混合动力汽车模型预测控制策略研究

Research on Model Predictive Control Strategies for Hybrid Electric Vehicles

针对传统混合动力汽车控制方法无法实现实时最优控制的问题,提出了基于简化混合动力汽车系统模型的预测控制智能优化策略。通过将3自由度的系统模型简化为1自由度的系统模型,并采用连续广义最小残量方法求解模型预测控制问题。运用MATLAB/Simulink与GT-POWER联合仿真平台进行仿真,实验结果验证了系统模型简化的有效性,以及所设计的模型预测控制算法大幅度提高混合动力汽车的燃油经济性的能力和实时控制性能。

This paper proposed model predictive control intelligent optimization strategies based on system model simplification for hybrid electric vehicles to deal with computation burden and on-line optimization problems in conventional control strategies. The 3 degrees of freedom system model was reduced to 1 degree of freedom system model. The model predictive control problem was solved using continuation/generalized minimum residual method. The simulation was conducted using MATLAB/ Simulink and GT-POWER co-simulation platform. The results showed that the proposed model simplification is effective, and the proposed model predictive control method can improve fuel economy significantly and can be implemented in real-time.