ISSN: 1829-9466 ©2008 Journal of the Indonesian Oil and Gas Community.

Published by “Komunitas Migas Indonesia”

THICK-WALLED PRESSURE VESSEL FAILURE AND ITS

ON-STREAM ASSESSMENT USING HIGH
TEMPERATURE CONVENTIONAL ULTRASONIC AND
TOFD METHODS

Farabirazy Albiruni1,*, Wildan Hamdani2, Joon-Hyun Lee1

1
Quality Engineering and Failure Prevention Laboratory/Basic Atomic Energy Research Institute,
Graduate School of Mechanical Engineering, Pusan National University
San 30 Jangjeon-dong, Gumjeong-Gu, Busan, 609-735, South Korea.
2
Technical Inspection Department, PT. Pupuk Kalimantan Timur
Jl. Ir. James Simanjuntak No.1, Bontang, East Kalimantan, Indonesia.
*
Corresponding author: farabirazy@pusan.ac.kr

Abstract - A failure analysis of thick-walled pressure OD outside diameter

vessel which suffered hydrogen attack in two years
since its initial operating time was presented. It is
constructed from 2.25 Cr–1 Mo steel which is a popular
material for pressure vessel in petrochemical industry.
The analysis showed that the failure on closing seam
weld joint was due to local improper post weld heat
treatment which can introduce high residual stresses
particularly in thick-walled pressure vessel. This high
level of residual stress created inside weld joint is very
prone to hydrogen attack for any components which are
operating in hydrogen gas environment. The repair
procedures based on the principle to decrease the
residual stress for this thick-walled pressure vessel then
proposed. The repair procedures were controlled very
carefully by applying several nondestructive tests in the
each stage of repair. To assure the successful of the
proposed repair procedures, after one year since repair
time, high temperature conventional ultrasonic and
ultrasonic time of flight diffraction methods were
applied on-stream to this thick-walled pressure vessel in
order to evaluate its post repair condition. The two
methods showed good results on the repaired area.
Keywords: Thick-walled pressure vessel, 2.25 Cr-1Mo
steel weldment, Hydrogen attack, High temperature
conventional ultrasonic, Ultrasonic ToFD
Nomenclature
BHN brinell hardness number
CW closing seam weld
BM base metal
HA hydrogen attack
HAZ heat-affected zone
HTCUT high temperature conventional
ultrasonic testing
ID inside diameter
PWHT post weld heat treatment
ToFD time of flight diffraction
TWPV thick-walled pressure vessel
VHN vickers hardness number
WM weld metal
1. Introduction
1.1. Overview
The most generic types of process equipments
used in petrochemical plant are large TWPV to face the
harsh environment such as high pressure, high
temperature, and interaction with hydrogen gas as its
typical process characteristic. They are usually
constructed from low alloy steel materials such as
2.25Cr–1.0Mo steel and fabricated using complex
manufacturing stages including fusion welding process.
During fusion welding process, residual stresses are
induced into the material particularly around the WM.
The non uniform and highly localized heat input from
welding process, subsequent cooling after welding, and
also the non linearity of the material properties are
factors that contribute to the creation of residual
stresses. These residual stresses may lead to cracking
just after welding and sometimes later, during the
service time of TWPV.
Although residual stresses can be reduced
through PWHT, they are not completely removed. For
practical reason, PWHT of TWPV weld joint is
performing locally using heating elements attached on
OD shell of the TWPV instead of using large furnace.
This residual stress level, in case of TWPV, can be
introduced into material near to the ID surface of the
shell after PWHT as high as the yield stress because the
temperature gradient occurs across the wall during
PWHT. Heuser [1] modelled this phenomena and
showed that residual stress as high as yield magnitude order to give a clear picture about these successful
residual stress can be formed as a result. The existence methods.
of high residual stresses in weld has became a primary
concern in many welded structures since it increases the 1. 2. Chronology of Failure
crack driving force and reduces resistance to brittle The failure of TWPV occurred after two years
failure. The cracking tendency also increases for since its initial operation. The failure first was
TWPVs which are used in hydrogen bearing streams at recognized by the operator as leaky indication found by
high temperature and high pressure environment. regular explosive gas level measurement that reached
Several cracking problems caused by interaction 100% explosive level at top portion of this TWPV. It is
with hydrogen gas in several industries such as really very fortune that there was no significant accident
petrochemical and oil and gas were outlined by Prescott happened concerning the explosive nature of hydrogen
[2] and Timmins [3]. The hydrogen induced cracking gas. Verification of this leak indication using infrared
susceptibility is usually related to hardness. In the oil camera found a hot spot indication around the south side
industry, cracking in sulfide containing streams is very thermowell nozzle as shown in Fig. 1. The plant then
aggressive on WMs or HAZs at hardnesses above 225 was shutting down and isolated for further examination
BHN. The maximum recommended hardness of and reducing the failure risk. The characteristic of this
pressure equipment operating in high and low failure TWPV is listed in Table 1 below.
temperature hydrogen service should be 250 VHN (230-
240 BHN) according to NACE standard MR-0175 [4] Table 1.
and the limit set in API 942 [5]. Thus, it is also normally Characteristics of the TWPV
recommended that all high temperature vessels
operating in hydrogen environment should have Design code A. D. Merkblatter
maximum hardness of WMs of 250 VHN . Fluid contained Synthesis Gas
The influence of PWHT on crack propagation of Design temperature 370 oC
2.25Cr–1Mo WMs in hydrogen environment has been Operating temperature 270 oC
studied by Tsay [6]. The study showed that PWHT gave Design pressure 155 kg/cm2 G
marked improvement in impact toughness even though Operating pressure 132 kg/cm2 G
change in microstructure between as welded and Shell material SA 542 B Cl. 4
tempered weld was limited. The work also indicated Thickness of shell 105.0 mm
that the tempered bainitic microstructure presents in Corrosion allowance 1.70 mm
WM or HAZ was more resistance to unstable crack Composition of synthetic gas H2 = 71.41 %wt.
growth than the annealed BM tested in gaseous
hydrogen. In contrast, another study conducted by
Fujibayashi [7] found that WM of 2.25 Cr–1 Mo steel
has low creep strength due to full bainitic microstructure 111.8° C
in which carbides (M2C) evolution can be accelerated.
HA is a degradation process occurring in steels 100

