Detecting Pipeline Buckling

As projects move into deep water, alternatives to traditional in-line methods become necessary.

Limitations and problems with traditional in-line pipeline buckle detection methods are prompting offshoreoperators and installation contractors to seek alternatives, especially as the search for oil and gas moves into deep water.

In order to detect buckles during installation that appear near the touch down point, it is common practice to use a buckle detector. Traditional buckle detectors usually involve pulling a buckle detector (gauge plate) beyond the touch down point, after the installation vessel, with a wire connected to the vessel (Figure 1). As long as a buckle is instantly detected, the buckled section can normally be recovered to the installation vessel and repaired without any significant delay or high cost. However, if the buckle is not detected before thepipeline installation is completed, the repair cost and time can be immense. It is therefore very important that the probability for a buckle is kept as low as possible, and if it occurs, it is instantly detected.

In deep water, the probability for such a buckle to appear is higher than for more shallow water. In spite of this increased need for the buckle detector in deep water, there is a growing reluctance among installation contractors to use such in-line tools. This is caused by several incidents with broken wire, lost pipe, and lost buckle detectors.

In addition to the above challenges, for deepwater pipelines, the reliability of such tools is also questioned. Deepwater pipelines therefore need alternatives to the traditional buckle detector. The goal here is to discuss some of the problems related to traditional buckle detection methods, and present some alternatives that may be particularly suited for deepwater pipelines.

Traditional methods

The traditional buckle detector has been successfully used for years. The device relies on monitoring the load increase in the wire pulling a relatively thin gauge plate of aluminium. This plate is normally split radially into sections, in order to ease the sizing of possible buckles (Figure 2).

The possible load increase in the wire due to a buckle, at least if it is relatively small, will be quite limited. For relatively shallow water, this has not been a problem, since the load increase is normally big enough to overcome the alarm settings.

However, for relatively large-diameter pipelines, the confidence associated with traditional buckle detection methods is increasingly being questioned. For deepwater pipelines, the confidence may be even lower, since most of the pulling force is due to the friction between wire and pipeline.

Due to the required high settings of the alarm level to avoid false alarms, the reliability of the traditional buckle detector will decrease proportionally with water depth. For ultra-deepwater pipelines, it is doubtful if local buckles, other than a full collapse, can be detected with the traditional buckle detector. The additional load to bend the buckle detector will be marginal compared with the very high force in the wire due to friction and self weight.

Other problems

In addition to the lower confidence in buckle detection in deep water, installation contractors have experienced other problems with traditional buckle detectors. If the buckle detector gives too many false signals, a vessel crew may turn the whole alarm off, out of frustration. This is not an unknown phenomenon, and it is not expected that this problem would be less in deep water.

In order to avoid this problem, some contractors have started to locate the buckle detector just after the stinger, in order to detect possible buckles that are created on the stinger only. This means, however, that buckles that take place in the sag bend will not be detected.

Another problem that has been frequently noticed is broken wire. The "fishing activity" required to get it out again can be difficult and time consuming. The consequences are even larger if the installation vessel loses the pipe for some reason. If the pipe is dropped, seawater will flow into it, and can push the buckle detector into the pipeline. This happened during a pipeline project offshore Norway not too long ago, in approximately 400 meters of water. The wire broke, and the water pushed the buckle detector and several hundred meters of wire into the pipeline. Crews tried to fish out the wire and buckle detector, but without success. Then they attempted to pig them out, also without success.

Ultimately, the pipe became blocked, and the whole pipeline had to be recovered. The buckle detector and the wire were found squeezed together in a one-meter section of the pipe. Fortunately, the above incident happened in a relatively early stage of the pipelay, and thus only a short distance was affected. The repair was, nevertheless, very costly. If this had happened during the installation of a long deepwater pipeline, the consequences could be very high. As a consequence, many installation contractors have become reluctant to use the traditional inline buckle detection methods, especially in deepwater operations.

Alternatives

Yet even with thorough installation analysis and careful adherence to procedures, there will still be a need for some kind of buckle detection, especially in deep water. If traditional methods are insufficient, what are the alternatives? Possibilities include:

• Increased safety against buckling
• Visual buckle detection by ROV
• parameter control in order to detect buckles by change of pipe string configuration and/or loss of submerged weight.
• Detection by escaping air
• Detection by acoustic noise.

