Computational methods have proven to be essential in the design of three-dimensional (3D) photonic crystals [1] . They have allowed the prediction of the properties of the photonic crystal design without expensive manu-facturing steps. Additionally, computations are used to interpret experimental results from real crystals. Interest-ingly, the fabrication of photonic crystals is necessarily never perfect as structural differences from the design are inevitable [2] , [3] . This structural mismatch between design and realization means that experimental results (trans-mission, reflection) are not faithfully interpreted, as the physical model (plane-wave expansion) does not have the true structural realization as input.