基于改进支持向量机算法的光伏发电功率预测  

Prediction for photovoltaic generation power based on improved support vector machine algorithm

作　　者：郑思达 刘影 杨磊 杨晓坤 ZHENG Si-da;LIU Ying;YANG Lei;YANG Xiao-kun(Electric Power Research Institute,State Grid Jibei Electric Power Co.Ltd.,Beijing 102208,China;Qinhuangdao Power Supply Company,State Grid Jibei Electric Power Co.Ltd.,Beijing 102208,China)

机构地区：[1]国网冀北电力有限公司电力科学研究院,北京102208 [2]国网冀北电力有限公司秦皇岛供电公司,北京102208

出　　处：《沈阳工业大学学报》2022年第4期378-382,共5页Journal of Shenyang University of Technology

基　　金：国家自然科学基金项目(51807051);国网浙江省电力有限公司科技项目(5211TZ170006).

摘　　要：针对当前光伏发电间断性和随机性对电流影响较大的问题,基于改进支持向量机算法,提出光伏发电功率预测方法.通过灰色关联度优化气象影响因素参数,改进支持向量机算法,建立多气象因素协同约束的光伏电池等效模型,得到光伏电池的串联和并联电阻;将训练数据的总数映射到特征空间中,筛选数据并定义特征空间中的超平面,计算光伏发电的有效功率并降低误差.结果表明,预测结果与实际结果符合程度高,功率预测的绝对误差小,有效性和应用性得到验证.Aiming at the discontinuity and randomness of photovoltaic generation with great influence on electric current,a prediction method based on the improved support vector machine algorithm for photovoltaic generation power was proposed.Through the grey correlation degree,the parameters of meteorological factors were optimized,the support vector machine algorithm was improved,and a photovoltaic cell equivalent model with the collaborative constraints by multiple meteorological factors was established to obtain the series and parallel resistances of photovoltaic cells.The total number of training data was mapped to the feature space,the data were screened and the hyperplane in feature space was defined,the effective power of photovoltaic generation was calculated for error reduction.The results show that the predicted results are in good agreement with the actual values,with the absolute error of power prediction getting smaller and the validity and applicability being verified.

关 键 词：支持向量机算法 光伏发电 发电机功率 灰色关联度 精准度 气象因素 特征空间 有效功率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TM614[自动化与计算机技术—计算机科学与技术]