混合动力汽车匀速下坡再生制动模型预测控制

Model Predictive Control of Regenerative Braking for Hybrid Electric Vehicle Cruising Downhill

基于车载导航系统(GPS/GIS等)所提供的未来一段预测路线上的汽车运行状态信息,建立中度混合动力汽车再生制动能量回收的全局优化动态规划模型;采用模型预测控制方法,将动态规划的全局优化控制策略转化成预测视距内的局部优化算法,实现滚动优化控制;为解决动态规划中的维数灾问题,确定了电池荷电状态和温度的可达区域;对模型预测控制策略、全局优化控制策略和瞬时优化控制策略进行了计算比较,在不同坡度、不同坡长的匀速下坡工况下的仿真表明:模型预测算法的计算效率显著高于全局优化策略的;应用模型预测控制策略的再生制动能量回收效率明显高于瞬时优化控制策略的,相比全局优化策略的降低不到1.31%,且采用档位提示的模型预测控制策略能量回收效果更好。

Based on the driving states of vehicles equipped with GPS/GIS on board in the future predictive route,a global optimal dynamic programming model of regenerative braking energy recovery for medium hybrid electric vehicle was established.In order to realize a receding horizon optimal control,the global optimal dynamic programming algorithm was converted into a local optimal algorithm within prediction horizon using with the model predictive control method.To overcome the curse of dimensionality of dynamic programming,the rechargeable ranges of SOC and temperature of battery were determined.The calculation comparison among the control strategies of model prediction,global optimization and instantaneous optimization was performed.The simulation under the conditions of different gradients and slope lengths,and cruising downhill was performed.The results show that（1） the computational efficiency of model predictive control strategy is higher than that of global optimal control strategy;（2） energy recovery efficiency of regenerative braking of model predictive control strategy is greater than that of instantaneous optimal control strategies,and not deduce to 1.31% compared with that of global optimal control strategy,the energy recovery is better with shift reminder.