PHEV转矩的自适应神经模糊推理系统分配策略  

PHEV Torque Distribution Strategy Based on Adaptive Neural-Fuzzy Inference System

作　　者：郭翠玲[1] GUO Cui-ling(Department of Mechanical and Electrical Engineering,Shangqiu Polytechnic,He’nan Shangqiu476000,China)

机构地区：[1]商丘职业技术学院机电工程系,河南商丘476000

出　　处：《机械设计与制造》2020年第6期157-162,共6页Machinery Design & Manufacture

基　　金：国家自然科学基金(25486855)。

摘　　要：为了减小混合动力汽车的燃油消耗和减少尾气中有毒气体含量,提出了迭代动态规划与自适应神经模糊推理系统相结合的能量管理方法.建立了并联式混合动力汽车动力系统模型,以车速和加速踏板开度为依据给出了换挡策略;建立了能量管理问题模型,使用迭代动态规划方法求解了最优控制律,但是此控制律无法应用于车辆的实时控制;提出使用自适应神经模糊推理系统探索最优控制律与车辆状态间的非线性映射关系,根据车辆状态确定最优控制律,实现了最优控制律的实时控制.使用CCBC工作作为验证工况,与基于规则控制策略、等效燃油消耗最小方法相比,控制方法的百公里油耗分别减少了18.56%、3.93%,且有毒气体含量明显少于另外两种方法,证明了控制方法的优越性.In order to decrease fuel consumption and toxic gas content of HEV,energy management method combining iteration dynamic programming and adaptive neural-fuzzy inference system is proposed.Dynamic system model of PHEV is built,and gear-shifting strategy is made depend on vehicle speed and opening degree of accelerator pedal.Energy management model is built,and optimal control law is solved by iteration dynamic program,but the control law cannot used to control vehicle real-time.Adaptive neural-fuzzy inference system is used to build nonlinear mapping relation between optimal law and vehicle state,so that optimal control law can be curtained by vehicle state,which means optimal control law can control the vehicle real-time.Using CCBC working condition as testing condition,compared with control strategy rule-based and ECMS,fuel consumption per 100km of the control method in this essay decreases by 18.56%and 3.93%,and toxic gas content is less than the other two methods,which can prove superiority of control method in this essay.

关 键 词：并联式混合动力汽车 转矩分配 换挡策略 自适应神经模糊推理系统 

分 类 号：TH16[机械工程—机械制造及自动化] TP273[自动化与计算机技术—检测技术与自动化装置]