Short Term Load Prediction based on LSTM Network for Iraqi Thermal Power Plant

Electricity generation must satisfy the demand for electric loads in order to optimize the functioning of the power system. Load prediction can assist power companies in safely and efficiently operating their electrical systems. Load prediction is a process employed by power providers to predict the quantity of power or energy always demanded for balancing supply and demand. Short-term load prediction (STLP) with high accuracy is crucial to the seamless operation of the power system and the improvement of economic benefits. An approach for predicting short-term electrical demand utilizing long short-term memory (LSTM) based on actual data collected from Wasit Thermal Power Plant in Iraq is proposed in this paper. MATLAB software is utilized to implement the data used in this work. The assessment metrics employed were mean absolute error (MAE), mean absolute percentage error (MAPE), root mean square error (RMSE) and the coefficient of determination (R-squared) to assess the precision of load prediction. The findings demonstrate that the LSTM model is highly effective at forecasting the random characteristics of an electrical demand.