Wind Speed Forecasting Method Based on Nondominated Sorting Genetic Algorithm and Machine Learning

As a result of technical advances and cost reduction efforts, wind power has become prevalent as a competitive alternative energy sources. Due to the strong volatility of wind speed, it the extensive wasting of resources on wind power is not uncommon. Improving the accuracy of wind speed prediction can avoid this situation and improve the economic efficiency and utilization rate of wind power generation. Therefore, this paper proposes a hybrid predictive model that combines multiple optimization algorithms. Firstly, based on the envelope entropy, the particle swarm optimization improved by NSGA-II is used to select the parameter of variational modal decomposition. Then, in order to improve the machine learning model’s structure, the sparrow search algorithm is used to determine the optimal parameters of the deep extreme learning machine. Finally, use SSA-DELM to predict each decomposed component to obtain forecasting results with higher accuracy. The example analysis shows that this method has good performance.