MATLAB® 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 (R) 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 (R) to facilitate the forecasting process.