In September 1997, Enbridge Pipelines Inc. initiated a crack-detection program on the 283 km, Cromer-to-Gretna, Man., section of Enbridge's Line 3.
Its purpose was to assess the integrity of this section and identify anomalies that might affect future operation of the pipeline.
A second objective was to evaluate the performance of a Pipetronix (now PII) Ultrascan CD in-line inspection tool and determine its potential role in Enbridge's integrity program. Part 1 (OGJ, Apr. 30, p. 64) of this two-part series discussed a similar inspection program in a 36-in. section of TransCanada PipeLines Ltd.'s natural gas system.
Enbridge's program led to 29 excavations based upon the data analysis of the Ultrascan CD tool.
In order to promote confidence in the feature-discrimination capabilities of the technology, Enbridge selected a wide variety of features for excavation even though not all of the features were suspected of being injurious defects.
Of the 18 features identified as potentially injurious during the data analysis process, two significant stress corrosion cracking (SCC) colonies were identified during field inspection.1
Neither colony was considered an immediate threat to the integrity of the pipeline based upon a fracture-mechanics assessment. The remaining 16 were noninjurious reflectors including inclusions and slivers.
This article will focus on the tool performance requirements established by Enbridge before the inspection run. These requirements include specific defect type and size and defects at a maximum sensitivity of the tool.
In addition, information was obtained as a result of the excavation program and on site inspection and assessments. Information gathered from this program was useful in better understanding the tool tolerance in detecting such defects and in differentiating among them.
Enbridge Lines
Line 3 is an 860 mm (34 in.) OD, API 5LX Grade X-52 pipeline with 7.14-12.5 mm nominal WT. The 1,240-km Canadian portion of the line runs from Edmonton to Gretna, Man. It was constructed between 1963 and 1969 in a series of loops designed to increase the capacity of the Enbridge system and at that time moved light, medium, and heavy crude oils.
Until 1999 the pipeline operated in a looped configuration with neighboring 24 in. and 48 in. pipelines. Line 4 downstream of Kerrobert, Sask., began operating in straight 34-in. configuration in 1999 following completion of the first phase of the Terrace Expansion Project that connects the 48-in. loops with a new 36-in. pipeline.
The Cromer-to-Gretna section consists of approximately 235 km of double submerged arc welded (DSAW) pipe and 48 km of electric-resistance welded (ERW) pipe.
The Ultrascan CD in-line inspection tool was selected for the inspection run through the 283-km section that consists primarily of pipe manufactured with a DSAW long seam with short sections of pipe having ERW long seams.
In total, 29 excavations were conducted based upon the data analysis.
Project objectives
Previously, Pipetronix had stated that the CD tool was capable of identifying axially oriented crack-like indications in the pipe body and long seam regions having a minimum depth of 1 mm and minimum length of 30 mm.2 The technology uses shear-wave ultrasonic signals generated by an oblique angle of incidence of the transmitted energy into the pipe wall.
The system is designed so that the product flowing in the pipeline acts as the coupling medium between the sensors and the pipe wall. In gas pipelines, the CD tool can only operate in a liquid slug.3
To facilitate the level of detail involved in the data analysis for such small indications, the data assessment and reporting process typically divides into stages.
A fracture mechanics approach was used to identify the length of injurious SCC colony on Line 4 based upon the pipe steel properties and operating parameters. Following the assessment was a decision to set the desired reportable defect size for the inspection so that all indications greater than 50 mm long and 1 mm deep were reported.
In order to accomplish this task, Pipetronix required a year to complete the data assessment because of its workload at that time. So that Enbridge would not have to wait for 1 full year following the inspection run to begin its maintenance work, the analysis was broken into four stages:
- Stage 1. All indications within ±50 mm of the long seam, greater than 100 mm long, and 40% through wall (TW) within the first 10 km downstream of pump-station discharges. Completed in November 1997.
- Stage 2. All indications in the full pipe body, greater than 50 mm long, and 1 mm TW within the first 10 km downstream of pump-station discharges. Completed in January 1998.
- Stage 3. All indications in the full pipe body, greater than 100 mm long, and 1 mm deep on the remainder of the inspected pipeline. Completed in June 1998.
