Study On Time-frequency-domain Fractional Order Modeling And Parameter Current-rate Dependant Characteristics For Lithium-ion Batteries

As the main energy source of electric vehicles,the performance of lithium-ion batteries largely determines the development of electric vehicles.In complex practical conditions,lithium-ion batteries often exhibit strong nonlinear characteristics.It is important for parameter estimation,thermal performance analysis and dynamic system simulation to accurately simulate the external characteristics of the battery under different current stresses.Therefore,this paper takes ternary battery as the research object,which is the most widely used battery in electric vehicles,analyzes the characteristics of model parameters under different current-rate conditions,and establishes the model that can adapt to the response characteristics of battery under large current excitation.In this paper,based on the electrochemical impedance spectrum test which reflects more information inside the battery,the fractional order equivalent circuit model describing the reaction process of the battery at different time scales is established by comparing the impedance spectrum characteristics with the frequency response characteristics of the electrical components.On the basis of this model,the correlation between impedance parameters and current in high frequency,medium frequency and low frequency regions is analyzed by combining with electrochemical reaction mechanism and electrode process dynamics.Based on the identification results in frequency-domain,the change law of parameters with current-rate is summarized and the conclusion of correlation with current is verified.Considering the problem that it is difficult for fractional order model to identify the parameters online based on frequency-domain impedance spectrum,the time constant of the electrochemical reaction process is analyzed.The beginning and the stabilizing time scale of the different polarization process is concluded,based on which the simplified fractional order model is established in time-domain.Then,differential evolution(DE)algorithm is used to identify the parameters of the time-domain multi-rate pulse data,and the corresponding time scale of the charge transfer process in the frequency-domain impedance spectrum is found through the consistency of the change law for the sensitive parameter.In the process of building the model that can be applied to a wide range of current-rate,the sensitive characteristic parameter is obtained and coupled into fractional order model according to the change law of model parameters with current-rate.The results of variable parameter model simulation coupling current characteristics are obtained by applying sinusoidal excitation of 100 Hz,50Hz and 20 Hz with constant frequency but variable amplitude to the battery and collecting voltage response data through electrochemical workstation.At the same time,the short-time square wave pulse excitation is applied to the battery.The parameters are identified by DE algorithm,and the time-domain simplified fractional order model is simulated and analyzed within the range of 3C rate,and the corresponding time scale of charge transfer process is verified.The simulation results of fixed parameters and variable parameters are compared from the perspectives of constant frequency excitation,short term square wave excitation and short term pulse excitation,and the superiority of the fractional order model with variable parameters considering the current characteristics is verified under the condition of large current-rate.In addition,based on the model parameter characteristics under different current-rates,the equivalent stress of medium and low frequency AC/DC superposition condition on lithium-ion battery is analyzed.Finally considering the complex operation problem about calling function and parameter setting in the process of parameter identification and model simulation,based on the matlab GUI function,realize the visualization operation about parameter identification,external characteristic simulation and error analysis,which greatly saves the time of the model development and has a certain application value.