NDT Solution
Observations on Magnetic Wall
Measurements of Coiled Oilfield Tubing
This is an interesting article describing one of the practical applications of magnetic NDT methods. A brief history of using magnetic methods in measuring wall thickness of ferromagnetic tubes, along with a simple explanation of the techniques involved, provides good background information to the reader unfamiliar with the technique. The author presents data indicating the influence of material grade on magnetic properties.
G.P Singh
Associate Technical EditorFigures 1-3
Figures 4-5
INTRODUCTION
The testing of coiled oilfield tubing is now considered to be an important requirement prior to and during offshore servicing. It is now mandated in the Norwegian sector of the North Sea via the Norsok requirements (Norsok, 1998) and is being considered by the Minerals Management Service for US operations. A draft document from the American Petroleum Institute (API) now covers test techniques and a joint industry project is investigating the effect of surface damage on coiled tubing bending fatigue, along with the nondestructive testing (NDT) used to detect that damage. Much of this has been discussed in earlier papers (Stanley, 1996a; Stanley, 1998; Stanley, 2001).One interesting method is that of wall thickness measurement of ferromagnetic tubes using a noncontact direct current magnetic technique (Stanley, 1992; Stanley, 1994; Stanley, 1996b). This paper covers some results obtained with this for carbon steel coiled tubing.
The advantage of this technique
is that it is a noncontact technique
TEST REQUIREMENTS
A draft document, API Recommended Practice 5C8: Care, Maintenance and Inspection of Coiled Tubing, is being prepared under the API Resource Group for Coiled Tubing. This document will cover standard sizes, cleaning of the internal and external surfaces, protection during periods of storage, corrosion and its mitigation, tube to tube welding (treated as a maintenance issue), NDT and some possible rules for assessing inservice carbon steel coiled tubing.The test section covers the techniques that have been developed to date for continuous tubing testing (including skelp end welds and tube to tube welds), along with imperfection signal "prove up" techniques. For continuous testing, tubing is delivered from one reel, passed through the test head at speeds in the region of 24.4 m/min (80 ft/min) and collected on a second reel. Testing is thus effected from the outer diameter.
Inside the test head (Figure 1), the tubing is magnetized to saturation longitudinally and scanned with arrays of hall effect sensors for the ambient tangential magnetic flux Bx through the sensors and for variations in this from surface imperfections (magnetic flux leakage). Because tubing expands under pressure and becomes oval under repeated bending, a dilation/ovalness detector provides the third test and a low frequency differential eddy current system is added for the detection of longitudinal gouges and heavily cycled areas (Stanley, 1999).
Because of the general condition of the outer surface of carbon steel coiled tubing, ultrasonic methods have not, to date, been attempted for any of the required tests with any degree of success. The introduction of ultrasonic methods for new oilfield tubulars in steel mills and fixed test facilities in the 1980s provided a severe test for the NDT industry. The use of ultrasonic testing on small diameter, variable surface and relatively thin walled tubing for both wall thickness and imperfection testing will thus present many problems.
THE COILED TUBING CONSORTIUM
For several years, a joint industry project existed that was administrated at the University of Tulsa and which focused on the cycle fatigue of the various grades of carbon steel coiled tubing (Table 1). The joint industry project became the Coiled Tubing Mechanics Research Consortium in 1999 and began to focus its efforts on the effects of controlled discontinuities of standard sizes and shapes placed into the outer surface of coiled tubing. The idea was to compare the cycle life of the tubing at various pressures, with and without the discontinuities.Discontinuities of many shapes were added to several grades of carbon steel coiled tubing, using milling and electrodischarge machining. These were fatigue cycled on machines built from an earlier joint industry project in Houston and Tulsa and the results were used to generate the computer program that enables the coiled tubing operator to determine the effect of the discontinuity on the fatigue life of the tubing (Tipton et al., 2002)
In this program, the length, width and depth of the discontinuity are required to provide the assessment of the loss of fatigue life, which is of course critical to practical coiled tubing operations. This, in turn, requires magnetic flux leakage indications during the magnetic test to be investigated and measured.
