Digital Twin of a Generic Jack-Up Platform

Abstract A development of physics-based digital twinning of a generic jack-up platform is presented in this paper. Due to lack of field measurement data, a generic large-scale jack-up model was designed, fabricated and tested in TCOMS ocean basin at 1:30 scale under different configurations, with the objective to provide high-quality datasets to validate the proposed digital twin methodologies. The framework and the performance of the digital twin are demonstrated using a realistic and representative basin-scale model as a proof-of-concept. Fundamental to any physics-based digital twins is the establishment of numerical models capable of reproducing consistent behaviors and responses of the physical assets. For this digital twin development, a full order model (FOM) and a reduced order model (ROM) are established. In view of uncertainties associated with the physical asset and numerical modelling, e.g., foundation fixities, leg stiffness, leg-hull connection stiffness and hydrodynamic coefficients, model updating or system identification is performed using the ROM to identify the parameters with relatively large uncertainties. A mapping between the parameters and the associated responses of the FOM and the ROM is subsequently established. After the model updating is completed with the identified parameters, good agreement in terms of the structural responses between the model test and numerical results can be achieved. Both the FOM and ROM are able to reproduce structural responses with good accuracy when compared to physical measurements. The ROM, being a linear structural model based on modal responses, is unable to account for larger non-linear effects due to spudcan fixities, if any. Nevertheless, the ROM is suitable for fatigue evaluation considering fast computational speed and validity of the piecewise linear constraints as assumed for the foundation. The FOM, being less computationally efficient, is suitable for strength evaluation and able to account for any non-linear structural behaviors. The results of boundary displacements from the global dynamic response analysis can be mapped to a detailed local joint model to derive the hotspots stress for a more accurate fatigue evaluation. The digital twin framework for fatigue and strength evaluations based on measured wave loading is demonstrated for a better structural integrity management. As an emerging technology, digital twin will provide visibility of structural health condition to facilitate the transition from preventive to predictive and reliability-centered maintenance strategies. Although the digital twin framework presented in the paper makes use of a representative jack-up at model-scale, the proposed methodology can be potentially applied to full-scale operating jack-ups.