PSI - Issue 57

Matthias Hecht et al. / Procedia Structural Integrity 57 (2024) 581–588 Matthias Hecht et al. / Structural Integrity Procedia 00 (2019) 000 – 000

583

3

presented here, at a load ratio of R = 0.1. The fatigue life under constant and variable amplitudes was investigated. It was experimentally proven that the fatigue life under variable amplitudes is significantly higher compared to constant amplitudes. However, the actual damage sum, which describes this relationship via the linear damage accumulation according to Palmgren [11] and Miner [12], is subject to high fluctuations. Furthermore, it was found that under variable amplitude loading, the fullness of the spectrum and the sequence length have no significant influence on the cyclic fatigue life. ▪ Tests under multiaxial loading at constant amplitudes were carried out within project [10]. The test setup was similar to the tests discussed here. Whereby the clamping was slightly modified, resulting in slightly higher stresses in the adhesive layer at the same loading force in these tests. With an R -ratio of R = 0.1, two different ratios between the vertical a,Ax and the horizontal loading force a,M were used in [10]. At ratio a,Ax =0.3 ∙ a,M (3) multiaxial tests were performed without phase shift. Compared to pure vertical loading, the endurable loads in the a,Ax -direction dropped by 6 % for a fatigue life of N = 2 ∙ 10 5 cycles. At the ratio a,Ax =0.7 ∙ a,M (4) this fatigue strength dropped by 23 %. At this load ratio, additional multiaxial tests were performed with a phase shift of φ = 90°. Here, the phase shift resulted in a 24 % higher fatigue strength at a life of N = 2 ∙ 10 5 cycles, compared to the proportional loading. A combination between variable amplitudes (first point) and multiaxial loading (second point) has not been investigated so far, therefore this paper presents the experimental results on these investigations and thus combines the research from [8,9] and [10]. 2. Test procedure The tested bowl specimens consist of two steel sheets that were bonded together. A representation of the bowl The bowl specimen consists of a bottom sheet and the deep-drawn bowl, which are bonded together. The steel sheets of HC340LA+ZE75/75 have a thickness of 1.2 mm and the surface was electrolytically galvanized. The used adhesive is a one-component, thermosetting, toughened structural adhesive based on epoxy (BETAMATE ™ 1496V), which has been used in many previous research projects [7 – 10,13 – 15]. During specimen production, the joined parts are first cleaned with n-heptane. An adhesive bead is then applied to the preheated bowl. To ensure a defined and constant adhesive layer thickness of t a = 0.3 mm few glass beads of the appropriate diameter are scattered on the adhesive layer. The bowl is then pressed onto the bottom sheet using a positioning device and fixed with spring clamps. The adhesive that was squeezed out in the process is removed from the inside and outside of the bowl in the next step using a spatula. The specimens are then placed in a preheated oven and cured. During this process, the temperature on a sheet dummy specimen is measured with a thermocouple and recorded for each batch. Once the target temperature of 180 °C of this dummy is reached, this temperature is maintained for a period of 30 min. Then, the specimens are cooled with the oven door open. In the final step, the cooled specimens are removed from the oven and the specimens are stored at room climate for at least ten days before the experiment. 2.2. Test rig The prepared bowl specimens are subjected to cyclic loading under load control on a servo-hydraulic testing machine. Due to the complex specimen geometry, a multiaxial stress state is present even under uniaxial loading. Nevertheless, in order to be able to investigate the influence of two different loading modes and in particular the influence of the phase shift of φ = 90°, the specimens are loaded with two independent hydraulic cylinders, Fig. 1. A specimen is shown on Fig. 1. 2.1. Specimen preparation