Energy Efficiency Optimization And Control Of Range-extended Electric Vehicles For Low-temperature Environment

As the global greenhouse effect and energy issues become increasingly severe,promoting vehicle electrification,and developing hybrid electric vehicles(HEVs)has become a significant trend.Range-extended electric vehicles(REEVs)are series HEVs,and can be used as an effective transition solution before pure electric vehicle technology matures.As low temperatures lead to increased energy consumption for cabin heating,reduced battery performance and accelerated aging,and reduced engine efficiency,REEVs suffer from worsened energy consumption and serious performance degradation in cold environments,making them unable to meet market demand in vast areas of China.Considering the relationship between the thermal management system and the power system,studying the energy management strategy for REEVs in lowtemperature environments to optimize the distribution of vehicle power and heat sources can effectively improve vehicle energy efficiency,battery life,and adaptability to low-temperature environments.This thesis focuses on the energy management and control issues of the vehicle power system and thermal management system of REEVs in low-temperature environments.Research is performed from three perspectives:single power source(engine)thermal management,multiple power sources(enginebattery)coupled thermal management,and power system-thermal management system collaborative energy management.The specific contents are:Firstly,an adaptive optimization control strategy is proposed for the temperature control and actuator energy consumption optimization problems of the engine thermal management system.To compensate for system uncertainty,model error,and time delay,the disturbance estimator and state predictor are designed to improve tracking control performance and anti-interference performance.The energy consumption optimization module is designed to solve the desired state trajectory corresponding to the optimal actuator energy consumption under the condition of satisfying the engine’s internal temperature consistency.Then,combined with the state tracking controller,the tracking control of the desired state trajectory and optimal cooling system energy consumption are achieved,and,the input to state stability of the closed-loop system is guaranteed.This control strategy effectively reduces temperature overshoot and steady-state error,and is more adaptable to complex conditions.Secondly,an engine-battery coupled thermal management(CTM)control strategy based on switched nonlinear model predictive control(NMPC)is proposed for power battery thermal management under cold start conditions.A liquid-cooled engine-battery CTM system structure and fuzzy logic control strategy are designed to verify the energy-saving potential of the CTM solution.Then,the system structure is improved to avoid heat loss caused by external airflow,and improve fuel consumption and battery aging damage.Considering the complex coupling characteristics and over-driving characteristics of the system,a temperature control strategy based on switched NMPC is designed.By analyzing the core optimization goals of the system at various stages,this strategy simplifies the original system model and optimization problems.In addition to ensuring optimal performance,it also improves real-time performance by reducing the complexity of optimization problems.Compared with the centralized NMPC strategy,the proposed strategy effectively reduces the online computing burden and achieves similar comprehensive optimization performance.Thirdly,aiming at the energy management problem considering the cabin thermal demand under cold start conditions,an energy management strategy based on double deep Q learning(DDQL)is proposed.By analyzing the coupling characteristics of the power system and the engine-cabin CTM system,a multi-objective optimization problem for optimizing fuel consumption and state of charge(SOC)stability is constructed.The optimization problem is transformed into a Markov decision process,and an energy-thermal collaborative optimization(E-TCO)control strategy based on DDQL algorithm is designed,which overcomes the dependence on accurate future traffic information and system mathematical models.As compared to the traditional MPC strategy,this strategy effectively reduces fuel consumption and is close to the global optimal fuel economy.In addition,this strategy has a small online computing burden and has good adaptability to different driving speed trajectories that conform to the random characteristics of the training conditions.Finally,considering the thermal demand of the whole vehicle during cold start,and the coupling characteristics of the engine-battery-cabin CTM system and power system,an E-TCO control strategy based on twin delayed deep deterministic policy gradient(TD3)is proposed to achieve multi-objective optimization of battery aging damage,fuel consumption and SOC stability.A complex state-action space results from too many system state variables and control variables.The TD3 algorithm is selected to process the complex and continuous state-action space,which ensures the optimization performance of the system and avoids performance degradation due to discrete state-action space.In addition,an energy management strategy based on forward dynamic programming algorithm is designed to obtain the global optimal solution.This serves as a best performance benchmark to analyze the optimized performance of other strategies.This strategy improves fuel economy and battery life,and achieves comprehensive optimization performance close to the global optimal solution,and overcomes the traditional MPC strategy’s dependence on accurate future traffic information and system mathematical models.