PSI - Issue 75

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

547

2

intensive and represent 10% of the capital expenditure of a wind farm project cost, see Santhakumar (2022) or Stehly (2022). The increase of offshore wind energy requires installing wind farms in more remote locations, deeper water and with larger generator power. The design of steel foundations for wind energy converters is often fatigue driven, with the welded joints being the critical locations in the structure, see Glienke (2024)) or Schaumann(2021). The jacket type of foundation is the preferred one for difficult soil conditions, combined with a large generator power. In a jacket structure, the main loads are passing through continuous members, the chords, with braces connecting them. In a typical jacket for a WEC, the angle between the braces and the chord is close to 45°. A significant stress concentration is present at the brace / chord connection, making these positions susceptible to fatigue cracks. The fatigue check of these connections compares the fatigue loading to the T-curve, which is the relevant S-N curve for tubular joints (DNV, ABS, API, ISO). The T-curve was established after many large-scale tests were performed in different countries. A literature review of these tests and some more recent ones was performed by the Health & Safety Executive (1999). Most of the tests were performed under axial loading on 90° joints. Tests were also performed in out-of-plane bending or in-plane bending. However, no test was performed in in-plane condition on a 45° joint. As the crown heel is the critical position for these conditions, it is striking that no test was performed in this case. The crown heel is also the position where the largest variation in weld geometry can be expected, because the accessibility for the welder is limited by the small angle, and it is then a common practice to weld on the outer side of the tube and no bevel on the tube. In the experiments establishing the T-curve, the failure criterion was a breakthrough crack. The loading was determined by calculating the structural stress (first principal stress) using measurements from two strain gauges located at two different distances from the weld toe, like given in Hobbacher (2016). In an offshore truss structure, like a jacket, it is expected that the dominant loads will be axial loading due to the action of the wind; the waves and current can also generate a significant out-of-plane bending loading. Both loading cases lead to failures from the saddle points, as these locations have the largest stress concentration factors. However, there is a clear lack of information regarding the fatigue strength for in-plane bending and failure at the heel. A recent Joint Industry Project called Jaco has tested tubular joints in fatigue with the purpose of comparing the fatigue strength of joints produced by robots compared to joints that were made by manual welding. In that test program, all the joints had the same dimensions. Most of them were tested in out-of-plane bending, but some were tested in-plane. The general results of the out-of-plane tests were reported by Thibaux et al. (2025). In this paper, the fatigue tests (one in-plane bending and one out-of-plane bending) of two joints that were produced by the same fabricator using manual welding are presented. The objective is to fill partially the gap in knowledge regarding the fatigue strength of 45° joint failing at the heel due to in-plane bending loading.

Nomenclature SCF

Stress Concentration Factor

WEC Wind Energy Convertor DPI Dye Penetrant Inspection

2. Experimental methods Large-scale tests were performed on X-joints with an angle of 45°. The dimensions of the joint were chosen to be representative of a Y-joint, with a chord of diameter 1219mm and thickness 50mm, connected to two braces / stubs of diameter 610mm and 25mm thickness (see figure 1). Two samples were produced by manual welders using flux cored arc welding at an offshore yard. The tubes were produced from plates of steel grade EN10025:4 S355ML+Z35. The samples were loaded in a test bench specifically designed for the fatigue testing of tubular joints. The sample is designed such that both resonance modes (in-plane and out-of-plane) have very close resonance frequencies. The