New article in Composite Structures

© 2024 EPFL

© 2024 EPFL

Combined computational-experimental investigation of residual stresses and pre-cracking in mode I behaviour of thick adhesively bonded GFRP composite joints

This paper presents a novel Finite Element (FE) simulation approach to examine the mode I fracture of thick adhesive joints used particularly in the trailing edge of the wind turbine blades. The approach involved FE models of the DCB specimens focusing on aspects overlooked in the existing literature. There has been limited investigation on residual stresses caused by thermal mismatch between composites and adhesives. Similarly, the impact of generating notches/pre-cracks in the adhesive layer during the preparation of Double Cantilever Beam (DCB) specimens on residual stresses has received minimal attention. Additionally, the Cohesive Zone Model, commonly used for simulating elastoplastic adhesives, may be inadequate due to its inability to account for the plastic deformation of the adhesive. In the present work, the pre-cracks were virtually generated in DCB FE models so that their effect on the stresses within the joint could be examined, making it a novel contribution to the field. The components were assigned with appropriate thermal expansion coefficients, and a simulation of the cool-down process was conducted to determine the thermal residual stresses. Furthermore, the Drucker-Prager plasticity criteria were used to capture the elastoplastic behaviour of adhesives in the FE simulations. Concurrently, the T-stresses were assessed through numerical investigations. For validation, experiments were conducted on DCB specimens made of two cross-ply composite laminates bonded with a ∼ 10 mm thick layer of an epoxy-based adhesive. A good agreement between computational and experimental results was observed, confirming the effectiveness and reliability of the proposed approach.

Funding

Lead Agency scheme from the Research Foundation - Flanders (FWO Vlaanderen) through the project grant G031020N and the Swiss National Science Foundation (SNF) through the project grant 200021E_18944/1