Converting laser scans of tubular joints to finite element models for detailed stress analysis

Jacket type substructures for offshore wind turbines began to gain prominence in the 2010s, particularly in locations where monopiles were no longer feasible due to increasing water depths and turbine sizes. As these turbines approach their design lifetime, there are three options for their substructures: decommissioning, lifetime extension, or repowering. The latter two options are economically and ecologically more attractive, but require an accurate assessment of the remaining fatigue life. The FlexWind project, supported by the Belgian Energy Transition Fund, aims to develop a novel method for evaluating the remaining fatigue resistance of offshore jacket substructures based on 3D laser scans of critical joints. Previous research at Soete Laboratory has shown that the real weld geometry has a strong influence on local stress concentrations and, consequently, the fatigue life of the structure. In the design phase of offshore jacket substructures, the fatigue governing stresses are determined assuming perfectly cylindrical tubular members without any surface degradation. Welds are idealized or only implicitly considered in the calculations. This research aims to improve the fatigue assessment by incorporating the as-built global joint geometry (ovality and out-of-roundness), surface irregularities (e.g. corrosion) and the real weld geometry. In the FlexWind project, four full-scale tubular joints were scanned. The joints originate from a decommissioned gas platform jacket. The platform was installed in the North Sea in the 1980s and was decommissioned in 2023. A handheld 3D laser scanner was used to capture the global and local geometry of the joints. Because only the external surface of a tubular joint can be scanned with a light-based scanner, algorithms were developed to automatically reconstruct the inaccessible areas of the joint. A key innovation in this research is the automatic generation of solid finite element models from the 3D surface scans. In this work, the different steps and algorithms incorporated in the reconstruction framework will be explained. Attention will also be drawn to the computational cost of the reconstruction and how the framework was optimized to reduce it.