Data-driven electrode parameter identification for vanadium redox flow batteries through experimental and numerical methods. (1st December 2020)

• Record Type: Journal Article
• Authors: Cheng, Ziqiang; Tenny, Kevin M.; Pizzolato, Alberto; Forner-Cuenca, Antoni; Verda, Vittorio; Chiang, Yet-Ming; Brushett, Fikile R.; Behrou, Reza
• Is Part Of: Applied energy. Volume 279(2020)
• Publication Date: 2020-12-01
• Subjects: Vanadium redox flow batteries (VRFBs) -- Interdigitated flow field (IDFF) -- Electrode parametric study -- Data-driven modeling -- Numerical modeling -- Experimental validation; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2020.115530
• Languages: English
• ISSNs: 0306-2619

Abstract

Graphical abstract: Highlights: A computationally expensive 3D VRFB model is reduced to a 2D model. The 2D model is experimentally and numerically validated. A full parametric sweep is performed based on 2D simulation results. Quantitative correlations are established from system non-linearities. Correlations provide the basis for electrode design improvements. Abstract: The vanadium redox flow battery (VRFB) is a promising energy storage technology for stationary applications (e.g., renewables integration) that offers a pathway to cost-effectiveness through independent scaling of power and energy as well as longevity. Many current research efforts are focused on improving battery performance through electrode modifications, but high-throughput, laboratory-scale testing can be time- and material-intensive. Advances in multiphysics-based numerical modeling and data-driven parameter identification afford a computational platform to expand the design space by rapidly screening a diverse array of electrode configurations. Herein, a 3D VRFB model is first developed and validated against experimental results. Subsequently, a new 2D model is composed, yielding a computationally-light simulation framework, which is used to span bounded values of the electrode thickness, porosity, volumetric area, fiber diameter, and kinetic rate constant across six cell polarization voltages. This 2D model generates a dataset of 7350 electrode property combinations for each cell voltage, which is used to evaluate the effect of these structural properties on the pressure drop and current density. These structure-performance relationships are further quantified using Kendall τ rank correlation coefficients to highlight the dependence of cell performance on bulk electrode morphology and to identify improved property sets. This statistical framework may serve as a general guideline for parameter identification for more advanced electrode designs and redox flow battery stacks.