·Table of Contents
·Methods and Instrumentation
The Non-destructive Evaluation Method for Far-side Corrosive Type Flaws in Steel Plates Using Magnetic Flux Leakage Technique
Faculty of Engineering, Yokohama National University
Tokiwadai 79-5, Hodogayaku, Yokohama, JAPAN
Graduate Student of Yokohama National University;
Tokiwadai 79-5, Hodogayaku, Yokohama, JAPAN
Japan National Oil Corporation;
Uchisaiwaicho 2-2-2, Chiyodaku, Tokyo, JAPAN
This paper proposes a quantitative evaluation method for the size of far-side corrosive flaws in steel plates or in pipe-walls by means of the magnetic flux leakage technique. For this purpose, an approach that considers both experimental and theoretical modeling works concerning the simplified model flaws of local corrosion or pitting has been adopted. The experiment has been carried out for steel plate specimens containing artificial flaws with various depths and widths, such as circular holes and rectangular grooves in the backside of specimen plates. These experimental results have suggested that the depth of flaw or residual plate thickness can be estimated by the measurements of strength and distribution of vertical component in magnetic leakage flux density arising from far-side flaws. This is confirmed by analytical model calculations, which have been proposed by the authors concerning far-side magnetic leakage field. Finally, a practical evaluation scheme for size of flaws or residual thickness of metal loss area using this technique has been presented.
Keywords: NDT, magnetic flux leakage testing, oil storage tank, underground pipe, under-side flaw, Hall probe, local corrosion
In the petrochemical industry, there is an increasing need for effective NDT technique which is able to evaluate quantitatively the under or outer side metal loss area due to pitting or local corrosion in oil-storage tank floors or underground steel pipes. Concerning reliable inspection for such far-side metal loss flaws, the magnetic flux leakage technique(MFL technique)is now considered to be one of possible useful methods-.For applications of the MFL technique to the inspection of defects in steel components, there are two schemes. One of them is the usual method for detecting crack-like or another type flaws in "near-side" of the materials,, and the other may be referred as that for "far-side flaws" on the backside of the materials being tested-. Although there are a number of experimental and theoretical works on the usual MFL method for the near-side flaws, a little detailed work has been done on quantitative evaluation, especially of depth for far-side metal loss area in steel plates or pipe walls, by means of the so called "far-side MFL technique".
This investigation has been designed to obtain basic information for quantitatively evaluating the size of far-side corrosive flaws in steel plates or pipe-walls by mean of the magnetic flux leakage technique. An approach that considers both experimental and theoretical modeling works concerning the simplified model flaws of local corrosion or pitting has been adopted. First, the experiment has been performed for steel plate specimens containing artificial flaws with various depth and width, such as circular holes and rectangular grooves in the backside of specimen plates. Next, theoretical expressions concerning the far-side magnetic leakage field have been proposed for model flaws. Both of analytical model calculations and experimental results have suggested that the depth of flaw or residual plate thickness can be estimated by the measurements of strength and distribution of vertical component in magnetic leakage flux density arising from far-side flaws.
On the base of the results of this study, we propose an effective size evaluation procedure which has possible application to oil storage tank floor inspection.
II. Specimens preparation and experimental procedure
The main target of this study is the inspection for bottom floors of oil storage tanks. Therefore, in this study, we used plain carbon steel plates of 12mm and 22 mm in thickness as the test specimens. As the flaw models for pitting and local corrosion in tank floors, two kinds of artificial flaw in the bottom side were chosen as being extreme cases of naturally occurring local corrosion flaw profiles. These artificial flaws are a series of flat-bottomed cylindrical holes and rectangular grooves having various widths and depths. The magnetization characteristic curve for the specimen-plates have been examined and the value of saturated magnetic flux density (Bs) was estimated to be of 1.6 (T) for the plain carbon steel used.
The experimental setup consisting of an electrically magnetizing yoke, a magnetic sensor, a controllable sensor positioner and test specimen is shown in Fig.1. The measurements
of magnetic leakage flux density distribution have been done using a Gaussmeter and the sensor head for which consist of an axial type Hall probe which has an active area of 0.1 mm ´ 0.1 mm and a thickness of 0.1 mm. The specimens were magnetized by direct current at a constant average magnetic flux density of 1.4 (T). This has been checked by using a search coil and a flux meter. The Hall probe was scanned across the center of flaws by taking small precise step using a computer controllable positioning motor. At each step, the horizontal (Bx) and vertical (By) components of magnetic leakage flux density in the upper spatial domain have been accurately measured. Then, the acquired data of Bx and By were stored into a personal computer, for the correction of background noise due to stray field arising immediately from the magnetizing yoke. In this experiment, the lift off value, which is defined as a distance from the upper-surface of test specimen to the central position of Hall probe, was set to be constant. This "lift off" is denoted by "y" mm in this paper.
