基于小波包和改进纵横交叉法的风电预测模型

针对风电出力的随机性、季节性和波动性及一般方法预测精度不高的问题,采用改进纵横交叉(improved crisscross optimization, ICSO)算法,建立了一种基于小波包变换与改进纵横交叉算法优化Elman神经网络的风电预测模型。仿真结果表明:改进的纵横交叉算法不但克服了一般算法早熟收敛的缺陷,有效提高神经网络的泛化能力和预测精度,而且表现出良好的稳定性,适用于风电的预测。

Wind power forecasting model based on wavelet packet and improved crisscrossoptimization method

Aiming at the problems of the randomicity, seasonality and intermittency of the wind powerand the prediction accuracy of general method is not high, using improved crisscross optimization(ICSO)algorithm, establishes a new wind power forecasting model based on hybrid wavelet packet andimproved crisscross optimization (ICSO) algorithm to optimize Elman neural network. The simulationresult shows that ICSO algorithm not only overcomes the problem of premature convergence in thegeneral algorithm, effectively improves the generalization ability and prediction precision of the neuralnetwork, but also shows fine stability, the forecasting model is suitable for the wind power prediction.