Regional load-curve models: QUERI's model specification, estimation and validation. Final report

Technical Report · Sat Aug 01 00:00:00 EDT 1981

OSTI ID:6259366

Ramanathan, R; Engle, R F; Granger, C W.J.

The objective of this project is to construct a model that relates the hourly demand for electricity in a given region to the short-run characteristics such as weather at specific hours of a day, and long-run attributes such as changes in income, population, appliance stock, and industrial mix. A two-step model or electricity demand was estimated using hourly load data for 32 regions for the period 1962 to 1974 and data on a number of region specific variables such as the ones listed above. The first stage related hourly demand in each region to weather and other time-of-day (TOD) variables. The set of parameters obtained from this stage were then related to the longer run variables stated above. Both by measures of goodness of fit and by statistical significance of the estimated parameters, the results may be deemed quite successful. The model highlights the importance or economic and demographic variables in long-run load forecasting. A post-sample validation of the model was also carried out using the data for the period 1975 to 1977 for six test regions representing a variety of climatic conditions and socio-economic characteristics. For five of the six regions, the model had an average forecast error of 4.5% (in one region, a correction had to be made for non-reporting by an individual utility). For the sixth region, the error was 15%. Differences in growth rates of explanatory variables between the periods 1970 to 1974 and 1975 to 1977 account for 5% with 10% unaccounted for. On the whole, the model is robust and captures the load shapes quite well. The predicted peak was usually within ten percent of the actual peak and mostly below. With minimal data requirements QUERI's model may be adopted to generate forecasts for a sub-regional area such as a utility service are. This report outlines a method to achieve this. However, further work is necessary to examine how well it will succeed.

• Research Organization: Quantitative Economic Research, Inc., San Diego, CA (USA)
• OSTI ID: 6259366
• Report Number(s): EPRI-EA-1672(Vol.2); ON: DE81904252
• Country of Publication: United States
• Language: English

Related Subjects

29 ENERGY PLANNING
POLICY AND ECONOMY
ELECTRIC UTILITIES
LOAD ANALYSIS
ENERGY MODELS
FORECASTING
REGIONAL ANALYSIS
SIMULATION
CLIMATES
CONSUMPTION RATES
DEMAND FACTORS
ELECTRIC POWER
ENERGY ANALYSIS
PRICES
SOCIO-ECONOMIC FACTORS
USA
VALIDATION
WEATHER
INSTITUTIONAL FACTORS
NORTH AMERICA
POWER
PUBLIC UTILITIES
TESTING
296000* - Energy Planning & Policy- Electric Power