A Novel Vanadium Redox Flow Battery State of Charge Estimation Method Utilizing Convolutional Neural Network and Gated Recurrent Unit for Different Operating Conditions

Abstract

Accurate state of charge (SOC) estimation is crucial to improving the efficiency of vanadium redox flow batteries and extending their service life. However, previous methods generally rely on prior electrochemical knowledge and do not fully consider the flow rate.  This paper combines the advantage of convolutional neural network (CNN) in extracting complex features with the capability of gated recurrent unit (GRU) in modeling temporal relationships and proposes a novel CNN-GRU hybrid model. CNN is first used to extract features from test data that include voltage, current, and flow rate. Subsequently, GRUs are used to process temporal relationships and establish nonlinear mappings with SOC, enabling accurate and stable SOC estimation. This method accounts for the impact of dynamic flow rate variations and eliminates the need to build a battery model. Training and validation of the model use test data collected under varying flow rates and operating conditions. Comparisons with other methods demonstrate significant superiority, achieving an average RMSE of 2.79%, MAE of 2.28%, MAXE of 7.5%, and R² of 98.52%. Additionally, results show that the trained model can achieve SOC prediction at approximately 47us per sample with high computational efficiency, makes it particularly suitable for battery management systems.

Keywords: Vanadium redox flow battery, State-of-Charge, Hybrid data-driven model, Flow rate, Convolutional neural network, Gate recurrent unit