Research On State Of Charge Estimation Method For Lithium-ion Battery Of The Electric Vehicle

Over the past few years, with the global climate, energy and environment issuesbecoming more prominent, the electric vehicle has become research focus of automobileindustry for low pollution, low noise, high energy efficiency, simple structure,convenient using and maintenance. As the power source of the electric vehicles, thelithium-ion battery has gradually become the preferable choice for high energy density,small, lightweight and high output power; state of charge (SOC) estimation has been theemphasis and difficulty in the lithium-ion battery research. By combining thecharacteristics of the lithium-ion battery in the actual working condition, thestudy,which chooses the lithium-ion battery used in the electric vehicle as the researchobject, lays special stress on analyzing the lithium-ion battery modeling and SOCestimation algorithm.Firstly, the study introduces composition, production technology, structure andworking principle of the lithium-ion battery used in the electric vehicles and targetsresearch on the main performance parameters, then using self-developed lithium-ionbattery testing system to test the battery parameters. Based on the analysis the classicbattery model and characteristics in the actual working condition, the study adopt thePNGV improved model for the lithium-ion battery modeling, through analyzing thebattery discharge curve and recovery curve, the model parameters can be identified bythe least square method, and with the help of MATLAB/SIMULINK simulation model,the accuracy of the model is validated.Secondly, after analyzing the classic SOC estimation method, the study uses extendedKalman filter algorithm and ampere-hour integral method to set up the system spacestate model of the lithium-ion battery based on the PNGV improved model, in virtue ofthe open circuit voltage method, the initial value of the lithium-ion battery SOC can becalculated, with the algorithm of extend Kalman filter, the lithium-ion battery SOCestimation is achieved.Finally, by comparing MATLAB simulation and the experimental data on chargingand discharging processes of battery, we know that the battery characteristic, whichsimulated by PNGV improved model, is close to the truth, and the SOC estimationalgorithm based on PNGV improved model used to lithium-ion battery is accurate andreasonable.