Electrochemical Thermal Coupling Model Of Lithium-ion Battery And SOC Estimation Based On Simplified Electrochemical Model

The new energy vehicle is developing rapidly,the technology on Li-ion battery is one of the most important research objects.In order to improve the utilization,prolong the service life,and to deal with the risk of self-ignition and explosion under the extreme condition,the battery state monitoring and management has become the research hotspot.And the premise is to establish an accurate lithium-ion battery model,which is suitable for control system.In addition,the electrochemical model can predict battery performance more accurately than equivalent circuit model,and temperature is one of the most important factors affecting the performance.Therefore,in this paper,the electrochemical-thermal coupling model of lithium-ion battery has been deeply studied,the model is built in the control system by simplification and reduction.Then,on the basis of this simplified electrochemical model,a robust observer is designed to estimate the SOC.Firstly,the structure and operational principle of lithium-ion power battery are analyzed,and the electrochemical reaction process is described by some related theories.Then,based on the electrode-average theory,ignoring the change of Li-ion concentration in liquid phase,the electrode-average model is obtained.For the solid phase Li-ion diffusion,which is simplified into a set of state-space equations by adopting uniform discrete finite difference method.Furthermore,the thermal effect law is studied.To achieve the direct application of thermal model in control system,the radial thermal behavior with boundary conditions of convection heat transfer is considered only.Finally,the input and output of each two models are connected through the establishment of the model for heat generation rate.Secondly,an identification algorithm is proposed based on particle swarm optimization(PSO).According to the model structure,the identification is divided into two parts,one is related to the solid phase Li-ion diffusion,and the other is to identify the rest.The accuracy of the model is tested by experiments.In the five sets of cell discharge experiments of 0.2-2C,the maximum deviation does not exceed 0.1V.In the NEDC test,the model voltage is in accordance with the actual value,especially under the low speed,about 0.03 V.And under high speed,the error is maintained at 0.08 V,up to 0.2V.Then,the parameter identification of the thermal model is determined directly by minimizing the Euclidean norm of the experimental and simulation errors,and the parameter values are calculated by the fmincon function in the Matlab.Finally,the accuracy of the parameters is also verified by the NEDC test.Thirdly,the SOC estimation is researched based on the simplified average-electrode model.The calculation method of SOC is redefined first.Then a robust observer based on the simplified model is designed to estimate the Li-ion concentration,and the SOC estimation can be calculated.Finally,the estimation algorithm is verified by the cell pulse discharge experiment of 1C and the NEDC experiment,and the results show that the SOC value estimated by the robust observer is consistent with the actual discharge of the battery.The error is within the acceptable range,which proves that the algorithm can accurately estimate the battery SOC.Finally,a hardware-in-loop(HiL)test platform is built for the electric vehicle to check the operation effect of SOC observer in actual controller.Based on the Matlab/Simulink,the vehicle model is built using the vehicle driving equation.The battery package model is built using the model studied in this paper,and the engineering files are compiled and generated into the NI VeriStand.Then,on the D2P-MotoHawk,the SOC estimation algorithm and control strategy are downloaded to the controller.So the final tests show that the observer can estimate the state of SOC accurately.