These alternatives are valid for deepwater pipelines only, since the buckle is assumed to be larger and thus easier to detect. DNV has studied some methods which have shown promise, but need further development and comprehensive testing.

Increased safety

One alternative to keep the risk at an acceptable level could be to increase the safety level against buckling. The risk (R) can be defined (P = probability of a failure and C = Consequence of that failure):

R = P x C

The philosophy is that in order to keep a constant risk level, the probability of a failure must be reduced if the consequence is increased.

DNV-OS-FlOl is a pipeline code that is well suited for such a philosophy. The code categorizes the consequence of failure in three safety class levels: low, normal and high. The higher the safety class, the higher the safety factors against failure modes that should be used. The typical (normally used) safety class for pipelineinstallation is safety class "low." One alternative to reduce the probability of a buckle can be to use safety class normal or high.

The main parameter affected by increased safety class for local buckling in the sag bend is the horizontal tension. This will, however, have a negative effect on the bottom tension, since it will normally require higher bottom tension to achieve a larger sag bend curvature.

For many pipeline projects, it is normally desirable to use low bottom tension, because it is desirable to reduce the free span lengths, and thus reduce possible intervention work, short radius curves, etc. For deepwaterpipelines, this is even more of a necessity, since the cost of intervention in deep water is more difficult and costly than in shallow water.

One typical example of this is the Ormen Lange field development in Norway, where the water depth is approximately 1,000 meters, and the seabed is extremely uneven. This area is also exposed to very high deepwater current. The cost of increased tension is expected to be significant. However, for pipelines where higher bottom tension is acceptable (i.e. where the seabed is relatively flat and there is a straight route) this may be an option.

Increased safety class will reduce the probability of a buckle. However,pipeline analysis and sag bend utilization is not the only parameter that affects the probability of a failure. As can be seen from DNV-OS-F-IOl, sec. 2B Safety Philosophy, there will also be additional quality assurance requirements. This means that there will be a whole range of activities that should be put in place; i.e., organization, personnel competence, quality assurance, linepipe control, vessel qualification, procedures, installation parameter monitoring, etc.

Visual detection

Another alternative can be to use an ROV to continuously follow the pipe just after the TDP. The idea is that if the pipeline buckles, the buckle could be detected visually from the outside. With continuous video monitoring, it should be possible to visually detect possible buckles. Many installation contractors are already using continuous TDP monitoring.

The problem with this method is that it requires good visibility. If the pipe stirs up mud, visibility can be impaired. This method also requires that a person continuously view the video screen. This should in principle not represent any problem, but in practice, can we really trust that a person will be able to continuously stare at the video screen without break?

An ROV may also experience breakdown. In deep water, it takes a long time to recover, repair and swim down again. A deepwater installation vessel can be very expensive, and, in practice, it would be difficult to wait until the ROV is fixed before resuming installation. The ROV may also have to be used for other purposes during installation. This can of course be solved by an additional ROV, but this will increase the cost and complexity of the operation.

Parameter control

If a local buckling or cross-sectional collapse occurs in very deep water, it will most likely develop into a propagating buckle if the pipelines are installed air-filled. This is of course dependent on the D/t ratio of the pipe, water depth, the size of the buckle, etc. If a long propagating buckle takes places in the sag bend, the buckled section will lose its buoyancy and bending stiffness. This will affect the pipe configuration, top tension, top angle and other relevant parameters. This effect could in theory be utilized to detect a buckle.

The pipe configuration can be easily established by finite element analysis. The configuration can be verified during installation by checking the water depth, top angle, top tension (lateral and vertical) and the layback distance (the horizontal distance between the top point and the touch down point), direct monitoring, etc. If the parameters that describe the pipe string configuration are closely monitored during installation, a buckle could be identified. This will require that the parameters be continuously monitored, in order to detect rapid changes.

Of course, this method has a number of limitations. Very few installation vessels have the capacity to monitor all parameters. And even if all the parameters are measured, this method can still be difficult due to errors and tolerances of the monitoring system. There are many sources for error. Typically, this could be that the monitoring equipment is too coarse, or is out of calibration. For a J-lay vessel, the weight of the new pipe stalk that is installed on top will give a significant top load change, and dynamic loads will impose a natural variation of the parameters. In addition, strong and rapidly changing current (Figure 3), dynamic effects, free spans along the route (Figure 4), and sloping seabed will change the pipe configuration, and thus the monitoring parameters. In order to be able to detect a propagating buckle, it is therefore expected that the affected section must be relatively long.

