1ROSEN Group, Houston, USA
Traditional remaining life estimates based on anomaly "box" dimensions introduce significant uncertainty, particularly in pipelines with complex corrosion morphologies, often leading to conservative or inefficient repair prioritization. These challenges are further amplified when integrating multi-vendor and multi-technology ILI datasets. Reliable remaining life prediction in pipeline integrity management depends on two key factors: (i) accurate in-line inspection (ILI) sizing and (ii) representative corrosion growth rates. Novel machine learning-based Data Fusion (DF) process has demonstrated improved metal loss characterization by integrating axial and circumferential magnetic flux leakage (MFL) data. Leveraging DF outputs, which provide profile-based anomaly characterization, reduces the uncertainty inherent in conventional ILI box-based approaches. By combining axial and circumferential MFL data, a 3D metal loss depth map is obtained, enabling improved assessment when used with established profile-based assessment methods such as Remaining Strength (RSTRENG) and Plausible Profiles (Psqr). This paper presents a case study applying existing corrosion growth assessment (CGA) methodologies to a pipeline containing complex corrosion morphology. Signal-level analysis and box-to-box comparisons were performed to quantify the improvement in remaining life estimates obtained through profile-based corrosion characterization. The results demonstrate that DF is compatible with existing CGA methods and can be seamlessly integrated into integrity management frameworks. The combined application of DF and appropriate CGA enhances reliability in anomaly sizing and remaining life prediction, enabling more informed and optimized integrity decisions.
Available after the event



