Research On Energy Management Strategy For Plug-in Hybrid Electric Buses Based On Driving Pattern Recognition

In recently,the shortage of energy and the environmental pollution are factors which restrict the sustainable development of the global economy and society,and the automobile industry is one of the main sources of causes of this factor.Thus,it is strategically significant to research and develop the electrical vehicles that could protect environment and save energy.As an important technology to alleviate the energy crisis,the plug-in hybrid electric vehicle(PHEV)has become the most marketable and industrialized prospects of the ideal types.As the characteristics of fixed route,the plug-in hybrid electric bus(PHEB)is also widely used in urban bus driving conditions.Since the strong coupling relationship between energy management and required power,and the stochastic vehicle mass is also an important factor affecting the required power,the Pontryagin’s Minimum Principle(PMP)-based energy management should consider the noise of stochastic vehicle mass.However,if the vehicle mass is evaluated on-line,the control complexity will be greatly increased.Thus,a driving pattern recognition method is proposed to address the problem.The main research contents of this paper can be summarized as:First,the power system structural of the vehicle is determined and the vehicle model of PHEB is established to satisfy the requirements of energy management control,including the longitudinal dynamic model,the engine model,the motor model and battery model,which laid a foundation for the follow-up research.Second,the model of PMP and K-nearest neighbor(KNN)algorithm are established.The PMP algorithm model is established as underlying energy management strategy.The mathematical model of KNN algorithm is established based on the theory of KNN algorithm,in which the KNN algorithm model is an important part of the method of driving pattern recognition.Third,the problem of deterministic optimization design method is studied.Based on Multi-island genetic algorithm(MIGA)and PMP algorithm,a deterministic optimization design model is established by taking the vehicle mass as a constant,and the optimal design values of deterministic optimization is obtained.Then,the reliability verification analysis of the optimal design value is carried out based on Monte Carlo simulation(MCS).Fourth,the driving pattern recognition-based energy management strategy is studied.First,the robust design values can be obtained,based on the Design for Six Sigma(DFSS)method,and the reliability is analyzed based on the MCS.Then,based on adaptive control,the real-time control vector can be obtained and the optimal energy management control can be realized.Reliability verification shows that the optimal design value,obtained based on the deterministic optimization design,still has 32% failure rate even with the verification of the driving pattern recognition method;while,the robust design value obtained based on the DFSS design also uses the driving pattern recognition method without failure rate.This shows that the DFSS method is reasonable and can find the robust design value against noise.Thus,the influence of vehicle stochastic mass on adaptive control can be ignored.The simulation results show that the proposed strategy can save 34.36% of the average fuel consumption compared with the rule-based energy management strategy,and can achieve the goal of energy saving and emission reduction.