TNT NDT Technician Article: BP Alaskan Pipeline Inspection
In August of 2006, a major petroleum company experienced the second loss of containment incident within the year on the pristine and environmentally sensitive North Slope of Alaska. Both leaks resulted from internal pitting corrosion on or near the bottom half of 0.85 m (34 in.) diameter transit pipelines. These lines transport 400 000 barrels of petroleum per day across 11 miles of Alaskan tundra. The failure meant an immediate shutdown of approximately three percent of the petroleum supply to the lower 48 states. The potential for environmental disaster and the ensuing shutdown were quickly brought to the attention of environmental groups and jurisdictional authorities and, as quickly, to the attention of national media. Americans watched as the balance of environmental responsibility and energy dependence came to rest on the non destructive testing (NDT) community. This article describes the investigation and application of fast-screening NDT techniques to ensure pipeline integrity and increase inspection efficiency.
The task before the petroleum company inspection team was to investigate alternative NDT corrosion screening techniques that could be submitted to the USDOT for possible modification of the standing CAO. The fast-screening NDT techniques needed would have to detect 50% wall loss inside surface pits at a 3:1 aspect ratio. The consequences of another failure required 100% probability of detection (POD) of any discontinuity that met or exceeded the criteria.
Each pipeline was segmented into 0.3 m (1 ft) inspection intervals, creating approximately 52 000 discrete areas to be screened for corrosion. Areas with less than 25% wall loss were ultrasonically tested to record minimum and average wall thicknesses within the segment. The team of 108 UT technicians inspected an average of 283 segments per day. Automated UT rates were 4.5 to 6.0 m (15 to 20 ft) per crew, per shift — unusually low, requiring additional manual scanning. Unlike automated UT, manual UT provides no ultrasonic image or permanent record of thickness measurement. Unless a successful alternative NDT technique could be found and accepted by the USDOT, the inspection of 52 000 areas was going to take 184 days. Neither manual nor automated UT in its current configuration could inspect remaining wall thicknesses at welds, supports, or anchor points. Internal corrosion in these areas, however, was not considered preferential and inspection of them was deferred to the intelligent pig (robotic pipeline inspection gage inserted into the line) run that would follow external tests.
Axially orientated electromagnetic acoustic transducer (EMAT) technology was trial tested for pipe support touch point corrosion and was determined to detect greater than 30% wall loss from 0.5 m (20 in.) away from the support. Tape stripping was later suspended in lieu of performing automated UT through 4 mm (0.16 in.) thick anti-corrosion tape, 8 mm (0.31 in.) at overlaps (Fig. 1). Ultrasonic performance on tape wrapped pipe was found to be fully equivalent to bare pipe inspection provided the tape was bonded and uniform. Areas where the tape wasn’t properly bonded were infrequent and were marked for tape removal and reinspection with automated UT. Absolute ultrasonic thickness measurements could be obtained by applying time of flight (TOF) delay correction factors of –10 mm (–0.39 in.) for single tape layers and –23 mm (–0.91 in.) for double tape layers. Ultrasonic echo-to-echo coating compensation mode was not used; pitting responses could interfere with proper UT signal gating. Ultrasonic testing amplitude sensitivity was established by 6 mm (0.25 in.) flat bottom hole (FBH) response on a bare calibration block followed by an applicable dB transfer value.
At the end of preliminary assessments, four electromagnetic techniques were favored for fast screening of isolated pitting. Electromagnetic techniques do not require direct surface coupling, allow for real time inspection of large areas without labor-intensive surface preparation, and can speed up inspection without sacrificing test sensitivity or data quality. Of the four techniques considered as automated UT alternatives, only two were selected as short-term solutions. EMAT. Electromagnetic acoustic transducers (EMATs) use a permanent or electromagnetic driver/coil arrangement to create various ultrasonic wave modes within carbon steel. Figure 2 demonstrates compression wave mode. In guided wave UT mode, EMAT typically generates a 5 cm (2 in.) wide sound beam that averages material volume and detects localized wall loss across the span of two permanent or electromagnetic driver/coil sensors. A mechanized scanner moves axially at a scan rate of roughly 75 to 150 mm (3 to 6 in.) per second. LFET. Low Frequency Electromagnetic Technique (LFET) uses an electromagnetic driver/coil arrangement to create magnetic lines of flux through the volume of carbon steel material. Corrosion causes changes to nominal conditions of the field. Signals produced by these changes are received by a pickup coil measuring magnetic flux amplitude and phase (Fig. 3). EMAT technology is well established in industry and recognized in ASTM document E 18161 whereas LFET technology is newer. Similar to magnetic flux leakage in its sensor arrangement and usage, it offers electronic phase analysis and intuitive data interpretation.
