Forecasting residential electric loads with time-of-day rates. Final report

Technical Report · Fri Aug 01 04:00:00 EDT 1980

DOI:https://doi.org/10.2172/5166226· OSTI ID:5166226

A model is developred and applied to predict daily and hourly residential electricity demand, allowing for the effect of time-of-day rates. The model structure was developed to allow maximum generalization of results from time-of-day rate experiments and to allow reestimation with new experimental data as they become available. The model structure reflects a two-stage process: consumers first choose the level of expenditure on daily consumption of electricity as a function of an index of the price of electricity and the price of related goods; then the expenditure is allocated to different time periods within the day. The model is estimated and evaluated using data from the Connecticut Peak Load Pricing Experiment. Although the nature of the data imposed limitations on the model, the estimated demand equations generally yielded acceptable statistical results. The model is evaluated by comparing forecasts for 20 observed households with actual demand measured for these households. Since the validation database is relatively small, this validation exercise probably represents an overly restricted test of forecasting capability. Even so, the results of this validation exercise indicate an acceptable level of forecasting accuracy.

• Research Organization: Charles River Associates, Inc., Boston, MA (USA)
• OSTI ID: 5166226
• Report Number(s): EPRI-EA-1321
• Country of Publication: United States
• Language: English

Related Subjects

29 ENERGY PLANNING, POLICY, AND ECONOMY
296000* -- Energy Planning & Policy-- Electric Power
ELECTRIC POWER
ELECTRIC UTILITIES
FORECASTING
LOAD MANAGEMENT
MANAGEMENT
MATHEMATICAL MODELS
PEAK-LOAD PRICING
POWER
POWER DEMAND
PRICES
PUBLIC UTILITIES
RESIDENTIAL SECTOR
TIME-OF-USE PRICING