Joint Estimation Of Lithium Battery SOC And SOP Based On H∞ Filtering

With the change of the world energy pattern,lithium-ion battery has become a significant role of the field of new energy.Accurate battery state estimation plays a key role in the efficient use of electric energy and the extension of battery life,and the battery state of charge and power are the key parameters related to the effectiveness and safety of lithium-ion battery in the use process.The purpose of this thesis is to obtain high-precision battery equivalent model parameters and optimize the battery state estimation method,and conduct experimental verification and analysis.The main research contents of this thesis are as follows.(1)The working principle and characteristics of lithium-ion battery are analyzed and equivalent modeling is established.The test experiment of battery operating characteristics is designed to explore the change rule of open circuit voltage,pulse response characteristics and internal resistance effect of power lithium-ion battery under different test conditions,providing a theoretical basis for the construction of an equivalent circuit model.According to the working characteristics of the battery obtained in the experiment,considering the dynamic characteristics of the lithium-ion battery during charging and discharging and the unidirectional conductivity of the diode,an improved PNGV equivalent circuit model is constructed.The concentration polarization and electrochemical polarization effects of the lithium-ion battery are reflected through a dual RC parallel circuit,which provides a basis for subsequent parameter identification and state estimation.(2)The forgetting factor least square(FFRLS)method is optimized based on dynamic particle swarm optimization algorithm.By analyzing the internal reaction and external influencing factors of lithium-ion batteries,Particle Swarm Optimization(PSO)algorithm was used to optimize Forgetting Factor Recursive Least Square(FFRLS).The optimal forgetting factor was found in each iteration.The inertia weight in dynamic PSO was used to improve the parameter identification speed of lithium-ion battery,and a dynamic particle swarm optimization least-square parameter online identification method of forgetting factor was formed.Experimental results show that the improved algorithm is more accurate than the traditional forgotten factor least square algorithm,and the error is less than 0.02 V through voltage simulation test.(3)The high precision SOC estimation strategy based on adaptive H∞(H-infinity)filtering algorithm is designed.In order to improve the robustness of the traditional State of Charge(SOC)estimation algorithm,an adaptive H∞ filtering algorithm based on Sage-Husa is designed.By adaptive matching the artificially set fixed value error covariance matrix,To make the noise characteristics match the current state of the system,it is necessary to improve the estimation accuracy and reduce the estimation error as much as possible under the premise of the stability of H∞ filtering algorithm.The simulation results show that,compared with Extended Kalman Filter(EKF)and the H∞ filtering algorithm,the improved algorithm has a higher estimation accuracy,and the SOC estimation error of the three complex working conditions is within 2%,which provides a new idea for SOC estimation.(4)Based on multi-parameter constraints,the State of Power(SOP)estimation is carried out.Aiming at the low accuracy of traditional power state estimation methods,a multiparameter estimation strategy was designed to estimate SOP for Li-ion batteries under the conditions of voltage,current and SOC constraints,respectively,considering the variable temperature.In order to estimate the SOP of Li-ion batteries,a multi-parameter constrained SOP estimation algorithm was designed,and the effectiveness of the constructed SOP estimation strategy was verified through the verification analysis of Dynamic Stress Test(DST)and Beijing Bus Dynamic Stress Test(BBDST)operating conditions at different temperatures.According to the experimental results,the multi-parameter constrained SOP estimation strategy constructed in this thesis can track the instantaneous and different duration of SOP,and the maximum estimation error of SOP is less than 84 W.In summary,starting from the factors affecting battery characteristics,this thesis established a dynamic particle swarm optimization forgetting factor least-squares online identification method to improve the accuracy of battery parameter identification process,adaptive improvement of the H∞ filtering algorithm through the Sage-Husa method aims to reduce the impact of noise and increase estimation accuracy.On this basis,through working condition experiments at different temperatures,it proved that the improved algorithm has a high precision,which provides a theoretical basis for the safe and efficient operation of batteries.