Session: 03-03-06 Feature Assessment Case Studies Cracking II

Paper Number: 134039

134039 - Enhancing Stress Corrosion Cracking Feature Identification Using EMAT Data and Field Learnings 

Abstract: 

Abstract

The primary objective of the Electromagnetic Acoustic Transducer (EMAT) In-line Inspection (ILI) is to detect, identify, and size axially oriented surface-breaking cracks, predominantly Stress Corrosion Cracking (SCC), as well as other crack-like features such as fatigue, corrosion fatigue, and weld-toe cracks. It is important to note that not all features detected by EMAT are critical to pipeline integrity, especially in the absence of aggressive pressure cycling, which is typically observed in gas transmission pipelines. Some detected features, such as stable manufacturing defects that have passed hydrostatic testing, may not pose an immediate threat to the integrity of the gas transmission pipeline. With the increasing costs of excavation and non-destructive examination (NDE) inspections, prioritizing the mitigation of injurious features is vital to optimize integrity resource allocation and streamline the excavation schedule, while maintaining the required level of pipeline safety and reliability. This paper aims to identify which EMAT feature calls are most likely to be injurious crack-like features. To accomplish this, the study integrates learnings and findings from SCC susceptibility assessment, EMAT ILI campaign, and the Stress Corrosion Cracking Direct Assessment (SCCDA) program. To quantify the probability of an unexcavated feature being injurious following an EMAT call, a variety of influencing factors are considered, including SCC susceptibility characteristics, EMAT ILI tool performance, pipeline operating parameters, environmental factors, and NDE inspection results. This work explores and discusses several methodologies, such as Bayesian updating, logistic regression, and machine learning classification, highlighting the advantages and disadvantages of each. The insights gained from this analysis can be directly used in pipeline risk and reliability assessments to make risk-informed decisions. A case study is included to demonstrate the practical application of these methods in refining the excavation decision-making process for gas transmission pipelines.

Key words: SCC, EMAT ILI, feature identification, crack-like feature, Bayesian update machine learning.

Presenting Author: Gabriel Langlois-Rahme Enbridge GTM

Presenting Author Biography: Gabriel Langlois-Rahme is a Consulting Engineer who helps Enbridge with budget and safety decisions. He has experience harmonizing inspection data, environmental data, cathodic protection data, and finding benefits from data with risk and machine learning principles. He has developed risk software, user-friendly dashboards around ALARP assessments, and tools for corrosion and SCC defect assessments.

Authors:

Gabriel Langlois-Rahme Enbridge GTM
Oleg Shabarchin Enbridge GTM
Alexander Foster Enbridge
Martin Di Blasi Enbridge GTM
Xueze Chen Enbridge GTM
Gordon Fredine Enbridge GTM