in hydrogen-rich environments at elevated temperatures

(> 200 oC). This phenomena is regarded as a complex 80

mechanism which involves chemical aspect in 60

competition with mechanical phenomena, are also
playing a major role. Hydrogen atoms can easily diffuse 40
into steel, where at discontinuities, they react with the
carbon in steel to form cavities filled by methane. 20

Review of the mechanism which is involved in HA can

be found in Timmins [3] and Louthan [8].
HA is a notorious failure mode in the
petrochemical industry in TWPV for hydrocracking,
hydrotreating, and another refinery application. Most of
the TWPV are often manufactured by circumferential
welding of forged belts made of 2.25Cr–1 Mo steel.
Since WMs of this bainitic steel possess various
microstructures that may be more susceptible to HA,
WMs are often deduced as a critical location for
hydrogen attack.
The failure analysis of CW joint of TWPV which
operated in hydrogen gas environment is presented in
this paper. The procedure of repair and its on-stream
assessment using HTCUT and ToFD are also outlined in
6.7° C
Fig. 1. Infrared camera display on leak indication
(white circle)
2. Failure Analysis Procedure and Results
2. 1. Procedure
The failure analysis techniques employed consist
of visual observation, positive material identification
using x-ray fluorescence (XRF) method, dye penetrant
testing, hardness testing, and conventional ultrasonic
testing. All these techniques performed in field only
from OD surface when the TWPV was shutted down
after leak found. Before data were taken, the CW joint hydrogen service. The hardness measurement was
at OD surface was grinded around 0.5 to 1.00 mm to performed very carefully on shell’s vessel because the
remove surface corrosion product that can affect the repair procedure is made based on hardness value. Table
observation’s results. Beside those above, there were 3 lists the average BHN in different zones of the
also performed comprehensive review concerning the pressure vessel on its OD surface.
TWPV fabrication’s record since the fabrication
preparation until its installation and its two years Table 2
operating condition history such as operating Chemical composition analysis of the shell material
temperature and pressure record. (wt %)
The visual observation performed together with
penetrant testing through all the CW joints where the Sample Mn Cr Mo Cu Ni V
leak occurred. These two methods were intended to BM 0.35 2.27 1.02 0.10 0.11 0.01
characterize and found all surface flaw indications. WM 0.33 2.24 0.99 0.08 0.12 0.01
Positive material identification using portable
XRF analyzer was performed for quick chemical Table 3
analysis of shell’s material. This method applied both Average hardness number of different zones of the
for BM and WM. vessel
Hardness measurement using brinell portable
hardness tester was applied for BM, HAZ and WM. Zone BM WM HAZ
There were 12 testing points for 360 degrees CW where
BHN 140 290 240
for each point three reading were taken and the average
found.
Conventional ultrasonic testing was performed Conventional ultrasonic testing found one linear
using longitudinal wave normal beam and shear wave indication around the mid-wall of the shell thickness as
angle beam transducers from three orientations, two shown in Table 4 below. This linear indication
orientation sides that perpendicular to the WM and one amplitude reached 30 % height of standard distance
orientation along the WM. For shear wave, two angles amplitude correction (DAC) curve using basic
were used, 45 and 60 degree angle beam transducers. calibration block with same material and thickness with
Ultrasonic testing was intended to find volumetric flaw the vessel’s shell. Linear indication sizing performed
indications that exist in vessel shell’s weld joint. using 6 dB drop technique.

