Bayesian Network‐Based Reliability Analysis of Suction Anchor System

ABSTRACT The suction anchor system is a critical anchoring foundation for offshore platforms due to its cost‐effectiveness, ease of maintenance, and reliable performance, making it widely utilized in offshore engineering. However, its complex structure and harsh operational environment present significant challenges. Despite its importance, targeted reliability analyses of suction anchor systems under complex conditions remain limited. This study addresses this gap by employing the “four‐unit” decomposition method, segmenting the suction anchor system into four subsystems: the anchor, emergency release system, pump unit, and control system. Failure Mode, Effects, and Criticality Analysis (FMECA) is conducted at both the subsystem and component levels. Based on these analyses, a fault tree model is developed, where system failure is defined as the top event, subsystem and component failures as intermediate events, and fundamental failure causes as base events. The reliability of the suction anchor is subsequently assessed using the fault tree model and Bayesian network mapping. Results reveal that among all subsystems, the control system exhibits the highest fault contribution rate and the fastest reliability degradation. Furthermore, considering seismic influences, system failures are primarily attributed to anchor settlement, slippage, and deviations in inclination angle. To ensure the stability of the suction anchor during installation, excessive deviations in inclination angle, directional angle, and installation depth must be effectively mitigated.