Institutional Electricity Load Forecasting Using Classical and Intelligent Forecasting Techniques

An accurate forecasting of Institutional Electricity load can proved to be useful asset for efficient utilization of the infrastructure available in terms of future demand and supply. Time series method of forecasting has got very wide applications like sales forecasting, yield prediction and Supply Chain Monitoring (SCM) system etc. This paper presents a classical time series models available for predicting the future demand. The classical models used to predict the future load and demand assumes the linear relationship between input and output but in the real world this doesn't seems to be practical. The intelligent and self-learning models like neural network has the lead to approximate any kind of non-linear function and can fit into these situations. Classification and prediction capabilities of Neural Network have also shown a great potential in forecasting. A neural network based time series forecasting model is also developed for electricity load forecasting. Behavioral pattern and trend of the experimental data are being studied and analyzed for accurate forecasting of electricity load.