COILED TUBING CONSORTIUM DATA
Data taken at the University of Tulsa are shown in Figure 2 during calibration of a magnetizing coil that is part of a mock up of a coiled tubing test unit. Photos of the mock up can be seen on the Coiled Tubing Mechanics Research Consortium Web site at <www.coiledtubingutulsa.org>.However, while the Coiled Tubing Mechanics Research Consortium is investigating the magnetic flux leakage from carbon steel coiled tubing imperfections, with a view of the extraction of signal information content, there was a need to investigate the magnetic wall thickness measurement. With a hall effect gaussmeter, the axial magnetic field strength versus current was measured as follows (Moran et al., 2002). First, the center of the coil with no tubing present was measured. This magnetizing field should be given by
where
N = the number of turns
I = the current
d = the average diameter of the coil.The number of turns on the coil N was not known, so the data were taken with the gaussmeter at 0.25 A intervals from 0 to 4 A, which was the peak direct current available from the power supply. This is given by the leftmost straight line data set (Figure 2), which calculates at 4.8 mT/A (47.8 G/A).
Then the hall probe was moved by 22.2 mm (0.88 in.) radially away from the central axis, which is where the outer diameter of the 44.5 mm (1.75 in.) tubes used in the discontinuity study would be if they were present. These data (the "no tube" set in Figure 2) lie on the top of the first data set, so it can be seen that the field is uniform in the air over the diameter of the tubing used in this test. This is to be expected in this situation.
Then data were collected at the same location with two grades (HS90/CT90 from one manufacturer and QT1000/CT100 from another) and two wall thicknesses of straight sections of coiled tubing present axially inside the coil (the tangential field just outside the surface of the tubing with the hall element).
OBSERVATIONS
First, it can be seen that the tangential field strength Btan (measured in tesla or gauss) at the pipe surface is always lower than the field at the same point in air, for the same applied current in the coil. This observation is critical to magnetic wall thickness measurement.Second, it can be seen that the field at the pipe surface saturates at about 2.5 A of coil current. This represents a field of about 9.5 kA/m (120 Oe) to saturate these relatively thin walled tubes. This low value should also be easily obtainable using suitably designed permanent magnets.
The lower, curved parts of the graphs occur when the tube wall is not saturated. However, it can also be seen that they are still lower than the leftmost (air) line. The fact that the data sets (with the tubing present) appear to run almost parallel to the air curve above 2.5 A is important in determining how well magnetic wall thickness gages will calibrate, that is, how insensitive they are to changes in magnetizing current after magnetic saturation has been reached.
Third, for 44.5 mm (1.75 in.) HS90/CT90 and QT1000/CT100 tubing, the tangential surface field readings decrease as the wall thickness increases. This illustrates the general principle of wall thickness measurement by the direct current magnetic method. The thicker the wall, the lower the surface field. This phenomenon is well known and has been observed on earlier tests with drill pipe.
Fourth, the results for the two grades are a little different. This is because the chemical composition of the two materials is different and this difference affects the saturation flux density of the respective steels. (This fact is, of course, well known from studies of B versus H curves for various materials. B versus H properties are very dependent upon carbon content. One typical B versus H curve, taken on QT-90, is given in Figure 3).
MAGNETIC WALL GAGE
Data at 2.97 A
The data obtained at 2.97 A (tubes saturated) are shown in Table 2. Plotting the data as field strength versus wall thickness, Figure 4 is obtained. With the pipe absent (wall = 0), the open coil field is used. Figure 4 illustrates an almost linear plot for 100 grade tubing and a curved plot for 90 grade tubing, over the wall thickness interval investigated. Looking at the data from wall thicknesses of 0 to 3.1 mm (0 to 0.1 in.), there is a drop of 0.92 mT/mm (232 G/in.) of pipe wall thickness for the 100 grade (upper curve) and 1 mT/mm (264 G/in.) for the 90 grade material (lower curve). Thus, there may be a grade effect on wall thicknesses measured in coiled tubing testing units using this technique, but it could easily be dealt with during calibration of a test system using tubes of the same grade, chemistry and heat treatment.Looking beyond to 3.96 mm (0.16 in.) wall material, the CT-100 grade material shows a drop of 0.93 mT/mm (237 G/in.), indicating a slight increase as the wall gets thicker. This is clearly seen in the 90 grade data, where the line curves more sharply.
