Dynamic Equivalent Circuit Modeling of Flow Batteries for Real-Time Grid Support Applications

Abstract

This paper introduces a dynamic equivalent circuit modelling framework of flow batteries with an aim of facilitating real-time grid support applications. A multi branch Thevenin equivalent circuit model is deduced to describe precisely the electrochemical and transient nature of flow battery system across different load conditions. In order to provide real time adaptability, recursive least squares (RLS), and the expanded Kalman philtre (EKF) algorithms are utilised in online estimation of the parameters, which allow the resonance parameters of the internal resistance and polarisation to be continuously updated. The model is developed under dynamic grid conditions such as load variations and disturbance conditions and is confirmed by the simulations using MATLAB/Simulink. Modelling accuracy metrics of performance evaluation show that significant improvement is achieved with the RMSE being lowered by about 32-45 % in comparison to traditional fixed-parameter models and there is good correlation between measured and predicted voltages (R²> 0.96). The analysis of dynamic response indicates that convergence is achieved more rapidly and the settling time is shorter which serves to prove that the strategy is equally suitable in real-time application. In addition, grid support performance implies better voltage control and tracking of power. The findings emphasise the fact that the presented adaptive modelling framework offers a consistent and computationally effective answer that will integrate flow batteries into the realm of contemporary smart grids.