Probabilistic Integrity Assessment of Offshore Pipelines Using Intelligent Pigging Data to Support Risk-Based Repair and Inspection Planning (Published)
Offshore pipelines transporting hydrocarbons and injection water are high-value, high-risk assets whose failure can inflict catastrophic safety, environmental, and economic consequences. Rigorous integrity management is therefore mandatory, yet must be executed economically over decades of operation. Conventional deterministic assessment codes (ASME B31G, DNV-RP-F101 Part A) compress multi-dimensional intelligent pigging (ILI) data into single “worst-case” defects evaluated with fixed safety factors. These procedures do not propagate inspection uncertainty, corrosion-growth variability, or operational fluctuations, producing binary “dig/no-dig” decisions that are either prohibitively conservative or unknowingly risky.This paper presents an integrated probabilistic framework that couples high-resolution MFL/UTCD ILI data with operationally-conditioned corrosion-growth models to estimate time-dependent pipeline reliability. A Bayesian hierarchical model calibrates defect-specific growth rates using successive ILI runs while accounting for tool sizing error and detection probability. Posterior predictive distributions of depth and length feed a Monte-Carlo limit-state analysis based on the DNV-RP-F101 Part B burst equation to compute annual probability of failure (PoF) for every defect. PoF is combined with consequence of failure (CoF) categories specific to offshore gas export and water-injection services to quantify risk. An expected-cost minimisation algorithm optimises the next inspection date and generates a risk-ranked repair list under ALARP constraints. Applied to a 20-inch, 85 km wet-gas export line and a 16-inch, 45 km water-injection pipeline in the North Sea, the framework revealed bimodal corrosion-rate distributions driven by slug-flow CO₂ excursions in the gas line and negligible growth under oxygen-controlled conditions in the water line. The probabilistic schedule extended the gas-line inspection interval by 18 months and reduced immediate repairs from 67 to 14 defects compared with deterministic DNV Level-1, cutting forecast expenditure by 58 % while lowering system-level PoF by an order of magnitude. The water line qualified for a 12-year interval versus 5 years deterministically, deferring USD 2.4 M in unnecessary interventions. Sensitivity analysis shows corrosion-rate uncertainty dominates PoF variance, guiding operators to prioritise repeated high-resolution surveys over marginal gains in tool accuracy.The study delivers a traceable, data-driven decision-support tool that transparently links raw ILI signals to risk-optimal inspection and repair actions, enhancing both the economic and operational efficiency of offshore pipeline integrity management programs while demonstrably maintaining safety margins.
Keywords: ALARP, Bayesian updating, Monte-Carlo Simulation, Offshore pipeline, Risk-based inspection, corrosion growth modeling, failure probability, intelligent pigging, probabilistic integrity, time-dependent reliability
Development of a Predictive Corrosion Threat Index (PCTI) for Offshore Pipelines Using Multi-Source NDT Data Integration (Published)
Offshore hydrocarbon pipelines are subject to complex internal and external corrosion mechanisms, yet integrity management remains largely reactive due to the fragmentation of high-resolution inspection data across multiple non-destructive testing (NDT) methods. Intelligent pigging (ILI), subsea ultrasonic thickness measurements, cathodic protection (CP) surveys, ROV visual inspections, and production chemistry analyses each generate detailed but isolated evidence of degradation, preventing operators from constructing a unified, forward-looking threat picture. This data-silo problem routinely leads to missed threats, inefficient resource allocation, and elevated risk of catastrophic failure. This research introduces the Predictive Corrosion Threat Index (PCTI), a novel, quantitative, and spatially explicit composite metric designed to overcome these limitations. The PCTI integrates multi-source NDT and operational data through a weighted-sum algorithm comprising an Internal Corrosion Score (ICS), External Corrosion Score (ECS), and Trend Severity Score (TSS). Inputs are spatially aligned to a common kilometre-post reference, normalised, and fused using weights calibrated against 173 confirmed corrosion-driven interventions across a 1,840 km fleet of 42 deepwater and shallow-water flowlines and export trunks in West Africa and the Gulf of Mexico.Retrospective validation and blind prospective testing demonstrated that the PCTI captures 94 % of actual failures within the top 100 ranked segments (versus 61–72 % for traditional ILI-only or semi-quantitative matrix methods) while flagging only 8–10 % of total pipeline length for heightened scrutiny. Segments exhibiting accelerating external corrosion beneath disbonded coatings—previously ranked low by ILI severity alone—were correctly elevated to top-tier urgency up to 15–36 months in advance. Predictive performance metrics on hold-out data yielded sensitivity of 0.94, specificity of 0.98, and AUC-ROC of 0.987.The PCTI transforms offshore pipeline integrity management from a fragmented, schedule-driven activity into a genuinely proactive, risk-prioritised discipline. By delivering diagnostic threat breakdowns and accurate forward projections, it enables precise repair scoping, dynamic optimisation of inspection intervals, and substantial capital savings. Implemented as an automated dashboard tool, the PCTI establishes a scalable blueprint for digitalised, data-driven integrity programmes that simultaneously enhance safety, environmental protection, and operational efficiency
Keywords: Cathodic protection surveys, Corrosion growth prediction, Data-driven integrity management, Deepwater flowlines, In-line inspection (ILI), Multi-source NDT data fusion, Offshore pipeline integrity, Predictive Corrosion Threat Index (PCTI), Proactive threat prioritization, Risk-based inspection