Buckling analysis of cosine functionally graded plates: Synergy between finite element and Gradient boosting machine learning algorithm
Tayeb Si Tayeb,Mohamed-Ouejdi Belarbi,Tan Ngoc Nguyen,Smain Bezzina,Mohammed Sid Ahmed Houari
Abstract
Cosine Functionally Graded (CFG) plates offer a smoother material transition and better stiffness-toweight ratios than traditional grading profiles. In this study, we investigate the buckling behavior of these plates using a recently developed Higher-Order Zigzag Theory (HOZT). The high-fidelity data generated from our HOZT model served as the training set for a Gradient Boosting Machine (GBM) algorithm. We used key parameters such as skew angle (Φ), boundary conditions, aspect ratio (a/b), side-to-thickness ratio (a/h) and gradient index (ℜ) as input features to predict the critical buckling load (𝑃̅𝑐𝑟 ). Our numerical results, show that the GBM model is remarkably accurate. A parametric dataset is generated considering key variables, including skew angle (0°-60°), aspect ratio (a/b), side-to-thickness ratio (a/h=5-100), boundary conditions, and material gradient index (ℜ). The results show that the non-dimensional critical buckling load significantly increases with skew angle (up to ~150% increase from 0° to 60°) and decreases with increasing a/h, with the most pronounced variation observed for a/h≤20. The GBM model demonstrates excellent predictive capability, with deviations from finite element results remaining negligible (typically less than 1% across all tested cases). Overall, this work demonstrates that coupling physics-based simulations with data-driven models is a highly effective way to reduce costs while maintaining the precision needed for the rapid stability assessment of advanced structures.
Tayeb Si Tayeb — University of Ahmed Ben Yahia El Wancharissi, 38000 Tissemsilt, Algeria; Laboratory of Geomatics and Sustainable Development, Ibn Khaldoun University of Tiaret, Algeria
Mohamed-Ouejdi Belarbi — Laboratoire de Recherche en Génie Civil, LRGC, Université de Biskra B.P. 145, R.P., Biskra 07000, Algeria
Tan Ngoc Nguyen — Department of Architectural Engineering, Sejong University, Seoul, 05006, Republic of Korea
Smain Bezzina — Deanship of Scientific Research, King Abdulaziz University, Jeddah, Saudi Arabia
Mohammed Sid Ahmed Houari — Laboratoire d’Etude des Structures et de Mécanique des Matériaux, Département de Cénie Civil, Faculté des Sciences et de la Technologie, Université Mustapha Stambouli, B.P. 305, Mascara, Algeria
PDF Viewer
Preview is limited to the first 3 pages. Sign in to access the full PDF.