Fig 1: Sketch of experimental set-up for
the far-side MFL inspection
III. Experimental results
Measurements of the magnetic flux density (Bx and By) at the lift off value y=1.0 mm in the upper spatial domain have been done by scanning with the Hall probe across the center of flaws. After the corrections of background noise, the distribution profiles of Bx and By vs sensor's position were obtained. Some examples of the results on the horizontal component of magnetic leakage field (Bx) are shown in Fig.2 and Fig.3 for various sizes of flaws.
Fig 2: Magnetic flux density distribution from circular holes of a 8mm depth (horizontal component)
Fig 3: Magnetic flux density distribution from rectangulare grooves of 50mm width(horizontal component)
Figure 2 shows the profiles of Bx for various opening diameters ranging 4 mm to 12 mm in the specimens having circular hole of 8 mm in depth. The geometrical feature of Bx distribution profile for far-side flaws of the relatively smaller width is such that the maximum value of Bx is taken at the center of flaw. On the other hand, in the case of the larger opening width of flaws, the distribution of Bx becomes to be that as shown in Fig.3, which is remarkably distinguishable from that shown in Fig.2. This suggests that the nature of Bx profiles depends on the ratio of flaw depth to its width.
In contrast to the Bx, the vertical component of magnetic field By, is the quantity of interest, since it gives important information for flaw size evaluation. The examples of measured vertical component By are shown in Fig.4 and Fig.5 for the 12mm thickness specimens containing various sizes of circular hole and rectangular groove, respectively. As can be seen from a sets of results, the profiles of By are symmetric with respect to the origin where is at the center of flaw, and a maximum and a minimum points of By curves appear at the vicinity of flaw edges. Fundamentally the amplitude of flaw signal, for instance DBy denoted in By profiles of Fig.4, may depend significantly on two parameters of flaw width and depth, when the lift off is fixed. Thus, we first made an attempt to estimate quantitatively opening diameter or width of flaws, regarding the interpeak distance denoted by the "DP" in the By profile of Fig.4 as an evaluation parameter. The plots of the distance between the two peaks (D
P) vs the width or diameter of flaw are shown in Fig.6 for different flaw depths. It is clear from Fig.6 that the dependence of DP on the real flaw width or diameter is approximately linear, and it is possible to evaluate the flaw width from only the DP value of By distributions for hole and rectangular types of flaws.
Fig 4: Magnetic flux density distribution
from circular holes of 8mm depth
Fig 5: Magnetic flux dcnsity distribution from
rectangular grooves of 50mm with(vertical component)
Figure 7 shows some examples of the dependence of the flaw signal amplitude DBy on the depth of flaws for different flaw widths. These plots suggest that the depth of far-side flaw or residual thickness of metal loss area can be evaluated for the case a constant lift off value and a constant specimen thickness, because the flaw width or the opening diameter would be estimated from Fig.6. From the approximately linear dependency of ΔBy on the flaw depth for various flaw width, it can be expected that the strength of ΔBy would be related to the two dimensional section area of flaws. This fact is confirmed in Fig.8 and it provides important information for flaw sizing procedure. Essentially, it is also considered that the flaw signal strength of DBy is more sensitive to the plate thickness . Thus, the experiments for the specimen with a different thickness of 22 mm have been performed by the same way as on the specimen of a thickness of 12 mm. Apart from the fact that the increase in the plate thickness gives rise to weaker signals, the experimental results on the 22mm thickness plate exhibited a similar tendency as that shown in Fig.7 and Fig.8.
Fig 6: Relationship between the real flaw width and the
Fig 7: Relationship between the real flaw depth and
IV. Analytical expressions for far-side magnetic leakage fields and its suitability for the application to flaw sizing
In the previous section, the behavior and the nature of far-side magnetic leakage field, especially of the signal strength D
By, with the flaw depth, width and plate thickness have been discussed based on the detailed measurements. This section is concerned with the model calculations using the mathematical expressions which have been already proposed for a two-dimensional far-side flaw by Zhang and Sekine ,, in order to confirm the evaluation scheme experimentally obtained for far-side flaws.
