Research On SOC Estimation Of Electric Vehicle Power Lithium Battery

At present,the rapid increase in the number of fuel vehicles brings convenience to people's life,but also aggravates the situation of environmental pollution and the shortage of nonrenewable resources.The use of electric vehicles can effectively solve the environmental and resource problems,which is in line with the era background of energy conservation and emission reduction.Battery management system is the key component of electric vehicle power battery pack,which can carry out real-time data acquisition,balance management,state estimation,over charging and over discharging protect.State of charge represents the remaining capacity of the battery.Accurate estimation of SOC has great reference significance in improving user convenience,avoiding battery overload and improving battery life.Therefore,this paper takes Li Fe PO4 power battery as the research object,relying on the tutor's practical engineering BMS technology research project,and estimates the SOC of lithium battery.This paper first introduces the basic principle and characteristic parameters of lithium battery,and then builds the experimental test platform,carries out the charge and discharge experiments under different working conditions,modifies the capacity,selects the second-order RC battery equivalent model to simulate the actual working condition of the battery,and constructs the corresponding state space equation.The experimental results are used to identify the model parameters,and the fitting curve between open circuit voltage and SOC is obtained through the charge discharge experiment.Then a complete battery model is established.The accuracy of the battery model is tested by DST condition.The results show that the battery model can meet the requirements of battery state estimation.Secondly,to solve the problem of system noise limitation when using Kalman filter to estimate SOC,particle filter algorithm is used to update the parameters of battery state model,which can effectively suppress the influence of non-Gaussian noise in SOC estimation process.Aiming at the problem that bat algorithm itself is easy to fall into local optimum or slow convergence,an improved bat algorithm is proposed by adding speed learning mechanism and second-order oscillation mechanism.Then,the IBA-PF algorithm is proposed by using IBA to optimize pf creatively.By adjusting the particle distribution,the diversity degradation problem of traditional PF is solved.The simulation experiment is carried out in MATLAB to test the accuracy of IBA-PF algorithm,which verifies the effectiveness and applicability of IBA-PF algorithm.Finally,IBA-PF algorithm is applied to the SOC state estimation model.In the simulation environment,the SOC estimation accuracy of IBA-PF algorithm and PF algorithm under three different test conditions is tested.The results show that IBA-PF has higher accuracy and stability than PF,and it can still maintain within 2% under the most complex DST conditions,which verifies the effectiveness of the method.Using the built battery cycle test platform,the relevant model parameters are written into the BMS system to test the estimation performance of the algorithm in the actual BMS.The results show that the IBA-PF algorithm can still maintain high accuracy and stability,and the maximum error is not more than 4%.It has good adaptability for battery SOC estimation.