Physics-informed CoKriging model of a redox flow battery

Vanadium redox flow batteries (VRFBs) offer the capability to store large amounts of energy cheaply and efficiently, however, there is a need for fast accurate models charge–discharge curve VRFB potentially improve battery capacity performance. We develop multifidelity model predicting VRFB. In model, we use Physics-informed CoKriging (CoPhIK) machine learning method that trained on experimental data constrained by so-called “zero-dimensional” physics-based model. Here demonstrate shows good agreement with results significant improvements over existing zero-dimensional models. show proposed robust as it not sensitive input parameters in also only small amount high-fidelity datasets are needed predictions range considered parameters, which include current density, rate, initial concentrations.