NN based support system for renewable energy forecasting

Abstract

This paper proposes to prosumers a NN based decision support application for selecting an optimal forecasting tool for energy produced from renewable sources. The exploration and the assessment of criteria used for choosing a forecasting tool are made in the neural network (NN) framework. Firstly, the criteria for selecting the best forecasting tool are addressed. Secondly, the identified criteria are integrated in an object oriented software application, built using Matlab-Guide User Interface. In order to underline the effects of the users' decision making, in the third part, the forecasting performances of feed forward neural networks (FF-NN) are tested and evaluated.