CC BY
13
Keywords:
stress corrosion, corroded pipelines, crack, stress test, rehabilitation.
УДК [620.194.22:622.691.4.053]:620.172.25
Approach and experiences regarding the rehabilitation of stress-corroded pipelines by the stress test
C. Günther1*, U. Marewski1, M. Steiner1
1 Open Grid Europe GmbH, Bld. 5, Kallenbergstr., Essen, 45141, Germany * E-mail: christina.guenther@open-grid-europe.com
Abstract. Experiences in the laboratory and in the field prove that the stress test is suitable for rehabilitating pipelines in operation with axial stress corrosion cracking. The stress test is the only integral strength test in which strength-reducing axial faults are eliminated and remaining faults are mitigated through stress displacement. Hence, this test is not only suitable for checking but also for optimising pipelines. If after the stress test the conditions for stress corrosion cracking still remain, the stress test need to be repeated at defined intervals based on the growth of the cracks. As a qualifying statement, it must be said that faults in the circumferential direction or at an angle to the longitudinal axis of a pipe can be eliminated only to a very small extent with this method.
Basics of the stress test
The stress test was introduced in Germany in the 1970s. It is a water pressure test in which a pipeline is being filled with water section by section and then is stressed to its yield strength. The implementation and evaluation of this procedure are regulated in the VdTÜV (germ. Verband der Technischen Überwachungs-Vereine, e.V.)1 data sheet «Pipelines 1060 -Guidelines for conducting the stress test»2 and in the germ. Deutsche Verein des Gas- und Wasserfaches e.V3 code of practice «DVGW G 469:2010. Pressure testing methods for gas transmission/gas distribution». The stress test is an integral strength test in which prior stresses, such as installation stresses, are relieved, and unstable strength-reducing defects, such as cracks, can be eliminated. Consequently, this test method is aimed not only at checking but also at optimising the pipeline and has been used for more than 40 years on new and operating pipelines [1].
In the stress test, the test pressure should be as high as possible but should not cause inadmissible deformation. Therefore, the maximum test (operating) pressure p is determined by the flow through the weakest pipe in the test section, as otherwise inadmissible widening would occur. However, at the same time, it must be ensured that the pipe with the highest value of wall thickness times yield strength is stressed to at least 85 % of its yield strength to fulfil the second requirement, the elevation profile of the pipeline and/or the hydrostatic pressure of the water must be taken into account. In practice, this often limits the length of the pipeline section that can be tested.
Course of the stress test. A quantified stress test with «training effect» is suitable for rehabilitating a defective active pipeline, for example, one damaged by stress corrosion cracking (fig. 1). First, the maximum test pressure is applied to a tested section by a fast increasing of the hydrostatic pressure followed by an immediate small reduction. After a holding time of 60 to 90 minutes at this reduced test pressure, the pressure is relieved to at least 2 bar at the highest point of the pipeline. This must be held for 30 minutes so that stress displacement can occur at the crack-tips. The pipeline is once again subjected to pressure, almost to the maximum test pressure. A holding time is also defined for this pressure period (fig. 2). Because of material effects, such as plastic reformation or the Bauschinger effect, it is quite possible that faults will be eliminated only during this second holding time. Due to the resulting elimination of unstable cracks, this method is especially good for rehabilitating pipelines damaged by stress corrosion cracking [1]. As a result, the
1 Association of Technical Inspection Agencies.
2 See: germ. VdTÜV MB ROHR 1060. Richtlinien für die Durchführung des Stresstests. April 2018.
3 German Technical and Scientific Association for Gas and Water.
Fig. 1. Pressure versus pumped volume Fig. 2. Course of a stress test [1]
for a quantitative stress test [1]
stress test is also mentioned in various international guidelines4 for handling pipelines damaged by stress corrosion cracking as the best tool for removing critical and almost critical axial cracks.
Calculating the remaining defect geometries in the pipeline. Barlow's formula is used to dimension the pipeline; hence, the following formula is applied for calculating the required wall thickness t:
t =
p • oD
2 a ;
where oD is the external diameter and c is the maximum permitted circumferential stress. If an axial defect exists with a length c and a depth d, the strength of the material that has been assumed previously during dimensioning is reduced accordingly. The effect of a defect on the load-bearing capacity of a pipe can be calculated using the Battelle concept [2]. This concept assumes that the failure of a pipe with a semi-elliptical crack depends only on the remaining bearing cross section. The resulting failure pressures of a pipe (L485, oD 1000, t = 16,8 mm) with various crack geometries are shown as examples in fig. 3.
4 See: UKOPA/GP/009. UK Onshore Pipeline Operators' Association - Industry good practice guide. Near neutral pH and high pH stress corrosion cracking [online]. 2015. Available from: http://www.ukopa.co.uk/documents/ UKOPA-GPG009.pdf;
CEPA. Recommended practices for managing near-neutral pH stress corrosion cracking. 3rd ed. Calgary, Alberta, 2015.
