See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/226760687

Detection of Carburization in Ethylene Pyrolysis Furnace Tubes by a C Core

Probe with Magnetization

Article  in  Journal of Nondestructive Evaluation · September 2010

DOI: 10.1007/s10921-010-0075-3

CITATIONS READS

18 884

5 authors, including:

Naoya Kasai
Yokohama National University
89 PUBLICATIONS   474 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Call for good quality NDT&E papers: Special Issue of Metals of MDPI (Impact Factor 2.117, Q1 Category of Metallurgy & Metallurgical Engineering, Web of Science) View
project

All content following this page was uploaded by Naoya Kasai on 07 May 2015.

The user has requested enhancement of the downloaded file.

*Manuscript
Click here to download Manuscript: Naoya Kasai ver.2.doc Click here to view linked References

1
2
3 Title
4
5
6
Detection of carburization in ethylene pyrolysis furnace tubes by a C core probe with magnetization
7
8
9
10
11
12
13 Authors
14
15
16 Naoya Kasai*a (Corresponding Author)
17
18
19 Graduate School of Environment and Information Sciences, Yokohama National University, 79-7
20
21
22 Tokiwadai, Hodogaya-ku, Yokohama 240-8501, JAPAN.
23
24
25
26 Tel:+81-45-339-3979 Fax:+81-45-339-4011
27
28
29 e-mail:[email protected]
30
31
32 Syujib Ogawa
33
34
35 Toshiyuki a Oikawa
36
37
38 Kazuyoshi Sekine a
39
40
41
Katsunobu Hasegawac
42
43
44
45
46
47
48 Affiliations
49
50
a
51 Graduate School of Environment and Information Sciences, Yokohama National University, 79-7
52
53
54 Tokiwadai, Hodogaya-ku, Yokohama 240-8501, JAPAN.
55
56
57 b
Graduate School of Engineering, Yokohama National University, 79-5 Tokiwadai, Hodogaya-ku,
58
59
60
61
62 1
63
64
65
 1
2
3 Yokohama 240-8501, JAPAN.
4
5
6 c
Idemitsu Engineering Co., Ltd, Makuhari Technogarden B 23F, 1-3 Mihama-ku, Chiba, Chiba 261-8501,
7
8
9
10 JAPAN
11
12
13
14
15
16 Abstract
17
18
19 A technique was developed for detection of the magnetic carburization layer generated inside a
20
21
22 ethylene pyrolysis furnace tube with magnetic oxide and nitride layers at the external surface. A C core
23
24
25
26 probe was used for detection of the carburized layer through a magnetic layer at the external surface of
27
28
29 the tube. Both preliminary experiments and finite element analysis for the measurement properties of
30
31
32 the probe were conducted using test pieces that simulated a tube with magnetic oxide and nitride layers at
33
34
35 the external surface. The results indicated successful detection of the carburized layer through the
36
37
38 magnetic layer, due to the magnetic circuit between the probe and the test piece. Furthermore, eddy
39
40
41
current measurements using the C core probe with magnetization to cancel the external magnetism were
42
43
44
45 conducted, because the cancellation effect of the magnetic oxide and nitride layers at the external surface
46
47
48 on the magnetic flux generated from the C core probe was still large. Impedance plots were measured
49
50
51 for used tubes with the C core probe and with magnetization, and the carbon and nitrogen concentrations
52
53
54 of the tubes were analyzed using infrared absorption spectroscopy. The impedance plots corresponded
55
56
57 with the analytical results for the carbon concentrations.
58
59
60
61
62 2
63
64
65
 1
2
3
4
5
6
Key Words: ethylene pyrolysis tube, carburized layer, eddy current method, C core probe, magnetization
7
8
9
10
11
12
13 1. Introduction
14
15
16 Ethylene is a key product in the petrochemical industry, and is produced by the thermal cracking of
17
18
19 complex hydrocarbons, such as naphtha. The surface temperature of the pyrolysis tube used for the
20
21
22 thermal cracking of hydrocarbons is approximately 1100 °C. Ethylene pyrolysis tubes are fabricated
23
24
25
26 using nickel-chromium alloy for high temperature ductility and carburization resistance. However,
27
28
29 during operation in petrochemical plants, carburization layers are still generated inside the tubes [1-7],
30
31
32 which causes a loss of ductility and introduces microcracking [8-11]. The generated carburized layer
33
34
35 becomes progressively more magnetic [11-16], although the tube is initially nonmagnetic. Therefore,
36
37
38 magnetic measurement as a nondestructive testing method was developed for the detection of
39
40
41
carburization in pyrolysis tubes to ensure safety in petrochemical plants [12,16]. In recent years, HP
42
43
44
45 type tubes with a nominal composition of 40%Fe, 35%Ni, and 25%Cr have been commonly used as a
46
47
48 requirement for increased carburization resistance. For the non-destructive inspection of HP type tubes,
49
50
