Testing and optimisation using stress tests

From 2010, plans provide for up to 55 billion cubic metres of natural gas – roughly half of Germany’s present annual consumption – to be transported annually through the 1,200- kilometre Nord Stream pipeline under the Baltic Sea from Russia to Germany. Together with Moscow-based Gazprom subsidiary DOAO Orgenergogas, Munich-based TÜV SÜD Industrie Service GmbH has already carried out stress testing on 150 kilometres of the pipeline under construction. This stress test is a water pressure test which is performed at elevated test pressures and allows integral strength testing, thus enhancing pipeline quality.

Applications for stress testing range from pipelines to conventional plants and systems, series production of pressure vessels and systems in nuclear power stations. In Russia, TÜV SÜD carried out its stress test on two of Gazprom's export pipelines to date: the Jamal-Europe gas pipeline which runs from Russia via Belarus to Poland and the Blue Stream pipeline leading under the Black Sea from Russia to Turkey.

For over 40 years, this method has been used to test and optimise both new and existing, in-service pipelines. Over 1,500 tests have been performed to date on approx. 20,000 kilometres of pipeline with diameters ranging from DN 50 to DN 1600. In the service period following stress testing, no rupture or leakage has occurred on these pipelines – exempt from damage caused by external influences, e.g. excavator bucket teeth, corrosion etc. Some pipelines with extensive defects were thought to put out of service; however, after they had been subjected to a stress test, perfect service performance was restored. In some cases, the question whether reliable a operation of an existing pipeline can be guaranteed over a longer period of time can even be answered in the run-up to the test, and may render construction of a new pipeline superfluous.

Benefits offered by stress tests
As a genuine strength and load test, the stress test imposes a load approximating the actual yield point on all pipe sections of a pipeline. In a quantified (interrupted) stress test with "training effect" (pre-stressing to increase fatigue strength), steps are taken to prevent unacceptable permanent deformations in the pipes. This is achieved through statistical evaluation of the actual values of pipe strength and wall thickness, measuring the volume of water pumped into the pipe, and what is known as the 'training effect of the material", which refers to the observation that the fatigue strength of a material can be increased through the correct application of pre-stressing.

Scientific analyses and field experience demonstrate that when performed at a sufficiently high pressure, the water pressure test, as an integral test method, is suitable for checking pipe walls – including their welds – for unacceptable faults. In a stress test, faults in pressurised components are eliminated and localised deformations and other strength-reducing faults diminished.

Build up water pressure
The water pressure test carried out according to the stress test method is performed on pipeline sections laid in the ground which have been filled with water to exclude air if possible. During the filling process, pigs (pipeline inspection gauges) run against a certain supply pressure. The water or air supply pressure must be high enough to ensure that the water column between pig and filling point remains dense throughout the entire filling process and does not break at local peaks.

Subsequently, the pressure is increased with the help of high-pressure pumps offering smooth and shock-free operation. The water volume pumped into the pipeline is measured on the suction side of the pumps using an ultrasonic or other measuring device. Pressure measurement using pressure scales is carried out parallel to volume measurement. Upon reaching the individual pressure levels, both the total water volume pumped into the pipeline and the water volume pumped into the pipeline between two pressure levels are determined.

When the pressure throughout the water volume pumped into a pipeline section is plotted, a straight line is obtained in the elastic zone. Initial deviations from this straight line indicate that the proportionality limits of individual pipes have been reached. Major deviations from this line show that individual pipes or pipe assemblies are subject to stresses approximating the actual yield points.

The maximum permissible volume of water pumped into the pipeline can be calculated on the basis of the pipeline's rate of elasticity (elastic pipe expansion and compression of water in the pipes) and an integral permanent expansion of its circumference. Applicable acceptable limits have been defined depending on the materials used and are underpinned by extensive materials analysis. The rate of pressure increase varies between 1.5 and 4 bar/minute. Electronic measuring instruments are used to display the measured values (pressure and volume pumped into the pipeline) which are drawn in a schematic diagram (Figure 1).

Stress test including "training effect"
A quantified (interrupted) stress test with "training effect" can be seen in Figure 2 . This type of stress test is used when the objective is to achieve particularly high pressures while minimising permanent expansion in circumferential direction in order to eliminate defects. Further areas of application refer to pipelines made of materials with specific expansion properties or pipeline sections involving major height differences. The figure shows that this method allows higher pressures to be reached with lower permanent pipe expansion.

Figure 3 shows pipeline stress testing as a function of time. To improve the elimination of so-called unstable faults, the pressure must be maintained over two holding times of 30 to 90 minutes each. By slowly increasing pressure in the upper pressure range, faults and the material in their vicinity are allowed enough time to start yielding or creeping. The period of pressure relief between these two holding times should be at least 30 minutes. During this time, the pressure should be as low as possible, so that it is just above 0 bar at the highest point of the pipeline.

Pressure relief will cause a reversal of plastic deformation (buckling) around crack-tips and other faults. Buckling occurs at points subject to high tensile stresses and strain during pressure relief due to pipeline movement and the Bauschinger effect. Buckling too needs time to develop fully. Leaks and ruptures are still relatively frequent during the second pressure cycle and holding time. Leaks occurred with a certain frequency after a pressure of 70 per cent of the intended maximum test pressure had been reached, while ruptures do not occur until the maximum test pressure had nearly been reached or before the second holding time in the majority of cases.

