The pipeline industry has traditionally used either steel reinforcement sleeves (Type A), or pressure containing ‘welded on’ steel sleeves (Type B). This basic technology was augmented by the introduction of composite repair techniques.

In the 1990s, steel compression repair was developed (the PetroSleeve 1) that is able to repair defects such as corrosion, denting and cracking, including stress corrosion cracking (SCC) without welding to the pipe or taking the pipe out of service.

With this technology, the steel sleeve applies compression to the carrier pipe in order to reduce the hoop stress. By reducing hoop stress (actually putting the carrier pipe into compression), any defect, including cracking, is prevented from continuing to grow and thus avoiding failure.

Installation procedure

To install a sleeve, the two sleeve halves are installed around the pipe and held in place using a specialised jacking system. Engineering software, incorporating the physical field conditions and the design post-repair stress requirement, is used to determine the installation parameters.

After heating the sleeve halves to attain the design installation temperature, the final fillet welds are completed creating a steel encirclement around the pipe. The quality assurance measurements then are applied into the software to confirm the stress state of the sleeve and pipe.

Testing pressure cycling in pipe

In order to test these compression repair systems, a testing program was carried out by the Canada Centre for Mineral and Energy Technology (CANMET), a Canadian Government-funded research department. CANMET was selected to partake in the program as an impartial third party.

The chosen compression repair system underwent a comprehensive testing regime in 2000 to justify that the system could meet the requirement to permanently restore the serviceability of the pipe.

In setting the critera, the views of various recognised industry groups were utilised. The intent was to produce, if possible, a failure in the testing assembly. Failure was expected to occur either by causing the defect, such as a crack, to extend; fracturing the sleeve; fracturing the attaching fillet weld.

The testing criteria included:
• 36,500 pressure cycles representing a complete pipeline shutdown each day for 100 years;
• Pressure cycling range from 689 kPa to 8050 kPa, simulating an R Factor of 0.085;
• 1950s vintage low-frequency non-normalised electrical resistance weld (ERW) pipe having low charpy impact values;
• Severe anomalies in the ERW non-normalised seam (50 per cent crack; 70 per cent through wall corrosion defect calculated to fail at 2000 kPa); and,
• Test vessel (sleeve and pipe) fully strain gauged to record any variance in strain.

This criteria were extremely severe. The pipe used in the testing had been installed and in operation for approximately 50 years. This meant that the pipe subjected to testing had been exposed to 50 years of operational pressure cycling, then subjected to an additional 100 years of severe pressure cycling. The sum of these two situations would represent a total of 150 years of 'operational pressuring' history.

The pressure variance selected for the testing was from 689 kPa to 8,050 kPa. This represents a pressure range from maximum operating pressure (MOP) to effectively zero (R factor of 0.085). Because of the large pressure variance, the stress intensity factor (delta K) at the crack tip was maximised. This maximises the crack tip driving force, which in turn promotes crack extension.

This testing criteria (R = 0.085) greatly exceeded the usual operating conditions in pipelines. Gas pipelines usually operate with high R ratios. At any specific location in the pipeline, such as at a defect location, the pressure probably varies by 500 kPa, representing an R value of approximately 0.95. This equates to a small delta K.

For refined product pipelines delivering product at intermediate locations, the R value could be within the 0.4 range. This could represent a pressure variance of 7,000–2,800 kPa, equating to a large delta K. However, the delta K for an operating liquids pipeline would still not approach the sever delta K specified for the test situation.

The pipe used for the test was 1950s vintage low-frequency non-normalised ERW pipe. The steel in this 1950s pipe had low charpy impact values. This pipe had exhibited the tendency to develop cracking in the ERW weld. In addition, cracking in the fillet welds associated with welding pressure containing Type B sleeves had been identified.

For the two separate cyclic pressure tests, the serious defects were placed in the ERW seam weld of the pipe. For the first test, a 300 mm long parabolic slot, 70 per cent through wall, was electrically machined (EDM) in the seam. For the second test, a single 50 per cent through wall fatigue crack was created in the ERW seam.

Both pipe pups were fabricated into vessels so that sleeves could be installed while the vessels were pressurised. The parabolic defect was predicted to have a maximum safe operating pressure of approximately 2,000 kPa.

Recording strain in pipelines

Both test vessels were equipped with strain and temperature gauges. During sleeve installation, all strain and temperature indications were recorded. The strain gauges were located to measure strains at the critical locations identified by the finite element analysis.

Following sleeve installation, the two vessels were transported to the CANMET research laboratory at Ottawa, Canada, to be pressure-cycled.

At the CANMET facility, 16 recording channels were used to monitor the strains in the vessel during the 36,500 pressure cycling period. The 16 data indications were recorded every two seconds. One pressuring cycle was completed every 42 seconds.

The cyclic pressuring procedure commenced with a 24-hour high-pressure hold period (8050 kPa), and the 36,500 pressure cycles (689 to 8,050 kPa) followed. At the completion, a second 24-hour 8,050 kPa hold period was performed. The pressure then was reduced to zero and the sleeves removed. During the sleeve removal process, all data channels were recorded.

The purpose to record data every two seconds was to have a record if a failure occurred, as well as to record if relaxation (stretching) of any part of the sleeving system occurred.

At the conclusion of the two tests, the vessels and sleeves (including fillet welds) were magnetically particle-inspected for body cracking. No cracking was found.

The variance in strain readings between the initial and final 24-hour high-pressure hold periods was within testing parameters. This illustrated that relaxation in the sleeving system had not occurred.

The EDM machined and crack specimens were removed from the test vessels and sent to DNV Dublin in Ohio for analysis. The metallographic examination of the machined defect in the ERW seam showed no evidence of crack initiation. The metallographic examination of the 50 per cent crack in the ERW seam also showed no evidence of fatigue crack extension.

Cyclic testing conclusions

The conclusions were:
• No failures occurred;
• The longitudinal crack was permanently arrested;
• The compressive and tensile longitudinal stresses in the pipe near the sleeve ends were insignificant; and,
• Relaxation of the initial clamping force over time is inconsequential.

The consequences of the above conclusions are that by using this technology, the pipe is repaired by a method that can permanently restore the serviceability of the pipe, as shown by reliable engineering tests and analyses.

Sharing technology across the globe

At the conclusion of the engineering testing and analysis, it was accepted that this technology, by reducing stress levels, prevented crack growth. Thus it permanently restored the serviceability of the pipe.

This technology has been used in several countries, including Australia, to repair various pipe defects including cracking. By using this technology, pipeline companies have saved significant amounts of money. The time and expenses associated with nitrogen purges and pipe cutouts have been avoided. The inherent risk in welding to an operating pipeline has been avoided. Finally, companies have been able to maintain deliveries at near-normal levels throughout the repair program.

Robert Smyth is Vice President of Engineering, Petro-Line Group, Nisku, Alberta Canada. For the past 20 years, Robert has been directly involved with the rehabilitation of oil and gas pipelines. Mr Smyth received a BS (1965) in mechanical engineering from Queens University, Kingston, Ontario, Canada. He is a registered professional engineer in Alberta.