基于改进果蝇算法优化的SVM风电功率短期预测

Wind Power Short-Term Prediction Based on SVM Trained by Improved FOA

由于风力发电功率预测的准确性直接关系到电网的供需平衡,直接影响着并网系统的运营成本,因此风电功率预测的准确性非常重要。对于预测精度不高的问题,提出了一种改进的果蝇算法优化的支持向量机的预测方法。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,因而利用改进的果蝇算法对支持向量机参数进行优化,用优化好的参数进行建模训练,然后把建好的模型应用于功率预测,最后对数据进行评估。预测结果表明:改进的果蝇算法优化的支持向量机对风力发电功率预测有更好的准确性。

The forecast accuracy of the wind power directly affects the operating cost of the network system,which is directly related to the supply and demand balance of the grid. Therefore, the forecast accuracy of wind power is very important. Considering the prediction accuracy is not high, we propose an improved predictive method that is based on MFOA-SVM. Since penalty factor and kernel parameters of SVM have a great impact on the prediction accuracy, the improved FOA optimizes the parameters of support vector machine and trains model with a good parameter optimization. Finally, the built model is used to the power prediction to evaluate the data. The prediction results show that the improved MFOA- SVM can produce better accuracy for wind power prediction.