Energy Management Strategy Of Fuel Cell Hybrid Electric Vehicle Based On Fuzzy Rule Learning

With people’s attention to clean energy and the upgrading of automobile technology,the cross fusion of the two points out a new direction for the new development and new road of the future automobile,and the new energy vehicle emerges at the historical moment.Fuel cell hybrid electric vehicle is the outstanding representative of new energy vehicle,it is composed of fuel cell and battery.Because the fuel cell hybrid electric vehicle(FCHEV)uses a variety of energy sources,the problem of energy distribution among energy sources has become a research hotspot.This paper takes Dongfeng X37 as the research object,aiming at improving fuel economy and engineering practical application,aiming at the shortcomings of fuzz y control too much rely on subjective experience,aiming at finding a method of parameter learning and making rules,and combining the global optimal results of dynamic programming and fuzzy rule control to make energy management strategy.The main contents of this paper are as follows.The topological structure of power system of fuel cell vehicle is analyzed and determined.The vehicle dynamics model,drive motor model,fuel cell model and battery model were established by ADVISOR software.The nonlinear complex model is expressed in the form of MAP in MATLAB.According to the actual data of the X37 vehicle,the parameters are matched to further develop the energy management strategy.Dynamic programming global optimal algorithm is introduced as the theoretical standard and rule making method of fuzz y control energy management strategy.Firstly,the dynamic programming algorithm is introduced.The algorithm flow was established to minimize hydrogen consumption,and the solution process was partially simplified based on the accuracy of state variables.The global optimal results of self-built combination conditions are used as the optimal database.Then,the specific parameter fuzzy controller rule set.First,the input and output sets of the fuzzy controller.Then,in the global optimal hierarchical cluster database,the average value after interval clustering is set as the standard to determine the position of membership function of the fuzzy controller.Finally,the main rules of fuzzy controller rule base are constructed through decision tree learning method,and the rule tree of total rules is learned after analysis as fuzzy control rule base,and the multiple linear regression model is used to calculate the corresponding fitting coefficient of each rule,and the output result of each rule.The fuzz y controller after rule learning is simulated and analyzed on computer and intelligent vehicle test bench.The computer standard operating condition test shows that the fuzzy controller based on rule learning has better fuel economy than the power consumption preserving rule strategy which is widely used in engineering.In order to reflect the adaptability of the energy management strategy,both drivers have different styles of testing the relevant terms of the energy management strategy in the intelligent vehicle in the same road test rig simulation results show that the fuzz y power management strategy based on rules of learning has a better effect on fuel economy than the minimum control strategy based on rules and on-line calculation.