Carburization detection

Russell NDE Systems is pleased to announce the development of a new Tool for the detection of
Carburization, typically found in refinery and gas plant furnace tubes.

Definition:
Carburization is the absorption of carbon atoms into the granular structure of steel tubes used
inside furnaces. Carburization makes the steel tubes harder, less ductile, and more prone to
brittle failure.
Carburization can permeate through the steel tubes from either the inside or outside surface and
because it is temperature sensitive it is usually worst on the hot side of the tubes.
The new probe can be scanned around the tube. It is low profile, and requires only 1.5” of
clearance between the tube and the refractory liner. Scale on the tube should be removed if it is
loose, but it has little effect on the technique so long as the scan is smooth.

The probe, known as CIRC

E-PIT, is light and portable
and can be secured to the
tube with a simple Velcro
strap.

The internal coils send a

signal into the tube surface
which reflects from the
inside surface and is
affected by the amount of
carburization within the
wall.

The technique uses a non-

contact eddy current, pitch-
catch arrangement of coils
at frequencies that are
suitable to the application.

Typical furnace tube

materials such as 5-chrome
can be inspected at rapid
speed, with minimal
training.

Fig-1: CIRC E-PIT probe on furnace tube

1
 In the tube pictured at left,
the carburization has
permeated from the inside of
the tube which was in a
“coker” service. The lighter
area is carburized.

The coke build-up on the

inside of the tube, on the hot
side, further accelerates the
permeation by providing a
ready source of carbon, and
making the tube locally
hotter.

Fig-2: 5-Chrome tube with internal carburization (light area)

Fig-3: 5-Chrome tube with external carburization (light area)

The tube pictured above has been in a gas-fired furnace where the carburization (light area) has
permeated the tube from the outside-in. In this case, the carbon atmosphere was created by the
fuel that was firing the furnace. This tube shows carburization that has penetrated 100% through
the wall on the hot side (centre of the photo), decreasing to 50% or less on the cold side.

The CIRC E-PIT probe connects to the Ferroscope 308 ET instrument and can operate in eddy
current or RFT mode. The measurements taken by the probe as it is rotated around the pipe are
similar to those obtained by the ultrasonic TOFD (time of flight diffraction) technique (which in
our experience, works well for detecting the internal carburization case, but not as well for
external carburization detection).

2
Fig-4: Ferroscope and lap-top computer used with CIRC E-PIT probe

The latest Ferroscope 308ET

model pictured above with
internal heat exchanger probes,
and a lap-top running EasyLog
Pro software.

The image on the left is a

TOFD circ. scan of an
internally carburized tube. The
depth of carburization can be
accurately measured with
TOFD.

The data from the Ferroscope

technique looks similar to a
TOFD scan for ID
carburization (see next page).

Fig-5: Time of flight diffraction (TOFD circ. scan signal) for internal carburization

3
Fig-6: CIRC E-PIT scan of tube with 100% external carburization on hot side and 50% on cold
side.

Notes:
1) A scan of a tube with internal carburization would be similar to the above.
2) Frequencies between 10 and 80 Hz can be used, depending on tube properties
3) The amplitude of the signal, or the Y-component can be used to image the damage
4) The values of amplitude or “Y” can be used to calibrate the data

Research and development

During the months of December 2017 and January 2018, finite element modeling of the
electromagnetic signal from tubes that had internal and external carburization were performed.
The FEA models were supplemented with physical measurements made by the CIRC E-PIT
probe, with good agreement; however, the number of samples available was small. We had:

1) Several 4” diameter x 0.237” wall thickness tube samples from a “coker furnace” with
varying amounts of internal carburization, starting at 0% and reaching 60% maximum
2) One sample from a 4” diameter x 0.237” wall thickness from a CCR (Catalytic Cracking
Regenerator), having external carburization to a depth of 50% for most of its
circumference and 100% through wall on the hot side.
3) The photos of Nital etched tubes from these samples are shown on page-2
4) Several arrangements of coils, frequencies, drive voltages, scan speeds and lift-off
distances were trialled empirically.

4
FEA Results
The following graphs show the expected signals from samples that have carburization at two
depths: 25% and 50%. We have modelled the carburization as a local area that is scanned past
with a detector coil. Perturbations in the EM field are modelled, and the carburized layer is
modelled as “non-ferromagnetic”, “weak ferromagnetic” and “strong ferromagnetic”. Until we
have a much larger sample, we are not sure which model is best; however, the “weak
ferromagnetic” matched the field data quite well.

Fig-7: Magnetic field pattern for 50% ID carburization modeled as non-ferromagnetic

5
Fig-8: Magnetic field pattern for 50% OD carburization modeled as non-ferromagnetic

6
Fig-9: Magnetic field pattern for 50% ID carburization modeled as weak ferromagnetic

7
Fig-10: Magnetic field pattern for 50% OD carburization modeled as weak ferromagnetic

8
Fig-11: Magnetic field pattern for 50% ID carburization modeled as strong ferromagnetic

9
Fig-12: Magnetic field pattern for 50% OD carburization modeled as strong ferromagnetic

10
Fig-13: ID and OD carburization modeled as non-ferromagnetic material
11
Fig-14: ID and OD carburization modeled as ferromagnetic material weaker than base tube
12
Fig-15: ID and OD carburization modeled as ferromagnetic material stronger than base tube

