In fuel cell systems, a multitude of coupled physical and chemical processes take place within the assembly: fluid flow, diffusion, charge and heat transport, as well as electrochemical reactions. For design and optimisation purposes, direct numerical simulation of the full three-dimensional (3D) structure (using CFD tools) is often not feasible due to the large range of length scales that are associated with the various physical and chemical phenomena. However, since many fuel cell components such as gas ducts or current collectors are made of repetitive structures, volume averaging techniques can be employed to replace details of the original structure by their averaged counterparts. In this study, we present simulation results for SOFC fuel cells that are based on a two-step procedure: first, for all repetitive structures detailed 3D finite element simulations are used to obtain effective parameters for the transport equations and interaction terms for averaged quantities. Bipolar plates, for example, are characterised by their porosity and permeability with respect to fluid flow and by anisotropic material tensors for heat and charge transport. Similarly one obtains effective values for the Nernst potential and various kinetic parameters. The complex structural information is thereby cast into effective material properties. In a second step, we utilise these quantities to simulate fuel cells in 2D, thereby decreasing the computation time by several orders of magnitude. Depending on the design and optimisation goals, one chooses appropriate cuts perpendicular or along the stack axis. The resulting models provide current densities, temperature and species distributions as well as operation characteristics. We tested our method with the FEM-based multiphysics software NMSeses, which offers the flexibility to specify the necessary effective models. Results of simulation runs for Sulzer HEXIS-SOFC stacks are presented. (C) 2003 Elsevier Science B.V. All rights reserved.