APPENDIX E
A BRIEF DESCRIPTION OF DOE AND EIA MODELS1

PROJECT INDEPENDENCE EVALUATION SYSTEM

The Project Independence Evaluation System (PIES), later named the Mid-term Energy Forecasting System (MEFS), was initially developed in 1974 by the Federal Energy Administration (FEA), a predecessor organization to EIA. PIES/MEFS was a static model, solving for one designated future year at a time. For its initial use in the Project Independence Blueprint, it provided forecasts for 1977, 1980, and 1985. For the 1981 Annual Report to Congress, the system projected 1985, 1990, and 1995.

The core model of PIES/MEFS was a single linear program of regional fuel supply, transportation, conversion, and end-use demand activities. This linear program optimized by solving for the configuration of fuel production and transportation to meet demand at the least cost to consumers. PIES/MEFS solved for a supply-demand equilibrium in energy markets by iterating between the linear program and a reduced-form representation of end-use demand models. From the linear program were derived the marginal, or shadow, prices for fuel delivered to the end-users by sector and region. The reduced-form demand representation was evaluated at these prices, the new end-use demands entered into the linear program, and the program reoptimized. This process continued until the end-use prices and demands were not changing between iterations, within a specified tolerance.

1  

 This section draws from papers written at the committee's request. Susan Shaw, Energy Information Administration, wrote the material on PIES, MEFS, IFFS, LEAP and STIFS, and Eric Petersen, Office of Policy, Planning and Analysis, wrote the material on Fossil2.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 129
The National Energy Modeling System APPENDIX E A BRIEF DESCRIPTION OF DOE AND EIA MODELS1 PROJECT INDEPENDENCE EVALUATION SYSTEM The Project Independence Evaluation System (PIES), later named the Mid-term Energy Forecasting System (MEFS), was initially developed in 1974 by the Federal Energy Administration (FEA), a predecessor organization to EIA. PIES/MEFS was a static model, solving for one designated future year at a time. For its initial use in the Project Independence Blueprint, it provided forecasts for 1977, 1980, and 1985. For the 1981 Annual Report to Congress, the system projected 1985, 1990, and 1995. The core model of PIES/MEFS was a single linear program of regional fuel supply, transportation, conversion, and end-use demand activities. This linear program optimized by solving for the configuration of fuel production and transportation to meet demand at the least cost to consumers. PIES/MEFS solved for a supply-demand equilibrium in energy markets by iterating between the linear program and a reduced-form representation of end-use demand models. From the linear program were derived the marginal, or shadow, prices for fuel delivered to the end-users by sector and region. The reduced-form demand representation was evaluated at these prices, the new end-use demands entered into the linear program, and the program reoptimized. This process continued until the end-use prices and demands were not changing between iterations, within a specified tolerance. 1    This section draws from papers written at the committee's request. Susan Shaw, Energy Information Administration, wrote the material on PIES, MEFS, IFFS, LEAP and STIFS, and Eric Petersen, Office of Policy, Planning and Analysis, wrote the material on Fossil2.

