基于敏感气象特征因子筛选与PSO-SVM模型优化的新能源功率预测特性研究  

Research on Power Prediction Characteristics of New Energy Based on Sensitive Meteorological Feature Factor Screening and PSO-SVM Model Optimization

作　　者：巩伟峥 GONG Weizheng(East Subsection of State Grid Corporation of China,Shanghai 200120,China)

机构地区：[1]国家电网有限公司华东分部,上海200120

出　　处：《电器与能效管理技术》2025年第3期38-45,共8页Electrical & Energy Management Technology

摘　　要：基于新能源电力系统的不断建设,研究新能源功率与气象间的关联特性迫在眉睫,提出一种基于敏感气象特征因子筛选与粒子群优化支持向量机(PSO-SVM)模型调优的新能源功率滚动预测算法。首先基于皮尔逊相关系数、互信息熵分析气象因子与功率的关联特性,并基于D-S证据理论计算优化组合后的相关性指标筛选敏感气象特征因子,利用粒子群优化(PSO)算法对支持向量机(SVM)新能源发电预测模型进行参数全局调优。然后结合新能源运行数据,建立滚动预测模型。最后通过实验验证分析,结果表明所提预测模型可有效提升新能源发电预测精度。With the continuous construction of new power systems,it is extremiy urgent to study the correlation characteristics between new energy power and meteorology.A new energy power rolling prediction algorithm based on sensitive meteorological factor feature screening and PSO-SVM model optimization is proposed.Firstly,based on the Pearson correlation coefficient and mutual information entropy,the correlation characteristics between meteorological factors and new energy power are analyzed.Based on the D-S evidence theory,the optimized combination of correlation indicators is calculated to screen sensitive meteorological feature factors.The particle swarm optimization(PSO)algorithm is used to globally optimize the parameters of the support vector machine(SVM)new energy power generation prediction model.Then,combined with massive new energy operation data,a rolling prediction model is established.Finally,through experimental verification and analysis,the results show that the proposed prediction model can effectively improve the accuracy of new energy generation prediction.

关 键 词：新能源 敏感气象特征因子 特征筛选 PSO-SVM模型 滚动预测 

分 类 号：TM73[电气工程—电力系统及自动化]