A data-driven approach for instability analysis of thin composite structures

Abstract

This paper aims to propose a data-driven computing algorithm integrated with model reduction technique to conduct instability analysis of thin composite structures. The data-driven computing method was originally introduced by Kirchdoerfer and Ortiz (2016), whose basic idea lies in directly employing the stress and strain sets to drive the mechanical simulation, thus eliminating the material modeling error and uncertainty. By introducing the Euler–Bernoulli beam theory into data-driven computing, the one-dimensional reduced beam model is adopted by the herein proposed approach, namely structural-genome-driven (SGD) computing. In this manner, not only the integration points number but also the database phase space dimensions will be decreased, thereby enhancing the computational efficiency for structural analysis. Besides, the weight coefficient settings in data-driven penalty function are determined by the locally tangent linear material behavior of the data sets and are updated for each integration point during data-driven iterations. Several demonstrative numerical tests are performed to validate the robustness and effectiveness of the proposed method in predicting buckling path and bifurcation point.