基于网格化负荷预测的智能电力控制技术研究

Research on intelligent power control technology based on grid load forecasting

针对配电网负荷预测准确性偏低、负荷与新能源协同度不足的问题,文中提出了改进初始聚类中心的选取方法,通过引入相关的权重系数来缩小不同网格之间的负荷差异,以提高预测的准确性。同时还构建了配电网多目标电力控制模型,该模型综合考虑了运行成本和弃风弃光量最小、负荷与新能源匹配度最大的优化目标,并采用NSGA-II算法来进行求解。仿真结果表明,所提算法能够有效提高负荷预测的准确度,相比于传统算法,其平均预测误差由6.01%降低为2.99%。而所设计的多目标电力控制模型不仅能够提升配电网运行的经济性,还可以保障新能源消纳,且优化控制后配电网的运行成本及弃风弃光量分别下降了7.6%和8.9%,负荷与新能源匹配度则提升了3.8%。

Aiming at the problems of low accuracy of load forecasting in distribution network and insufficient coordination between load and new energy,this paper proposes to improve the selection method of initial clustering center and introduce load weight coefficient to reduce the load difference between different grids and improve the accuracy of load forecasting.The multi-objective power control model of distribution network is constructed,and the optimization objectives of minimum operation cost,minimum wind and light waste,and maximum matching degree of load and new energy are comprehensively considered,and NSGA-II algorithm is used to solve the model.The simulation results show that the proposed algorithm can effectively improve the accuracy of load forecasting.Compared with the traditional grid load forecasting algorithm,the average prediction error is reduced from 6.01%to 2.99%.The proposed multi-objective power control algorithm can improve the operation economy of the distribution network and ensure the absorption of new energy sources.After the optimization control,the operation cost of the distribution network is reduced by 7.6%,the amount of wind and light discarded is reduced by 8.9%,and the matching degree of load and new energy is increased by 3.8%.