PSI - Issue 57

Cristian Bagni et al. / Procedia Structural Integrity 57 (2024) 598–610 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

601

4

integral which does not distinguish between tensile and compressive loads. Furthermore, this approach requires the meshes of the adherends and of the adhesive to be congruent and consequently, additional meshing efforts. To optimise the design of adhesively bonded structures and reduce the risk of in-service fatigue failures, the transportation industry needs efficient, robust, and easy-to-use approaches for the modelling and fatigue life estimation of adhesively bonded joints. This work presents a practical methodology for estimating the fatigue life of adhesively bonded joints that can be easily adopted by companies in the transportation industry. In Section 2, FE modelling guidelines are proposed to recover the required stresses along the bond lines. Furthermore, it is also described how the results of FE analyses can be used in nCode DesignLife to estimate the fatigue life of adhesive joints using the standard SN analysis engine and bespoke SN curves obtained through testing of adhesively bonded specimens, representative of the joints in the production parts, as described in Section 3. 2. Finite element modelling and fatigue analysis The proposed methodology requires an FE model of the structure to be solved in order to recover the required stresses along the bond lines. In particular, the proposed methodology uses the peel (or opening) stress to estimate the fatigue damage and fatigue life, since, as mentioned in Section 1, these are the most detrimental stresses for adhesively bonded joints. The first part of this Section describes an FE modelling strategy to recover the peel stresses along the bond lines. The proposed strategy requires reasonably little changes to the typical FE modelling strategies currently used, especially in the automotive industry. Furthermore, it provides FE models that are not computationally too onerous, with a good level of mesh insensitivity and that do not require congruent meshes. Finally, at the end of this Section, there is a brief description of how the DesignLife standard SN Analysis Engine can be used to predict the fatigue life of adhesive joints using the results from the FE analysis. The proposed FE modelling strategy consists of modelling the adherends using mid-surface linear shell elements and the adhesive using linear solid elements. Furthermore, the exposed faces of the solid elements are wrapped with linear shell elements with membrane behaviour only (Fig. 2). The membrane shell elements are modelled using the properties of the adhesive and with a negligible thickness (e.g. 10 -6 mm) to avoid introducing any noticeable contribution to the behaviour of the joint. The solid elements and the shell elements (both adherends and membrane shell elements) are connected using ‘bonded contacts’ (or similar tools, depending on the FE software used). With this modelling strategy, the adhesive is modelled from mid-surface to mid-surface of the adherends and therefore, it is thicker than in the physical joint. As a consequence, the Young’s modulus of the adhesive might need to be corrected to account for the larger thickness and preserve the real stiffness of the joint. Finally, the peel stresses are recovered from the centroid of the membrane shell elements.