If c = 0 mm, failure occurs at the limit load-bearing capacity of the pipe. For longer cracks, the pressure at which failure occurs is very dependent on d [3]. In this example, a crack depth of 2 mm results in a minor reduction in failure pressure (approx. 5 %), while a 10 mm deep crack considerably reduces the pressure at which failure occurs (more than 50 %). Because of the effects on safety, the type of failure (leak or break) is very important and can also be calculated according to Battelle concept. If the failure of the pipe is above the amber-coloured curve (see fig. 3), it will occur as a break, and below this curve, as a leak. Two examples of pipeline breaks are shown in fig. 4. For most of the crack geometries shown in fig. 3 the calculated failure pressures are above the leak-before-break curve, which explains why in these cases a break occurs in the event of failure. For the chosen example, without a prior stress test, the pipe would fail with a break at 100 bar operating pressure if a very long (c > 420 mm) and deep (d > 10 mm) crack existed. This and other crack geometries would have been eliminated with a prior stress test.
The stress test stresses the pipeline well in excess of p. If p reaches the failure pressure of the corresponding defect geometry, the defective pipe fails and is replaced. Because of this, after the stress test it can be assured that pipes with remaining faults will not fail at operating pressure. Fig. 5 shows a qualitative sketch of the limit geometries for longitudinal cracks that remain after various levels of stress in a pipeline. Assuming the same c value, the more shallow faults would
Fig. 3. Failure pressures of various faults and leak/break behaviour on a pipe
Fig. 4. Examples of pipeline ruptures during stress tests [4, 5]: a - failure due to a groove caused by mechanical damage; b - longitudinal weld of an HFW-pipe is split open
Fig. 5. Qualitative sketch of remaining crack sizes after different types of pressure loadings [1]
be eliminated with a stress test than with a pressure test at 1,3p. Consequently, after a stress test lesser faults remain in the pipeline than would be the case with a usual pressure test or with no pressure test. Shorter cracks cause leaks in case of failure, while pipelines with longer cracks fail due to a break. Due to the internal pressure of the stress test, the stresses for faults in the circumferential or diagonal directions are much less than for longitudinal cracks. Therefore, very high pressures must be chosen to eliminate existing circumferential cracks. Even under these conditions, only very long and deep circumferential cracks would lead to failure. This is why in practice application of stress tests for elimination of circumferential cracks is limited [6].
With the cracks that were not eliminated during the stress test, the stress occurring during the test causes stress displacement, and at the crack tips there is a yielding in the material, which is associated with residual compression stresses. Therefore, after the stress test these faults usually exhibit delayed growth of cracks as a result of load cycles if there is no additional driver like stress corrosion cracking.
An example of the stress test on a gas pipeline damage by stress corrosion cracking
At the end of the 1980s a gas leak occurred as the result of a rupture in a high pressure city gas pipeline that was constructed for an operating pressure of 55 bar. The 76 km long pipeline (oD 600, t = 10,2 mm, material St 52-3 according to DIN 17100) was previously tested with air at 1,1p during commissioning. The detailed damage analysis showed that the damage was caused by hydrogen-induced corrosion cracking. The conditions for this were provided by the moist city gas and its components (CO, CO2, O2) and, in addition, water remained in the pipeline after the construction phase [7]. As a result of the investigation, the pressure in the pipeline was limited to 10 bar. However, due to increased supply requirements the pipeline was still to be used to transport natural gas with no reduction in pressure. For the necessary rehabilitation of the pipeline, in addition to other measures, such as cathodic corrosion protection, it was decided to carry out a stress test to eliminate or at least mitigate stress corrosion cracking inside the pipeline. After that the product ought to be changed to dry natural gas to prevent further crack growth.
Laboratory tests. First, laboratory tests were conducted to prove that the stress test was suitable
to rehabilitate this pipeline [8]. Dismantled pipe sections damaged by stress corrosion cracking were used to simulate the stress test in the laboratory (fig. 6).
The course of the investigation programme was defined by an expert for pipeline construction in coordination with VdTUV. A maximum test pressure of 138 bar was applied to the dismantled sections. In accordance with the procedure for a quantified stress test with «training effect», the components were stressed, relieved and stressed again with the holding times taken into account. The dismantled sections did not fail during this test. Based on ultrasound investigations it was shown that the stress test had triggered a stress conversion so that the investigated cracks had not grown when stress was applied the second time. In subsequent metallography investigations plastification of the crack base was also shown (fig. 7), associated with rounding of the crack tips and the development of residual compression stresses. These residual compression stresses have a stress-reducing effect, i.e. with local stresses, they mitigate the crack areas. After simulation of the stress test, sustained internal pressure tests (test facility shown in fig. 8) were carried out on the stressed components to investigate long-term behaviour of the stressed cracked pipeline sections. The components did not fail during this test either. These results proved that the stress test was suitable for rehabilitating the damaged pipeline, and the stress test was decided to be carried out on the pipeline [8].