51 the thickness of the carburized layer in the tubes cannot be accurately evaluated by conventional
52
53
54 techniques, because the magnetic oxide and nitride layers (with maximum thicknesses of several
55
56
57 millimeters) formed at the HP tube surface strongly affect the signals.
58
59
60
61
62 3
63
64
65
 1
2
3 between the poles was 50 mm. The excitation frequencies of both the coil and probe (10 to 800 Hz)
4
5
6
were selected on the basis of the skin depth effect, and the excitation voltage was 10 V. Transient
7
8
9
10 analysis with FEA was conducted for three periods of excitation current with an excitation frequency of
11
12
13 10 to 800 Hz. For the FEA, test plates (Fig. 1) were divided into an external magnetic surface, sound
14
15
16 and carburized layers in the direction of depth to simulate the conditions of the pyrolysis tubes. The
17
18
19 surface layer simulated both the oxide and nitride layers of a pyrolysis tube. A magnetic material was
20
21
22 used for the surface and carburized layers and a nonmagnetic material was used for the sound layer in the
23
24
25
26 analysis. The surface layer was 0.05 mm thick and the total plate thickness was 10 mm. The ratios
27
28
29 between the thicknesses of the carburized layers and the test pieces used were 0, 25 and 50%.
30
31
32 Fig. 2 shows the magnetic flux density amplitude obtained for the surface of a carburized layer
33
34
35 using the C core probe with an excitation frequency of 100 Hz at 29.5 ms. A large magnetic flux was
36
37
38 induced from the C core probe, and the flux induced by the probe flowed from pole to pole, deeply
39
40
41
penetrating toward the plate interior by formation of a magnetic circuit. The eddy current amplitude
42
43
44
45 with the C core probe on the carburized layer was predictably large compared to that with the pancake
46
47
48 coil (analytical results not shown), although the magnetic surface layer induced a large magnetic flux.
49
50
51 A pancake coil and C core probe with the same sizes as those used for the FEA were then used for
52
53
54 experimental measurements. The excitation and detection coils of the pancake coil had 700 and 1137
55
56
57 turns, respectively, whereas the excitation and detection coils of the C core probe both had 500 turns. A
58
59
60
61
62 5
63
64
65
 1
2
3 In the present study, a C core probe was considered for detection of the carburized layer through
4
5
6
the oxide and nitride layers at the tube surface. The probe properties were analyzed by finite element
7
8
9
10 analysis (FEA). The detection properties of the C core probe were experimentally verified.
11
12
13 Furthermore, because the cancellation effect of the magnetic oxide and nitride layers on the magnetic flux
14
15
16 generated from the C core probe was still large, eddy current measurements were conducted using the C
17
18
19 core probe with magnetization. The C and N concentrations were analyzed at each position of the
20
21
22 internal surface of ethylene pyrolysis tubes for comparison with the experimental results.
23
24
25
26
27
28
29 2. FEA and preliminary experiments to confirm the C core probe properties
30
31
32 A pancake coil and a C core probe were used to detect the carburized layers of pyrolysis tubes and
33
34
35 the properties of both were analyzed using FEA. The pancake coil was comprised of an excitation coil
36
37
38 and a detection coil with 100 turns, respectively, as shown in Fig. 1(a). The excitation and detection
39
40
41
coils of the pancake coil were 7.5 mm thick with internal and external diameters of 24 and 35 mm,
42
43
44
45 respectively. The C core probe consisted of a silicon steel core, two excitation coils, and two detection
46
47
48 coils, as shown in Fig. 1(b). The two excitation coils of the C core probe were connected in series to
49
50
51 obtain a large inductance and the two detection coils were similarly connected in series. The excitation
52
53
54 and detection coils of the C core probe had external and internal diameters of 35 and 25 mm, respectively,
55
56
57 and were 10 mm thick. The cross-sectional area of the C core was 15×15 mm2, and the distance
58
59
60
61
62 4
63
64
65
 1
2
3 magnetic surface layer (0.05 mm thick) was formed on silicon steel, medium carbon steel was used to
4
5
6
model the carburized layer, and austenitic stainless steel was used for the sound layer, as shown in Table 1.
7
8
9
10 The ratios between the thicknesses of the carburized layers and the test pieces were the same as those
11
12
13 used for the FEA model, and the results are shown in Fig. 3. The change of the signals in (y axis in
14
15
16 Fig.3) was calculated using the following equation.
17
18
19
20
21
Signal voltage
22 Signal change  1
