Recent Survey of Electric Load Forecasting Techniques PDF

Keywords

Electric Load, Forecasting Techniques, Soft Computing, Swarm intelligence.

Abstract

Electric load forecasting is considered nowadays as one of the key issues for electricity utilities to ensure both good planning and design in long-term and efficient operation and management in short and medium terms. In fact, predicting as accurately as possible the electric load and peak-load can contribute to avoiding electricity feeding disturbances and allow thus a high quality of service provided to consumers. This paper reviews a selected set of papers on electricity demand forecasting techniques, published from 2008 to 2016. This review is intended to classify the proposed models and methods. The results presented in the surveyed papers show that a wide range of models and methods have been implemented and tested. Depending on the forecasting horizon, the study area and the used tools, the accuracy of the obtained forecasts are case-sensitive. Although classical time-series continue to be used, combined approaches involving more than two tools (models and methods) have started to attract more attention. More particularly, artificial neural networks (ANN) and swarm intelligence techniques are being increasingly used for their universal character and for the fact they do not need particular regularity of the load records.