A

FAILURE ANALYSIS natural gas production field be-

gan operation in 2003. Natural

Failure of
gas produced from the well is
processed through different
stages. During gas processing, a mixture
of mono-ethylene glycol (MEG), salt, and
water is separated from the produced gas.

a Column
This separated mixture is subjected to
further processing to recover MEG be-
fore recycling it again. The MEG recov-
ery unit includes a column reboiler where

Reboiler
the MEG mixture is heated. The column
reboiler is designed to increase the effi-
ciency of the distillation process for the
liquid phase downstream of the distilla-
tion column.

in a MEG
The tube bundle of the column re-
boiler consists of 134 U-bend tubes fixed
in a horizontal position. The tubes were
made from SA 179 steel with 19.05 mm

Recovery Unit
(0.75 in) outer diameter, 2.11 mm (0.084
in) thickness, and 3.66 m length. The
diameter of the column reboiler is 500
mm (19.69 in). Table 1 shows the operat-
ing conditions.
A. EL-BATAHGY AND M. AMIN, After only seven months of operation,
Central Metallurgical R&D Institute, Cairo, Egypt some of the column reboiler tubes leaked
G. FATHY, Burullus Gas Co., Cairo, Egypt in the front tube sheet zone.

Inspection
On-site inspection of the failed tube
After only seven months of operation, a column bundle showed that several tubes had
leaked around the front tube sheet (Figure
reboiler in a mono-ethylene glycol recycling unit at 1[a]). A close-up investigation of the leak-
age zone showed abnormal clearance
between the leaking tubes and the tube
a natural gas production plant failed because of sheet on the backside of the tube sheet
(Figure 1[b]). Leakage was confined only
localized corrosion on the portion of the tubes in the to the tube sheet zone where perforation
of some tubes occurred.
tube sheet. Tis attack was caused by abnormal Dimension measurements indicated
that abnormal clearance found between
clearance between the tubes and tube sheet, due to failed tubes and the back side of the tube
sheet was related to a wider diameter of
tube sheet holes on the back side in com-
poor manufacturing quality, which allowed parison with the front side. Holes for
nonleaking tubes had a uniform and
deposits to collect and initiate crevice corrosion. proper diameter through the whole thick-

66 MATERIALS PERFORMANCE August 2008

 M AT E R I A L S S E L E C T I O N & D E S I G N

TABLE 1
Operating conditions of the column reboiler
Fluid allocation Shell side Tube side
Fluid circulated MEG Heating medium
Total fluid 5.032 13.262
entering (kg/h)
In Out In Out
Liquid (kg/h) 5.032 3.960 13.262 13.262
Vapor (kg/h — — 1.072 23.7 — — — —
mol. wt.)
Liquid Vapor Liquid Vapor Liquid Vapor Liquid Vapor
3
Density (kg/m ) 977 — 970 1.054 828.6 — 885 —
Viscosity (cP) 0.554 — 0.619 0.014 0.578 — 0.972 —
Temperature 135 (275) 156 (312.8) 275 (527) 200 (392)
[°C (°F)]
Operating 0.25 (35.7) 0.25 (35.7) 0.75 (107) 0.7 (100)
pressure
[MPa (psi)]
Velocity (m/s) 1.14
Design temp. 200 (392) 350 (662)
[°C (°F)]
Design pressure 0.6 (85.7) 0.6 (85.7) 1.1 (157) 1.0 (142.9)
[MPa (psi)]

