Short Term And Medium Term Electrial Load Forecast

Abstract

,The electrical load forecasting finds its application in several federal policy related matters, network expansion, and suitable allocation of energy resources within masses. The planning institutions of power utilities in Pakistan are making use of traditional statistical methodologies for electrical load forecasting purpose, which are not capable of incorporating system non-linearities effectively. The modern day deep neural network based non-linear parametric modeling techniques are more suitable to handle the system dynamics and non-linearities effectively, rather than traditionally employed statistical methodologies. In this research work, Long Short-Term Memory based Recurrent Neural Network model (RNN-LSTM) is developed and implement for the load forecasting of Pakistan. The temporal and climatic factors are also embedded as input parameters in these forecasting models after thorough exploratory data analysis. The results of RNN-LSTM are compared with different linear and non linear parametric modeling techniques. The qualitative and quantitative comparison among all linear and non-linear parametric methodologies reveals that the proposed RNN - LSTM outperforms among all other forecasting models.