Research On The Control Strategy Of Plug-in Hybrid Electric Vehicle Based On Multi-condition Optimization

As an important mean to solve global issues such as energy crisis and environmental pollution,plug-in hybrid electric vehicle(PHEV)has become the focus and hot spot of today's automotive industry.At present there are still some potential can be tapped of fuel saving by PHEV energy management strategy,and the PHEV energy management strategy is the key technologies of PHEV.In order to develop coupling energy management strategy to adapt the daily travel(ECMS-SDP)around a specific driver for personality characteristics,this paper took optimize the energy management strategy as a mean,took improve the PHEV fuel economy as the goal.It has proved the coupling energy management strategy can enhance the fuel economy of the vehicle through the simulation experiment.In this paper,the parallel PHEV was the main research object,and the research is based on the actual trip statistics of a specific driver.First,establishing construction of a specific travel condition database driver,developing the typical travel specific driver system based on ‘analysis-dimensionality-reduction-clustering'.In order to build the driver cycle condition by correlation coefficient analysis,principal component analysis and K clustering analysis.And providing experimental data,research and verification method for subsequent development of offline simulation and control strategy.Secondly,On the basis of forward simulation model of PHEV vehicle by MATLAB/Simulink,it has developed energy management strategy based on rules based on the improvement of control logic,and according to the structure of the former,established the equivalent consumption minimization strategy(ECMS).Depend on the theory and simulation to verify the rationality and real-time of ECMS the energy management strategy,as well as provide the basis for constructing coupling control strategy.At the same time,It provided the criterion by the comparison of other energy management strategy with the depth research and study of dynamic programming(DP)algorithm principle.Further,the driver's personality travel conditions as the Markov chain,and according to the maximum likelihood estimation method to construct the state transition matrix of the specific driver.Solving the law of online stochastic dynamic programming based on ‘optimal mathematical expectation' off-line.Verified the fuel saving capacity of energy management strategy of stochastic dynamic programming with the typical driver travel conditions.Finally,condition identification criterion used random forest algorithm.Combining the ECMS control algorithm and stochastic dynamic programming of energy management strategy to establish the coupling energy management strategy based on multiple conditions optimization(ECMS-SDP).Compared with rule based control strategy,equivalent fuel consumption minimum control strategy and dynamic programming control strategy by multiple sets of simulation.Taking comparison,evaluation and analysis of the interval of the work efficiency of the engine power,battery SOC consumption curve,torque distribution,fuel economy and comprehensive cost.The results shown that,the optimization of the coupling control strategy saving 8-15% than energy management strategy based on the rules,saving 4-7% than ECMS strategy.Further,under the condition of obtaining mileage through related software and intelligent transportation system,an improved coupling adaptive energy management strategy(ASDP-ECMS)is developed to make the vehicle's economic performance increase compared to ECMS-SDP,which coincides with the development direction of intelligent transportation in future.