LAUSR

Abstract Accurate wind speed forecasting is useful for large-scale wind power integration, which can reduce the adverse effects of wind power on the power grid. However, due to the randomness and uncertainty of wind speed, accurate wind speed forecasting becomes very difficult. To improve the forecasting accuracy, an adaptive hybrid model based on variational mode decomposition (VMD), fruit fly optimization algorithm (FOA), autoregressive integrated moving average model (ARIMA) and deep belief network (DBN) is proposed. First, the original wind speed is decomposed into some regular and irregular components by VMD and FOA. Second, ARIMA model is built to forecast the regular components, while DBN is used for irregular components forecasting. Third, the final forecasting results is obtained by summing the forecasting results of each component. The effectiveness of the proposed model is verified by using data from two different wind farms in China. To demonstrate the performance of the proposed model, some well-recognized single models and some latest published hybrid models are selected as the comparison models. Empirical results show that the accuracy of the adaptive model is more higher than the other models.