基于BP神经网络的纯电动汽车动力传动系统效率建模及分析  

Modeling and Analysis of the Drivetrain Efficiency for Pure Electric Vehicles based on BP Neural Network

作　　者：陈柯序 李海波 赵小娟 阴晓峰 孙超 窦畅 CHEN Kexu;LI Haibo;ZHAO Xiaojuan;YIN Xiaofeng;SUN Chao;DOU Chang(Institute of Automotive Engineering,Xihua University,Chengdu 610039,China;Sichuan Provincial Tianxun Technology Co.,Ltd.,Chengdu 610057,China;Zhitong Testing Technology Co.,Ltd.,Changzhou Jiangsu 213000,China)

机构地区：[1]西华大学汽车工程研究所,成都610039 [2]四川省天循科技有限责任公司,成都610057 [3]中质智通检测技术有限公司,江苏常州213000

出　　处：《机械设计与研究》2021年第5期180-185,共6页Machine Design And Research

基　　金：国家自然科学基金资助项目(51375402);四川省科技计划资助项目(2021YFQ0052、2019YFG0367);西华大学研究生创新基金项目(YCJJ2020072)。

摘　　要：精确的动力传动系统效率模型是优化纯电动汽车换挡规律以充分挖掘整车性能潜力的前提。基于电机和机械式自动变速器(Automated Manual Transmission,AMT)效率的主要影响因素制定了动力传动系统效率实验方案,在采集的电机和AMT效率数据的基础上,应用反向传播(Back Propagation,BP)神经网络建立了电机和AMT各挡效率模型,进而获得动力传动系统效率模型,分别分析输入转矩、输入转速、润滑油温度、挡位对动力传动系统效率的影响。所建立的效率模型,能准确揭示动力传动系统效率随工况的变化规律,为考虑效率变化的换挡规律优化奠定了基础。Accurate drivetrain efficiency model is the premise of optimizing the shift schedule of pure electric vehicle to fully exploit the performance potentials of the vehicle.The experiment scheme for the drivetrain efficiency collection is formulated according to the main influencing factors of efficiency of the motor and automated manual transmission(AMT).Based on the motor and AMT efficiency data,the efficiency models of motor and AMT are established using back propagation(BP)neural network.These models are combined to formulate the drivetrain efficiency models for each gear position.The influence of input torque,input rotate speed,oil temperature,and gear position on the efficiency of the drivetrain are analyzed,respectively.The built efficiency model can accurately reveal variations in the drivetrain efficiency due to changes in working conditions,which lays a foundation for the shift schedule optimization considering the change of efficiency.

关 键 词：动力传动系统 BP神经网络 效率模型 电机 机械式自动变速器 

分 类 号：U469.72[机械工程—车辆工程]