WRLS-ARMAX-Based Parameter Identification for System Frequency Response Model with Battery Energy Storage System

With the increase of the primary frequency regulation provided by battery energy storage systems (BESSs), the frequent parameter variations pose challenges in modeling and dynamic parameter identification. To address these issues, this paper proposes a parameter identification for system frequency response model with the BESSs based on the weighted recursive least squares (WRLS) method and autoregressive moving average with exogenous inputs (ARMAX) model. First, focusing on the dynamics of the BESSs, an ARMAX model is established for the active power output under different frequency disturbances using the bilinear transformation method. Then, to mitigate parameter identification biases arising from the varying ARMAX model orders across different energy storage units, Bayesian information criterion (BIC) is employed to determine the model order, and the WRLS method is utilized for parameter identification. Finally, a case study is conducted to validate the accuracy of the proposed scheme.