- Stage 4. All indications in the full pipe body, greater than 50 mm long, and 1 mm deep on the remainder of the inspected pipeline. Completed in September 1998.
The first two stages were considered highest priority because as the operating-pressure spectrum would be the highest immediately downstream of the discharge side of the pump stations along this pipeline section.
Field inspection methods
Enbridge established the inspection guidelines for the field excavation in conjunction with the non-destructive examination (NDE) technologists at Pipetronix. In essence, the methodology followed the guidelines in the CEPA Recommended Practices for inspection of SCC colonies with procedural suggestions from Pipetronix for ultrasonically examining indications.1
Wet fluorescent magnetic particle inspection (WFMPI) was used to identify surface breaking indications. External indications found were subsequently reinspected with black-on-white magnetic particle inspection (BWMPI) to facilitate photography.
Internal and mid-wall indications were located and the depths of all indications initially assessed with ultrasonic testing (UT).
Pipetronix suggested that a 6.25 mm diameter, 45°, 10-Mhz ultrasonic transducer be used when assessing defects because such a transducer mimics the transducers incorporated in the CD tool. In the event that an external surface indication was found, the UT depth estimate of the indication could be confirmed by grinding.
Analysis results
Since this was as much a validation run of the tool as an assessment of the condition of Line 3, a wide variety of features was selected for excavation including 11 that were assessed to have had a low probability of being injurious defects.
The focus of the excavation program was to repair all defects while assessing the tool's capabilities for both characterizing and sizing reflectors in order to refine the feature selection process for future inspections.
The following abbreviations4 were used by Pipetronix to characterize reflectors:
cf = crack field
cl = crack-like
ml = metal loss
nl = notch-like
il = inclusions and laminations
ihg = inhomogenity
int = internal
ext = external
Feature depths are grouped into ranges based upon the estimated percentage of WT as follows:
- < 12.5%
- 12.5% - 25%
- 25% - 40%
- > 40%
Feature lengths were provided in millimeters.
Pipetronix did not report any features beyond 25% WT. Out of the indications excavated, only the two SCC colonies were considered to be noteworthy. Neither feature, however, appeared to be near failure at the time of the excavations.
They were the first SCC colonies found with an inspection tool on the Enbridge system that met the CEPA definition of significant. In the first instance the SCC colony was correctly classified as a crack field with associated corrosion. In the second case, however, the indication was identified as an external surface irregularity.
While external surface irregularity may not be an incorrect ultrasonic description of the feature, it did not imply that the feature was potentially an SCC indication. It was selected for excavation based upon its depth estimate and its proximity to the upstream pump station
In both cases, the depths of the colonies were determined by overlaying both crack tool and metal-loss tool data.
Field measurements reported the two colonies were found with depths of 28% and 20% which included the depth of the external wall loss in addition to the depth of the cracks. The external corrosion was on the order of 10% WT in each case.
If 10% of the total depth is removed from the field measurements. the estimated crack depth would have been 18% for one colony and 10% in the second, while Pipetronix predicted that both were <12.5% WT.
Fig. 1 identifies the morphology of one of the SCC colonies.
On the left hand side of the picture, the irregular surface profile resulting from the metal loss can be seen surrounding the cracks. Towards the right-hand side of the photo is a smooth area on the surface of the pipe that is the result of the grinding process used to confirm the crack depth.
An assessment of the remaining indications labeled as "possible crack fields or external surface irregularities" was undertaken to try to identify those that could potentially be SCC. The two colonies previously found were close to the discharge of pump stations where the mean hoop stress would normally be high.
The remaining features occurred closer to the suction side of pump stations where the mean hoop stress would be low. Each of the remaining indications was in pipe from the same manufacturer. Previous excavations on this pipe suggested that the pipe had a fairly high ultrasonic reflector population as a result of noninjurious, nonmetallic inclusions.
Nine more "possible crack field or external surface irregularities" were scheduled for excavation, although based upon the companies' assessment, the probability that any of the features were SCC was deemed to be low. All nine were identified in the field to be inclusions or slivers and therefore noninjurious.
