INTELLIGENT APPROACH FOR SHORT TERM ELECTRICAL LOAD FORECASTING

Abstract

Load forecasting play very important role in the operation of electricity companies. It helps the electric utility to make unit commitment decisions, efficient energy planning, reduce spinning reserve capacity and schedule device maintenance plan properly. It is therefore necessary that the electricity companies should have prior knowledge of future electricity demand with great accuracy. This dissertation focuses on study of short term load forecasting using two different types of computational intelligence methods. It includes fuzzy logic and artificial neural network based approach. In this dissertation, the daily demand of Shahpura, Jaipur, India has been collected from Rajasthan Electricity Board ( Shahpura Sub-station), India. To avoid the convergence problems, the input and output load data are scaled down such that they remain within the range of (0.1-0.9). The inputs of the fuzzy logic and ANN based models are the electrical demand during the day for the four consecutive and the output or forecasted value is the demand of the fifth day. The results obtained from fuzzy logic and ANN models have been validated with the actual value and found accurate. The mean absolute percentage error (MAPE) in the fuzzy logic model is 2.515% and using ANN 2.165%.