Printer Friendly Version
Making the "Smart Pig" Smarter
Addition of nonlinear harmonic sensors improves the discovery of dangerous pipeline defects
By Al Crouch and Graham Chell, PhD
Buried typically three to five feet below ground, pipelines silently carry natural gas, gasoline, crude oil and other products from their sources in the nation's oil- and gas-producing regions to processors and consumers in the country's population centers. Not until there is an anomaly in the line's operation does the average citizen become aware of a pipeline's presence. In the rare event of a pipeline failure, the aftermath is widely reported and the public is made aware of its presence and vulnerability.
Outside force, such as excavation from construction equipment, is the leading cause of these incidents, according to statistics compiled by the U.S. Department of Transportation's Office of Pipeline Safety (OPS). Even though pipeline rights of way are marked, sometimes a digging crew will not realize a pipeline is in its path until someone hits it.
Pipelines that fail from this type of external damage most often fail at the time damage is inflicted. The result can range from a nuisance leak - requiring a brief pipeline shutdown and repair - to a major event with significant environmental damage, fire and even injury or loss of life.
Although only about 5 percent of mechanical damage incidents results in a delayed failure, which can come months or years after the initial damage occurred, the potential for disaster is great enough that detection and remediation of mechanical damage defects carries a high priority with the OPS and also with pipeline operating companies.
Southwest Research Institute scientists and engineers have developed a new nondestructive evaluation technology based on nonlinear harmonics (NLH) to improve traditional in-line inspection (ILI) methods for pipelines. This method will provide pipeline operators with additional information about mechanical damage resulting from outside force, improving their ability to detect and characterize these defects and allowing them to take remedial action to avoid potentially dangerous pipeline failures.
Mechanical damage defects consist largely of dents and gouges. A gouge is more serious than a simple smooth dent because of its propensity for developing cracks. Although such defects are produced by action external to the pipe, once the pipe is underground, they can be detected only from the inside by devices called "smart pigs" that are propelled through the pipeline.
Smart pigs, or ILI devices, carry sensors, data processing electronics and data storage for the typical 50-mile run through an operating pipeline. After the pig has been retrieved from the pipeline, engineers gather the data and analyze it to reveal the condition of the pipeline.
There are two types of ILI systems that respond to damage defects. The so-called caliper pig, or geometry pig, uses mechanical or electronic feelers to sense distortions of the pipe's inner surface. This system can reliably detect and measure distortions on the inside pipe wall such as dents, but will not necessarily indicate whether a gouge accompanies the dent. The magnetic flux leakage pig, designed for corrosion detection, will detect gouges that are favorably oriented but, in general, is neither a reliable detector of dents and gouges nor of the severity of such defects.
Given the shortcomings of the existing ILI systems for detecting and evaluating mechanical defects, industry and government were interested in developing more reliable technology. GTI, and later the OPS, sponsored SwRI's development of new sensing technologies and analysis procedures to better quantify and assess the severity of mechanical damage.
SwRI has pioneered the use of NLH sensors for detecting stress and strains in ferromagnetic materials. Because a mechanical defect is expected to produce stress and strain anomalies in the pipe wall, detection and even quantification of these should be achievable with NLH sensors.
Nonlinear harmonic method
The NLH method developed at SwRI consists of impressing an alternating magnetic field onto a magnetic material such as steel and sensing the amount of magnetism produced in the part. To do this, a transformer is used to cause a magnetic field to pass through two electrical coils and into the pipe wall. One of the coils is connected to a source of electrical current that oscillates thousands of times per second. Because of the nature of the pipe wall material, the magnetism in the pipe wall does not oscillate with the same pure waveform provided by the electrical source. Instead, its oscillation pattern is distorted so that it contains frequencies that are several times higher than the frequency of the electrical source. The NLH method takes advantage of these higher frequency oscillations, sometimes called the "harmonics of the excitation frequency. "A secondary electrical winding on the transformer core responds to the oscillating magnetism and produces an electrical signal that can be filtered to remove the source frequency and retain the harmonics. The amplitudes of these remaining harmonics are considered related to the level of stress and strain in the steel pipe wall.
Testing of mechanically damaged pipes
The bulk of the project work performed by SwRI for the OPS and GTI has made use of artificial defects put into 24-inch-diameter pipes by Battelle Columbus using a machine designed especially for that purpose. Special indenters were forced into the surface of a pressurized pipe under controlled loading conditions. Battelle placed 18 such defects in six 6-foot-long pipe segments for SwRI use.
The first step in characterizing the mechanical damage in the pipe specimens was to document the profile of each defect on the inside surface of the pipe specimen. This was done with a laser range-finding sensor. Later, similar measurements were also made on the outside surface of the pipes to directly characterize the dimensions of the defects.
