Vehicle Power Lithium Battery Research On State Estimation And Balance Control

Under the background of energy crisis,although the clean-energy vehicles have been developed very fast in recent years owing to the country's subsidy policies,the annual sales in last year were still only 2.1% of fuel vehicles and the reasons are safety and endurance problems partly.The research on the state of charge(SOC)and remaining useful life(RUL)of the battery can support the safety and security of the battery management system of clean-energy vehicles.At the same time,the research on the battery control system can maintain the safe operation of the battery system to ensure the safety of vehicles.The paper consider the lithium-ion battery as the object of study.Firstly the paper determined the mapping relationship between the battery open voltage and the state of charge through the battery charge and discharge test date and chose the Thevenin model as battery model through the analysis on the battery rebound voltage stage.The battery model was established and the simulation was carried on.The simulation results were compared with the real data and maximum error is 33 m V,which can prove the high reliability of battery model.The paper chose the standard particle filter to estimate state of charge to overcome the negative effect when the noise distribution of Kalman filter were Gaussian distribution.The standard particle filter and the extended Kalman filter was combined with battery model and the paper chose the main vectors and the concrete equation form to estimate the state of charge,whose results indicated the standard particle filter was more accurate and error is less than 4%.The standard particle filter was improved on the aspect of importance density function,which can be called auxiliary paticle filter algorithm.The simulation of the auxiliary particle filter to estimate SOC was carried out in MATLAB/Simulink and the maximum error didn't exceed 3.5%,which is more accurate than the results of auxiliary particle filter.The paper chose the standard particle filter used random resampling algorithm to estimate the RUL,which based on the analysis of the commonly RUL estimation algorithms.To overcome the weaknesses of the random resampling algorithm,the standard particle filters based on multinomial resampling,residual resampling and systematic resampling were carried out to estimate RUL.the results showed that the the standard particle filter based on systematic resampling go t the most accurate result,whose estimation error was 34 charge and discharge cycles.To overcome the particle degradation phenomenon,the paper improved the standard particle filter and got the regularized particle algorithm.The simulation results showe d the estimation accuracy based on regularized particle filter improved 3.64% compared with that of standard particle filter,whose estimation error was 16 charge and dicharge cycles.The paper chose the one-way energy transfer active balance based on inductance as the balance scheme after analyzing the general balance methods.Then the paper calculated the parameters of the inductive module and the flyback transformer module and established the balance model based on PSIM.The balance simulation was accomplished when the voltage values of the battery arranged on the positive order.To overcome the weakness that the balance failed to start when the voltage values arranged out of order,the paper combined with the flyback transformer and achieved the closed-loop topology of the balance circuit.The simulation results showed that the active balace scheme based on the combination between the inductance and the flyback considered both energy utilization and engineering practicality and the simulation results was satisfactory.