Evaluation of hybrid forecasting approaches for wind speed and power generation time series

Abstract

Forecasting of wind speed and wind power generation is indispensible for the effective operation of a wind farm, and the optimal management of its revenue and risks. Hybrid forecasting of time series data is considered to be a potentially viable alternative compared with the conventional single forecasting modeling approaches such as autoregressive integrated moving average (ARIMA), artificial neural network (ANN), and support vector machine (SVM). Hybrid forecasting typically consists of an ARIMA prediction model for the linear component of a time series and a nonlinear prediction model for the nonlinear component. In this paper, we systematically and comprehensively investigate the applicability of this methodology based on two case studies on wind speed and wind power generation, respectively. Two hybrid models, namely, ARIMA–ANN and ARIMA–SVM, are selected to compare with the single ARIMA, ANN, and SVM forecasting models. The results show that the hybrid approaches are viable options for forecasting both wind speed and wind power generation time series, but they do not always produce superior forecasting performance for all the forecasting time horizons investigated.