A highly efficient self-consistent clustering analysis method with field refinement capability for the mesoscale damage behavior of 3D woven composites

To effectively balance the accuracy and efficiency in solving high-dimensional damage problems, a self-consistent clustering analysis framework with field refinement capability (RESCA) incorporating a mesoscale damage model, is developed to investigate the mesoscale failure behavior of 3D woven composites (3DWCs). The RESCA method includes three stages: offline stage, online stage and field refinement stage integrating damage information. In the third stage, a damage-related field refinement framework is proposed to achieve cluster-based field dehomogenization and efficiently reconstruct the voxel-based field information. The results indicate that the RESCA method can accurately predict the local stress concentration, the voxel-based damage field distribution and the damage accumulation process, which are not available with the traditional SCA method. Importantly, the RESCA method can improve the computational efficiency by 25∼55 times compared to the finite element analysis (FEA) method. The RESCA method has double advantages in the efficiency and accuracy for the damage analysis of 3DWCs.