This paper introduces a new methodology to predict power usage in irrigation system, using smartgrid data and Random Forest, in order to improve energy efficiency of these systems. The proposed methodology is able to predict energy consumption of a given timestamp based on previous information, using Random Forest Regressions. Then using a Random Forest Classifier, is able to classify that timestamp in either an ideal or not period to irrigate, based on network capacity and energy price, with the main goal of reducing the costs of energy to the client. Besides the methodology, this paper includes its implementation and experimental results. It was possible to achieve a 0.0468 Wh error in the prediction and a 87% accuracy in the classification.

--

Machine learning,Random forest,Smartgrid,Sustainability,Smart irrigation,Energy prediction