Machine Learning Enhancement of Micro-CT Based Micromechanics of Composite Materials

1) Prediction of effective mechanical properties of random composite materials based on structure tensor analysis of high-resolution micro-CT images. 2) Construction of high-quality micro-CT images of the microstructure of composite materials using machine learning methods to make scanned volume representative: a) by reconstructing a high-resolution image from several images of lower quality, b) by expanding and predicting possible patterns using only high-quality images. 3) Prediction of advanced mechanical properties (strength, progressive damage, fatigue) of materials based on reconstructed micro-CT images. Analysis of weak periodicity in textile composites based on micro-CT imaging. 4) Implementation of the developed methodology of a real problem from industry.