基于EEMD⁃PSO⁃SVM的短期风电功率预测

Short⁃term wind power prediction based on EMD⁃PSO⁃SVM

针对风电功率预测精度不高的问题,建立基于EEMD⁃PSO⁃SVM的短期风电功率组合预测模型。利用集合经验模态分解法将不平稳的原始风电功率序列分解为不同复杂度的子序列,对这些子序列分别建立相应的SVM预测模型。由于支持向量机的惩罚因子和核函数参数选择对预测精度有很大影响,因此应用粒子群算法对支持向量机参数进行寻优,用优化好的参数进行建模训练,通过叠加各分量的预测结果得到最终风电功率预测值,最后对仿真结果进行分析。预测结果表明,基于EEMD⁃PSO⁃SVM的组合模型和单一模型相比可有效降低误差,在风力发电功率预测中有更好的可靠性和准确性。

In view of the low prediction accuracy of wind power prediction,a short⁃term wind power combination prediction model based on EEMD⁃PSO⁃SVM is established.With the ensemble empirical mode decomposition(EEMD)method,the unstable original wind power sequence are decomposed into subsequences with different complex rates,and the SVM(support vector machine)prediction models corresponding to these subsequences are established.As the selection of penalty factor and kernel function parameter of SVM has great influence on the prediction accuracy,the PSO(particle swarm optimization)algorithm is applied to the parameter optimization of the SVM.The optimized parameters are used for modeling training.The final wind power prediction value is obtained by superposing the prediction results of each component.The simulation results are analyzed.The prediction results show that the combined model based on EEMD⁃PSO⁃SVM can efficiently reduce the error in comparison with the single model,and has better accuracy and reliability in the wind power prediction.