A new approach using artificial neural network and time series models for short term load forecasting

Abstract

This paper presents a new approach for short-term load forecasting (STLF). Artificial neural network and time series models are used for forecasting hourly loads of weekdays as well as weekends and public holidays. In addition to hourly loads, daily peak load is an important data for system's operators. Most of the common forecasting approaches do not consider this issue. It is shown that the proposed approach provide very accurate forecast of the daily peak load. The input variables of the models have been selected based on their correlation coefficients. In addition, a new technique for selecting the training vectors is introduced. The valuable experience of expert operators is included in the modeling process. The model is simple, fast, and accurate. Obtained results from extensive testing on Ontario load data confirm the validity of the proposed approach. The mean percent relative error of the model over a period of one year is 2.066% including holidays.