PSI - Issue 75

Philippe Thibaux et al. / Procedia Structural Integrity 75 (2025) 546–554 P. Thibaux et al. / Structural Integrity Procedia (2025)

548

3

sample is excited close to the chosen resonance frequency by two shakers rotating in opposite directions and with equal excitation amplitude, resulting in a force in one plane. Using this method, it is possible to force the excitation of the sample either in-plane or out-of-plane. The supports are located at the positions of the nodes of the resonance modes. The test is performed at the load ratio R=-1. The locations of failure and strain gauges are referenced by “h ourly positions ” . The saddle points are at 3 and 9h; the crown toes are at 12h on the left side and 6h on the right; the crown heel at 6h on the left side and 12h on the right (figure 1). The failure criterion was a breakthrough crack. The braces were filled with water with a slight over-pressure, while a membrane was installed inside the chord and put under vacuum. A breakthrough crack is detected by a pressure change in the considered member.

Chord

Crown heel

Crown toe

12h

12h

Saddle

Left Brace

Right Brace

9h

9h

6h

6h

Fig. 1. Picture of the sample tested in-plane.

Rosette or bi-axial strain gauges were installed close to the weld toe on the brace and on the chord close to the expected failure locations. Beside the strain gauges to monitor the strain evolution close to the weld toe, uniaxial strain gauges were installed far from the weld to monitor the global loading of the sample. The strain arrangement for the out-of-plane test is described in Thibaux and Thiele (2025). For the in-plane test, biaxial strain gauges were installed at 18 and 49mm (chord) or 56mm (brace) to extrapolate to the weld toe at the crown heel. At the crown toe, only uniaxial strain gauges were used. Two rosette strain gauges were also installed at 30° / 1h from the crown heel on the chord and on the brace. 3. Expected failure mode The stress concentration factors (SCF) were calculated using the formulas from Efthymiou (Efthymiou 1988) for the cases of balanced out-of-plane bending, balanced in-plane bending and axial loading. The SCF were 3.33 and 2.73 for respectively the chord and brace saddle points (out-of-plane bending), and 2.19 and 2.98 for respectively the chord and brace crown positions (in-plane bending). In axial loading, the highest SCF is at the chord saddle. In the in-plane test, one should expect a failure through the brace, while the failure will be through the chord in the out of-plane test. Finite element simulations with a solid-model were performed for a Y-joint with the same angle and tube dimensions. The geometry of the weld was neglected as a simplification. While the results of the out-of-plane bending simulations were as expected, the in-plane bending simulations indicate that the highest stress concentration is not necessarily at the crown heel. In figure 2, the evolution of the stresses measured at 49mm or 56mm (red curve), 18mm (blue curve) and the extrapolated stresses (green curve) are presented for the brace (left) and the chord (right). The crown toe is at the 90° position and the crown heel at 270°. On the brace, the maximum SCF is