Self-Consistent Clustering Analysis for Efficient Prediction of Material Strength: A Case Study of C/SiC Composites

This study employs the Self-consistent Clustering Analysis (SCA) method to achieve rapid prediction of the macroscopic tensile strength of carbon/silicon carbide (C/SiC) composites, based on periodic unit cells at the mesoscale. The SCA framework comprises offline and online stages. During the offline stage, the strain concentration tensor and interaction tensor are computed to generate a reduced-order model (ROM). The online stage leverages this ROM to solve the discretized Lippmann–Schwinger equation to obtain the mechanical response. Utilizing the established ROM, SCA completes a longitudinal tensile strength calculation within 6 s. This represents a computational speedup of approximately 31,000 times compared to the Finite Element Method (FEM). The maximum deviation from experimental measurements was 6.64%. The computational efficiency of the SCA method presents a viable approach for the rapid design of future aeroengine structures.