OCR for page 129
The National Energy Modeling System As the model evolved over time, a number of special features were added to reflect regulatory policies or to ensure that certain end-use prices from the linear program were average or regulated prices, rather than strict marginal prices. Other features were added to ensure intertemporal consistency between the time periods of the model. These capabilities were incorporated in either the linear program that controlled the equilibrium and determined convergence. As a modeling system, PIES/MEFS encompassed a host of satellite models--coal, oil, natural gas, synthetic fuels, refinery, electric utility, end-use demand, and macroeconomic. Each of these satellite models produced the necessary coefficients and objective function costs for the linear program and incorporated sector-specific features as required. Other models used the results of PIES/MEFS to perform macroeconomic and distributional analysis. This structure served to organize the data and allocate responsibilities for the modeling activities. INTERMEDIATE FUTURE FORECASTING SYSTEM The current EIA integrated modeling system is the Intermediate Future Forecasting System (IFFS), developed in 1982. IFFS partitions the energy system into fuel supply, conversion, and end-use demand sectors and then solves for a supply-demand equilibrium by successively and repeatedly invoking the modules that represent these sectors. The model solves annually, currently with a forecast horizon of 2010. The supply-demand equilibration is performed one year at a time, stepping forward to the next forecast year when the equilibrium for one year is complete. Fundamental assumptions for the modeling system are the assumptions for the world crude oil price and a baseline macroeconomic forecast. The fuel supply modules of IFFS encompass all the activities necessary to produce, import, and transport the fuel to the end user, thus computing the domestic production and the regional end-use prices necessary to meet end-use demand. Each of the end-use demand modules compute the fuel requirements for the sector by region, based upon the regional end-use prices of all competing fuels, as well as other factors such as economic variables and technology characteristics. The electricity module, as a conversion module, simulates an input of fuel based on relative prices and technology characteristics as well as generates and prices electricity as an output. Within IFFS, the primary interfaces between the modules are the regional end-use prices and demands for each fuel. Each fuel supply or end-use demand module is called in sequence and solves assuming all other variables in the energy markets are held constant. That is, the coal module solves for the production and end-use prices of coal, assuming a slate of demands for coal and assuming that all other sectors are fixed. Any module that uses the coal prices would then use these new prices to compute demand the next time the module is executed. As the model solves, the modules are called in sequence and percentage differences between iterations for all end-use, regional prices and demands are computed. When differences are within the user-specified tolerance, convergence is declared for that forecast year and the solution of the next forecast year begins.

OCR for page 129
The National Energy Modeling System Some attempt is made in IFFS to utilize convergence characteristics of particular modules. For example, the sensitivity of the natural gas price to the level of demand and the concomitant sensitivity of gas demand in certain sectors to the price is well recognized. So the electricity module computes a derived demand curve for natural gas, explicitly representing the demand for gas at a variety of prices, for both the electricity and gas modules to deal more effectively with convergence. Due to the partitioning of the energy markets and the specific implementation of the modules, it is straightforward to execute any subset of the modules or a single module in IFFS or to substitute a module that meets a minimum interface requirement. In addition, the modular nature of IFFS readily allows each sector of the energy market to be represented with the methodology deemed most suitable to that sector, allowing for a more natural representation of each market. IFFS currently contains a mix of simulation, process, econometric, and optimization methods within the various sectors. It also allows each module to vary the depth and breadth of its coverage of the sector. As examples, there is a variation in whether the fuel supply modules explicitly include a representation of the fuel transportation and in what regional definition is used for fuel supply. Like the PIES/MEFS system, IFFS utilizes a number of satellite models or analytical procedures in order to represent the energy market. Although IFFS encompasses more of the modeling directly, there still remain exogenous components. This is due either to the initial design of the exogenous models or to the necessity for increased flexibility of the component. Examples include a world coal trade model that provides export projections, an oil market model for world crude oil price forecasts, and a procedure for deriving natural gas import forecasts. LONG-TERM ENERGY ANALYSIS PROGRAM From 1979 to 1981, EIA utilized the Long-Term Energy Analysis Program (LEAP) for long-range forecasting to the year 2020. LEAP was EIA's configuration of the Generalized Equilibrium Modeling System (GEMS), originally developed by the Stanford Research Institute (currently, SRI International) and now with Decision Focus, Inc. This is utilized by a number of organizations, configured to suit their particular purposes. LEAP/GEMS solves for a supply-demand equilibrium in a way fundamentally similar to IFFS with prices and quantities of the various types of energy being computed by modules that represent production, raw material transportation, conversion, final product transportation, and end-use energy consumption. Within the overall system, the order of solution is directional, prices flowing from supply to end-use demand and quantities flowing in reverse. Thus, it solves for an equilibrium by a recursive technique. One difference between LEAP/GEMS and IFFS is the method of solving through time. Each module solves for all forecast years at a time in LEAP/GEMS, attaining an equilibrium for all years simultaneously and enabling some modules to utilize perfect foresight over the forecast horizon.