2. 2. Results Table 4
Visual observation found a leaky transverse Linear indication characteristic found by ultrasonic test
crack with 40 mm in length on CW 2 as shown in Fig.
2 below. The crack position was 100 mm in distance Length Depth from
Position Height (mm)
(mm) surface (mm)
after south side thermowell nozzle (180o orientation).
This result agreed with the result from infrared camera 30o 30 5 43
observation and explosive gas level measurement that
showed the area around the south side thermowell The manufacturing records indicated that the CW
nozzle suffered leakage. By looking the crack 2 was subjected to local PWHT using heating elements
characteristic, it is obvious that hoop stress was the around the OD of the vessel. The local PWHT was
major contribution for crack propagation. For TWPV performed for 10 hours at 660oC.
which is applied internal pressure, the maximum hoop
stress occurs on ID surface. Then it can be deduced that
crack start from ID surface of CW joint. Maximum
hoop stress, high temperature and direct contact with
hydrogen gas are the major factors for crack initiation
and propagation.
Liquid penetrant testing found another crack
indication on HAZ of CW2 at 100o orientation with 20
mm in length, as shown in Fig. 3 below. This crack had
irregular characteristic and completely different with the
leakage one. This indication disappeared after grinding
up to 4 mm depth from surface.
Positive material identification using XRF
method shows that the WM and BM are made from
2.25Cr–1Mo steel. This result agreed with the vessel’s
material specification and it is shown in Table 2.
As outlined in introduction above, hardness is the
primary indication to assess the WM condition for
CW-2

N
(0o)

100 mm

(180o)
S

Fig. 2. A leaky transverse crack position on CW 2

 surface of the vessel. The CW 2 temperature then
increased into 400 o C and held in this temperature for 5
hours and finally decreased into 200 o C. In this
temperature, leak’s gouging was conducted and to
assure there was no crack left during leak’s gouging,
high temperature penetrant testing was applied. The
result of leak’s gouging in CW 2 is shown in Fig. 5.

Surface Crack

60 mm
Height of the Hole
After Leak’s
Gouging
Fig. 3. Another surface crack found by penetrant testing

3. Repair Procedures and Results

Fig. 5. Leak’s gouging result
3.1. Procedures

The repair procedure flow chart is shown in Fig.

The 1st preheat was performed by increasing the
4 below.
temperature of CW 2 into 225 oC ± 10 oC. The 1st stage
welding was conducted using SMAW and fill the leak’s
Hydrogen 1st NDT Gouging the gouging result up to 50 mm height of weld deposit (half
removal leak
the shell thickness). After that PWHT was applied to
this repair zone by increasing temperature into 450 oC ±
10 oC and held for 4 hours, then into 720 oC ± 10 oC and
held for 3 hours, finally the temperature decreased into
200 oC. In order to attain the uniform temperature
distribution on repair zone, 4 thermocouples were
1st PWHT 1st stage 1st stage installed on it as shown in Fig. 6.
welding preheat

Heating 2 Incoloy
2nd NDT Final stage Final stage elements Plate
preheat welding 3 Monitoring
4 Thermocouple
Controlling
Thermocouple

