Root Cause Failure Analysis and Predictive Risk Modeling of Lower Riser Package Leak Events in Ultra Deepwater Oilfields

Leak events in Lower Riser Packages (LRPs) pose significant risks to ultra-deepwater oilfield operations, threatening equipment integrity, environmental safety, and production continuity. This study presents a comprehensive investigation into the root causes of LRP leak failures and introduces a predictive risk modeling framework for early detection and prevention. A structured methodology was employed, beginning with the acquisition and preprocessing of operational, maintenance, and failure data from multiple offshore assets. Root Cause Analysis techniques, specifically Fault Tree Analysis and Failure Mode and Effects Analysis, were applied to identify the principal drivers of leak events, including design flaws, material degradation, manufacturing defects, and procedural lapses. These findings were used to inform the development of a predictive model using tree-based machine learning algorithms, capable of classifying risk states and estimating time-to-failure without reliance on simulation data. The model achieved high precision and recall, with environmental variables, component age, and operational stress emerging as key predictors. The integration of predictive insights with empirical diagnostics enables a shift from reactive to proactive maintenance strategies. This dual approach enhances equipment reliability, informs design improvements, and supports safer, more efficient ultra-deepwater operations. Recommendations for expanding the model to other subsea systems and incorporating real-time monitoring are also discussed.