Research On Intelligent Energy Management Strategy For Hybrid Electric Bus Based On Vehicular Network Information

With the application and promotion of the vehicle networking technology in new energy bus,the networking and electrification have become an important development direction of the new energy bus.At the same time,the vehicle networking technology also provides a new way for the development of the optimal energy management control in hybrid electric bus.The hybrid electric bus is a complex nonlinear system with multi-power sources.How to extract the available information based on the vehicle networking platform,adopt the effective intelligent control method to optimize the energy management strategy to realize the efficient and reasonable work in each power source,and deeply improve the optimality and adaptability of the energy management strategy are the key to the research of the intelligent networking hybrid electric bus.Besides,they are also the industrial demand of the integration development in networked,electrification and intelligentize.This paper is carried out based on the new energy bus networking platform.In view of the contradiction between the optimality and adaptability of the energy management strategy in hybrid electric bus,the research on information data mining and intelligent energy management strategy based on the driving conditions from vehicle networkingare carried out.There are three key issues in vehicle networking platform,which are the data mining methods,utilization degree of driving condition information by the energy management strategy and the optimality and adaptability of energy management strategy.Around the above issues,this paper utilizes the advanced data mining method,the extremum principle of the optimal control problem,the dynamic programming mathematical tool,and the advanced intelligent learning control theory and simulation test to carry out the research.First,in order to obtain the available driving condition information from the vehicle networking platform,a data mining method for the bus driving conditions with fixed route is established based on the vehicle networking information.The driving conditions of the bus are obtained from a vehicle networking platform.By analyzing the characteristics of the driving conditions data from the vehicle networking platform,the problems existing in the data are determined.And then,the method for processing the missing data and data noise was established and verified.Then,the relationship between the characteristic of bus driving conditions and energy consumption characteristics is analyzed.Considering the historical and future dimensions of driving conditions data,a synthesis method of fixed driving condition based on and an intelligent prediction method of future driving condition are proposed based on the energy consumption characteristics and driving route characteristic parameters.Thus,the data mining of bus driving conditions is fully carried out.Second,in order to improve the utilization degree of driving condition information by the energy management strategy,based on the data mining results of driving conditions,a hierarchical optimization adaptive intelligent energy management strategy based on the working condition information is proposed.Combined with the data mining results of the history and future dimensions of the driving conditions information,a hierarchical optimization adaptive intelligent energy management strategy framework based on the driving condition information is innovatively designed,which realizes the global planning and local real-time optimization for the driving condition information.The upper level plans the optimal SOC trajectory from the perspective of global optimization based on the typical synthetic operating conditions.The lower level adaptively distributes the torque from the perspective of local optimization based on the predicted future conditions.The strategy effectively improves the optimality and adaptability of the energy management strategy.Third,considering the strong adaptability of learning intelligent algorithms,a deep reinforcement learning energy management strategy based on fixed routes driving information is proposed.Combining with the data mining results from fixed driving conditions information of vehicle networking a deep reinforcement learning energy management strategy framework based on the fixed driving information is innovatively proposed.This strategy effectively utilizes the vehicle networking platform to realize global optimization for hybrid electric bus with fixed route,thereby obtaining an approximately optimal mode switching rules.Besides,based on the deep Q-Learning algorithm,the power source torque is distributed in the hybrid mode,which improves the optimality and adaptability of the energy management strategy.Finally,to verify the real-time performance of the proposed intelligent energy management strategy,this paper builds the HIL test platform of the planetary hybrid system based on dSPACE / simulator to test and verify the algorithm.Under synthetic driving condition,the effectiveness and real-time performance of the two intelligent energy management strategies are verified.The results show that the algorithm can achieve a comprehensive improvement on system economy and working condition adaptability.