Semi-submersible structural design

Abstract

Semi-submersibles are common vessels in the offshore sector. These vessels usually have a bracing structure to restrain floater movement and to support the deckbox structure. The bracing configuration influences strength, fatigue and other semi-submersible parameters. The first assignment objective was to study the structural design impact on a twin-pontoon semi-submersible when applying different bracing configurations. The influence of bracings on characteristic responses, defined as loading and accelerations governing for the strength and fatigue of semi-submersibles, was studied first. Generally, an increase in bracing diameter results in an increase in characteristic response. However, bracings do not affect characteristic responses much, as differences below 11% are observed. Global strength assessments in the ultimate and accidental limit states were performed using finite element analyses (FEA) for different bracing configurations. Each bracing configuration differentiates itself being beneficial for certain load cases, or is beneficial regarding fatigue sensitive locations. The structural design of semi-submersibles with different bracing configurations were modified to have similar structural performance compared to a reference semi-submersible. The bracing configurations were evaluated based on payload, structural centre of gravity, structural redundancy and fatigue. Generally, the presence of transverse horizontal bracings affects structural performance most significantly, since a payload reduction of 22% is observed when not present due to the dominant splitting force load case and ineffective load path. Adding diagonal bracings in the horizontal or vertical plane, reduces column and deckbox loading for the longitudinal shear, torsion moment and inertia load cases, resulting in a payload increase up to 5%. Omitting braces results in the lowest amount of fatigue sensitive locations. However, since the columns and deckbox structure needs strengthening, welding volume at other fatigue sensitive locations increases, which affects fatigue negatively. The bracing configuration selection should be merely based on the semi-submersible’s requirements. Therefore, the designer should first rank the requirements after which a bracing configuration can be designed.

Fatigue of semi-submersibles remains a challenge in today’s practice, since service cracks are frequently found during inspections. Fatigue resistance is usually determined by S-N curves based on tested small-scale specimens (SSS). Large-scale specimen (LSS) and full-scale specimen (FSS) fatigue tests are performed less frequent. Semi-submersible tubular brace-column and brace brace fatigue resistance is typically determined by LSS tubular joint tests. The fatigue resistance similarity between SSS planar joints and LSS tubular joints was examined by the hot spot stress and averaged effective notch stress concepts. Five different LSS tubular joint fatigue tests, derived from literature, were studied by shell FEA and volume FEA. Weld geometry in tubular joint shell FEA can affect results significantly. When not present, bending stresses are overestimated up to 208%. For most LSS tubular joints, similarity with respect to SSS planar joints is increased for the average effective notch stress concept, compared to the hot spot stress concept. Most LSS tubular joints fit inside the average effective notch
stress SSS planar joint fatigue resistance data. Dissimilarity for divergent LSS tubular joints can be related to different modelling assumptions compared to actual conditions, where differences in boundary conditions and weld geometries is most likely. Fatigue resistance similarity, expressed as the strength scatter index and intercept log10(𝐶), of LSS tubular joints is increased compared to hot spot fatigue resistance. Differences in slope 𝑚 are similar. A SSS planar joint based design S-N curve is therefore applicable for tubular fatigue sensitive locations of semi-submersibles and other structures. Moreover, this study increases the applicability of the average effective notch stress as fatigue assessment concept. More LSS tubular joints should be studied to demonstrate similarity with higher confidence. 
To accurately estimate fatigue lifetime, a detailed fatigue assessment is performed of a tubular brace-brace connection. Multi-axial loading was studied at global and local level, however at the critical saddle location, mode-I stress dominates. Therefore, multi-axial fatigue was not considered. The structural stress,𝑆𝑠, average effective notch stress,𝑆𝑒, and hot spot stress,𝑆ℎ, were applied as fatigue assessment concepts. The fatigue assessment of the simple tubular joint concluded insufficient fatigue lifetime. Stress intensities at the critical saddle location were reduced by implementing internal ring-stiffeners, classifying the connection as a complex tubular joint. Stress reduction factors of 3.9 and 4.1 were achieved, resulting in fatigue damages below 1, thus acceptable. Compared to common fatigue assessment concepts, the detailed 𝑆𝑠 and 𝑆𝑒 fatigue assessments reduce the possibility of service cracks and maintenance and inspection work can be planned more precise. However, DNV-GL and IIW guidelines state a fatigue resistance slope change is present above 107 cycles (N), which is not accounted for in 𝑆𝑠 and 𝑆𝑒. To study the presence of a slope change and to possibly establish a more accurate fatigue damage estimation for 𝑆𝑠 and 𝑆𝑒. a recommendation for further research is to include
more fatigue tests for 𝑁 > 107, from which a design S-N curve can be derived.