神经遗传增程式电动汽车控制策略  

Neural networks and genetic extended-range control for electric vehicle

作　　者：张袅娜[1] 李昊林 李彦阳[1] 周长哲[1] 

机构地区：[1]长春工业大学电气与电子工程学院,吉林长春130012

出　　处：《长春工业大学学报》2017年第1期48-52,共5页Journal of Changchun University of Technology

基　　金：吉林省省级经济结构战略调整引导资金专项项目(2014Y126);长春市科技局重大科技攻关计划基金资助项目(14KG095)

摘　　要：结合恒温控制和功率跟随控制策略,在增程器开关启动时以动力电池的SOC和APU的需求功率作为判断条件。采用模糊规则,依据汽车荷电状态和驾驶员踏板开度建立APU需求功率模型。需求功率和功率阈值判断条件与荷电状态和电量阈值判断条件相结合,实现了不同状态下的增程器开关控制。通过Cruise仿真在FTP75工况下得出在不同功率阈值、电量阈值情况下电动汽车油耗的数值,采用神经网络遗传算法实现对燃油消耗率函数值的寻优。Constant thermostatic control is combined with power follow control to turn on a range extenderswitch with battery SOC and APU required power as judgement conditions.Fuzzy rule is applied to build the APU required power model based on the chargestate of vehicles and pedal opening.The required power power threshold is combined with charge state charge threshold to control the range extender switch for different operations.Both the power threshold and charge threshold are obtained with Cruise simulation under FTP75 conditions,and Neural Network Genetic Algorithm is used to optimize the fuel consumption function.

关 键 词：增程式电动汽车 控制策略 神经网络遗传算法 燃油消耗率 

分 类 号：U463.6[机械工程—车辆工程]