基于粒子群优化支持向量机的短期负荷预测  

Research on short term load forecasting based on particle swarm optimization-support vector machine

作　　者：杨小明[1] 崔雪[1] 周斌[1] 彭政[1] YANG Xiaoming;CUI Xue;ZHOU Bin;PENG Zheng(School of Electrical Engineering,Wuhan University,Wuhan 430072,China)

机构地区：[1]武汉大学电气工程学院

出　　处：《武汉大学学报（工学版）》2018年第8期715-720,共6页Engineering Journal of Wuhan University

基　　金：国家自然科学基金项目(编号:51177112)

摘　　要：随着我国电力行业体制改革的不断深入,对负荷预测精度的要求也不断提高.为提高短期负荷预测的精度,针对单一核函数最小二乘支持向量机（LS-SVM）模型在实际使用中的问题,将高斯（RBF）核函数与多项式（Poly）核函数进行组合,得到了新的混合核函数,从而提高了SVM模型的学习能力与泛化能力;采用基于双种群的粒子群算法（DP-PSO）寻求混合核函数LS-SVM模型的最优参数;结合实际电网进行日平均负荷预测,算例结果表明,采用该负荷预测模型的平均误差仅为1.22%,预测精度较高.With the development of the process of power market in China,it put forward higher requirements of the accuracy and intelligence of load forecasting.In order to improve the accuracy of short-term load forecasting of power system,and overcome the defect of single kernel function of least squares support vector machine（LSSVM）model,this paper builds a mixed function kernel support vector machine with Gauss kernel function（Radial Basis Funtion,RBF）and polynomial kernel function（Poly）based on the weight of a certain combination,which can effectively improve the learning ability and generalization ability of LSSVM prediction model;through the particle swarm optimization（PSO）algorithms based on double population particle swarm optimization（DP-PSO）to get the optimum parameter values of the LSSVM model.The results of actual power grid of daily average load forecasting show that the average error of the load forecasting model is only 1.22%,and it has a certainpractical value.

关 键 词：支持向量机 负荷预测 粒子群优化算法 最小二乘 

分 类 号：TM715[电气工程—电力系统及自动化]