Prediction Of Vehicle Power Lithium-ion Battery Remaining Useful Life And SOC

Power lithium-ion battery has the advantages of high energy density,high discharge voltage,long cycle life and so on,which is widely used in new energy electric vehicles.The remaining useful life?RUL?of power lithium-ion battery characterizes discharge performance?or the maximum discharge capacity of the battery?compared to the state of 100%new battery,and the state of charge?SOC?of power lithium-ion battery characterizes the ratio of the remaining discharge capacity of the battery to the maximum discharge capacity under the current charge and discharge cycle,both are important parameters in the lithium battery management system,and accurate prediction of the lithium battery RUL can provide a important reference for SOC estimation.The particle filter and its improved algorithm in the existing lithium battery RUL prediction method are widely used because of high prediction accuracy and can provide probability distribution to the result,but there are also difficulties such as particles easy to degration,large amount of calculation,and poor adaptability of working conditions;The existing SOC prediction methods have complex mathematical modeling,large computational complexity,difficulty in developing and implementing single-chip C program,and easy to overflow.For the above problems in the remaining life prediction and SOC estimation of power lithium batteries,the paper is as follows:research work:?1?The charging and discharging characteristics,basic degradation mechanism and equivalent model of the power lithium-ion battery are introduced.The correlation between the state of charge and the remaining life of the lithium battery and the influencing factors of various characteristics are analyzed.The next step is to study the lithium battery RUL and The SOC provides a basic theoretical basis.?2?In order to improve the prediction accuracy of lithium battery RUL,reduce the amount of calculation and enhance the adaptability of working conditions:firstly,relying on the efficient search ability of firefly algorithm and the basic working principle of particle filter,using optimized firefly algorithm?FA?to improve particle filter,and apply it to RUL prediction of lithium battery under fixed conditions;then,based on the degradation mechanism of lithium battery and the degradation of battery discharge performance under severe conditions,establish a modified model of residual life of lithium battery under severe working conditions:the adverse conditions affect the lithium battery RUL all the number of cycles and statistics,through dichotomy method to obtain the severe conditions on the battery impact degree coefficient,record the characteristic factor value and duration,obtain the lithium battery's life correction value;finally,the predicted life of the lithium battery under fixed conditions is linearly added to the life correction value under severe conditions,and the accurate lithium battery RUL is obtained,which provides a reference for the subsequent lithium battery SOC estimation.?3?The Ah integral method is the mainstream lithium battery SOC estimation method in engineering application,it is mainly affected by SOC initial value?SOC₀?,battery calibration capacity?CN?,discharge coulombic efficiency and current acquisition error,in order to improve prediction accuracy and speed up system response of SOC estimation:firstly,combined with different battery usage conditions,the relationship between the open circuit voltage,temperature,discharge current of the battery,cycle number?related to battery's life?and the complex nonlinear characteristics of the SOC is made into a corresponding two-dimensional array table,according to the use condition,judge whether to update SOC₀,CN by dichotomy method,obtain dynamic adaptive SOC₀,CN,to reduce the amount of calculation;then,by calling BP?Back Propagation?neural network training under offline state,to obtained corrected coulomb efficiency values;finally,the accumulated charge of the battery charge and discharge is obtained in real time by using the bq76930 charge collection chip of Texas Instruments,and the influence of system energy consumption is considered to further improve the SOC estimation accuracy.Using the lithium battery cycle life test data of the University of Maryland Advanced Life Cycle Engineering Center and a provincial key laboratory for power lithium batteries and materials,the simulation test was carried out,the results show that the optimized lithium battery RUL prediction method which is adopted by this paper,has higher prediction accuracy,the maximum relative error is 4.6%,and the calculation amount is greatly reduced;The lithium battery charge and discharge test is carried out under the simulated urban road cycle conditions,the results show that the SOC prediction error of the optimized ampere-time integration method which is adopted by this paper,the relative error is within 4.3%,the system response speed is faster,the stability is good,and the comprehensive prediction effect is obviously improved.It will provide a good reference value for the design and optimization of battery management systems.