Research On A-ECMS Of Hybrid Electric Vehicles For Complex Operating Conditions

In recent years,China’s crude oil imports and crude oil dependence have been high,and the world’s energy demand continues to grow.China is a big country of automobile production and sales,and the proportion of automobile fuel consumption in our country’s total fuel consumption has always been high.It is therefore important to study and design energy management control strategies for new energy hybrid electric vehicles.The EMCs is able to reduce the vehicle’s energy consumption while meeting the vehicle’s power requirements by distributing the drive power between the different powertrains.In this thesis,using a hybrid vehicle as the research object.Three energy management strategies,the deterministic Rule-based energy management strategy（Rule-based strategy）,the Dynamic Programming global optimal energy management strategy（DP Strategy）and Adaptive Equivalent Consumption Minimisation Strategy based on a deep learning approach for complex applications（A-ECMS for complex conditions of application）are designed based on deterministic rules,dynamic programming（DP）algorithm and Pontryagin’s Minimum Principle（PMP）.The DP strategy is the optimal strategy and serves as a benchmark for comparison between the strategies;Rule-based strategy is the base strategy;and A-ECMS for complex conditions of application is the target strategy of this paper.In addition,three composite test conditions are constructed based on various standard cycle test conditions,and the A-ECMS for complex applications is investigated under one of the composite test conditions,while the fuel economy performance of the three energy management strategies is compared under the three composite test conditions.The main elements are.（1）Complete the modeling of engine,motor,power battery,vehicle driving dynamics,AMT transmission and other system components of the vehicle,complete the design of Rule-based energy management strategy based on deterministic rule strategy,and verify Rule-based strategy based on WLTC conditions.The simulation results show that the fuel economy of Rule-based strategy is good under WLTC conditions,and the engine runs in the high efficiency region as expected.（2）The Global Optimal Energy Management strategy based on Dynamic Programming（DP）algorithm is studied.The Dynamic Programming algorithm is used to model the optimal control decision of the hybrid vehicle,and the DP strategy is validated based on WLTC conditions and analysed to compare the optimisation results of the rule-based strategy with the DP strategy.Simulation results show that the fuel economy performance of DP strategy is 16.5% higher than that of Rule-based strategy under WLTC conditions.（3）The design of A-ECMS based on deep learning methods is completed.The optimal equivalent factor dataset is calculated from the DP strategy operating conditions data,and the BP neural network is trained based on this dataset and some of the operating conditions data of the DP strategy,and the fitted BP neural network is used to predict the equivalent factor values to obtain A-ECMS for complex operating conditions.The results of the optimization of the A-ECMS strategy are compared with those of Rule-based strategy,DP strategy and A-ECMS.The simulation results show that the A-ECMS strategy follows DP strategy well,and the fuel economy performance of the A-ECMS achieves 95.4% of that of DP strategy,which is 11.9% better than that of the rule strategy.（4）Three composite test conditions were constructed based on the combination of multiple standard cycle test conditions,and A-ECMS for complex condition of application was designed based on one of the composite test conditions.Rule-based strategy,DP strategy and A-ECMS for complex condition of application were verified in different initial state of charge（SOC）of power battery under three composite test conditions,respectively,and the optimization performance of A-ECMS for complex condition of application was analyzed and compared.The simulation results show that the difference between the starting and ending values of the state of charge of A-ECMS for complex condition of application is within 0.1,and the variation curves of SOC and power battery demand follow DP strategy well.The average equivalent fuel consumption rate for the A-ECMS is 3.9668L/100 km,and the average fuel economy performance of the A-ECMS is 92.1% of DP strategy,and 14.9% higher than that of Rule-based strategy.