Application of support vector machine models for forecasting solar and wind energy resources: A review

Abstract

Conventional fossil fuels are depleting daily due to the growing human population. Previous research has proved that renewable energy sources, especially solar and wind, can be suitable alternatives to the conventional energy sources that could satisfy global demand and protect the atmospheric environment. There are many factors that influence the performance of solar and wind energy predicting tools. The accurate forecasting of solar and wind energy resources is highly needed for the optimum utilization of these resources. Different methods have been applied to forecast solar and wind energy resources. Prediction performance of the support vector machine modeling approach found to be better than other modeling approaches. The support vector machine is fast, simple-to-use, reliable and provides accurate results. Findings based on critical analysis suggests that the hybrid support vector machine models can reach much higher accuracies than other models for both solar and wind energy predictions for most of the locations. This investigation highlighted main problems, opportunities and future work in this research area. Novel hybrid models are proposed for further investigation for more accurate predictions of solar and wind energy resources.