Simulation of multidimensional binary random fields with application to modeling of two-phase random media

This paper proposes a methodology for simulation of binary random fields with application to the problem of generating sample realizations of two-phase random media. The methodology is based on the concept of nonlinear transformations with memory of Gaussian random fields. The simulation is performed according to the autocorrelation function of the binary field which contains considerable information about the microstructural characteristics of the medium. The determination of the probabilistic characteristics of the underlying Gaussian field is achieved through an iterative procedure that was introduced in a previous paper by the same authors in one dimension and is extended here to multiple dimensions. Limiting cases and alternative mappings are also presented. The capabilities of the methodology are demonstrated in a series of examples.