Thus, now we have a wall gage that measures at about 0.1 mm (4 ´ 10-3 in.) of wall thickness per 0.1 mT (1 G) at the surface of the tubing, at least in the region of wall thicknesses used in this quick test. Obviously data need to be taken for walls that are thicker than 3.96 mm (0.16 in.) and since there is no reason to suppose the data over a wider range are linear, we can expect to have to put a transfer curve into any wall gage that uses this method.
Data at 3.47 A
A further plot at 3.47 A is shown in Figure 5. As with Figure 4, the upper curve is for CT100/QT1000 and the lower one is for CT90/HS-90. The same general properties are exhibited as are shown in Figure 4.
A SIMPLE EXPLANATION OF THE TECHNIQUE
The magnetic field lines from current I in the coil have a pattern similar to that shown in Figure 1. All the field lines encircle some part of the total current in the coil. Some lines enter the tubing, while others do not. At the points where these field lines enter and leave the tubing (areas P1 and P2, which are known as poles in older magnetic texts), the value of the magnetic flux density B in the field line is continuous, but the value of the magnetizing force (or magnetic field intensity) H is discontinuous; this manifests itself as poles at these points. These poles create a demagnetizing field Hd within the tube wall system, which is in the opposite direction to the applied coil field. The actual magnetizing field at any point in the tube wall is therefore not the original H (air) field, but rather a field that is affected by the presence of this demagnetizing field Hd. In fact, even the value of the field at the location Hc (far away from the tube) will be less when the pipe is present because the effect of Hd spreads throughout the magnetizing system.The number of poles at locations P1 and P2 increases with the wall thickness of the tubing. Clearly, there are no poles at these locations when there is no tubing and the value of Hd, the demagnetization field, rises with t. The largest value of Hd would be where there is a rod of diameter 44.5 mm (1.75 in.) present, that is, t = 22.2 mm (0.88 in.).
Location of Hall Element
It has been discovered that the hall element does not have to be right up against the tube surface for this method to work and, in fact, in one system, the hall sensors are placed just inside the magnetizing coil so they are well protected. (Note the comment on Hc above.)
A LITTLE HISTORY
In the early 1980s, a method for measuring the average wall thickness in tubes (or the thickness of steel rods, such as sucker rods) was developed using a magnetizing coil and an encircling coil connected to an integrating circuit. In order not to infringe on that patent, it was supposed that if the encircling coil was broken up into a series of elements, then the patent would not be infringed (Stanley et al., 1985). The hall element proved to be the ideal solution since it does not require the electronic integration that the encircling coil requires. Tests were performed on some used drill pipe at the International Pipe Inspectors Association. The end result of these tests was that magnetic wall measurement systems have been added to some commercial drill pipe test units and also form the basis of wall thickness measurement variations on coiled tubing test units.
ADVANTAGES AND LIMITATIONS
The advantage of this technique is that it is a noncontact technique. The hall element can be lifted away from the pipe surface and the wall thickness method still works. This is particularly useful and should be investigated more thoroughly.Unfortunately, the sensors used for discontinuity detection rather than wall thickness measurement do have to be as close as possible to the tube surface because a serious and well documented liftoff effect occurs.
For determination of the magnetic properties of carbon steel coiled tubing used, it was essential to determine the B versus H curve properties in both the longitudinal and circular directions. The curve shown in Figure 3 is one taken in the longitudinal direction. The sample used was 305 mm (12 in.) long and 25.4 mm (1 in.) wide, cut from a section of 73.7 mm (2.9 in.) coiled tubing.