ness of the tube sheet, thus preventing any FIGURE 1

clearance between tube and tube sheet.
Figure 2 shows enlarged and close-up
views of the outer surface of two of several
failed tubes. The most important factor
noticed was that localized attack was
confined only to the outer surface of the
tube part inserted in the tube sheet.
Except for the tube part inserted in the
tube sheet where leakage had occurred,
visual inspection indicated that the out-
side surface of the failed tubes generally (a) General view of column reboiler showing leaked zone at backside of tube sheet,
(b) close-up view of leaked zone showing abnormal clearance between failed tubes
was intact and free from indications of and tube sheet.
corrosion. Dye penetrant testing of failed
tubes showed no indications for surface
cracks, either around or away from the FIGURE 2
leakage zone.
Figures 3 and 4 show stereoscopic
photographs of the outer surface of the
failed zone of two of several failed tubes.
Stereoscopic investigation of both outer
and inner surfaces of attacked or leakage
zones confirmed that thinning was con-
fined to attacked zones. Another impor-
tant observation was the absence of any
cracks around leaked or attacked zones.
Enlarged (a) and close-up (b) views of outer surface of failed zone of two of several
It was confirmed that localized corrosion failed tubes. Note localized attack on outer surface of tube part inserted in tube
attack had started on the tube outer sheet.

August 2008 MATERIALS PERFORMANCE 67

M AT E R I A L S S E L E C T I O N & D E S I G N Failure of a Column Reboiler in a MEG Recovery Unit

FIGURE 3

Optical micrographs with different

magnifications of cross sections were
taken from perforated (leaked), partially
attacked, and non-attacked zones. One
important factor was the absence of
cracks around leaked or attacked zones.
No microstructure difference was ob-
served for leaked, attacked, and non-
attacked zones. The microstructure of all
zones was a normal ferrite-pearlite struc-
ture and is typical for SA 179 steel.
A survey of hardness measurements
indicated no remarkable difference in
Stereoscopic photographs of outer surface of failed zone of one of several failed hardness values of leaked, attacked, and
tubes showing no cracks around the localized attacked zone. non-attacked zones. Hardness tests on
several tubes showed hardness values of
FIGURE 4 about 158, 161, and 160 Vickers hardness
(HV) for leaked, attacked, and non-
attacked zones, respectively.

Cause of Failure
The premature failure of the column
reboiler was attributed to localized cor-
rosion attack confined to the tube part
inserted in the tube sheet. Localized at-
tack was encouraged by the existence of
the abnormal clearance between tube
and tube sheet. This clearance was cre-
ated by a wider diameter of tube sheet
holes on the backside of the tube sheet
than on the front side. These uneven
holes were the result of poor manufactur-
ing quality of the reboiler. This clearance
encouraged deposits on the outer surface
Stereoscopic photographs of the outer surface of the failed zone of one of several
leaked tubes showing no cracks around the leaked (perforated) zone. of tubes that in turn acted as a shield and
helped in creating a crevice as well as a
stagnant condition, leading to localized
surface and penetrated through thickness leakage or attacked zone. crevice corrosion attack on the tube outer
where some attacked zones had been Except for the localized attacked surface.
perforated (Figure 4) and others had not zones, the wall thickness of both failed Crevice corrosion is localized corro-
(Figure 3). and non-failed tubes experienced almost sion of a metal surface at, or immediately
Visual inspection indicated that the no reduction from the original wall thick- adjacent to, an area that is shielded from
inner surface of the leakage zones gener- ness of ~2.1 mm (0.83 in). full exposure to the environment because
ally was smooth, clean, and free from Table 2 shows results of chemical of close proximity between the metal
indications of corrosion, except for local- analysis together with the specified range surface and the surface of another metal
ized perforation, which had allowed for SA 179 steel. Chemical composition or material such as deposits.
leakage to occur. Thinning of the failed of the tubes was within the specified range The main factors that influence local-
tube wall was confined to the perforated for SA 179 steel. ized corrosion include crevice geometry,

68 MATERIALS PERFORMANCE August 2008

 M AT E R I A L S S E L E C T I O N & D E S I G N

TABLE 2
Chemical analysis of failed tubes, together with the nominal chemistry of SA 179 steel (wt%)
Element
Material C Si Mn P S Cr Mo Ni
Failed tube 0.119 0.159 0.49 0.003 0.0005 0.100 0.033 0.082

