Systematic Development of Short-Term Load Forecasting Models for the Electric Power Utilities: The Case of Pakistan

Load forecasts are fundamental inputs for the reliable and resilient operation of a power system. Globally, researchers endeavor to improve resulting forecast accuracies. However, lack studies detailing standardized model development process remains major issue. In this regard, paper advances knowledge systematic short-term load (STLF) electric utilities. The proposed has been developed by using hourly (time series) five years an utility in Pakistan. Following investigation previously models, study addresses challenges STLF utilizing multiple linear regression, bootstrap aggregated decision trees, artificial neural networks (ANNs) as mutually competitive forecasting techniques. also highlights both rudimentary advanced elements data extraction, synthetic weather station development, use elastic nets feature space upscale its reproducibility at global level. Simulations showed superior prowess ANNs over other techniques terms mean absolute percentage error (MAPE), root squared (RMSE) R2 score. Furthermore, empirical approach taken underline effects recency, climatic events, cuts, human activities, public holidays on model’s overall performance.