Forecasting abnormal load conditions with neural networks

Abstract

The authors present a new approach to power load forecasting under abnormal weather conditions using artificial neural networks (ANN). Accurate forecasting for cold fronts and warm fronts is of special importance to utility companies for monetary reasons and planning reasons. Temperatures below 50 degrees F are treated as cold fronts and temperatures above 90 degrees F are treated as warm fronts in the area of interest. The architectures take into account some inherent characteristics of these days. The results obtained by using ANN have been found to give better results than other conventional techniques.<>