Application of Artificial Neural Networks for Hourly Load Forecasting in Distribution Systems

Abstract

Today's electricity power industry is in the process of up-gradation and modernization. The power system
operation has become more competitive in the open market environment. The accuracy of load forecasting is important
because it has a direct influence on the planning of the operation schedule of power generation plants. A major part of the
operating cost of power generation units depends on the amount of electricity production. To minimize the total operation
cost, unit commitment scheduling is used to determine the optimal commitment schedule of power generation units to satisfy
the forecasted demand. Load forecasting is an important requirement for unit commitment planning. Hour-ahead load
forecasting is necessary for optimally controlling the online resources to supply the next hour load. The objective of this
study is to prove the application of Artificial Neural Networks for performing hourly load forecasting in distribution systems.
The forecasting results indicate that ANN based load forecasting systems can be used for load forecasting in distribution
systems with a good accuracy.