CPSO-LSSVM算法在车载电池SOC预测中的应用  

Application of CPSO-LSSVM Algorithm in Vehicular Battery SOC Prediction

作　　者：谈发明 王琪 TAN Faming;WANG Qi(Information Center, Jiangsu University of Technology,Changzhou 213001,Jiangsu,China;School of Electrical and Information Engineering, Jiangsu University of Technology,Changzhou 213001,Jiangsu,China)

机构地区：[1]江苏理工学院信息中心,江苏常州213001 [2]江苏理工学院电气信息工程学院,江苏常州213001

出　　处：《实验室研究与探索》2018年第8期110-114,共5页Research and Exploration In Laboratory

基　　金：江苏省高等学校自然科学研究面上项目(17KJB470003);江苏高校自然科学基金(15KJB470004)

摘　　要：针对车载电池SOC难以精确预测的问题,提出以CPSO算法优化LSSVM模型参数,避免了参数选择的盲目性,提高了测量精度及泛化能力。利用ADVISOR软件采集车载电池各项性能参数,其中,电流、电压及温度数据作为CPSO-LSSVM预测模型的输入,SOC作为预测模型的输出。验证结果表明:CPSOLSSVM相比PSO-LSSVM预测模型预测最大绝对误差降低了3.06%,平均相关误差降低了0.35%,为车载电池SOC的预测提供一新方法。In order to solve the problem that the estimation of vehicle battery SOC is difficult to be accurate,an algorithm is proposed to optimize LSSVM with CPSO.The blindness of parameter selection is avoided,and the measurement accuracy and generalization ability are improved.The performance parameters of the vehicle battery are collected by software ADVISOR,the current,voltage and temperature data are used as CPSO-LSSVM predictive model inputs,and SOC is used as predictive model output.Compared with the PSO-LSSVM prediction model,the verification results show that the maximum absolute error of CPSO-LSSVM is reduced by 3.06%and the average correlation error is reduced by 0.35%.Hence,it provides a new method for the prediction of vehicle battery SOC.

关 键 词：混沌粒子群优化 最小二乘支持向量机 荷电状态 预测 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]