Highly Efficient Short Term Load Forecasting Scheme Using Long Short Term Memory Network

Load forecasting procedure is one of the most vital factors to make the selection of unit commitment, economic dispatch, relaxation, spinning reserve capacity and schedule device maintenance. Energy providers, related organizations in power generation, transmission system and policy makers of a country are dependent on the scale of future electricity demand. Short term load forecasting (STLF) procedure is considered one of the most complex and accuracy defining factors of the electrical power system. However, conventional forecasting strategies may not always fulfill the precise result demand in the issue of STLF. In this research work, a viable strategy for the issue of STLF procedure using long short term memory network (LSTM), a special type of recurrent neural network is presented. The forecasting outcomes on time series electrical load data of Bangladesh represent that the developed strategy has the ability to obtain higher precision and accuracy in STLF. Evaluation parameters of proposed network are also compared with the outcomes of conventional regression model and obtained superior outputs.