Study On Battery-health Conscious Energy Management Optimization Of Plug-in Hybrid Electric Delivery Vehicle

Plug-in hybrid electric vehicle(PHEV)has become a hot research topic in the field of new energy vehicles because it can improve the structure of energy use and has a long range of electric driving.The development of plug-in hybrid delivery vehicles can reduce logistics industry pollution and promote long-term development of the city.However,how to coordinate the driver’s power demand between the engine and the electric driving system when driving is the key to taking advantage of energy saving and reducing the life cycle cost of PHEV.Vehicle life cycle cost includes energy use cost and purchase cost.At present,the optimization objective of vehicle energy management control strategies mostly is to minimize the overall fuel consumption in order to reduce energy use cost.During complicated driving conditions,the battery inevitably ages due to frequent deep charge-discharge,which shortens the battery life.Second replacement of the battery will greatly increase the purchase cost of the vehicle.Therefore,from the perspective of vehicle energy management optimization,it is necessary to proactively extend battery life and reduce the life cycle cost of PHEV.On top of that,this paper conducted research on real-time multi-objective energy management optimization control that balances overall vehicle fuel consumption and battery performance degradation.First,control-oriented models of key components in the hybrid electric system were established with relative accuracy and low calculation load.At the same time,in order to achieve the purpose of optimizing battery life,the mechanism of battery performance degradation was studied,the factors that affect the battery life attenuation was explored and finally the battery life model was built,which lays the foundation for the calculation and verification of the energy management optimization strategy described later.Then,a multi-objective optimization control strategy for energy management that takes into account the overall fuel consumption and battery performance degradation of vehicle was carried out and the weight coefficient was introduced to convert to single-objective optimization.Given the cycle conditions in advance,dynamic programming(DP),a representative of global optimization algorithm,was used to solve this problem.By rationally selecting the weight coefficient,the life cycle cost of the vehicle can be effectively reduced,which is of good control effect.However,the global optimization control strategy that is dependent on working conditions was solved under specific operating condition which also need to be known in advance,thus it cannot be applied in real time and can only provide research support and evaluation benchmarks for other types of energy management strategies.To solve the problem that the DP algorithm depends on the working conditions and cannot be applied in real time,this paper further developed a real-time control strategy for energy management based on working condition identification.Multi-objective global optimization energy management was performed on three representative driving routes of the delivery vehicle.Then the global optimization results of the three routes were respectively trained for the support vector machine(SVM)model,which divides the electric and hybrid modes,and the neural network model,which realizes the energy management of the hybrid electric mode to develop the real-time optimization strategy for energy management under the corresponding route.Next,a random forest model was trained to identify the operating conditions which aims to determine which route the current vehicle is traveling on,so as to adopt the energy management control strategy under the corresponding driving route.Finally simulation results show that the energy management control strategy based on operating condition recognition developed in this paper greatly improves the calculation speed and optimizes the overall performance of the vehicle.