skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Regional load curve models: specification and estimation of the DRI Model. Final report. [Forecasts of electric loads in 32 US regions]

Technical Report ·
DOI:https://doi.org/10.2172/6699775· OSTI ID:6699775

The DRI Model of hourly load curves is developed in this report. The model is capable of producing long-term forecasts for 32 US regions. These regions were created by aggregating hourly system load data from 146 electric utilities. These utilities supply approximately 95% of all electricity consumed in the continental US. The model forecasts electricity demands for each hour of the year for each of the 32 regions. Model output includes forecasts of peak demands, megawatt hour demands, load factors, and load duration curves. The DRI Model is estimated in two stages. In the first stage, for each region and month, hourly electricity demands are parameterized into load components representing the effects of lifestyles and weather on regional loads through a time-series model. In the second stage, the variation in these parameterized load components across months and regions is modeled econometrically in terms of energy prices, income levels, appliance saturation rates, and other variables. The second-stage models are essentially models of electricity demand which are estimated using estimated first-stage parameters as dependent variables, instead of observed demands. Regional price and income demand elasticities are implied by the second-stage models. Moreover, since the dependent variables refer to particular hours of the day, these estimated elasticities are hour-specific. (Since prices did not vary over the day in years when hourly load data were available, hour-to-hour, cross-price elasticities were not estimated.) Integrated system hourly load forecasts are obtained combining the influences of individual customer classes. Finally, approximate customer class hourly load shapes can be produced for each region, though these series may be useful only in research endeavors since they lack the precision available through survey methods.

Research Organization:
Data Resources, Inc., Lexington, MA (USA)
OSTI ID:
6699775
Report Number(s):
EPRI-EA-1672(Vol.1)
Country of Publication:
United States
Language:
English