PSI - Issue 19

Jan Presse et al. / Procedia Structural Integrity 19 (2019) 423–432 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

430

8

Fig 5 shows that the load case has the most influence on the fatigue strength of adhesively bonded connections. Additionally, under quasi-static loads, it has been observed that the stiffness of the structure is driven by different parameters for the two load cases. Under shear loads, the stiffness in longitudinal direction dominates the stiffness of the structure and is driven by the Young's modulus of the assembled sheets. The bending stiffness is driven by the thickness of the assembled sheets, which leads to higher stiffness values of material symmetric aluminum connections with = 1.5 compared to steel as the adherend with = 0.9 for the CP load case. Since the failure initialization of adhesives is a result of local increased stresses arising from deformation of the structure, the yield strength of the assembled sheets and the stiffness of the structure have the main impact on the resulting strength. Similar behavior can be observed under cyclic loads. Under shear loads, the fatigue performance of the material symmetric connections with DP600 is higher than for connections that include the thicker EN AW-6016 sheets. It should be noted that with undercutting a specific load range for the lap shear configuration failure occurs in the aluminum sheet instead of the adhesive. Since this approach assumes failure in the adhesive layer all diverging failure modes have been excluded. For the coach peel configuration, the bonded aluminum sheets show the highest fatigue performance due to their high bending stiffness, while the less stiff material symmetric steel connections show lower values. 5.2. Effective stress approach and modification A widely used approach to estimate the fatigue life of structural adhesives is the calculation of effective stresses in a local stress field around the joint. The effective stress enables a comparison of geometrically different specimens to the endurance limit of an un-notched specimen. To extract the relevant stress values the adhesive needs to be modeled in an FE-simulation as solid elements that are coupled with RBE2- and RBE3-elements to the modeled shell representation of the assembled sheets. For this approach, a path orthogonal to the location of the highest stress at the edge of the adhesive needs to be defined, to read out several values. Schmidt [9] considered two approaches to calculate the effective stress value. First, the critical distance concept, based on the works of Tayler [10]. In this approach, a stress value at a certain dis tance (“critical distance”) along the above-mentioned path is set as effective stress. Another approach to derive a relevant effective stress value is the averaging of the stress over a micro-structural length. This approach is based on the works of Neuber [11]. Both approaches have the aim to find a representative stress definition that leads to a reference SN-curve for a specific adhesive, with the lowest possible scatter. The above-mentioned method of searching for the best fitting effective stress value needs to be performed for every adhesive separately. Since the method was formerly successfully used on the structural adhesive Betamate 1496V it seems reasonable to adapt it also on Teroson EP5089 from Henkel. But, for the investigated Teroson EP5089 the minimum of scattering of the resulting SN-curve is still high and occurs for the maximum major principal stress 1, and von Mises stress , on the edge of the notch. As the stress value is extracted from a modeled solid element, the elements need to be very small to gain stress values that describe the failure initiation on the edge for each load case configuration adequately.

◼ Sheets: shell elements

◼ Adhesive:solid elements ◼ Additional shell-layer

Fig 6. Schematic representation of the additional shell layer for the LS configuration.