The design and manufacture of electrodes for use in SOFCs is one of the greatest challenges to the commercialization of fuel cell technology. Composite SOFC electrodes mix three phases (ion conducting, electron conducting, pore phase) in order to improve performance by increasing the amount of triple-phase boundaries (TBPs)-meetings of the ionic and electronic pathways with the percolating gas network-where the redox reaction takes place. The electrode microstructure is critical since electrode performance is directly dependent on the abundance of TPBs and the transport properties of the three phases. A fundamental understanding of the quantitative effects of microstructure on electrode performance is required. However, electrode models commonly neglect heterogeneity and assume effective values for key parameters. In contrast, we present a computational framework that can readily be linked to experimental studies of microstructure, thereby providing crucial insight into the conditions and competing processes in the porous microstructure, insight that can be used to design future generations of electrodes. In the proposed methodology, a virtual electrode is generated by randomly placing spherical particles in a packed bed. The particles are then expanded to simulate sintering to ensure large contact surfaces between the different phases. Once the porous structure is obtained, we can analyze the porosity and percolation of the various phases and the amount of triple-phase boundary and its percolation throughout the electrode. Furthermore, the transport and redox phenomena are also modeled to determine the potential, current, and chemical distribution throughout the different phases. We are then able to predict electrode performance based on fundamental properties of the underlying microstructure. These results are used to relate microstructural properties to electrode performance. The microstructural properties can include porosity, particle radii, and radius ratio and the effect of graded electrodes. The method is tested on model systems and used to demonstrate the effect of particle size on performance.