Study On Fractional Order Modeling And SOC Estimation Of Ternary Lithium-ion Batteries

In recent years,electric vehicle has become one of the important directions for the development of automobile industry because of its advantages of energy saving and environmental protection.As the energy storage unit of electric vehicles,the performance of batteries has an important impact on the vehicle’s power performance,economy and reliability,and the modeling of batteries and the estimation of State of Charge(SOC)are the two hotspots of battery-related research.In this paper,based on fractional calculus theory and battery impedance spectrum analysis,the fractional modeling and parameter identification of ternary lithium ion battery were studied.On the basis of the model,offline and online estimation of battery SOC were realized.The main research contents are as follows:(1)Based on the battery charge-discharge test platform,the charge-discharge test of soft-pack Ni-Cr-Mn ternary lithium battery and the test condition test of hybrid impulse capability characteristics were designed.Through the analysis of experimental data,the actual capacity of the battery and the relationship curve between SOC and open-circuit voltage were obtained.The integer order equivalent circuit model of ternary lithium battery was established,and the battery model parameters under different SOC were obtained.(2)In order to describe the polarization process and dispersion effect of lithium ion battery more accurately,the electrochemical impedance spectroscopy test of Ni-Cr-Mn ternary lithium battery was designed based on the battery electrochemical test platform.The electrochemical impedance spectrum was obtained and the fractional equivalent circuit model of ternary lithium battery was established by using the constant phase element with fractional order characteristic.(3)In order to accurately establish the off-line battery model,particle swarm optimization was used to study the parameter identification of fractional and integer order models respectively.Based on the dynamic stress test condition and the driving condition data of federal cities in the United States,the influence of the parameters of particle swarm optimization algorithm on the model accuracy and the calculation amount of the algorithm is studied,and the optimal value of the parameters is determined,and it is verified that the accuracy of the fractional model is higher than that of the integer model.At the same time,the adaptive extended Kalman filter algorithm is used to study the SOC estimation of the fractional-order model and the integer-order model respectively.The results show that the accuracy of the fractional-order model is higher than that of the integer-order model.(4)In order to accurately establish the online battery model,the recursive least square method was used to study the online parameter identification of the fractional-order model and the integer-order model respectively,and the accuracy of the fractional-order model was verified to be better than that of the integer-order model.Based on the online parameter identification,a joint estimator of model parameters and SOC is constructed by using adaptive extended Kalman filter algorithm.The results show that the SOC estimation accuracy of the second-order fractional model with high SOC segment is higher than that of the second-order integer model,and the SOC estimation accuracy of the second-order integer model with low SOC segment is higher than that of the secondorder fractional model.The SOC estimation accuracy of the first-order fractional-order model is higher than that of the second-order fractional-order model and the second-order integer-order model,but the model accuracy is lower than both.