Strain localization in reduced order asymptotic homogenization
By: Harpreet Singh, Puneet Mahajan
Potential Business Impact:
Predicts how strong materials break and bend.
A reduced order asymptotic homogenization based multiscale technique which can capture damage and inelastic effects in composite materials is proposed. This technique is based on two scale homogenization procedure where eigen strain representation accounts for the inelastic response and the computational efforts are alleviated by reduction of order technique. Macroscale stress is derived by calculating the influence tensors from the analysis of representative volume element (RVE). At microscale, the damage in the material is modeled using continuum damage mechanics (CDM) based framework. To solve the problem of strain localization a method of the alteration of stress-strain relation of micro con- stituents based on the dissipated fracture energy in a crack band is implemented. The issue of spurious post failure artificial stiffness at macroscale is discussed and effect of increasing the order to alleviate this problem is checked. Verification studies demonstrated the proposed formulation predicts the macroscale response and also captures the damage and plasticity induced inelastic strains.
Similar Papers
Characterizing failure morphologies in fiber-reinforced composites via k-means clustering based multiscale framework
Computational Engineering, Finance, and Science
Predicts how strong materials will break.
Heterogeneous multiscale methods for fourth-order singular perturbations
Numerical Analysis
Makes materials stronger by understanding tiny details.
E$^2$-TFA based multiscale analysis of failure in elasto-plastic composites
Computational Engineering, Finance, and Science
Predicts when strong materials will break.