Parametrization and Validation of a Vanadium-Redox-Flow Cell Model

Performance improvement of electrochemical energy storage systems is a continuous process. Every application has its unique requirement profiles. In the field of stationary energy storage the cyclability is an important feature, flow batteries are beneficial for applications with high numbers of cycles. This study focuses on Vanadium-Redox-Flow Batteries (VRFB) to improve an existing cell design. Production of prototypes is expensive, but with a good model optimized geometries can be predicted. A key task is the precise validation of the model to minimize the errors. In this study a model of a VRFB is developed in COMSOL Multiphysics and it is evaluated with measurements on the cell itself. The model is based on a laboratory cell. It consists of graphite felts as electrodes, separated by an ion exchange membrane and the cupper current collectors are protected from the corrosive electrolyte by bipolar plates. The electrodes are in rectangular shape and squeezed to about 80% of their thickness. The porosity of the electrodes and the specific area were determined with optical measurement methods with a visual light microscope and image processing. The diffusion coefficients inside the electrodes were determined with electrochemical impedance spectroscopy. The diffusion coefficients were determined at different state of charge (SOC) to evaluate the difference in ion transport. The SOC was determined in two different ways, with UV-VIS spectroscopy, measuring the amount of Vanadium species in the electrolyte additionally to potential measurement on cell level before taking the sample. In addition to the data sheet information the commercial vanadium electrolyte was investigated with ICP-OES and the amount of vanadium and sulfuric acid was determined more precisely. The model of the cell is a 2D implementation using the here determined parameters of geometry and kinetics of the active material inside the cell. For future purposes it is modeled as the 3D problem, accepting the additional computing time, to increase the precision of the model. The model includes the porous electrodes, the membrane and the bi-polar plates. These are all active parts of a VRFB and the bi-polar plates are included as they might have an effect of the electrochemical reactions too. The evaluation shows that the model does represent the measured potentials the selected SOCs very well. With the confirmation of the evaluation, it is possible to withdraw information from the inside of the