SA 179 steel 0.10 max. 0.50 max. 0.50 max. 20.0 - 23.0 8.0 - 10.0 58.0 min. 3.15 - 4.15 5.0
(nominal) max.

deposits on the metal surface, and fluid Barer, R.D. and B.F. Peters. Why Metals Fail.
velocity. 6th ed. New York, NY: Gordon and
Breach Science Publishers, 1991.
Need reprints of MP ads,
The tighter the crevice is, the greater articles, or covers?
Brooks, C.R. and A. Choudhury. Metallurgi-
the propensity for crevice corrosion. De- cal Failure Analysis. New York, NY:
position of salts on the outer surface of McGraw-Hill, 1993.
REPRINTS ARE A GREAT
the tube will help to establish concentra- Colangelo, V.J. and F.A. Heiser. Analysis of INVESTMENT!
tion cells and permit corrosive species to Metallurgical Failures. 2nd. ed. New York,
collect and concentrate under deposits. NY: Wiley, 1987. Professionally printed reprints and
Localized corrosion will be accelerated El-Batahgy. A. and B. Zaghloul. “Practical photocopied reprints of all MP
Solution for Repairing Absorption ads, articles,
when stagnant no-flow conditions exist.
Tower at a Nitric Acid Plant—Replac- and covers are
These provide the environments most ing welding with Mechanical Joining.” available for
suitable for build-up of concentration International Conference on Fracture
purchase.
cells that in turn increase the tendency for ICF 11. Turin Italy: 20-25 March 2005.
Reprints can
localized corrosion. Fontana, M.G. Corrosion Engineering. 3rd ed. be customized
New York, NY: McGraw-Hill, 1987.
with your
Solution to the Problem A. EL-BATAHGY is head of the Manufacturing
company’s
The problem was solved by manufac- Technology Dept., Central Metallurgical Research logo, addi-
turing a new reboiler with a perfect qual- and Development Institute, El-Felezzat St., El- tional product information, or the
ity, so that a tight contact between the Tabbin, PO Box 87, Helwan, Cairo, Egypt. He has 28 magazine cover—with no limits
tubes and tube sheet was confirmed along years of basic and applied research work in the on creativity!
field of arc, resistance, and laser beam welding of
the whole thickness of the tube sheet.
ferrous and nonferrous alloys, and 25 years of
Absence of clearance between the tube experience in failure analysis and troubleshooting
and tube sheet will minimize the possibil- for various industry sectors, including petroleum,
Order your reprints today; it simply makes good
ity of stagnant and crevice condition, and petrochemicals, fertilizers, and power generation.
business sense! For reprint information and
in turn increase the lifetime of the re- He received a distinguished member award from rates, call +1 281-228-6219.
boiler. “Marquis Who’s Who in the World” (1999) and
“Marquis Who’s Who in Science and Engineering
for Outstanding Achievement” (2000). He received
Conclusions an International Welding Engineer Diploma in 2002.
An important question that should be
considered is, “In the long run, which will M. AMIN is a researcher at the Welding Technology
cost more, perfect manufacturing quality & NDT Lab at the Central Metallurgical Research
or less-than-perfect quality with increas- and Development Institute. He has 20 years of
experience in the field of welding consumables,
ing possibility for longer-term shutdown?”
repair welding, hard facing, solid state welding,
In this case, poor manufacturing quality and resistance welding of ferrous and nonferrous
created clearances between the tubes and alloys. He received an International Welding
tube sheets. These clearances led to crev- Engineer Diploma in 2002.
ice corrosion of the tubes and subsequent
tube failure. G. FATHY is an environment assistant general
manager at Burullus Gas Co., 294 St., New Maadi,
Cairo, Egypt. He has 25 years of experience in the
Bibliography field of corrosion control and chemical treatment.
ASM Handbook, Failure Analysis and Preven-
tion. Vol. 11. Materials Park, OH: ASM
International, 1996.

August 2008 MATERIALS PERFORMANCE 69