Most deepwater pipelines, where the risk of propagating buckling is relevant, will have buckle arrestors in order to stop such propagating buckles. Usually, these will be installed at a distance of 500 to 1,000 meters if the pipeline is installed by use of tensioners. In such a case, the length of the deformed section will be limited to the section between two arrestors. If the distance between the arrestors becomes very short, the affected section will also be very short, and thus less efficient to use as a buckle detector. Thus, for installation vessels that use lay collars in order to hold the pipe during installation, this method will have limited value. A typical distance between the lay collars is 50 to 80 meters. The lay collars will act as arrestors, and thus limit the distance of the section with reduced buoyancy and stiffness.

In Table 1, the effect of loss of buoyancy and bending stiffness of a 24-in. pipeline installed in 2,000-meter water depth is presented. Three different cases are studied: case 1-100 meters buckled section; case 2-200 meters buckled section; and case 3-500 meters buckled section. The buckled section starts at static TDP, and propagates upward the catenary. Note that case BC is the static configuration of an intact case. No other parameters are included like current effects, free spans, etc.

As can be seen from the results, the buckled section must be quite extensive before it will be possible to detect, even under optimal conditions and a very good monitoring system. One negative aspect of this method is that it is not effective in detecting smaller buckles that do not initiate a propagating buckle.

Air and noise

It has been argued that there have been cases where buckling has been detected by air that is pushed out through the pipeline at the installation vessel. Buckles that occur on ultra-deepwater pipelines may initiate a propagation buckle. If this occurs, and the propagation goes upwards along the catenary, air will be pushed out through the pipeline at the top. If the vessel crew is observant, this method could in theory detect a possible buckle. The propagation will go quite fast, and the escaping air will move quickly.

However, this method will have the same or even more limitations than the methods mentioned above. If the propagation is going outwards from the vessel, no air will escape at the vessel. If only a short section is propagated, the amount of air coming out will be relatively limited. If installation collars are used, the amount of air will be limited.

Another problem is that there will normally be air blowing in and out of the pipe all the time anyway. Anyone that has been on a deepwater lay vessel has experienced this phenomenon. It can thus be difficult to distinguish between a normal situation and a buckling situation.

It has also been argued that a buckle will create significant noise that can be detected and used as buckle detection. It may be true that a buckle makes significant noise. However, the level of noise is not documented, and, in any case, it will be difficult to distinguish between normal noise on a lay vessel, and noise from a buckle. This method may however, be studied and developed further, especially through the installation of special listening devices.

Conclusion

Pipeline buckling is a significant risk that must be addressed in any deepwater installation activity. The consequence of an undetected buckle in an ultra-deepwater pipeline can be very high. Traditional in-line buckle detection methods have significant shortcomings, and can represent a significant hazard in and of themselves. If at all possible to avoid, nothing should be put inside the pipeline during installation. This raises the need for alternative methods.

There are an array of alternatives, some of which have been used in the field, and some of which are still in development. As indicated, most have significant shortcomings, so there is still a significant need for the industry to develop improved methods.

As long as there is no single method that can detect buckles with confidence, it is recommended to reduce the probability for local buckling as much as possible. This may require increasing safety factors, and conducting a comprehensive and thorough analysis to establish all relevant monitoring parameters. Then, one would need to monitor as many installation parameters as possible, set forth detailed installation procedures, and conduct thorough qualification and testing of all installation and monitoring equipment. Additional measures would include training of personnel, and increased focus on quality control from fabrication to installation.

Even after the probability of a buckle has been reduced as much as possible, some kind of buckle detection method should still be put in place, especially for ultra-deepwater pipelines. A combination of the methods described herein could, in many cases, give relatively good confidence. To a large extent, however, these methods rely on personal skills, understanding, and continuous and thorough attention during installation. This means that thorough training would be required for the methods to have any confidence.

In any case, the method(s) selected has to be project-specific. Many parameters will affect this selection, such as water depth, D/t ratio, distance between arrestors, pipelay vessel capabilities, seabed topography, and environmental loads

Sumber : Asle VenÃ¥s; Collberg, Leif. "Detecting Pipeline Buckling". 27 Januari 2014. http://search.proquest.com/docview/214452600?accountid=31562