基于K_means++聚类与RF_GRU组合模型的电力负荷预测方法研究

Research on Power Load Forecasting Method Based on K_means++Clustering and RF_GRU Combined Model

短期负荷预测是电力系统对负荷进行规划的重要依据之一,为了进一步提升短期负荷预测的精度,提出一种基于K_means++聚类与RF_GRU组合模型的电力负荷预测方法。首先,采用K_means++聚类算法将负荷群体分成负荷特性相近的群体,然后利用改进后的CSO算法优化随机森林中的相关参数使其性能最优,接着根据聚类情况采用随机森林选择结构不同的多层GRU网络分别对各组负荷群体进行预测,最后将所有组的预测结果相加得出最终预测值。算例结果表明,聚类算法的归纳整理功能为预测方法节省了预测时间,而采用组合模型又进一步提高了预测精度。

Short-term load forecasting is one of the important basis for power system load planning.In order to further improve the accuracy of short-term load forecasting,a power load forecasting method based on K_means++clustering and RF_GRU com-bined model is proposed.First,the K_means++clustering algorithm is used to divide the load groups into groups with similar load characteristics,and then the improved CSO algorithm is used to optimize the relevant parameters in the random forest to achieve the best performance,and then the random forest is used to select multiple groups with different structures according to the clustering situation.The hierarchical GRU network separately predicts each group of load groups,and finally adds up the prediction results of all groups to obtain the final prediction value.The results of the calculation examples show that the induction and sorting function of the clustering algorithm saves the forecasting time for the forecasting method,and the use of the combined model further improves the forecasting accuracy.