50 mm
Weld
Final NDT Final PWHT 3rd NDT

Fig. 4. Flow chart of repair procedure Fig. 6. Local PWHT process

3. 2. Results
Hydrogen removal is intended to remove
absorbed hydrogen atoms in CW 2. This is done by
using heating elements which were attached from OD
The final preheat was performed by increasing
the temperature of CW 2 into 250 oC ± 10 oC. The final
stage welding was performed up to full thickness of the
shell. Then, the final PWHT was applied to this repair
zone by increasing temperature into 740 oC ± 10 oC (for 4. On-Stream Assessment
WM) and 650 oC ± 10 oC (for BM) for 4 hours, and
finally the temperature decreased into 200 oC.
Non destructive testing such as hardness testing, HTCUT performed after one year since the
high temperature penetrant testing, magnetic particle repair time. This testing conducted on-stream to vessel
testing, conventional ultrasonic testing, and TOFD were with outer surface temperature reach 240 oC. The gain
applied during and after weld repair. setting for this testing was 80-85 dB. In this gain range,
Final hardness test result is presented in Table 5 this testing is not suitable for detection of defects that
while the non destructive test results in Table 6. Fig. 7 may occur within the outer 0-25 mm range due to high
shows the shell repair after weld up to 50 mm and its level noise during testing. The accuracy of the defect’s
ToFD result respectively. depth measurement is likely within ± 10% accurate
meanwhile the height of defect is greater than actual.
Table 5 High temperature ultrasonic ToFD was also
Average hardness number of different zones of the performed in order to solve the weaknesses of HTCUT
vessel after repair for estimating the defect’s height. Same as conventional
one, high temperature ToFD is also not suitable for
Zone BM WM HAZ detection of subsurface defect and give high level noise
up to 20 mm due to gain adjustment as shown in Fig. 9
BHN 140 170 175 below.
HTCUT and high temperature ToFD results
Table 6 shown that all indications found were acceptable. It
Non destructive result during vessel repair means that CW2 in good condition. The results are
shown in Fig. 6 and 7 respectively. The indication found
Stage PT MT UT ToFD from HTCUT is summarized in Table 7.
1st NDT --- --- Acc ---
2nd NDT Acc Acc --- Acc
3rd NDT Acc --- Acc Acc

(a) Fig. 8. Indications found by high temperature

conventional ultrasonic

Outer surface

Depression

Indication
Inner surface

(b)