23 Signal voltage measured with the test piece of the 0% medium carbon steel layer
24
25
26 (1)
27
28
29 The pancake coil was not sensitive to changes in the thickness of the carburized layers; however,
30
31
32 the C core probe measurements successfully detected the changes. The properties of the C core probe
33
34
35 for detection of the carburized layer were in good agreement with reference [15].
36
37
38
39
40
41
3. Magnetization experiment using the C core probe
42
43
44
45 Impedance analysis, which is typically used for the analysis of eddy currents, was employed to
46
47
48 measure the inductance and resistance of the excitation coil with an LCR meter, as shown in Fig. 4. A
49
50
51 total of 16 test pieces with different magnetic surface layer thicknesses from 0.3 to 2 mm were prepared.
52
53
54 Selected test pieces with a magnetic surface layer of 2 mm in thickness and without a magnetic surface
55
56
57 layer are shown in Table 2. The total thickness, length and width of the test pieces were approximately
58
59
60
61
62 6
63
64
65
 1
2
3 10, 200 and 150 mm, respectively. The surface and carburized layers were modeled using medium
4
5
6
carbon tool steels, and the sound layer was modeled using austenitic stainless steel. Characteristic
7
8
9
10 results measured for selected test pieces are shown in Fig. 5. When a magnetic surface layer was present
11
12
13 in the test piece, the measured inductance and resistance increased. In addition, increasing inductance
14
15
16 was measured with increasing thickness of the carburized layer. Fig. 5 also shows the large effect of
17
18
19 increasing carburized layer thickness for a test piece without a magnetic surface layer. However, the
20
21
22 effect was small for a 2 mm thick magnetic surface layer. The measurement results are dependent on
23
24
25
26 two parameters; the magnetic surface and the carburized layer. Therefore, the effect of the magnetic
27
28
29 surface layer must be reduced for accurate evaluation of the carburized layer. A DC magnetizer was
30
31
32 used to reduce the effect of the magnetic surface layer. The results were dependent on the carburized
33
34
35 layer thickness, as shown in Fig. 6; therefore, the detectability of the carburized layer was improved by
36
37
38 magnetizing the magnetic surface layer. Moreover, for the purpose of quantitative comparison, the
39
40
41
tangential magnetic field at the center of poles was also measured three times during the experiment, and
42
43
44
45 the results are shown in Table 2. The average tangential magnetic field was 1 to 13 mT, because of the
46
47
48 different magnetic circuits between the magnetizer and the test pieces.
49
50
51 10 ethylene pyrolysis tubes with similar thickness and diameter were obtained for experimental
52
53
54 detection of the carburized layer using the C core probe and magnetizer. These pyrolysis tubes had
55
56
57 undergone 5-10×104 h of use in petrochemical plants.
58
59
60
61
62 7
63
64
65
 1
2
3 Figure 7 shows the selected experimental results measured using the C core probe with
4
5
6
magnetization. The experimental tube sizes are shown in Table 3. During the measurement, the
7
8
9
10 magnetizer attachment was used so that the poles of the magnetizer contacted the tube surface. The
11
12
13 tangential magnetic field at the center of the poles was also measured three times during the experiment,
14
15
16 and the average tangential magnetic field was approximately 20 mT for all tubes, as shown in Table 3.
17
18
19 The inductances obtained for tubes I, II and IV were large, while that for tube III using the C core probe
20
21
22 with magnetization was small. The cancellation effect of the nitride layer on the magnetic flux
23
24
25
26 generated from the C core probe is considered to be larger than that of the oxide layer at the tube surface,
27
28
29 because the nitride layer is thicker than the oxide layer. Therefore, the carbon and nitrogen
30
31
32 concentrations of tubes I, II, III and IV were analyzed using infrared (IR) absorption spectroscopy, and
33
34
35 the results are shown in Fig. 8.
36
37
38 Although the nitrogen concentration of tube III was highest, the inductance was smallest, which is
39
40
41
good agreement with the carbon concentration and indicates the reduction effect of the surface
42
43
44
45 ferromagnetic layer. The large inductance of tube II indicates a large carburization concentration, and
46
47
48 tubes I and IV were estimated the similar carburization concentration because of the similar impedance
49
50
51 behavior as shown in Fig.8(a). From these experimental findings, the results measured with the C core
52
53