Anticipating USDOT acceptance of data along with formal approval of the technology, advance NDT crews began to implement the alternative NDT techniques immediately upon approval by the NDT development team. Trial results were presented to USDOT officials and an independent NDT subject matter expert from the U.S. Department of Energy. Many officials (including the U.S. Secretary of Transportation) were in attendance to personally witness NDT technicians apply the alternative NDT techniques in repeated field performance trials. The alternative techniques were accepted by the USDOT and the CAO was modified after three weeks.
By using an automated UT crawler, NDT technicians could now spend 90% of their time monitoring data collection from the comfort of a truck parked nearby. Areas that had been inaccessible because they were over bodies of water or at extended elevations could now be inspected without scaffolding. The potential for injuries or accidents related to fatigue and exposure was now greatly reduced.
LFET is highly sensitive to sensor liftoff from the carbon steel surface yet can tolerate up to 6.35 mm (0.250 in.) in non-conductive coatings while maintaining discontinuity sensitivity. Figure 8 shows a mechanized 160-channel LFET axial scanning array capable of continuous and autonomous inspection. Each of the 20 sensor cars is equipped with a separate wheel system and is spring tensioned into a surface conforming array to control liftoff. The LFET scanner can test 18 m (60 ft) of pipe and provide real time data in just 6 min. The 50% wall loss 3:1 pit aspect ratio discontinuity criteria mandated by the CAO worked in favor of EMAT and LFET screening tools. Each had a tendency to ignore pits below this criterion, thus reducing the time needed for data analysis. For example, both the LFET data shown in Fig. 8c and the automated UT data set in Fig. 7c are from the same section of pipeline. Only a few of these pits in the 3 m (10 ft) automated UT scan were of interest and those showed up in the LFET scan.
1. ASTM E 1816, Standard Practice for Ultrasonic Examinations using Electromagnetic Acoustic Transducers (EMAT) Techniques. West Conshohocken, PA: ASTM International (2002).
This article is from TNT The NDT Technician vol 6 number 3 July 2007. [PDF] Version including photos.https://asnt.org/tnt
Figures Referenced in the article are Property of TNT and are displayed in the PDF as part of their publication.
If you would like to learn more about our inspections, our systems, or how they can be used on your site, please contact us at 412.798.8990 or click here.
Background
After shutdown, the U.S. Department of Transportation (USDOT) issued a legally binding Corrective Action Order (CAO) that mandated exclusive use of automated UT to examine the 4 to 8 o’clock sectors (radially designated) of all pipelines throughout the petroleum transit system. Inspection with UT would require the removal of polyurethane insulation panels and preparation of pipe surfaces throughout the system. A machine applied anti-corrosion tape coating half of the pipeline created further difficulty. The number of insulation workers and ultrasonic technicians needed to accomplish the removal and inspection tasks was initially thought to be beyond what the North Slope could provide for in terms of temporary housing and travel logistics. At its height, NDT work alone would require 108 UT technicians working in alternating 12 h shifts.The task before the petroleum company inspection team was to investigate alternative NDT corrosion screening techniques that could be submitted to the USDOT for possible modification of the standing CAO. The fast-screening NDT techniques needed would have to detect 50% wall loss inside surface pits at a 3:1 aspect ratio. The consequences of another failure required 100% probability of detection (POD) of any discontinuity that met or exceeded the criteria.
Ultrasonic Method
During the inspections, UT was acknowledged as the only NDT method that could measure absolute remaining wall thickness within localized corrosion areas. All other methods were considered screening techniques subject to ultrasonic validation and measurement.Each pipeline was segmented into 0.3 m (1 ft) inspection intervals, creating approximately 52 000 discrete areas to be screened for corrosion. Areas with less than 25% wall loss were ultrasonically tested to record minimum and average wall thicknesses within the segment. The team of 108 UT technicians inspected an average of 283 segments per day. Automated UT rates were 4.5 to 6.0 m (15 to 20 ft) per crew, per shift — unusually low, requiring additional manual scanning. Unlike automated UT, manual UT provides no ultrasonic image or permanent record of thickness measurement. Unless a successful alternative NDT technique could be found and accepted by the USDOT, the inspection of 52 000 areas was going to take 184 days. Neither manual nor automated UT in its current configuration could inspect remaining wall thicknesses at welds, supports, or anchor points. Internal corrosion in these areas, however, was not considered preferential and inspection of them was deferred to the intelligent pig (robotic pipeline inspection gage inserted into the line) run that would follow external tests.
