Short-term forecasting for distribution feeder loads with consumer classification and weather dependent regression

Abstract

Short term load forecasting (STLF) for feeder loads is critical for risk management of distribution companies in a competitive market. In this paper, weather variables and load profile classifications were investigated and their relative effects on the feeder load are reported. Moreover, Forecast techniques including time series models and ANN constructs were used as forecasting tools. Finally, risk assessment on load forecasting errors by using time domain and frequency domain respectively were proposed.