Physics-guided data-driven failure identification of underwater mooring systems in offshore infrastructures

Many offshore infrastructures have been developed to explore vast marine resources over the past several decades. In addition to the conventional fixed-type offshore infrastructures, a new class of offshore infrastructures, the so-called floating offshore infrastructures, have gained dramatically increasing applications owing to their flexible deployment and enhanced capacity in renewable energy exploitation in deep seawater. As the key functional component of the floating infrastructure, the underwater mooring systems are subject to sustained dynamic loads pertinent to marine waves and currents, which are prone to different types of failures. Identifying those mooring system failures timely and reliably thus plays a vital role in offshore infrastructure health management and maintenance. This study aims to achieve this objective by developing an integrated numerical framework that seamlessly synthesizes the physical mooring system modeling and data-driven analysis. Specifically, a high-fidelity physical model that takes into account the sophisticated fluid-structure interaction is established to mimic the underlying behavior of the mooring system. The mooring line failures are incorporated into the model to generate the respective dynamic responses. With the aid of data-driven modeling, the causative relationship between mooring line failure scenarios and dynamic responses can be characterized. Given the sensor measurement in actual practice, this framework offers a feasible solution for the failure identification of underwater mooring systems. The results clearly demonstrate the feasibility of the proposed methodology.