1DNV, Columbus, USA. 2CenterPoint Energy, Evansville, USA
In-line inspection (ILI) run-to-run comparisons are a cornerstone of pipeline integrity management. By tracking changes in measured metal loss (corrosion) anomalies between successive inspections, integrity personnel can pinpoint which anomalies are accelerating toward criticality and which may no longer warrant immediate excavation. This targeted insight not only helps to re-prioritize digs—removing lower-risk anomalies in favor of those growing the fastest—but also maximizes limited excavation budgets and minimizes overall system risk. Corrosion growth rates (CGRs) from ILI run-to-run comparisons can be derived using various workflows. For this study, the following methods were utilized:
- Statistical segmentation approach
- Establish CGRs on statistically prominent sections of the pipeline by grouping anomalies into regions of similar susceptibility, pipe-tally, and engineering judgment and assigning each section a single, characteristic CGR (typically the maximum historical CGR).
- Joint-based approach
- Establish CGRs for each individual pipe joint by directly comparing its reported anomaly depths between ILI surveys, calculating a unique CGR per joint based on the change in wall-loss over the inspection interval.
Both the statistical segmentation and joint-based approaches are regularly used within the industry to establish CGRs. The statistical segmentation method simplifies integrity planning by assigning one characteristic CGR per corrosion-susceptible region. However, because the approach is based on a chosen population statistic, it can veer into over-conservatism (triggering unnecessary excavations) or, conversely, under-conservatism if localized peak growth is masked. By contrast, the joint-based approach calculates a unique CGR for each pipe joint by directly comparing depth measurements between the subsequent ILI surveys. The joint-based approach often yields more realistic, localized corrosion growth estimates, albeit with greater scatter from measurement noise and stochastic behavior.
In this case study, the authors applied both approaches to ILI datasets from two consecutive inspections on the same pipeline asset. First, the authors generated characteristic segment CGRs by clustering anomalies via statistical models based on anomaly density, integrating pipe-tally transitions (grade, diameter, coating, welds, etc.), rationalizing the resulting sub-segments, and assigning each segment a CGR based on maximum or 95th percentile values. The authors then computed joint-specific CGRs by directly measuring wall-loss changes at every inspected joint between the subsequent ILI surveys. When both approaches were applied to evaluate remaining lives of ILI-reported metal loss anomalies, the statistical segmentation approach indicated a need for excavations at nearly 33% of ILI-reported corrosion anomalies, whereas the joint‐based method cut that down to just 3%. Moreover, the joint‐specific analysis uncovered several localized corrosion hotspots that the segmented approach smoothed over. This demonstrates that using segment‐level rates for system‐wide budgeting and inspection planning, paired with joint‐specific CGRs for precise excavation prioritization, achieves the best balance of conservatism, risk mitigation, and resource allocation.
Available after the event



