A section of submerged arc welded 864 mm diameter API 5L grade X65 gas pipeline containing known stress corrosion cracks (SCC) was selected for fatigue testing on the basis of a successful in-service in-line inspection (ILI) by an ultrasonic intelligent pig. The purpose of the fatigue test was to determine whether SCC shallower than a certain depth could be recoated without the need for grinding and returned to service without the risk of later failing by fatigue.
The SCC containing pipe section was removed from service, instrumented with strain gauges, fitted with proprietary glass reinforced plastic (GRP) repair sleeves (Clocksprings) at the location of some of the cracking, and hydrostatically pressure cycled under automatic control. After pressure cycling, SCC cracks were extracted from the pipeline and examined metallographically. The extensive SCC colonies contained SCC cracks up to 6.4 mm deep (71 per cent wall thickness) and the metallographic examination showed excellent agreement in terms of both location and crack layout with the ILI results.
The fatigue testing was terminated by a small leak after the equivalent of approximately 8,000 years of daily and maintenance related pressure cycling for the pipeline concerned. The equivalent number of 0 to 72 per cent SMYS pressure cycles required to grow the 6.4 mm deep crack to failure was 8,800. This life exceeded the fatigue life of defect free as-welded joints for pipelines and structures according to the fatigue design rules of e.g. DIN and DNV standards respectively.
The amount of fatigue crack growth extension caused by pressure cycles equivalent to 50 years of service proposed for the pipeline concerned was small on bare pipe and less under the GRP sleeves. Although the GRP repair sleeve was applied at a pressure of 36 per cent SMYS, it nevertheless was effective in reducing the experienced stress range and the fatigue crack growth rate.
The measured fatigue crack growth rates were compared to a computer model based on BS7910 and it was found that the model tended to overestimate fatigue crack growth. The reason for this overestimation is because the shape of the SCC cracks was far from ideal and because of complex crack interaction effects. The SCC cracks were inclined away from the perpendicular and had a range of configurations and crack aspect ratios different to typical fatigue crack characteristics. This added further complexity to the issue of predicting fatigue crack interaction and growth.
In summary, if the SCC causing environment can be excluded from the pipe surface by effective re-coating, it is likely that SCC cracks small enough to be safely left in the pipeline under static pressure will not need to be removed by grinding and will not grow to critical levels by fatigue in normal gas pipeline service. This work has also shown that BS7910 can provide a conservative fatigue crack growth estimate for a gas pipeline containing SCC cracks rendered dormant. The results are of great economic significance in reducing the cost of repairs and extending the life of pipeline assets.