 Roger Naill, "A System Dynamics Model for National Energy Policy Planning,”
System Dynamics Review, Winter 1992.
Naill and Roger Naill, Sharon Belanger, Adam Klinger and Eric Petersen, "An Analysis
Belanger of the Cost Effectiveness of U.S. Energy Policies to Mitigate Global
1992 Warming, System Dynamics Review, Summer 1992.
SSI 1988 Systematic Solutions Inc., "Introduction to Energy 2020," Jeffrey Amlin, George
Backus, and Randal Heeb, Vandalia, Ohio 45377, January 1988.
Zepeda Enrique Zepeda, "The Capacity Expansion Process in the Electricity Supply
1975 Industry", Dynamica, published by the University of Bradford System Dynamics
Research Group, 1975.
73
