论文部分内容阅读
功率预测对于接入大量风电的电力系统运行具有重要意义。文章对提前4 h的风电机组出力预测进行了研究,分别采用BP神经网络法直接预测输出功率,以及时间序列法间接预测输出功率,并将两种方法组合以提高预测精度,组合权系数的选取以方差最小为目标函数。研究结果表明,不同方法的预测精度不同,尤其是在个别预测点处不同模型的误差差别较大,组合预测可减小预测系统的误差,提高预测精度。
Power prediction is important for the operation of power systems that access a large number of wind power. In this paper, the forecast of output of wind turbines 4 h ahead is studied. The output power is directly predicted by BP neural network and the output power is indirectly predicted by time series method. The two methods are combined to improve the prediction accuracy. The selection of combination weight coefficient The least variance is the objective function. The results show that the prediction accuracy of different methods is different, especially in different prediction points, the errors of different models are quite different. Combined forecasting can reduce the error of prediction system and improve the prediction accuracy.