CFD modeling of residence time distribution and experimental validation in a redox flow battery using free and porous flow

Abstract

Abstract The successful use of clean energy sources will rest, in part, on the availability of efficient energy storage systems. Thus, the present work examines two critical factors (flow patterns and residence time distribution, RTD) affecting the performance of one of the most promising energy storage systems: redox flow batteries (RFBs). The goal is to develop a methodology for analyzing the hydrodynamic and mass transport behavior within a commercial redox flow battery (CRFB). This methodology includes modeling of the flow behavior as well as its experimental validation with RTD. The CRFB consisted of a square geometry with an interdigitated flow field design. Experimental residence time distribution curves for this device were obtained at four different flow rates using the stimulus-response technique with step signal. Theoretically, the flow behavior within the CRFB was approximated by an axial dispersion model (ADM), and a plug dispersion exchange model (PDE), as well as by solving the hydrodynamic equations (Navier–Stokes and Brinkman equations) and the mass transport equation (convection-diffusion) using COMSOL Multiphysics® 5.3. The calculated RTD curves are in good agreement with the experimental RTD curves. Our theoretical and experimental analysis resulted in better approximations of the axial dispersion, and the presence of stagnant zones, and channeling and by-pass (i.e., preferential flow) effects at low and intermediate Reynolds numbers (Re). The experimental RTD curves show that the liquid flow pattern in the CRFB deviates considerably from the axial dispersion model at low Re, conditions under which the CRFB exhibits significant channeling effects, as well as stagnant and dead zones. The PDE model was able to describe the deviation from ideal flow pattern caused by channeling, recycling, or by the presence of stagnant regions in the CRFB. The mass transport in the CRFB was determined by measuring the distribution of tracer molecules at different time points after the initial (step) tracer injection. The interdigitated flow field geometry exhibited a homogeneous flow distribution, featuring an increased mean electrolyte velocity inside the electrode. This simulation allowed locating stagnant zones that would locally affect the mass transport properties and globally lead to a reduction of the RFB conversion efficiency.