Cooperative Optimization Method For Global Vehicle Speed And Energy Management Of Plug-in Hybrid Electric Vehicle

Plug-in hybrid electric vehicles have much higher fuel saving rate than hybrid electric vehicles,and there is no widely criticized "mileage anxiety" problem in pure electric vehicles,so they are widely concerned by the industry in recent years.However,due to the existence of two larger power systems,the acquisition cost of PHEVs is still high,and the fuel economy of PHEV needs to be further enhanced to offset the increase in its acquisition cost.With the development of Internet of vehicles and the improvement of vehicle intelligence,it is possible to accurately predict road and traffic information.By integrating the road and traffic information,the vehicle control system can not only plan the most fuel-saving driving speed profile,but also realize predictive energy management,thus further improving the operation efficiency of the power system.Obviously,vehicle speed planning and predictive energy management are two coupled problems.Therefore,collaborative optimization of vehicle speed and power system energy management is needed which may lead to a sharp increase in computing costs and cannot meet the needs of real-time vehicle applications.In order to solve this problem,this paper takes an extended-range electric school bus as the research object,and proposes a framework for collaborative planning of vehicle speed and electricity based on two-layer iterative optimization.The two key issues of "predictive energy management" and "double-layer iterative optimization" in the framework were studied.First,through a large number of numerical simulation experiments,it is found that the global optimal state of charge(So C)has an approximately linear relationship with the accumulated drive energy.This feature is of great significance for the rapid planning of the global So C trajectory.In order to find the theoretical explanation behind the global optimal feature,Pontryagin’s Minimum Principle(PMP)is used to analyze and derive the solution of the global optimal energy management problem.The results show that: i)the engine generator set(APU)will only start when the DC bus demand power exceeds the threshold,and the threshold is related to the cumulative demand energy of the entire trip;ii)the output power of the APU is constant when it starts.Based on these two conclusions,this paper compares and analyzes the ratio of So C variation to driving power in hybrid and pure electric modes,and demonstrates the basic reason why the global So C has an approximately linear relationship with accumulated driving energy.Based on the linear characteristics of the global optimal So C trajectory and the conclusions derived from the above theory,two new predictive energy management strategies are proposed: i)A regular energy management strategy based on total energy prediction is proposed;ii)An energy management strategy based on fast global SOC trajectory planning and analytic optimization of instantaneous power allocation is proposed.The simulation results under different driving cycles show that the control effects of the two strategies are close to the global optimal solution obtained by offline PMP.Although the calculation efficiency of the former is better than the latter,its control performance has a strong dependence on the accuracy of total demand energy prediction,while the latter only needs to predict traffic flow and terrain information,and is more robust to prediction errors.Combining the above-mentioned second predictive energy management strategy,a collaborative optimization method of global vehicle speed and power system energy management is proposed.This method is divided into two layers.The upper layer uses the forward dynamic programming algorithm to solve the vehicle speed planning problem,and the lower layer uses the above-mentioned second method to solve the energy management optimization problem.The upper and lower layers iterate repeatedly until the difference in energy consumption is less than the set value.The simulation results show that the proposed method can realize the collaborative optimization of vehicle speed and So C under the premise of ensuring real-time performance,and has similar fuel economy to the results obtained by using dynamic programming offline optimization.Finally,a corresponding model-in-the-loop testing(MIL)platform was built.The test results show that the two predictive energy management strategies proposed in this paper can achieve near-global optimal fuel economy,and the iterative process can improve the fuel economy of the global vehicle speed and energy management collaborative optimization method by nearly 1%.