Online Estimation of Battery Equivalent Circuit Model Parameters Using Decoupled Least Squares Technique

The equivalent circuit models (ECMs) have been widely used for real-time battery management applications because of offering acceptable modeling accuracy with ease of parametrization and implementation. The online estimation of ECM parameters can estimate parameters in real-time and effectively improve the identification accuracy, therefore draw much more attention. However, the classical recursive least squares (RLS) method widely used often encounters numerical problems and low accuracy, thus the improved method is needed. This paper presents a decoupled recursive least squares (DRLS) method based on the traditional RLS method to overcome its shortcomings. The DRLS method devices the battery's dynamics into the fast dynamic and slow dynamic, and estimate the parameters of each part separately. Compared with the traditional online RLS method, the DRLS avoids the simultaneous estimation of all parameters and effectively reduces the computational difficulty and complexity by decoupling. The experimental results demonstrates that the DRLS approach can not only increase the accuracy of parameter identification by 83.3% when compared to RLS but also solved the numerical problems of parameter identification results.