ANN Based Short Term Load Forecasting Paradigms for WAPDA Pakistan

This paper discusses a comprehensive approach for Short Term Load Forecasting (STLF) for Water and Power Development Authority (WAPDA), Pakistan. Keeping in view, non-linear power systems load characteristics; two different Artificial Neural Network (ANN) based architectures have been devised for STLF. Proposed architectures were trained and tested using previous five years actual load data obtained from WAPDA i.e., year 1991-95. Several back propagation based training algorithms were applied to both architectures. Data engineering approach was applied for thorough comparison and identification of the most appropriate approach that yields accurate forecast