Axially orientated electromagnetic acoustic transducer (EMAT) technology was trial tested for pipe support touch point corrosion and was determined to detect greater than 30% wall loss from 0.5 m (20 in.) away from the support. Tape stripping was later suspended in lieu of performing automated UT through 4 mm (0.16 in.) thick anti-corrosion tape, 8 mm (0.31 in.) at overlaps (Fig. 1). Ultrasonic performance on tape wrapped pipe was found to be fully equivalent to bare pipe inspection provided the tape was bonded and uniform. Areas where the tape wasn’t properly bonded were infrequent and were marked for tape removal and reinspection with automated UT. Absolute ultrasonic thickness measurements could be obtained by applying time of flight (TOF) delay correction factors of –10 mm (–0.39 in.) for single tape layers and –23 mm (–0.91 in.) for double tape layers. Ultrasonic echo-to-echo coating compensation mode was not used; pitting responses could interfere with proper UT signal gating. Ultrasonic testing amplitude sensitivity was established by 6 mm (0.25 in.) flat bottom hole (FBH) response on a bare calibration block followed by an applicable dB transfer value.
Alternative NDT Corrosion Screening
Alternative corrosion screening techniques to complement or replace ultrasonic techniques had to maintain discontinuity detection thresholds while increasing NDT production tenfold. All commercial techniques were considered for application but, because of the highly isolated nature of material damage in the petroleum transit lines, the extreme consequences of another failure, and the inspection opportunities afforded by complete removal of the polyurethane insulation panels, only those techniques using a small, localized energy field were chosen. Real time data analysis was also a consideration as was the need to minimize further preparation of pipe surfaces.At the end of preliminary assessments, four electromagnetic techniques were favored for fast screening of isolated pitting. Electromagnetic techniques do not require direct surface coupling, allow for real time inspection of large areas without labor-intensive surface preparation, and can speed up inspection without sacrificing test sensitivity or data quality. Of the four techniques considered as automated UT alternatives, only two were selected as short-term solutions. EMAT. Electromagnetic acoustic transducers (EMATs) use a permanent or electromagnetic driver/coil arrangement to create various ultrasonic wave modes within carbon steel. Figure 2 demonstrates compression wave mode. In guided wave UT mode, EMAT typically generates a 5 cm (2 in.) wide sound beam that averages material volume and detects localized wall loss across the span of two permanent or electromagnetic driver/coil sensors. A mechanized scanner moves axially at a scan rate of roughly 75 to 150 mm (3 to 6 in.) per second. LFET. Low Frequency Electromagnetic Technique (LFET) uses an electromagnetic driver/coil arrangement to create magnetic lines of flux through the volume of carbon steel material. Corrosion causes changes to nominal conditions of the field. Signals produced by these changes are received by a pickup coil measuring magnetic flux amplitude and phase (Fig. 3). EMAT technology is well established in industry and recognized in ASTM document E 18161 whereas LFET technology is newer. Similar to magnetic flux leakage in its sensor arrangement and usage, it offers electronic phase analysis and intuitive data interpretation.
Technique Trials
Performance of EMAT and LFET equipment was assessed under actual field conditions. A meticulous effort was made to disregard expectations based on preconceptions or manufacturer’s data. The 0.75 m (30 in.) decommissioned pipeline selected for trials was subjected to preliminary testing with intelligent pigging to provide known pitting corrosion areas for study. Computed radiography provided images of pitting. Ultrasonic thickness measurements with tape coating thickness compensation provided pit depth and aspect ratio information. A wide range of pit sizes, depths and morphologies were used to establish discontinuity depth detection thresholds and minimum detectectable discontinuity aspect ratios for each method (Figs. 4-7).Field Trial Summary
EMAT. Performance attributes for EMAT testing were as follows (Fig. 5):- EMAT demonstrated 100% POD for 25% wall loss isolated pitting at a 3:1 aspect ratio in a 9 mm (0.375 in.) pipe wall [limited to T2 mode at 0.28 m (11 in.) probe spacing].
- EMAT can detect 30% wall loss at a 4:1 aspect ratio in a 9 mm (0.375 in.) pipe wall wrapped with anti-corrosion tape.
- Ten percent of anti-corrosion tape wrapped EMAT indications were false positives. False positives are not detrimental to the POD of EMAT testing but require rework with other NDT techniques.
- EMAT is susceptible to attenuation (as with all guided wave UT) with false calls due to outside or inside surface boundary conditions. Internal sludge may deaden EMAT responses.
- A two-man crew using EMAT can inspect 300 m (1000 ft) per day, 150 m (500 ft) per day on tape wrapped pipe.
- EMAT provides a permanent image of the entire inspection segment.
- EMAT Indications must be followed up with UT thickness measurement.
- LFET demonstrated 100% POD for 25% wall loss isolated pitting at a 3:1 aspect ratio.
- LFET performance remains unchanged on pipe wrapped with anti-corrosion tape.
- LFET has a false positive overall rate of less than 1%.
- LFET can verify false positive EMAT indications.
- LFET performs better than automated UT and EMAT on fluorocarbon resin repair tape.
- Inspection coverage for a two-man crew using handheld LFET instrument is limited to 60 m (200 ft) per day. With automation and improved probe fixtures, scanning production is increased to 3 m (10 ft) per min.
