www.ChinaAET.comComputerTechnologyandItsApplications计算机技术与应用基于levy飞行优化BOA-BP网络的电池SOC估计*李畅,王琪,姜佳怡(西安工业大学电子信息工程学院,陕西西安710021)摘要:目前电动汽车动力输出的来源主要是动力电池,其荷电状态(StateofCharge,SOC)表示电池的剩余电量情况,精确估算SOC对于电池的使用安全有重要意义。将蝴蝶优化算法(ButterflyOptimizationAlgorithm,BOA)进行改进并用于优化BP神经网络估算动力电池SOC,解决了普通BP网络估计SOC时遇到的训练时间长、收敛慢、精度较低、易陷入局部最优解的问题;同时提升了全局搜索速度,选取电压和电流为输入变量、SOC为输出变量,根据误差的大小调整神经网络的权值和阈值。仿真结果表明,优化后得到的SOC估计结果误差率控制在1.1%以内,该方法寻优速度快,具有更好的鲁棒性。关键词:荷电状态估计;蝴蝶优化算法;BP神经网络;Levy飞行中图分类号:TP13文献标志码:ADOI:10.16157/j.issn.0258-7998.222834中文引用格式:李畅,王琪,姜佳怡.基于levy飞行优化BOA-BP网络的电池SOC估计[J].电子技术应用,2023,49(4):88-91.英文引用格式:LiChang,WangQi,JiangJiayi.BatterySOCestimationbasedonLevyflightoptimizationofBOA-BPnetwork[J].ApplicationofElectronicTechnique,2023,49(4):88-91.BatterySOCestimationbasedonLevyflightoptimizationofBOA-BPnetworkLiChang,WangQi,JiangJiayi(CollegeofElectronicInformationEngineering,Xi′anTechnologicalUniversity,Xi′an710021,China)Abstract:Atpresent,thepoweroutputofelectricvehiclesismainlyderivedfrompowerbatteries,whoseStateofCharge(SOC)representstheremainingpowerofbatteries.AccurateestimationofSOCisofgreatsignificanceforthesafetyofbatteryuse.But‐terflyOptimizationAlgorithm(BOA)wasimprovedandusedtooptimizeBPneuralnetworktoestimateSOCofpowerbattery,whichsolvedtheproblemsoflongtrainingtime,slowconvergence,lowaccuracyandeasytofallintolocaloptimalsolution.Atthesametime,theglobalsearchspeedisimproved,voltageandcurrentareselectedasinputvariables,SOCasoutputvariables,andtheweightandthresholdofneuralnetworkareadjustedaccordingtothesizeoferror.SimulationresultsshowthattheerrorrateofSOCestimationresultsobtainedafteroptimizationiscontrolledwi...