3D reconstruction of two-phase random heterogeneous material from 2D sections: An approach via genetic algorithms

This paper introduces a method to reconstruct the three-dimensional (3D) microstructure of two-phase materials, e.g., porous materials such as highly irradiated nuclear fuel, from two-dimensional (2D) sections via a multi-objective optimization genetic algorithm. The optimization is based on the comparison between the reference and reconstructed 2D sections on specific target properties, i.e., 2D pore number, and mean value and standard deviation of the pore-size distribution. This represents a multi-objective fitness function subject to weaker hypotheses compared to state-of-the-art methods based on n-points correlations, allowing for a broader range of application. The effectiveness of the proposed method is demonstrated on synthetic data and compared with state-of-the-art methods adopting a fitness based on 2D correlations. The method here developed can be used as a cost-effective tool to reconstruct the pore structure in highly irradiated materials using 2D experimental data.