54 probe and magnetization corresponded well with the carbon concentration results. Thus, the proposed
55
56
57 eddy current method using a C core probe with magnetization was confirmed to be effective for
58
59
60
61
62 8
63
64
65
 1
2
3 measurement of the carburization layers in pyrolysis tubes.
4
5
6
7
8
9
10 4. Conclusion
11
12
13 The measurement properties of the C core probe for detection of carburized layers were confirmed to
14
15
16 be effective. A C core probe was used to measure test plates to simulate tubes with oxide, nitride, sound
17
18
19 and carburized layers. A magnetizer was also used to reduce the effect of the magnetic surface layer in
20
21
22 order to obtain more accurate results. Used pyrolysis tubes were measured using the C core probe with
23
24
25
26 magnetization and the results were in good agreement with the analytical results for carbon
27
28
29 concentrations. Improvement of the proposed method for the evaluation of carburized layers in
30
31
32 pyrolysis tubes is currently in progress.
33
34
35
36
37
38
39
40
41
References
42
43
44
45 1. C. M. Schillmoller. Alloys for ethylene cracking furnace tubes, Proc. 1st Int. Conf. on Heat-Resistant
46
47
48 Materials, 1991; pp. 469–76.
49
50
51 2. M. J. Bennett, J. B. Price. A physical and chemical examination of an ethylene steam cracker coke and
52
53
54 of the underlying pyrolysis tube, Journal of Materials Science, 1981; 16(1): 170-88.
55
56
57 3. A. Chauhan, M. Anwar, K. Montero, H. White, W. Si. Internal carburization and carbide precipitation
58
59
60
61
62 9
63
64
65
 1
2
3 in Fe-Ni-Cr alloy tubing retired from ethylene pyrolysis service、Journal of Phase Equilibria and
4
5
6
Diffusion, 2006; 27(6): 684-90.
7
8
9
10 4. S. Forseth, P. Kofstad. Carburization of Fe-Ni-Cr steels in CH4-H2 mixtures at 850-1000°C, Materials
11
12
13 and Corrosion, 1998; 49(4): 266–71.
14
15
16 5. A. Ul-Hamid, H. M. Tawancy, S. S. Al-Jaroudi, A. I. Mohammed, N. M. Abbas. Carburization of
17
18
19 Fe–Ni–Cr alloys at high temperatures, Materials Science-Poland, 2006; 24(2): 319-31.
20
21
22 6. A. A. Kaya , P. Krauklisb, D. J. Young. Microstructure of HK40 alloy after high temperature service in
23
24
25
26 oxidizing/carburizing environment: I. Oxidation phenomena and propagation of a crack, Materials
27
28
29 Characterization, 2002; 49(1): 11-21.
30
31
32 7. A. A. Kaya. Microstructure of HK40 alloy after high-temperature service in oxidizing/carburizing
33
34
35 environment: II. Carburization and carbide transformations, Materials Characterization, 2002; 49(1):
36
37
38 23-34.
39
40
41
8. H. M. Tawancy. Degradation of mechanical strength of pyrolysis furnace tubes by high-temperature
42
43
44
45 carburization in a petrochemical plant, Engineering Failure Analysis, 2009; 16(7): 2171-8.
46
47
48 9. X. Q. Wu, Y. S. Yang, Q. Zhanand, Z. Q. Hu. Structure Degradation of 25Cr35Ni Heat-Resistant Tube
49
50
51 Associated with Surface Coking and Internal Carburization, Journal of Materials Engineering and
52
53
54 Performance, 1998; 7(5): 667-72.
55
56
57 10. D. Jakobi, R. Gommans. Typical failures in pyrolysis coils for ethylene cracking, Materials and
58
59
60
61
62 10
63
64
65
 1
2
3 Corrosion, 2003; 54(11): 881–7.
4
5
6
11. Y. Saiga, A. Ohtomo. Damage due to carburization of austenitic alloys, Corrosion Engineering, 1977;
7
8
9
10 26(9): 515-26 (in Japanese).
11
12
13 12. I. C. da Silva, R. S. da Silva, J. M. A. Rebelloa, A. C. Brunob, T. F. Silveirac. Characterization of
14
15
16 carburization of HP steels by non destructive magnetic testing, NDT & E International, 2006; 39(7):
17
18
19 569-77.
20
21
22 13. I. Silva, J. Rebello, A. Bruno, P. Jacques, B. Nysten, J. Dille. Structural and magnetic characterization
23
24
25
26 of a carburized cast austenitic steel, Scripta Materialia, 2008; Vol. 59(9): 1010-13.
27
28
29 14. D. M. Stewart, K. J. Stevens, A. B. Kaiser. Magnetic Barkhausen noise analysis of residual stress and
30
31
32 carburization, Review of Progress in Quantitative Nondestructive Evaluation: Volume 20. AIP Conference
33
34
35 Proceedings, 2001;Vol. 557: pp. 1732-9
36
37
38 15. K. J. Stevens, A. J. Tack, C. W. Thomas, D. Stewart. Through-wall carburization detection in ethylene
39
40
41
pyrolysis tubes, Journal of Physics D: Applied Physics, 2001;Vol. 34: 814-22.
42
43
44
45 16. K. J. Stevens, W. J. Trompetter. Calibration of eddy current carburization measurements in ethylene
46
47
48 production tubes using ion beam analysis, Journal of Physics D: Applied Physics, 2004; Vol. 37: 501-9.
49
50
51
52
53
54
55
56
57
58
59
60
61
62 11
63
64
65
 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62 12
63
64
65
*Manuscript
Click here to download Manuscript: Figure captions.doc Click here to view linked References