To examine the defects while the pipes were under pressure, SwRI scientists designed and fabricated a scanner to deploy two NLH sensors against the inside surface of the pipe segments. One sensor had its sensitive axis aligned with the axis of the pipe segment and the other (the hoop sensor) was perpendicular to the first so that they would be sensitive to the axial and hoop components of pipe wall stresses and strains respectively. The sensors were gimbaled to facilitate their contact with the surface in the deformed area on the inside pipe surface under the defects. This arrangement of NLH sensors on the pig gives 360-degree coverage of the inside surface of the pipe.
After all the defects were laser-mapped, they were scanned while the pipes were under zero pressure with the two NLH probes to obtain signals that could be used to verify the ability of NLH to detect and characterize mechanical defects. The pipe segments were then outfitted with a hemispherical closure on one end and a flange ring on the other in preparation for pressurized testing in a test fixture. This fixture was filled with water to raise internal pressure to simulate fluid flow, first to 400 pounds per square inch (psi) and then to 800 psi. At each pressure level, SwRI scientists took further NLH data from defects in the specimens. Scientists again used axial and hoop sensors and recorded data for a 12-inch by 16-inch rectangular scan zone around each defect.
The final step in the test program was to pressurize the pipe specimens to failure to determine the order in which the defects failed as a first-order indication of the defect severity with respect to burst pressure. Pressure testing is not perfectly representative of the failure mode of such defects in the field, as the normal mode of failure involves time-dependent crack growth caused by pressure cycling and other mechanisms. However, it is the most representative test allowed by the schedule and budget available for this project.
Interpretation of NLH Signals
SwRI scientists constructed finite element models designed to simulate the gouging process to predict the displacements, stresses and strains in the damaged pipes used as test specimens. These finite element models were used in stress analyses to provide physical interpretations of the signals the NLH scanner detects in actual pipes. The Institute obtained the special elastic-plastic finite element software ABAQUS/Explicit, which uses dynamic re-meshing to compensate for the extreme deformations experienced by the pipe wall during gouging. Institute scientists validated the modeling procedures by successfully comparing residual gouge displacements obtained from the laser scanning and load-versus-time data measured by Battelle Columbus during the manufacture of the gouges with the results of the finite element analyses.
It was important to the greater goals of the project that the finite element analyses accurately predict the conditions in the damaged pipes to help understand the physical meaning of the NLH signals.
Preliminary comparisons between results of the NLH and finite element modeling verify that the NLH signals are related to stresses and strains generated on the inside surface of a pipe by external mechanical damage.
In addition, NLH signal contour plots, when compared with laser scans made on the outside of damaged pipes, demonstrate that these signals not only show the presence of a defect but possibly also contain information related to the length and width of the gouges. Investigations are under way to determine whether information related to the depths of the gouges is also contained in the NLH signals. These multiple capabilities are not presently available in commercially available ILI tools.
Defect severity criterion: the ultimate aim
Because detection of mechanical damage and the stresses and strains in the pipe wall resulting from it was demonstrated to be feasible using the NLH method, the important next step is to develop a defect severity criterion tied to NLH signal characteristics.
A severity criterion is critical for mechanical damage because a pipeline may contain hundreds of dents caused by rock impingement or other non-gouging sources. The vast majority of mechanically induced defects in pipes will never produce cracks or lead to pipeline failure. However, at present a pipeline operator who wants to remove all serious defects from a pipeline is faced with unearthing every defect site to determine if the dent is smooth or gouged. This process can be unjustifiably expensive. In addition, the unearthing process itself carries a small risk of introducing new defects into the line. The ideal solution is an inspection method that can detect mechanical damage and assign a severity ranking to each so that only the critical defects have to be investigated through digging.
SwRI engineers believe that NLH sensors can be developed for use with ILI equipment, such as smart pigs, and combined with the results of analytical investigations, such as stress analysis, not only to detect mechanical defects, but also to characterize their severity.
SwRI successfully demonstrated the feasibility of NLH technology to detect and characterize mechanical damage defects. This will lead to the availability of a new in-line inspection capability to pipeline operators as a component of integrity management.
Commercialization is already under way in a U.S. Department of Energy project, funded by the National Energy Technology Laboratory, in which NLH sensors are being designed and installed on a new ILI smart pig by Varco's Tuboscope Division in Houston. As prime contractor of that project, SwRI is designing NLH sensors and transferring technology to Tuboscope to manufacture and integrate them into the inspection system.
The authors would like to acknowledge Gary Burkhardt, who guided development of the NLH probes and electronics; Errol Brigance and David Jones, who designed and built the NLH scanner and sensors; Vic Aaron and Steve Clay, who were responsible for the pressure testing and failure testing of the pipe segments; Yi-der Lee, who led the advanced finite element modeling and analysis; and Bob Warke, who is directing fractographic analysis of the post-test defects.
Published in the Fall 2002 issue of Technology Today®, published by Southwest Research Institute. For more information, contact Joe Fohn.