中文摘要：

针对BP神经网络、RBF神经网络和粒子群BP神经网络在风电场风速预测中存在的问题，提出一种基于遗传算法优化神经网络的风速组合预测模型。该模型为单输出的3层前馈网络，将3种神经网络的预测结果与预测结果平均值作为神经网络的输入，将实际风速值作为神经网络输出，使学习后的网络具有预测能力。该模型能降低单一模型的预测风险，提高预测精度。仿真结果表明，所提出的组合预测模型的精度高于其中任一单一模型，也高于传统的线性组合预测模型。

英文摘要：

To solve the problems existing in the wind speed forecasting of wind farms by back propagation （BP） neural network, radial basis function （RBF） neural network and particle swarm optimization （PSO） neural network, a combined wind speed forecasting model based on artificial neural network （ANN） optimized by genetic algorithm （GA） is proposed. This model is a BP neural network with three layers and a one neuron output. The forecasted results of the three mentioned neural networks and the average of the forecasted results are used as the inputs of the combined forecasting model, and the actual value of wind speed is used as the output. The proposed model possesses a forecasting ability after training. The forecasting risk of single neural network can be reduced and the forecasting accuracy can be improved in this model The simulation results show that the accuracy of the proposed combined forecasting model is higher than any of its single network model and the traditional linear combined forecasting model.