Fig 8: Relationship between the two dimensional
flaw section-area and DBy
Fig 9: Two dimensional far-side flaw model
As a practical and usable analytical model for magnetic leakage field arising from a far-side flaw, the article  deals with the equivalent double current model whose basic geometry is shown in Fig.9. In this model, the influence of the boundary surface between the mediums is taken into consideration by the mirror image method . According to the , the magnetic field at a point(x,y), H(Hx, Hy) arising from the far-side slot in the medium uniformly magnetized can be described as the following forms,
where the equivalent current moment I2=(m-1)(d+g)/(m(d+g)+(1-m)d)×H0gd (H0 : applied magnetic field, : permeability) and the mirror image factor due to the presence of boundary between the and the is m2=2+g2/6h(h+d) . The other geometrical parameters in the above equations are denoted in Fig.9.
Fig 10: Magnetic flux leakage distribution from the flaw of a 8mm depth|
(vertical component ,equivalent current model)
Fig 11: Magnetic flux leakage distribution from the flaw of a 50mm width
(vertical component ,equivalent current model)
Fig 12:Reationship Between the flaw depth and
Figure 10 and Fig.11 show examples of the profiles for the vertical component of magnetic field (arbitrary unit) Hy/Ho calculated using the expressions (2) for given different sizes of far-side slot. The calculation was carried out under the condition of a constant lift off. The results of model calculation are identical with those experimentally obtained for far-side circular holes and rectangular grooves. For the calculated profiles of magnetic field distributions, the peak-to-peak amplitudes Hy/H0 (signal strength) are corresponding to the By of real flaws in Fig.4 and Fig.5. These two figures of Fig.10 and Fig.11 suggest that the Hy/H0 and depth of flaw are uniquely related for a constant flaw width , if the lift off is fixed. The signal amplitudes By estimated using the theoretical expressions are plotted as a function of flaw depth for various flaw widths as shown in Fig.12 In this case, the conversion from Hy/Ho to real signal strength By and the correction have been made by using some experimental data. We can see that the agreement with Fig.7(experimental) and Fig.12 (theoretical) is fairly good. The conclusion from this theoretical approach is that the analytical expression forms of Eq.(1) and (2) provide a valuable aid in evaluating the far-side corrosive type flaws.
We have proposed a basic scheme of quantitative evaluation for far-side corrosive type flaws using the MFL technique. It has been shown that the combined use of theoretical model expressions and some experimental data for the far-side model flaws in the reference test specimen, provides an practical estimation procedure for backside metal loss area in steel plates, such as oil storage tank bottom floors.
- D. M. Amos: Magnetic flux leakage as applied to aboveground storage tank flat bottom tank
floor inspection, Materials Evaluation, 54(1996), p.26
- E. Altschuler and A. Pignott: Nonlinear model of flaw detection in steel pipes by magnetic flux leakage, NDT and E International, 28(1995), p.35
- S.Mukhopadhyay and G.P.Srivastava: Characherization of metal loss defects from magnetic flux leakage signals with discrete wavelet transform, NDT and E Internatonal,33(2000), p.57
- K.Sekine and A.Iizuka :An improved method of magnetic flux leakage inspection for far-side
corrosion type defects of ferromagnetic specimen, Proc.14th WCNDT, New Dehli, Ed. by C.G.Kirshnadas Nair et.al, (1996), p.1627
- C. Edwards and S. B. Palmer: The magnetic leakage field of surface breaking cracks, J. Phys. D, 19(1986), p.657
- W. Lord and D. J. Oswald: Magnetic leakage field method of defect detection, Int. J. NDT,4(1972), p.249
- Y. Zhang and K. Sekine: Magnetic leakage field due to sub-surface defects in ferromagnetic
specimens, NDT and E International, 2895), p.67
- Y. Zhang: Ph. D. Thesis "A fundamental study on magnetic flux leakage testing" Yokohama National Univ. (1994), p.176
- U. Patel and D. Rodger: Calculation of magnetic field around far-side defects for nondestructive testing, IEEE Trans. on Magnetics, 31(1995), p.2170
- V.E. Shcherbinin and M.L.Shur: Calculating the effect of the boundaries of a product on the
field of cylindrical defect, Defectoskopiya, NO.6(1976),p.30