Skip to main navigation Skip to search Skip to main content

Revealing critical failure laws of composite curved beams through network-free renormalization and clustering algorithm

  • School of Civil Engineering, Harbin Institute of Technology
  • Harbin Institute of Technology
  • Commercial Aircraft Corporation of China, Ltd.

Research output: Contribution to journalArticlepeer-review

Abstract

Curved beams are irregular composite structures commonly used in aircraft and ships. The complexity of their material properties and irregular geometries make it challenging to define their failure accurately based on the failure phenomenon. Moreover, few full-size composite laminate curved beam tests have been reported in previous publications because of their high cost and difficulty. This work proposes a 3-order network-free renormalization method based on the thermodynamic-based failure definition to reveal the failure law of curved beams more accurately, which is suitable for characterizing the stressing state of full-size irregular composite structures. Applying 3-order network-free renormalization and clustering algorithms to eight full-size composite curved beams can reveal the elastoplastic branching (EPB), failure starting (FS), and progressive failure (PF) points. We can also use the phase transition definition of Wilson's theory to verify the stability of phase transition loads. Unlike failure loads based on buckling or fracture phenomena, phase transition loads in composite curved beams are based on catastrophes in the relative deformation distribution, which is more physically significant.

Original languageEnglish
Article number119857
JournalEngineering Structures
Volume330
DOIs
StatePublished - 1 May 2025

Keywords

  • Bending condition
  • Clustering algorithm
  • Composite curved beams
  • Network-free renormalization
  • Phase transition loads

Fingerprint

Dive into the research topics of 'Revealing critical failure laws of composite curved beams through network-free renormalization and clustering algorithm'. Together they form a unique fingerprint.

Cite this