Research On The Observer Design Based On The Battery State Of Charge Estimation

State of charge of batteries which indicates the usable charge of a battery, is an important parameter of the battery management system, as it can provide judgment basis to vehicle control strategy. So estimation of SOC is important to the development of electric vehicles. The paper studied an SOC estimation method used to estimate the SOC of Lithum battery used in electric vehicles via descriptor system observer.Battery state space equation was established, based on battery RC model, and there exit the linear constraints between the open circuit voltage of the battery and the SOC. To avoid the linear constraints, in some paper, they take the time derivative of the equation of the output voltage to get a normal system, then design synovial observer which has pulse error. We designed a normal observer of the normal system and carried out a simulation experiment using Matlab/Simulink, and found that the pulse error still existed. This proved that the approach is problematic. Then,We reduced order of the battery state space equations,and designed reduced order observer,and found that the estimation error was much smaller,but was not eliminated.The linear constraints is suitable to deal with descriptor system.We studied an SOC estimation method via descriptor system observer, and showed how to determine the descriptor system observer parameters. a simulation experiment using Matlab/Simulink proved that the descriptor system observer could eliminate dynamic estimation error, compared with the first two methods.Finally,We studied the robust H∞ observer to solve the problem of the system parameter uncertainty, and showed how to determine the noamal robust H∞ observer parameters and the descriptor robust H∞ observer parameters. a simulation experiment using Matlab/Simulink proved that the robust H∞ observer had berrer predicting performance when the system parameters changed. The robust H∞ observer is strong enough that it can keep a good estimation effect.