Carrying out the stress test. Before the quantified stress test with «training effect» was carried out, the general conditions for testing the pipeline had been defined. Seventeen test sections had been assigned with a maximum height difference of 73 m. Because of the height
Fig. 6. Stress corrosion cracking on the inner pipe surface [8]
Fig. 7. Expanded crack (left) and plastic deformation (right) after simulation of the stress test [8]
Fig. 8. Rupture pit to carry out sustained pressure tests
Fig. 9. Widening of a pipe after the stress test
differences, the lengths of the sections had varied between 530 and 12500 m. The test pressure at the low point of the pipeline had been not to exceed 137 bar. Based on the material, the remaining integral circumferential expansion had been defined as maximum 0,2075 %.
The pressure test was set up and carried out in accordance with VdTUV data sheet2 1060. While the sections were being filled with water, the flow, total volume, pressure, water temperature and the time were recorded. When the test pressure was applied, the pressure, volume of water pressed in and pressure increase rate inside the pipeline were recorded. During the specified holding times, the pressure and air temperature were being measured every hour. The temperature at the wall of the pipe was also being measured at least every four hours.
During the stress test, ten breaks occurred in four of the 17 pressure test sections. The rupture pressure varied from 59,2 to 130,7 bar. After each failure, the damaged pipe was replaced, and the stress test was repeated for that section.
After the stress tests, pigging was carried out to remove the remaining water, followed by geometry pigging to rule out inadmissible widening (fig. 9). Before the individual test sections were integrated, the pipeline had to be dried to rule out renewed corrosion damage. Air drying was carried out in accordance with a technically recognised method. On completion of the stress test, the respective rehabilitation and the measures described above, the pipeline was filled with natural gas. As a result of the stress test the pipeline could be described as rehabilitated and be operated at 55 bar with no further conditions imposed. Six years after the stress test the pipeline worked at an operating pressure of >50 bar with no problems or failures. A cost comparison between construction of a new pipeline and the rehabilitation process showed that the stress test, including rehabilitation, costed just 50 % of what a new pipeline would have cost. Consequently, the rehabilitation can be described as a success.
References
1. CAMP, H.J., de la. Prüfen und Optimieren mittels «Stresstest». In: BBR. 2006, no. 7-8, pp. 18-23. (Germ.). Available from: https://slidex.tips/ download/prfen-und-optimieren-mittels-stresstest
2. COSHAM, A., P. HOPKINS. The pipeline defect assessment manual. In: Proc. of IPC 2002. IPC02-27067.
3. KIEFNER, J.F. Role of hydrostatic testing
in pipeline integrity assessment [online]. Presented at Northeast Pipeline Integrity Workshop, Albany, New York, 2001. Available from: https://pdfs.semanticscholar.org/4ed9/b92944d8d6 fc1f32e7f44c1f6ee09d6b21af.pdf
4. LAIMMER, J. Schadensuntersuchung an einem beim Streßtest geborstenen Rohr. 2002. (Germ.).
5. EPLE: Druckprüfung und Abnahme
von Gashochdruckleitungen über 16 bar. 2003. (Germ.).
6. DECHANT, K.E. Streßtest-Prüfziel und bisherige Erfahrungen. 3R international. 1976, vol. 15,
no. 1, pp. 26-30. (Germ.).
7. BUSACK, V., U. HOFFMANN, K. PEIN, et al. Beispiel der Anwendung des Streßtests bei einer durch Spannungsrißkorrosionen geschädigten Leitung. GWF. 1998, vol. 139, no. 2, pp. 75-83. ISSN 0367-3839. (Germ.).
8. DELBECK, W., A. ENGEL, Z. KNOCINSKI, et al. Auswirkung des Stresstests auf spannungsrißkorrosionsgeschädigte Leitungsbauteile. GWF. 1993, vol. 134, no. 5, pp. 263-271. ISSN 0367-3839. (Germ.).
Принципы и опыт восстановления трубопроводов, подвергшихся коррозионному растрескиванию под напряжением, путем нагрузочного тестирования
К. Гюнтер1*, У Маревски1, М. Стайнер1
1 Open Grid Europe GmbH, Германия, 45141, г. Эссен, Каленбергштрассе, д. 5 * E-mail: christina.guenther@open-grid-europe.com
Тезисы. Практика лабораторных и полевых испытаний показала, что нагрузочные испытания позволяют восстанавливать работоспособность действующих трубопроводов со стресс-коррозионными дефектами. Установлено, что нагружение - это единственный вид комплексных прочностных испытаний, способный ликвидировать в трубах осевые дефекты, снижающие сопротивляемость материала труб деформации (разрыву), при этом влияние сохранившихся после процедуры трещин смягчается остаточными напряжениями сжатия. В связи с этим нагрузочные испытания предлагается применять не только с целью технической диагностики состояния трубопроводов, но и для их реставрации. Если по завершении соответствующих работ условия, благоприятствующие стресс-коррозии, сохраняются, тест необходимо повторить спустя некоторое время, определяемое скоростью роста трещин. Следует, однако, уточнить, что кольцевые трещины либо трещины, расположенные под углом к продольной оси трубы, лишь в малой степени поддаются уничтожению указанным способом.
Ключевые слова: коррозионное растрескивание под напряжением, дефектный трубопровод, трещина, испытание нагружением, восстановление работоспособности.
