Research On Energy Management Strategy Of Hybrid Vehicle Based On Adaptive ECMS With Working Condition Identification

With the increasing prominence of resource and environmental issues,the global automotive industry is moving in the direction of energy conservation and environmental protection.Hybrid vehicles are new energy models composed of two or more power drive systems,which can effectively reduce pollution emissions and save energy,and are the most suitable new energy models for expansion and popularization in the current period.The core of hybrid vehicles is the energy management strategy,which reasonably distributes the energy output and storage between power sources to meet the premise of power performance,so as to achieve the effect of energy saving and emission reduction.The current energy control strategy simulation only considers the typical working conditions,which differ greatly from the actual urban working conditions in China,and the performance of these control strategies cannot be optimized when they are applied to the actual urban working conditions.Therefore,this paper takes plug-in hybrid electric vehicles as the research object,constructs Fuzhou urban working conditions,develops adaptive ECMS energy control strategies based on working condition identification,and verifies its ability to improve fuel economy through simulation comparison.The specific work is divided into four parts as follows:(1)Firstly,Savitzky-Golay filtering,pre-processing and interpolation of complementary points are performed on the collected raw data,and the working condition segments are divided using the short-trip method,and the 16 characteristic parameters are dimensionalized by principal component analysis.Next,K-means++ clustering algorithm is used to divide the working condition segments into three categories,and the Markov chain transfer probability matrix and the proportion of each working condition are used to construct the urban driving conditions in Fuzhou.Comparing with the original data parameters,the error is within 10 %,and the constructed working condition reflects the driving condition of Fuzhou city more realistically.Finally,using the BP neural network toolbox,the recognition module with a comprehensive recognition rate of 88.7 % is obtained.The BP neural network online working condition recognition model is constructed by building the cycle fixed time jump,feature parameter extraction module and working condition recognition module.(2)Calculate and match the parameters of each power component based on the parameters of the whole vehicle,build the longitudinal dynamics modules of each power component,driver,and the whole vehicle using MATLAB/Simulink,and integrate the forward simulation model of the plug-in hybrid vehicle.(3)Firstly,develop a rule energy control strategy and build a state transfer model in Stateflow.Secondly,Using the variable step firefly algorithm,five dynamic logic threshold parameters are optimized.Finally,the variable-step firefly algorithm rule energy control strategy(VSFACD-CS)is compared with the normal rule energy control strategy and its fuel economy is improved by 7.24 %.(4)Firstly,the ECMS control strategy model is built on the plug-in hybrid vehicle forward simulation model based on the Pontryagin’s principle of minimal value and Hamiltonian function.Secondly,through simulation analysis,the relationship between the equivalence factor and the initial state of the battery,driving distance and driving conditions is explored,and the variable step firefly algorithm is used to optimize the equivalence factor in three different cases and establish the optimal equivalence factor library.Using the reference curve to construct the penalty function,the equivalent factor is dynamically adjusted in real time to maximize the battery power consumption,which is combined with the BP neural network online working condition recognition module to develop an adaptive ECMS energy management strategy based on working condition recognition.Finally,compared with the VSFA-CD-CS control strategy and ECMS control strategy,the fuel consumption based on the condition recognition adaptive ECMS energy management strategy is 3.027 L per 100 km,and the fuel economy is improved by 7.77 % and 4.48 %,respectively.Therefore,the strategy has better fuel economy.