AN ARTIFICIAL NEURAL-NETWORK HOURLY TEMPERATURE FORECASTER WITH APPLICATIONS IN LOAD FORECASTING

Many short term load forecasting techniques use forecast hourly temper atures in generating a load forecast. Some utility companies, however, do not have access to a weather service that provides these forecasts . To fill this need, a temperature forecaster, based on artificial neu ral networks, has been developed that predicts hourly temperatures up to seven days in the future. The prediction is based on forecast daily high and law temperatures and other information that would be readily available to any utility. The forecaster has been evaluated using dat a from eight utilities in the U.S. The mean absolute error of one day ahead forecasts for these utilities is 1.48 degrees F. The forecaster is implemented at several electric utilities and is being used in prod uction environments.