基于DP的随机模型预测控制能量管理研究

Research on Energy Management of Stochastic Model Predictive Control based on DP

基于模型预测控制的基本原理,对超轻度混合动力汽车的能量优化问题进行了研究。采用马尔可夫模型预测加速度变化过程,通过计算获得混合动力汽车未来需求转矩。建立以燃油经济性最优为回报函数的混合动力汽车能量管理优化模型,采用动态规划算法(DP)进行有限时域内的滚动优化求解。在MATLAB/Simulink平台上搭建仿真模型,仿真结果表明,提出的研究方法能够得到最优的转矩分配,可有效解决能量管理优化问题。

Based on the basic principles of model predictive control,the energy optimization problem of ultra-mild hybrid electric vehicles is studied.The Markov model is used to predict the acceleration change process,and the future required torque of the hybrid electric vehicle is obtained through calculation.A hybrid electric vehicle energy management optimization model with optimal fuel economy as the reward function is established,and a dynamic programming algorithm(DP)is used to perform rolling optimization in a limited time domain.A simulation model is built on the MATLAB/Simulink platform.The simulation results show that the proposed research method can obtain the optimal torque distribution and can effectively solve the energy management optimization problem.