Fig. 7 (a). Shell repair (b) ToFD result after 1st Fig. 9. Indications found by high temperature ToFD
PWHT stage
Table 7
Detail of indications found by HTCUT
Height
Indication no. Depth (mm)
(mm)
#1 50 2 6. Conclusion
#2 60 4
#3 75 10
#4 24 3 The failure of this TWPV is caused by HA. WM
#5 ID Surface <3
#6 ID Surface <6
5. Discussion
The TWPV had failed after 2 years in operation.
The failure occurred was a leaky transverse crack on
CW2 which had applied local PWHT during vessel
fabrication. The similar failure with short operating time
after put in service also happen in petrochemical related
facilities such as in heat exchanger. The repair had also
been done using temper bead welding techniques
combined with local PWHT such as outlined by Firt
[10].
Hardness test from OD surface shown that CW2
around the leakage zone have hardness value above the
maximum value after PWHT, i.e. more than 250 VHN
according to NACE standard MR0175 and API 942 [4,
5]. This is most caused by improper local PWHT during
fabrication, such as:
1. Local PWHT using heating elements only applied
from one shell surface (usually from OD surface).
2. Lack of thermocouple amount to control uniformity
of temperature during PWHT.
3. Insufficient temperature and or insufficient holding
time during PWHT.
The synthesis gas contained in this TWPV is rich
with hydrogen gas with operating temperature around
300 oC. With improper local PWHT applied, not all
bainite structure transformed into tempered bainite and
residual stress reduce significantly as shown by
hardness results. This makes the vessel still has high
sensitivity to high temperature HA and failure occurred
in two years operating time of this vessel.
During repair, re-PWHT had applied on CW2 in
order to decrease the sensitivity to high temperature HA
by transformed all bainite structure into tempered
bainite and also decrease the residual stress. For leakage
zone, two times PWHT had applied, those were after
weld up to 50 mm and after weld into full thickness 110
mm. The PWHT temperature 740 oC ± 10 oC was higher
than ASME recommendation on 677 oC [9]. This is
because from simulation using carbon steel plate 100
mm thickness found the different temperature of 100
o
C–120 oC between two contact and non contact
surfaces with heating elements. Because the repair only
attach the heating elements from OD surface of shell,
higher temperature with two stages PWHT was intended
to reach the ID surface of shell because this surface has
direct contact with synthesis gas and decrease its
sensitivity to high temperature HA.
is critical location due to various microstructures and
high level of residual stresses that created during
welding. The tendency for failure caused by HA can be
decreased with PWHT if it is done properly. Proper
PWHT can decrease the residual stress level inside
weldment. Repair procedure that had been performed is
based on this principle. From the on-stream vessel’s
assessment after one year since its repairing time using
HTCUT and high temperature ToFD showed the CW2
in good condition. This successful repair procedure is
very dependent on very controlled and careful step
during repair stage.
7. References
[1]. Heuser, Albert., ”Repair of a Thick-walled
Ammonia Synthesis Converter Containing a Leak”.,
AIChe., USA., 1991.
[2]. Prescott, G. R., Shannon, Brian., ”Process
Equipment Problems Caused by Intercation with
Hydrogen”., Ammonia Plant Safety & Related
Facilities p.p 237-252., USA., 2001.
[3]. Timmins, P. F., ”Solution to Hydrogen Attack in
Steels”., ASM International., USA., 1997.
[4]. ____., ”Petroleum and Natural Gas Industries—
Materials for Use in H2S-containing Environment in
Oil and Gas Production Parts 1, 2, and 3”, NACE
MR0175 / ISO 15156-1:2003 (E).
[5]. ____., ”Controlling Weld Hardness of Carbon Steel
Refinery Equipment to Prevent Environmental
Cracking”, API 942.
[6]. Tsay, L. W., Liu, C. C., Chao, Y. H., Shieh, Y. H.,
”Fatigue Cack Propagation in 2.25Cr – 1Mo
Steel Weldments in Air and Hydrogen”., Material
Science and Engineering A 299 (2001), 16 – 26
[7]. Fujibayashi, S., Kawano, K., Komamura, T.,
Sugimura, T., ”Creep Behaviour of 2.25Cr – 1Mo
Steel Shield Metal Arch Wedment”., ISIJ
International, Vol 44 (2004), No. 3, 581-590.
[8]. Louthan Jr, M.R., ”Hydrogen Embrittlement of
Metals: A Primer for the Failure Analyst”., Journal
Failure Analysis and Prevention (2008) 8:289 – 307.
[9]. ____., ”Rules for Construction of Pressure
Vessels”., ASME Sect. VIII Divison 1&2., 1998.
[10]. Firth, D. M., Keen, D., Jones, C., Kartensen, A.,
”Cracking and Repair of Closing Welds in
2.25 Cr1Mo Steel Vessels Operating in High
Temperature Synthesis Gas”., Ammonia Plant
Safety &Related Facilities pp. 161-169.,
USA., 2005.
 8. Biographies

Farabirazy Albiruni obtained his

B.Sc. from University of Indonesia
(UI), Indonesia in 2002 in
Metallurgy and Materials
Engineering. He joined PT. Pupuk
Kalimantan Timur in 2003 as a
Metallurgical and NDE Engineer.
Currently, he joined Basic Atomic
Energy Research Institute of Pusan National University,
South Korea, as Research Graduate Student under
supervision of Professor Joon-Hyun Lee. His research
interest is advance non-destructive evaluation
techniques for structural health monitoring.

Wildan Hamdani is a senior

inspection engineer at PT. Pupuk
Kalimantan Timur. He holds B.Sc
degree in Mechanical Engineering
from Diponegoro University
(UNDIP). He is currently pursuing
his Master Degree at Gadjah Mada
University in Maintenance
Engineering field.

Joon-Hyun Lee is a Professor in

Mechanical Design System
Engineering in Pusan National
University, South Korea. He
received his B.Sc degree in
Mechanical Engineering from
Pusan National University (PNU),
South Korea and his M.Eng and
Ph.D degrees in Mechanical Engineering from Tohoku
University, Japan, in 1985 and 1988. He was a Research
Assistant Professor at the Center for Quality
Engineering and Failure Prevention of the Northwestern
University, U.S.A, from 1988 to 1990. Currently, he is
an NDE Group Leader of Research Institute of
Mechanical Technology of the PNU. He is also a
Director of Basic Atomic Energy Research Institute and
research center for Failure Analysis and Reliability.
Professor Lee’ works and research interests include
structural integrity evaluation and quality assurance and
control based on nondestructive evaluation techniques,
application of NDE technique to characterize non-
conventional engineering materials, wave propagation
and NDE modeling and signal processing.