Physics-constrained deep neural network method for estimating parameters in a redox flow battery

In this paper, we present a physics-constrained deep neural network (PCDNN) method for parameter estimation in the zero-dimensional (0D) model of vanadium redox flow battery (VRFB). approach, use networks to approximate parameters as functions operating conditions. This allows integration VRFB computational models physical constraints learning process, leading enhanced accuracy and cell voltage prediction. Using an experimental dataset, demonstrate that PCDNN can estimate range conditions improve 0D prediction compared with constant operation-condition-independent estimated traditional inverse methods. We also approach has improved generalization ability estimating values not used training process.