基于工况识别的混合动力汽车预测控制

基于工况识别的混合动力汽车预测控制重庆大学硕士学位论文（学术学位）学生姓名：熊张林指导教师：舒红副教授专业：车辆工程学科门类：工学重庆大学机械工程学院二O一四年五月PredictiveControlforaHybridElectricVehicleBasedonDrivingcycleRecongnitionAThesisSubmittedtoChongqingUniversityinPartialFulfillmentoftheRequirementfortheMaster’sDegreeofEngineeringByXiongZhanglinSupervised:AssociateProf.ShuHongSpecialty:VehicleEngineeringCollegeofmechanicalengineeringofChongqingUniversity,Chongqing,ChinaMay2014中文摘要I摘要插电式混合动力汽车(Plug-inHybridElectricVehicle,PHEV)兼具纯电动汽车和传统燃油汽车的优点，在节能减排方面具有明显的优势。本文以插电式并联型混合动力汽车为研究对象，借助车载导航系统提供的汽车未来运行状态信息，以提高燃油经济性为目标对PHEV预测控制策略进行研究。主要研究内容如下：①建立了插电式混合动力汽车动力系统各部件数学模型，并采用粒子群算法获得了发动机燃油消耗率、蓄电池内阻和电动势的拟合计算公式，提高了动态规划计算效率。②将发动机油耗和电机耗电量用价格统一起来，提出了以汽车每次总行驶里程费用最低为价值函数的插电式混合动力汽车动态规划全局优化策略。并通过在价值函数中加入附加代价函数来避免发动机频繁起停。分别对UDDS、NEDC、1015工况进行动态规划全局优化仿真，获得汽车在一般工况电池荷电状态(SOC)最优变化规律。对长上坡、长下坡、怠速停车等特殊工况进行仿真分析，获得汽车在一般工况和特殊工况的SOC变化规律。③为了使预测控制逼近全局优化控制效果，将汽车行驶工况分为一般、过渡和特殊工况，提出汽车工况类型识别方法。制定了基于SOC最优轨迹的一般工况模型预测控制策略、基于模糊规则的过渡工况控制策略，以及基于简单规则的特殊工况控制策略。通过这三套控制策略的交替使用，实现混合动力汽车的预测控制。④采用C语言与MATLAB/Simulink建立了基于工况识别的插电式混合动力汽车预测控制仿真模型。在几种标准工况下模型预测控制仿真得到的SOC曲线能够较好跟随最优SOC轨迹。对特殊工况进行预测控制仿真分析，仿真表明各种运行模式能够合理切换，在过渡工况能够发挥提前充放电功能，发动机、电机大部分工作在低油耗、高效率区域，表明作者提出的基于工况识别的混合动力汽车预测控制能有效提高汽车燃油经济性。利用dSPACE实时仿真系统对预测控制算法的运算时间进行测试，结果表明所提出的基于工况识别的预测控制算法能够满足实时控制要求。关键词：混合动力汽车，动态规划，工况识别，预测控制，仿真重庆大学硕士学位论文II英文摘要IIIABSTRACTPlug-inHybridElectricVehicle(PHEV)hastheadvantagesofpureelectricsandconventionalvehicles,andobviousadvantagesinenergyconservationandemissionsreduction.EnergymanagementstrategyisthecoreofthedevelopmentofPHEV,itsqualityaffectsthevehicleperformancedirectly.ThepaperregardsPHEVastheresearchobject,theresearchonthecontrolstrategyofPHEVwasconducted,withtheobjectiveofimprovingbasedonthefuturestateinformationofvehiclesprovidedbyon-boardnavigationsystem.Themaincontentsofthispaperareasfollows.①ThemathematicsmodelofPHEV’spowertrainwereestablished,.Inordertoimprovecomputingefficiencyofdynamicprogramming,thecalculationformulaofenginefuelconsumptionrate,batteryinternalresistanceandelectromotiveforcewerefittingoutbasedonparticleswarmoptimization(PSO).②Thefuelconsumptionofengineandpowerconsumptionofmotorwereunifiedwithprice,theglobaloptimizationstrategyofPHEVwasputforwardtoachievethelowestcostineverywholemileage.Theadditionalcostfunctionwasaddintotheobjectivefunctiontoreducetheadditionalfuelconsumptioncausedbythefrequentstart-stopofengine.ThesimulationcalculationofglobaloptimizationunderdifferentconditionsofUDDS,NEDC,1015wereconductedrespectively.ThechangerulesofSOCcurveundergeneralconditionswereabletobeobtainedbythesimulation.thechangerulesofSOCundergeneralconditionsandspecialconditionswereobtainedbythesimulationandanalysisonthenewconstitutedcycleconditions,suchasthelonguphill,longdownhillandidlestopconditions.③Inordertomakepredictivecontrolapproximatetoglobaloptimizationcontroleffect,Thevehicleworkingconditionswereclassifiedtogeneralcondition,transientconditionandspecialcondition,recognitionmethodsaboutfutureconditionswereproposed.Controlmethodsaboutfutureconditionswereproposed.ThemodelpredictivecontrolstrategybasedonmajorizedSOCTrajectoryundergeneralconditions,thefuzzycontrolstrategyundertransientconditions,andthecontrolstrategybasedonrulesunderspecialconditionswerestudied.ThepredictivecontrolofHEVscanberealizedbythealternateuseofthesethreecontrolstrategies.④ThepredictivecontrolmodelofPHEVsbasedondrivingconditionreco-ngnitionbylanguageCandMATLAB/Simulinkwasestablished,Thesimulate重庆大学硕士学位论文IVdSOCcurveofmodelpredictivecontrolunderseveralstandardconditionscanfollowtheoptimaltrajectoryofSOCbetterintheglobaloptimization.Thesimulationanalysisofpredictivecontrolunderspecialworkingconditionsindicatedthatallkindsofmodescanbeswitchedreasonably,intheearlytochargeanddischargefunctioncanplayundertransientconditions,theengineandmotormostlyworkintheregionofthelowoilconsumptionandhighefficiency,whichsuggeststhatthepredictivecontrolofPHEVtheauthorputforwardbasedontherecognitionofworkingconditionscaneffectivelyimprovethefueleconomy.Finally,operationtimeofpredictivecontrolstrategybasedonthedSPACEreal-timesimulationsystemwastested,andtheresultshaveshownthattheproposedpredictivecontrolalgorithmcanmeettherequirementsofreal-timecontrol.Keywords:HybridElectricVehicle,DynamicProgramming,ConditionRecognition,Predictivecontrol,Simulation目录V目录中文摘要..........................................................................................................................................I英文摘要.......................................................................................................................................III1绪论..............................................................................................................................................11.1选题背景和意义.......................................................................................................................11.2插电式混合动力汽车概述.......................................................................................................11.2.1PHEV研究现状....................................................