Study On Parameter Identification Of The Equivalent Model In Lithium-ion Battery

Lithium-ion battery has been widely applied to the field of energy storage due to its large energy density,good cycle performance and other obvious advantages. In order to exert the performance of the battery, prolong its life and ensure safety work,it's necessary to establish a high precision battery model to monitor the running state of the battery. Therefore, it has significant importance to study the equivalent mechanism model of battery for the sake of the characteristics of battery. In this paper, there are several points carried out around the lithium-ion battery modeling and parameter identification:According to the experiments of the different charge-discharge rate and pulse discharge with lithium-ion batteries, the research has been conducted on the basic characterization of the battery. This paper discusses the characteristics of charge-discharge, the capacity based on the experiments of the different charge-discharge rate and pulse discharge with lithium-ion batteries. The open circuit voltage (OCV) calibration value is obtained by using the small current interpolation method to calibrate the OCV of SOC. The rebound voltage characteristic is analyzed last.The OCV and the time-varying parameters of the model are identified based on the second-order RC equivalent circuit model and the pulse discharge experiment.This paper uses genetic algorithm (GA) to estimate the initial value of model parameters identification, and then uses non-linear least squares (NLS) algorithm to estimate the final value of model parameters identification in order to obtain the OCV and the model parameters model parameters under different SOC. Comparing the OCV value and the calibration value, the validity of OCV identification based on GA-NLS algorithm is verified.The simulation model of the lithium-ion battery is established by using the OCV and the equivalent model time-varying parameters. The experimental results show that under the condition of pulse discharge and constant current discharge, the relative error between the output voltage curve and the actual test curve is less than 1.78%.The research results indicate that the proposed battery model parameter identification method is proved to have good convergence, the time-consuming experiments for model parameters before the first operation of the battery are avoided,and the obtained model can simulate dynamic voltage characteristics of the battery effectively.