Development of a SOFC Performance Model to Analyze the Powder to Power Performance of Electrode Microstructures

Abstract

A computational model that evaluates the triple phase boundary length, normalized effective transport coefficients, and charge production within solid oxide fuel cell electrode microstructures is described. Local charge production, within the electrode, is predicted by coupling the species transport equations with an expression for the electrochemical reaction rate. A particle placement model, employing the drop-and-roll algorithm, is used to generate the electrodes studied in this work. Parametric studies with the computational model are then preformed to investigate the influence of the anode porosity and solid volume fraction on charge production. It was found that charge production is jointly influenced by the reaction site (triple phase boundary) density, and ionic transport within the electrode structures. High current densities are observed in electrodes with low porosities and ionic volume fractions greater than 50%, for equal sized particles.

Progress in 3D electrode microstructure modelling for fuel cells and batteries: transport and electrochemical performance

“…Schneider et al [157] investigated the effect of the volumetric fraction of solid phases in synthetic microstructures, using a resistor network model, suggesting that the lowest polarisation resistance is expected for a composition close to the percolation threshold of the electron-conducting phase. Similar conclusions were reached by Gawel et al [170] for computer-generated Ni/YSZ anodes using FVM and taking into account gas concentration losses, predicting a maximum current production for a 40%/60% Ni/YSZ composition reducing it in favour of the electron-conducting phase close to the current collector, as investigated by Schneider et al [198]. Nevertheless, simulations show that the graded electrode does not perform significantly better than a composite electrode with optimized uniform composition [198] or indeed a scaffold structure [21].…”

“…Schneider et al [157] investigated the effect of the volumetric fraction of solid phases in synthetic microstructures, using a resistor network model, suggesting that the lowest polarisation resistance is expected for a composition close to the percolation threshold of the electron-conducting phase. Similar conclusions were reached by Gawel et al [170] for computer-generated Ni/YSZ anodes using FVM and taking into account gas concentration losses, predicting a maximum current production for a 40%/60% Ni/YSZ composition with 15% porosity. These indications were verified by Kishimoto et al [169], who tested, reconstructed and modelled Ni/YSZ anodes with different compositions, highlighting the importance of Ni percolation and active TPB density distribution along the thickness.…”

“…In the context of microstructure-resolved 3D modelling of electrochemical devices, FVM has been applied less frequently than FEM. In the majority of cases, an in-house FVM implementation is used [76,77,88,89,[164][165][166][167][168][169] but open-source [170] and commercial [37,109,[170][171][172] software is applied as well.…”

“…These approaches allow for a substantial reduction in the number of degrees of freedom, being more effective than voxel resampling [168]. On the other hand, even in FVM, when local structural details and sharp interfaces are explicitly resolved [76,77,109], i.e., by using the same meshing approach extensively discussed in the previous section, special attention is paid to refine the mesh where required, for example refining a cut-cell mesh at surfaces [37] and TPBs [170]. A specific analysis on the number of finite volume edges bordering three phases and the errors introduced by the voxel-based representation of TPBs is reported by Prokop et al [88].…”