13
Fig-16: Trace angle for ID and OD carburization modeled as non-ferromagnetic

14
Fig-17: Trace angle and signal strength for OD carburization modeled as non-ferromagnetic

15
Fig-18: Trace angle for ID and OD carburization modeled as weak ferromagnetic

16
Fig-19: Trace angle and signal strength for OD carburization modeled as weak ferromagnetic

17
Fig-20: Trace angle for ID and OD carburization modeled as strong ferromagnetic

18
Fig-21: Trace angle and signal strength for OD carburization modeled as strong ferromagnetic

19
 MAGNITUDE VALUES OF THREE CARBURIZED SAMPLES
1,900

1,800

1756.6 1760.6
1748.3
MAGNITUDE SIGNAL VALUE

1739.6 1734.9

1,7001679.7 1686.3
1675.8 1679.7
1672.6 1672.1

1643.2

1,600

1,500
1 2 3 4 5 6
DIFFERENT SPOTS ALONG PIPE

S1 M min S1 M max

Fig-22: Magnitude values taken on internal carburized sample#1 at 300mm intervals

Min=Cold side minimum value, Max=Hot side maximum value, arbitrary amplitude units

20
 MAGNITUDE VALUES OF THREE CARBURIZED SAMPLES

1,900

1812.3

1,800
1771.3
1762.3
1754.6
MAGNITUDE SIGNAL VALUE

1,700
1669.9
1657.9 1658.1 1660
1643

1,600

1527.4

1,500
1 2 3 4 5
DIFFERENT SPOTS ALONG PIPE

S2 M min S2 M max

Fig-23: Magnitude values taken on internal carburized sample#2 at 300mm intervals

Min=Cold side minimum value, Max=Hot side maximum value, arbitrary amplitude units

21
 MAGNITUDE VALUES OF THREE CARBURIZED SAMPLES

1,900

1,800
MAGNITUDE SIGNAL VALUE

1738.6

1706.6 1703.1 1700.3

1,700

1619.7 1622.7

1593.6
1,600

1556.2

1,500
1 2 3 4
DIFFERENT SPOTS ALONG PIPE

S3 M min S3 M max

Fig-24: Magnitude values taken on internal carburized sample#3 at 300mm intervals

Min=Cold side minimum value, Max=Hot side maximum value, arbitrary amplitude units

22
 MAGNITUDE VALUES OF TWO CARBURIZED SAMPLES
2,000 1964

1918

1,900 1872
MAGNITUDE SIGNAL VALUE

1,800
1738

1,700

1589
1,600

1,500
1440

1,400
0% 50% 100%
ID Carb 1872 1918 1964
OD Carb 1738 1589 1440
% CARBURIZED

ID Carb OD Carb

Fig-25: Values measured from the “cold (min) and hot (max) sides” of three tube samples which had 0% carburization on the
cold side and carburization up to 50% deep on the hot side. This illustrates the variations that are due to material properties of the
tubes (i.e. the cold side readings should all represent 0% carburization, so the spread in EM values is due to other properties of
the tube)

23
 Magnitude-min/max, from three tubes with "no" carburization
1850
1812.3

1800
1771.3
1762.3 1756.6 1760.6
1754.6
Raw Magnitude Values

1748.3
1750 1738.6 1739.6 1734.9

1706.6 1703.1 1700.3

1700 1686.3
1679.7 1675.8 1679.7
1669.9 1672.6 1672.1
1657.9 1658.1 1660
1643 1643.2
1650
1619.7 1622.7

1593.6
1600

1556.2
1550
1527.4

1500
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
M raw min 1812.3 1771.3 1762.3 1754.6 1669.9 1706.6 1738.6 1703.1 1700.3 1679.7 1739.6 1734.9 1748.3 1756.6 1760.6
M raw max 1657.9 1643 1658.1 1660 1527.4 1556.2 1619.7 1622.7 1593.6 1643.2 1686.3 1675.8 1672.6 1679.7 1672.1
Axis Title

M raw min M raw max Linear (M raw min) Linear (M raw max)

Fig-26: values of Magnitude (arbitrary) from a sample with ID carburization that varied from 0% on the cold side to 50% on the
hot side, and one with OD carburization from 50% (cold side) to 100% (hot side). Note 0% for OD carb and 100% for ID carb are
calculated, not actually measured.

24
Conclusions from Empirical tests
1) On all three sample tubes with internal carburization, there are clear differences in signal magnitude values between hot and cold sides
2) Cold side values on all samples range from 1669 to 1812
3) Hot side values on all samples range from 1527 to 1686.3
4) Field testing should include a “young tube” with a very short time in the furnace. This should be measured and the lowest magnitude value
recorded as 0% carburized.
5) A field operator should use two other sample tubes from the same furnace (or at least the same tube size, wall thickness and material, with
different operating hours and known amounts of carburization. From these known values a calibration curve can be constructed.
6) Note: because the tube wall is subject to small material property and thickness variations, the accuracy of carburization depth prediction will be
affected. In most electromagnetic measurements, the accuracy expected is +/- 15%, for these reasons.
The data presented in this report in confidential; is authored by D.E. Russell, and belongs to Russell NDE Systems. All rights are reserved to this material and it shall not be published or copied without the author’s express written consent.

25