1
2
3 Figure captions
4
5
6
Table 1 Thickness of each layer in the preliminary experiment test pieces.
7
8
9
10 Table 2 Test pieces used to simulate tubes with oxide, nitride, sound and carburized layers, including
11
12
13 the tangential magnetic field in the experiment.
14
15
16 Table 3 Various sized tubes measured, and the tangential magnetic field in the experiment.
17
18
19 Fig. 1. Analytical model used in the FEA.
20
21
22 Fig. 2. Magnetic flux distribution for the C core probe by FEA.
23
24
25
26 Fig. 3. Comparison of the pancake coil and C core probe properties.
27
28
29 Fig. 4. Experimental setup for measurement of the inductance and resistance of the excitation coil.
30
31
32 Fig. 5. Impedance diagram for the plate without magnetization. The measurement frequency was varied
33
34
35 from 40 to 8000 Hz.
36
37
38 Fig. 6. Impedance diagram for the plate with magnetization.
39
40
41
The measurement frequency was varied from 40 to 8000 Hz.
42
43
44
45 Fig. 7. Impedance plots measured with the used tubes. Only characteristic data are shown. The
46
47
48 measurement frequency was varied from 40 to 8000 Hz.
49
50
51 Fig. 8. Analysis of the carbon and nitrogen concentrations with respect to the distance from the internal
52
53
54 surface.
55
56
57
58
59
60
61
62 1
63
64
65
Figure
Click here to download Figure: Fig.1.ppt