- LFET indications must be followed up with UT thickness measurement.
Field Implementation
After three weeks of NDT production and development, the performance boundaries of the alternative NDT screening techniques were established, although not yet approved by the USDOT. Until this time, hundreds of insulation strippers, tape scrapers, and ultrasonic technicians had been working simultaneously, around the clock with only one acceptable surface preparation and inspection technique. Even in August, fatigued workers were enduring working conditions that included cold, frequent rain, mud and standing water.Anticipating USDOT acceptance of data along with formal approval of the technology, advance NDT crews began to implement the alternative NDT techniques immediately upon approval by the NDT development team. Trial results were presented to USDOT officials and an independent NDT subject matter expert from the U.S. Department of Energy. Many officials (including the U.S. Secretary of Transportation) were in attendance to personally witness NDT technicians apply the alternative NDT techniques in repeated field performance trials. The alternative techniques were accepted by the USDOT and the CAO was modified after three weeks.
Pipe Crawler Development
Upon USDOT approval, the company NDT lab in Houston, Texas began work on robotic multi-channel sensor arrays for LFET and automated UT. Most of the UT and LFET work done to this point had been done by hand. Automated UT had continued to be inefficient. To obtain the needed tenfold increase in inspection production, further mechanization of these techniques was necessary. Deep water NDE research and development projects already underway at the Houston lab included LFET and automated UT. Mechanical phases of the projects showed potential for application at the North Slope site and were hastened into service in an ambitious three-week pipe-crawler construction program; all other work at two NDT development firms suspended until the machines could be fabricated. Figure 7 shows an axial scanning automated UT crawler capable of continuous ultrasonic imaging of the 4 to 8 o’clock sectors of 0.85 m (34 in.) pipe. The two-piece clamshell assembly runs autonomously from pipe support to pipe support, a distance of about 18 m (60 ft). At which point, the crawler is removed and redeployed in a 5 min. procedure for the next scan segment. The system runs at a rate of 9.75 m (32 ft) per hour with a single UT transducer and can increase the rate to 30 m (100 ft) per hour by implementing a four-transducer array and data merging software. Figure 7c shows a typical automated UT crawler data sample for a pitted pipe area.By using an automated UT crawler, NDT technicians could now spend 90% of their time monitoring data collection from the comfort of a truck parked nearby. Areas that had been inaccessible because they were over bodies of water or at extended elevations could now be inspected without scaffolding. The potential for injuries or accidents related to fatigue and exposure was now greatly reduced.
LFET is highly sensitive to sensor liftoff from the carbon steel surface yet can tolerate up to 6.35 mm (0.250 in.) in non-conductive coatings while maintaining discontinuity sensitivity. Figure 8 shows a mechanized 160-channel LFET axial scanning array capable of continuous and autonomous inspection. Each of the 20 sensor cars is equipped with a separate wheel system and is spring tensioned into a surface conforming array to control liftoff. The LFET scanner can test 18 m (60 ft) of pipe and provide real time data in just 6 min. The 50% wall loss 3:1 pit aspect ratio discontinuity criteria mandated by the CAO worked in favor of EMAT and LFET screening tools. Each had a tendency to ignore pits below this criterion, thus reducing the time needed for data analysis. For example, both the LFET data shown in Fig. 8c and the automated UT data set in Fig. 7c are from the same section of pipeline. Only a few of these pits in the 3 m (10 ft) automated UT scan were of interest and those showed up in the LFET scan.
Findings
Necessity drives invention. The daunting task of large scale inspection in a remote area brought many NDT personnel together with just two goals. While under significant pressure, implement the NDT techniques that were already known to work; then come up with new NDT techniques that could do the job better and faster. In less than a month, both goals had been accomplished. As the new inspection tools were pressed into service, both transit line inspection rates and data quality steadily improved, as did USDOT confidence in their ability to perform. The western petroleum transit line was approved for production. Three weeks later, the eastern petroleum transit line was returned to service. Despite challenging workloads and difficult living conditions, NDT technicians and NDT engineers had presented a concerted effort. The open discussion and free exchange of ideas had facilitated solutions for the set of problems that appeared with each new day and ultimately to the timely and environmentally responsible restoration of a vital natural resourceReferences
1. ASTM E 1816, Standard Practice for Ultrasonic Examinations using Electromagnetic Acoustic Transducers (EMAT) Techniques. West Conshohocken, PA: ASTM International (2002).
This article is from TNT The NDT Technician vol 6 number 3 July 2007. [PDF] Version including photos.https://asnt.org/tnt
Figures Referenced in the article are Property of TNT and are displayed in the PDF as part of their publication.
If you would like to learn more about our inspections, our systems, or how they can be used on your site, please contact us at 412.798.8990 or click here.