State Of Charge(SOC) Estimation Methods For Power Battery Based On Variable Structure Filtering

Environmental pollution,energy crisis and energy security issues have become more and more serious.Energy vehicle is one of the most effective ways to solve the problems of environmental pollution for traditional vehicle.Power battery is one of the three core key technologies of new energy vehicles.Battery state of charge(State Of Charge,SOC)is the most important parameters for power battery,High-precision state of charge estimation for power battery which can prevent the battery that is not under overcharge and over discharge,and extend battery usage time,it is also the important basis for estimating electric vehicle driving status.The SOC is different from the voltage and current.Which cann't be measured by sensor,it only can be obtained by other indirect methods.SOC estimation methods for electric vehicle power battery are proposed in this paper,the main research work is as follows:(1)Thevenin model is selected as model of the power lithium battery.The DST data is used to identify the battery Thevenin model parameters by variable structure filtering(VFS),Voltage(OCV or Uoc)error of the model is corrected for the VFS gain,and use the same method to get the parameter identification of FUDS.Based on the OCV-SOC relationship curve through BP neural network,the battery SOC under FUDS condition is can be obtained.Based on the Arbin battery test platform,the simulation results show that the error estimation error is 0.94%.(2)The variable structure filtering algorithm can effectively guarantee convergence when estimating SOC.In order to improve the accuracy of variable structure filter estimation.the fuzzy-variable structure filtering algorithm is proposed to improve the adaptability of variable structure filtering,and achieve accurately SOC estimation.The simulation results show that the proposed method can effectively improve the accuracy of SOC estimation.The maximum absolute error of SOC estimation is 1.50% and the average absolute error is 0.09%.(3)In order to study the effectiveness of the algorithm under different Gaussian noise,adding Gaussian noise with the amplitude of 1%,2% and 3% to the system data to simulate the sensor error characteristics in the real industry.The first fuzzy-variable structure filter is used to eliminate the Gaussian noise interference and improve the accuracy of the battery SOC to estimate the input voltage of the terminal voltage.Another fuzzy-Variable structure filtering uses the filtered terminal voltage data for SOC estimation.The simulation results show that the double fuzzy-variable structure filtering algorithm can be suitable for industrial applications with low precision.In the simulation experiment,three different SOC initial values are selected to estimate battery SOC.The result shows that the algorithm has the low dependency on the initial value.