Detection C core
Excitation coil Magnetic
coil surface layer

Excitation
coil
Detection
coil

Sound layer

Carburized layer

(a) Pancake coil (b) C core probe

Fig. 1. Analytical model used in the FEA.

Figure
Click here to download Figure: Fig.2.ppt

Ａmplitude of
C core magnetic flux density

High
Excitation
coil

Magnetic
surface layer Detection Low
coil

Sound layer
Carburized layer

Fig. 2. Magnetic flux distribution for the C core probe by FEA.

Figure
Click here to download Figure: Fig.3.ppt

0.12
Pancake coil
0.1 C core probe
Signal change (-)

0.08

0.06

0.04

0.02

0
0 10 20 30 40 50 60
The ratio between the thicknesses of
the carburized layer and the test piece (%)

Fig. 3. Comparison of the pancake coil and C core probe properties.

Figure
Click here to download Figure: Fig.4.ppt

Magnetizer LCR meter

Test piece C core probe

Fig. 4. Experimental setup for measurement of the inductance and

resistance of the excitation coil.
Figure
Click here to download Figure: Fig.5.ppt
Test piece
3 A
B
2.5 C
D
2 E
F
G
L /L 0

1.5 H

0.5

0
0 0.1 0.2 0.3 0.4 0.5 0.6
R /( L 0)
Fig. 5. Impedance diagram for the plate without magnetization. The
measurement frequency was varied from 40 to 8000 Hz.
Figure
Click here to download Figure: Fig.6.ppt

Test piece
3
A
B
2.5 C
D
2 E
F
L /L 0

G
1.5
H

0.5

0
0 0.1 0.2 0.3 0.4 0.5 0.6
R /( L 0)
Fig. 6. Impedance diagram for the plate with magnetization.
The measurement frequency was varied from 40 to 8000 Hz.
Figure
Click here to download Figure: Fig.7.ppt

Tube
3
I

2.5 II

III
2
IV
L /L 0

1.5

0.5

0
0 0.1 0.2 0.3 0.4
R /（ L 0)

Fig. 7. Impedance plots measured with the used tubes. Only characteristic data are shown.
The measurement frequency was varied from 40 to 8000 Hz.
Figure
Click here to download Figure: Fig.8.ppt

4 1.2
I I
Carbon concentration（wt.%)

Nitrogen concentration (wt. %)

II II
3 III III
IV 0.8 IV

0.4
1

0 0
0 2 4 6 8 0 2 4 6 8
Distance from the internal surface (mm)
Distance from the internal surface (mm)

(a) Carbon concentration (b) Nitrogen concentration

Fig. 8. Analysis of the carbon and nitrogen concentrations with respect to

the distance from the internal surface.
Table
Click here to download Table: Table1.ppt

Table 1 Thickness of each layer in the preliminary experiment test pieces.

The ratio between the

Thickness of austenitic
Thickness of silicon steel Thickness of medium
stainless steel thicknesses of the medium
(mm) carbon steel (mm)
(mm) carbon steel layer and the
test piece (%)
0.05 10 - 0
0.05 7.5 2.0 20
0.05 5.0 5.0 50
Table
Click here to download Table: Table2.ppt

Table 2 Test pieces used to simulate tubes with oxide, nitride, sound and carburized layers, including
the tangential magnetic field in the experiment.

Magnetic surface layer Sound layer Carburized layer Tangential magnetic field
(mm) (mm) (mm) (mT)

A - 10 0 13.4
4.6
B - 7 3

2.9
C - 5 5

D - 3 7 2.3
E 2 8 0 7.9
F 2 5 3 3.0
G 2 3 5 1.7
H 2 1 7 1.2
Table
Click here to download Table: Table3.ppt

Table 3 Various sized tubes measured, and the tangential magnetic field in the experiment.

Internal diameter Thickness Tangential magnetic field

Test piece
(mm) (mm) (mT)
I 63 7.6 20.9
II 65 6.7 20.3
III 61 7.6 24.4
IV 61 8.0 21.9

View publication stats