Fault elimination
The pressure load imposed on the pipeline is intended to eliminate faults in the pipe wall. This is achieved if a load approximating the material's cohesive strength which corresponds to its tensile strength occurs at the edges or across the remaining cross sections of faults. There is negligible growth in the sheet's direction of rolling of localised faults which cause leaks when stress is imposed during the pressure test. Longer faults result in rupture when tearing.

Figure 4 shows the critical fault size in various pressure tests carried out on a 28" pipeline. This critical fault size was determined in the test and confirmed by calculations based on fracture mechanics.

If faults increase in size due to operating load cycles and if they reach a size critical for the pipeline operating pressure in question, leakage or rupture will result in the further course of pipeline operation. Higher test pressures cause higher circumferential stresses in the pipeline, thus eliminating less deep faults of equal length.

Surface faults, brittle cracking and martensite transformations even if only a few hundredths of a millimetre in depth, may also trigger pipeline rupture. Pipeline rupture may also occur in heat affected zones in the presence of high residual stresses, as these zones have been pre-damaged by excessive stretching in warm condition (points of repair, strip weld connection).

Elimination of faults in the girth weld depends on the Poisson's ratio, which increases from 0.3 to 0.5 only when the limit of proportionality has been reached and thus approximates the tensile strength of the material in the vicinity of the girth weld fault. Assuming natural, normal fault distribution, the number of eliminated faults increases disproportionately as circumferential stress increases (Figure 5). As strength – and thus the tensile strength ratio – rises, an increasing number of minor faults are eliminated in stress testing.

Manufacture effects
We know from experience that a 100 per cent circular pipe does not exist. Manufacturing causes certain deformations which,according to the standard, are still acceptable and, to some extent, may occur simultaneously. Soil pressure and the degree of compression of the pipeline after laying also cause pipeline deformations and stresses.

Stresses caused by pipe laying and residual stresses in the pipeline are reduced in circumferential direction during the stress test. They are added to the stresses imposed by the test pressure. In unevenly distributed residual stresses, the load increase exceeding the limit of proportionality results in relatively little plastic deformation. Similar to stress relief annealing, the circumferential (hoop) stresses applied during the stress test reduce residual stresses in components acting in the same direction to between 15 and 20 per cent of the yield point.

Around remaining fault locations, especially around crack-tips, the stress test causes stress displacement (Figure 6) which, in an unloaded pipe, results in the building up of high compression stresses at these points.

After stress test completion
The load cycle performance of any faults and deformations withstanding the loads applied during the stress test is clearly improved. Around crack-tips and locations with high stress concentrations in particular, yielding will occur due to high internal pressure loads. This, in turn, will improve pipe geometry leading to lower stress concentration factors (SCF) and negative compression stress during pressure relief.

If the test pressure was high enough, compression stresses may even still be negative during subsequent operating loads. This fact also explains why crack propagation initially observed came to a halt in cyclic load testing carried out with consistent dynamic loads on test specimens with cracks after a high individual load had been applied. A similar effect was noticed in tests carried out on large diameter pipes damaged by stress corrosion cracking. Tests also revealed the following association: the higher the individual test load and the longer the interval between this test load and subsequent dynamic service loads, the more crack propagation was decelerated or the more likely it was that crack propagation would be stopped altogether.

It has become evident that problems in pipelines in which faults must be reckoned with cannot be solved by either raising the safety coefficient or reducing pressure. Faults which, in continued operation, may result in leaks or ruptures must be eliminated in pipeline rehabilitation.

Prerequisites and limits
One of the prerequisites of stress test effectiveness is that test stresses act in the same direction as the stresses occurring during pipeline operation and that stress levels differ sufficiently. In pipelines with pitting corrosion, however, stress tests are not sufficient as the sole rehabilitation procedure. In this case, defects caused by corrosion must be tracked down with the help of other suitable test instruments such as "intelligent pigging systems" and removed if they overstep a critical size.

Since pressure tests performed according to the stress test method may move into critical ranges, experienced experts should be involved in the planning of these tests. Additionally, these tests should only be carried out under expert on-site supervision. The stress test can only be carried out if the deformation behaviour of the components in the pipeline section to be tested is almost identical and the pipeline appropriately fixed. If individual pipeline components have been designed with higher strength or subjected to different (lower) levels of internal pressure due to excessive differences in pipeline height levels, these components can no longer be referred to as "stressed".

Conclusion
Stress tests, as high-pressure tests, offer the possibility of optimising new pipelines and rehabilitating existing ones that display signs of stress corrosion cracking or groove corrosion.

In gas pipelines with remaining faults, cyclic stress testing may extend service life to more than 80 years. In some cases, these pipelines are operated at a higher service pressure than prior to the test. In oil-pipelines, depending on the mode of operation, the stress test may result in a remaining service life of up to 30 years – unless the pipeline is additionally harmed by internal and external corrosion or violently damaged.

Currently, the stress test is the only test method which directly eliminates dangerous faults. For this reason it is also referred to as an intelligent test method. Integral strength testing performed on each individual component has a preventive effect as removal of peak stresses improves pipe geometry. In contrast to most other test procedures, stress tests can therefore be applied to both pipeline testing and optimisation, enhancing both the operational safety and the costeffectiveness of the stress-tested pipelines.