MATLAB/spl reg/ implementation of neural & neuro-fuzzy approaches for short-term electricity demand forecasting

Electricity is a commodity that cannot be stored in bulk; its demand must be forecasted so as to ensure adequate supply while not generating far exceeded that required. In that, forecasting the short-term demand is increasingly vital for planning operations to ensure economic margin of electricity supply. This paper demonstrates the use of MATLAB/spl reg/ technical computing tools to implement neural and neuro-fuzzy approaches for short-term demand forecasting. Two types of forecasting models are illustrated herein, i.e. one-hour ahead and next-day forecasting. Firstly, the paper emphasizes on presenting various stages of the model-development process, from preliminary data analysis to deployment of the developed models. The second section of the paper focuses on introducing the neural and neuro-fuzzy approaches. The paper concludes by illustrating the deployment aspects of the forecasting models; with the use of graphical user interfaces (GUIs) developed using MATLAB/spl reg/ to facilitate the forecasting process.