Online monitoring of state of charge and capacity loss for vanadium redox flow battery based on autoregressive exogenous modeling

Accurate monitoring of state of charge (SOC) and capacity loss is critical for the management of vanadium redox flow battery (VRB) system. This paper proposes a novel autoregressive exogenous model for the vanadium redox flow battery, based on which the model-based monitoring of state of charge and capacity loss is investigated. The offline parameterization based on genetic algorithm and the online parameterization based on recursive least squares are investigated for the proposed model to compare the model accuracy and robustness. Leveraging the parameterized model, an H-infinity observer is exploited to estimate the battery state of charge and capacity in real time. Experimental results suggest that the proposed autoregressive exogenous model can accurately simulate the dynamic behavior of vanadium redox flow battery. Compared with the offline model based method, the observer based on online adaptive model is superior in terms of the accuracy of modeling, state of charge estimation and capacity loss monitoring. The proposed method is also verified with high robustness to the uncertain algorithmic initialization, electrolyte imbalance, and the change of system design and work conditions.