OCR for page 129
The National Energy Modeling System LEAP, or any GEMS-derived model, segments the energy system by separating all supply, transportation, conversion, and end-use processes. Each of these activities is defined as a node, and a network describing the flows of all information between nodes must be explicitly drawn. Every regional activity, such as coal supply by region, would also be a separate node. At all decision points in the network, there are allocation nodes that represent the fuel or technology choice or the regional mix for different supply options. The allocation nodes use market share algorithms with market share coefficients, price premiums, behavioral lag coefficients, and initial year market shares. One feature of the GEMS system is a library of generic models, from which one can choose in building a representation of the energy system. These generic models include a simple and a complex conversion process, an allocation process, a primary resource process, a end-use demand process, and a transportation process. In building a model using GEMS, a user draws the network by selecting a generic model for each node, defining all the input and output links to other nodes, and specifying all necessary data to characterize the node specifically. It is the data that distinguishes, for example, a residential electric heat pump node from an industrial machine drive node. If a model builder wishes to add new generic models to the system, it can be done. As an example, for LEAP, EIA created a separate coal supply module as distinguished from the oil and gas supply modules, believing that a single primary resource process could not adequately represent such dissimilar fuel supplies. SHORT-TERM INTEGRATED FORECASTING SYSTEM. A very different modeling system has been used by EIA since 1979 for short-term forecasting and related analysis. The short-term energy outlook of EIA presents a two-year, quarterly forecast of energy supply and demand, produced using the Short-Term Integrated Forecasting System (STIFS). Primary inputs to STIFS include assumed prices for crude oil and natural gas, macroeconomic indicators, assumed weather patterns, and current data on the energy system, including inventory levels. STIFS is a national representation of energy markets. Consumption for each of the major fuels is computed based upon relative end-use prices and recent trends. Domestic crude oil production is a function of the assumed crude oil price. The consumption forecasts and trends for fuel production, imports, and stock levels are integrated into an energy balance for each of the fuels. Thus STIFS does not account for all feedbacks of energy prices or consumption on production. Being a short-term system, STIFS also does not account for capital stock changes. FOSSIL2 The integrating analysis tool for the National Energy Strategy is a large-scale model of the U.S. energy system called Fossil2. Fossil2 is a dynamic simulation model of U.S. energy supply and demand designed to project the long-term (30 to 40 year) behavior of

OCR for page 129
The National Energy Modeling System the U.S. energy system. The model structure, which includes all energy producing and consuming sectors, simulates the marketplace in a series of dynamic stocks and flows; the stocks include energy production facilities (e.g., oil fields), energy transformation facilities (power plants) and energy consuming entities (e.g., houses, vehicles), while the flows include energy, prices and information. The Fossil2 model can be characterized as an equilibrium energy market model, as energy markets “clear” over time through feedback among such factors as prices, demand, conservation investments, production costs and production capacity. The model clears markets for each iteration period, which is a variable within the model that is typically set to one quarter year. The model uses System Dynamics, a methodology that represents system behavior through differential and integral equations. In Fossil2 the demand for energy is determined in a “least cost/energy services” framework of total U.S. energy demand. Following this approach, the model first projects the demand for energy services (heat, light, steam, shaft power) in each end-use category, and then calculates the share of service demand captured by end-use technologies. For most energy service categories, there are several fuels that can provide the required energy services--in addition to “conservation.” In a few categories (such as lighting or appliances), only electric energy services can be used. For those where there are choices of fuels, a least-cost algorithm based on consumer choice theory is used in Fossil2 to determine the market share for different fuel-using categories of new energy equipment. The energy supply sectors of the Fossil2 model represent the decisions that lead to the commitment to new production capacity, the operation of existing production capacity, and the setting of energy prices for oil, gas, coal and electricity. Energy producers choose to invest in production technologies that maximize the industry's rate of return (or minimize the average cost of production), subject to environmental constraints (for example, SO2 restrictions). The sectors keep account of production capacity and assets, and calculate energy prices in accordance with the rules estimated to be followed in each industry. These rates then feed into the demand sector, helping to determine current and future growth in energy demand.

OCR for page 129
The National Energy Modeling System This page in the original is blank.