Investigating the Mode-I Failure behavior of Thick Adhesive Joints using a coupled computational/experimental approach
     Topic(s) : Material and Structural Behavior - Simulation & Testing

    Co-authors​ :

     Akash SHARMA (BELGIUM), Ali SHIVAIE KOJOURI (BELGIUM), Jialiang FAN (SWITZERLAND), Anastasios P. VASSILOPOULOS (SWITZERLAND), Véronique MICHAUD (SWITZERLAND), Kalliopi-Artemi KALTEREMIDOU (BELGIUM), Danny VAN HEMELRIJCK , Wim VAN PAEPEGEM (BELGIUM) 

    Abstract :
    Wind turbine blades are manufactured by molding them in two halves and joining them using thick adhesive joints. The failure of these adhesive joints, particularly in the trailing edge of the blades, compromises the structural integrity of the wind turbine. Therefore, comprehending the mechanisms of failure in adhesive joints is critical to designing wind turbine blades efficiently. For this purpose, the present study proposes a novel approach that integrates computational and experimental methods to enhance the overall understanding of the factors that influence the failure of thick adhesive joints. The experimental specimens consist of two cross-ply glass fiber composite laminates bonded with a ~10 mm thick layer of an epoxy-based adhesive. The specimens are cured at 70°C. After curing, a pre-crack is generated within the adhesive layers of each specimen. The specimen is subjected to Double Cantilever Beam (DCB) tests at room temperature to induce mode I failure. The load-displacement curves of the DCB specimens are obtained. The strain in the adhesive layer is determined using the Digital Image Correlation (DIC) (Fig. 1a). FE models of the DCB specimens having virtually generated pre-cracks are created to predict the experimental load-displacement curves. So far, most researchers have employed the cohesive zone model for the adhesive in such numerical studies. However, epoxy-based adhesives typically exhibit plastic deformation. Hence, the Drucker-Prager plasticity criteria are utilized to model the mechanical response of the adhesive. Also, it is crucial to assess the influence of thermal residual stresses that arise from the thermal mismatch between composites and adhesives, an aspect that has not been adequately addressed in the literature. Thus, appropriate thermal expansion coefficients are assigned to both composites and adhesives. Furthermore, a cool-down is simulated before mechanical loads to mimic the temperature transition from curing to room temperature. Concurrently, the influence of mesh size on crack propagation is explored. A very good agreement is observed between the experimental and numerical results (Fig. 1b). A satisfying correlation is also observed between the bending strains obtained using FE analysis and the DIC, further verifying the efficacy of the proposed modeling strategy (Fig. 1c).
    Acknowledgment
    The authors acknowledge funding under the 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 with the title "Combined numerical and experimental approach for the development, testing and analysis of thick adhesive joints in large wind turbine blades”.
    a)
    b)
    c)

    Fig 1. a) Strain Contours from DIC in the DCB specimen; b) Load-Displacement Curves: Experimental Results versus Finite Element Analysis (FEA); c) Bending strains in the adhesive layer: DIC versus FEA.