The most difficult task for data analysts appeared to be correctly classifying clusters of external surface slivers, which provided signals similar to that of shallow crack fields. With the experience gained from the Enbridge field excavations, Pipetronix analysts have gained further insight into the structural differences between crack fields and slivers in the CD data and have said that they will be able to reduce false calls in future inspections.
With the exception of the two SCC colonies, the depth estimates based upon the tool data for the remaining features were conservative. The features located and sized in the field were all less than 10% WT even for the features with predicted depths of 12.5-25%.
As illustrated in Fig. 2, the length estimates taken from the CD data matched quite well with the field measurements for all but two features. The length of one of the SCC colonies was underestimated by 300 mm. It is likely that the cracks on the periphery of the SCC colony were too shallow for the tool to identify; thus the total length was underestimated.
For the remaining indications, the characterization from the tool data appeared to be relatively accurate. Internal and external notch-like reflectors and weld reflectors identified by the tool were typically confirmed in the field.
What was learned
One of the program objectives summarized in this article is to evaluate the performance of the Ultrascan CD tool in identifying anomalies that might affect the integrity of Line 3 and its future operation. It was found that the Ultrascan CD tool was effective in detecting and locating cracks and crack-like defects.
Also, several lessons were learned from this project:
- The data-analysis phase for the CD data was a long process when assessing defects at the maximum sensitivity of the tool. The staged approach to completing the analysis was appropriate, targeting the highest priority locations along the pipeline first. The process may be expedited further under less stringent defect-size detection requirements.
- Differentiating between sliver indications and shallow-crack fields appeared difficult based on CD data. Pipetronix has informed Enbridge that the information gathered from this project was useful in identifying means of differentiating between slivers and crack colonies.
- Other factors can be used to assist in selection of features having the highest probability of being SCC including coating type and quality, soil information, and water table and, SCC's effects on crack formation and growth.
- Depth estimates for most CD features appear to be conservative. Overlaying CD data with data from metal-loss inspections, however, can assist in identifying the total depth of multimode defects.
- Enbridge's experience with using the CD tool in this program was the first, and the life of the CD tool use in liquid pipelines is in the early stage.
References
- CEPA SCC Working Group, "Stress Corrosion Cracking Recommended Practices," Canadian Energy Pipelines Association, 1997.
- Willems, H.H., Hugger, A., and Barbian, O.A., "Results of In-Line Crack Inspection using the Ultrascan CD Tool", 6th International Colloquium on High Pressure Pipeline Reliability after Long Time Operation, Prague, Mar. 10-12, 1997.
- Willems, H., and Barbian, O.A., "Ultrasonic Crack Detection in Pipelines by Advanced Intelligent Pigging," proceedings of the Second International Conference on Pipeline Technology, Ostend, Belgium.
- Pipetronix Ltd., "Primus Data Interpretation Program Users Manual."
The authors
L. Blair Carroll is a project engineer with the Fleet Technology Ltd.'s Materials Technology Center, Kanata, Ont. Caroll is involved with damage tolerance and engineering assessment projects for clients in the pipeline industry. He is also involved in the evaluation and development of non-destructive testing (NDT) technologies. Prior to joining FTL, Carroll was a pipeline integrity engineer with Enbridge Pipelines Inc., where his duties focused on the selection and validation of in-line inspection tools to detect pipeline cracking mechanisms, the implementation of field inspection and repair programs, and overseeing the company's stress corrosion cracking (SCC) management program. Carroll obtained Bachelors and Masters degrees in mechanical engineering from Memorial University of Newfoundland. His graduate research focused on the assessment of an electromagnetic NDT technology for field inspection of SCC.
Moe Madi is currently responsible for managing engineering advisory and training projects, and coordination with client needs and consulting services, at Enbridge Pipelines Ltd., Edmonton, which he joined 4 years ago as a pipeline integrity engineer. Madi holds a BS (1990) in materials and metallurgical engineering from the University of Tripoli, a masters of engineering (1996) in engineering management from the University of Ottawa, and is a candidate for PhD in management of technology from North Central University, Prescot, Ariz. He is a member of NACE and the Engineering Association of Rome.
Based on a presentation to IPC 2000, the International Pipeline Conference, Oct. 1-5, 2000, Calgary.