It can be seen that the material does not saturate in the longitudinal direction until after the application of about 15.9 kA/m (200 Oe), but based upon prior work, where it was determined essential to obtain large amounts of magnetic flux leakage from small transversely oriented discontinuities in the walls of carbon steel coiled tubing, it appears that application of 9.5 kA/m (120 Oe) will provide sufficient field strength to obtain measureable amounts of magnetic flux leakage from such discontinuities. Indeed, substantial and measureable amounts of magnetic flux leakage occur at much smaller applied fields and it is well known that optimal magnetic particle testing may be performed at much smaller fields. Here, of course, the higher field is needed to provide magnetic flux leakage at a relatively high liftoff distance than is used for magnetic particle testing.
REFERENCE
Moran, D.W., C. Chinsethagid, J.R. Sorem, Jr., and S.M. Tipton, "Challenges Facing the Development of Coiled Tubing Inspection Technology," Proceedings from SPE/Icota Conference, Houston, Texas, April 2002.Norsok, Norsok Standard M650: Qualification of Manufacturers of Special Materials, Lysaker, Norway, Norsok, 1998.
Stanley, R.K., "Assessment of Tubing in Oil and Gas Wells by NDE Methods, with Profiles of Tubular Damage," Proceedings from the 1st ASME/NDE Engineering Division Topical Conference, San Antonio, Texas, April 1992.
Stanley, R.K., "Modern Methods of Magnetic Inspection," Proceedings from EPRI's 3rd Heat Exchanger Conference, Myrtle Beach, EPRI, June 1994.
Stanley, R.K., "Overview of the Nondestructive Inspection Techniques for the Evaluation of Coiled Tubing and Pipe," Materials Evaluation, Vol. 54, No. 11, 1996a, pp. 1245-1250.
Stanley, R.K., "Magnetic Methods for Wall Thickness Measurement and Flaw Detection in Ferromagnetic Tubing and Plate," Insight, Vol. 38, No. 1, 1996b, pp. 51-55.
Stanley, R.K., "Results from Inspections of Coiled Tubing," Proceedings from SPE/Icota Conference, Houston, Texas, SPE, April 1998.
Stanley, R.K., "Results of Recent Inspections Performed on Coiled Tubing," Proceedings from SPE/Icota Conference, Houston, Texas, SPE, May 1999.
Stanley, R.K., "Problems Associated with Coiled Tubing," Proceedings from the International Chemical and Petroleum Conference, Houston, Texas, June 2001.
Stanley, R.K., M. Milewits, R.C. Knauer, Jr., and J.E. Bradfield, "Magnetic Flux Method for Measuring Tubular Wall Thickness," US Patent 4,555,665, 1985.
Tipton, S.M., D.W. Moran, C. Chinsethagid and J.R. Sorem, Jr., "Quantifying the Influence of Surface Defects on Coiled Tubing Fatigue Resistance," Proceedings from SPE/Icota Conference, Houston, Texas, SPE, April 2002.
* Quality Tubing, Inc., 10303 Shelton Rd., Houston, TX 77049; (281) 456-0751; fax (281) 456-7620; e-mail <rkstanley@ndeic.com>.
Copyright © 2004 by the American Society for Nondestructive Testing, Inc. All rights reserved.
Copyright © 2008 by the American Society for Nondestructive Testing, Inc. ASNT is not responsible for the authenticity or accuracy of information herein. Published opinions and statements do not necessarily reflect the opinion of ASNT. Products or services that are advertised or mentioned do not carry the endorsement or recommendation of ASNT.
IRRSP, NDT Handbook, The NDT Technician and www.asnt.org are trademarks of the American Society for Nondestructive Testing, Inc. ACCP, ASNT, Level III Study Guide, Materials Evaluation, Nondestructive Testing Handbook, Research in Nondestructive Evaluation and RNDE are registered trademarks of the American Society for Nondestructive Testing, Inc.
ASNT exists to create a safer world by promoting the profession and technologies of nondestructive testing.