Multi-step ahead short-term wind speed forecasting approach coupling variational mode decomposition, improved beetle antennae search algorithm-based synchronous optimization and Volterra series model

Abstract

Abstract Accurate wind speed forecasting has become an indispensable part in the dispatching and management of wind power generation. However, the volatility and intermittency of wind speed impede the improvement of forecasting accuracy. To figure out the problem and strengthen prediction performance, a novel hybrid framework consisting of variational mode decomposition (VMD), phase space reconstruction (PSR), improved beetle antenna search (BAS) and Volterra series model is established for multi-step ahead short-term wind speed forecasting. To confirm the performance of the proposed compound model, three datasets from the Sotavento Galicia wind farm are used for multi-step short-term wind speed forecasting experiment. The prediction results indicate that: (a) the prediction precision of the proposed model is significantly improved by employing VMD based on central frequency observation approach, which can weaken the non-stationarity of the original series; (b) the proposed DEBAS algorithm based on hierarchical strategy achieves the best performance compared with other comparative algorithms on the benchmark functions, because DEBAS can expand population diversity in the later iteration; (c) the forecasting performance of the hybrid model can be further enhanced by the proposed synchronous optimization strategy based on the improved DEBAS.