Historical Data Based Online Optimal Control Strategy For Power-split Hybrid Electric Bus

Plug-in hybrid electric city bus is an important part of the new energy commercial vehicle in China.In the current various kinds of hybrid electric configuration,the planetary hybrid electric bus has achieved good promotion effect because of its excellent fuel economy.However,with the sales of planetary hybrid electric bus in different regions and cities,the energy management strategy based on artificial calibration is difficult to guarantee the optimal energy saving effect under different driving conditions.And the huge amount of calibration efforts increase the promotion cost significantly.On the other hand,planetary hybrid system is a low damping system with highly coupled power sources.And core power sources have complex time-varying response characteristics.The executive layer controller based on fixed parameters is difficult to adapt to the complex dynamic characteristics of the system.Then longitudinal shock may be easily caused.Therefore,solving the problem of optimal energy saving and high-robustness dynamic coordination control under complex driving conditions will provide an important guarantee for reducing vehicle promotion cost and enhancing vehicle competitiveness.Based on the in-depth survey of current online optimal control strategies,the online optimal control strategy should solve the contradiction of optimality,robustness and real-time performance.According to this,this paper presents a design method of online optimal control strategy based on historical data.(1)Parameter identification method and agent model technology are applied to build the hybrid electric system model based on historical data,which provides accurate computing environment for the development of online optimal control strategy,and also provide accurate input for the model-based controller design.(2)Representative driving cycle is synthesized based on historical data.The representative driving cycle ensures the robustness of the optimal control strategy.(3)Global optimization is carried out based on the representative driving cycles to derive the optimal control strategy.In order to avoid the "dimension disaster" of the global optimization,this paper designs a hierarchical optimization framework.Optimal allocation of engine power and battery power is achieved by using the dynamic programming algorithm,and then the instantaneous optimal control strategy is implemented to determine the engine operating point.(4)To ensure the real-time performance of the online optimal control strategy,the mapping relationship between the energy consumption characteristics and the global optimal control rules are established based on the in-depth analysis of the global optimization results.Then an automatic control rule extraction method is put forward to obtain the near-optimal control rules.(5)A data-driven based model predictive controller is designed for the mode switching process to ensure the driving comfort,and a controller combined with extended Kalman filter and model predictive controller is built for the stable operation stage of the system to ensure the implementation of the optimal energy management strategy.The validation of the online optimal control strategy design method proposed in this paper is carried out through offline simulation and hardware-in-loop simulation.And the optimal control strategy is validated both on the Chinese City Bus Cycle and the historical driving cycles of the hybrid electric bus.Simulation results verify the near-optimal energy saving effect and the real-time performance of the optimal control strategy.In addition,better driving comfort is achieved with the dynamic coordinated control strategy proposed in this paper,in comparison with the PID controller.And the optimal operation points are well controlled with the dynamic coordinated control strategy,which guarantees the implementation of the optimal control strategy.