Input Vector Comparison for Hourly Load Forecast of Small Load Area Using Artificial Neural Network

This paper presents an hourly load forecast of small load area using Artificial Neural Network (ANN). For this case-study duration of February-2010 to Januray-2011 is considered. In this study ANN is trained and tested for by providing two different input vectors. In this paper the input vector design and the data is mainly focused. Also, suitable ANN topology is also discussed. Further the training and testing process for ANNs of these months are explained. Back-propagation algorithm is employed in this process. Finally by comparing network performances for these two input vectors for each of the considered month, optimum vector is selected.