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

net/publication/257714166

Repair of Tube–Tubesheet Weld Cracks in a Cracked Gas/Steam Heat

Exchanger

Article  in  Journal of Failure Analysis and Prevention · December 2011

DOI: 10.1007/s11668-011-9482-8

CITATIONS READS

13 15,296

2 authors, including:

Hosni Ezuber
University of Bahrain
32 PUBLICATIONS   717 CITATIONS   

SEE PROFILE

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

CaSO$ scaling View project

All content following this page was uploaded by Hosni Ezuber on 20 November 2014.

The user has requested enhancement of the downloaded file.

J Fail. Anal. and Preven.
DOI 10.1007/s11668-011-9482-8

CASE HISTORY—PEER-REVIEWED

Repair of Tube–Tubesheet Weld Cracks in a Cracked Gas/Steam

Heat Exchanger
Shaker Adullah • Hosni M. Ezuber

Submitted: 10 November 2010 / in revised form: 14 June 2011

 ASM International 2011

Abstract Serious cracks were found at tube-to-tubesheet seamless tubes were made of SA 213-T11 steel with
welds during the start up operation of a cracked gas/steam 1.75 in. OD (44.45 mm) 9 7 BWG (4.57 mm) 9 21 ft
heat exchanger in a gas plant. The cause of the failure has (6.4 m) LG. The tube sheets were made from SA 182-F11
been analyzed, and the exchanger is repaired. The root CL2 steel with 1.5 in. (38 mm) thick., normalized at
cause of the cracks is mainly attributed to poor workshop in 1600 F for 8 h, tempered at 1325 F for 8 h, with air as
the tube–tubesheet welds. Failure analysis and repair pro- coolant. The inside diameter of the shell is 55.75 in.
cedure were reported in the present article. Several months (1416 mm) and was made of SA 516-70N steel. The tube-
of operation after the repair confirmed that the repair was to-tubesheet filler material was GTAW wire ER80S-B2.
successful. The exchangers have been delivered to the jobsite, in a
gas plant, after reportedly passing the dye-penetrant
Keywords Heat exchanger failures
Tube-to-tube joints
examination of each weld pass, a 25 psig air, and soap test
Weld cracks before expansion, dye-penetrant examination after rolling,
dye-penetrant examination after postweld heat treatment
(PWHT), and a 3031 psi shell side hydro test. During the
Background boil-out, the inlet was exposed to a maximum temperature
of 300 C, while the outlet reached 220 C. Of the nine
The object for repair in this study is a weldment tube– heat exchangers, leaks have been detected in two tube-to-
tubesheet joints on a cracked gas/steam heat exchanger in a tubesheet welds of two heat exchangers during the start-up
gas plant. The exchanger is one of the nine identical shell- operation. The cracks are present in the inlet, tube-to-
and-tube heat exchangers, fixed-tube exchanger type tubesheet weld.
(1 shell–1 tube pass type), made by the same manufacturer, It is generally a common practice where tube-to-tube
with cracked gas in the tube side being cooled down by sheet seal weld is used to supplement an expanded tube joint
process steam in the shell side. The design temperature and to insure leak tightness [1, 2]. The tube joints are commonly
pressure of the shell side are 653 F (345 C) and 1812 psi welded when both tube-to-tubesheet are of suitable mate-
(127.4 kg/cm2), respectively. rials [1]. In practice, most welds are multipass. Welded
The tube bundle of the heat exchanger consists of 484 joints are often heat treated (preheat and postweld heat
tubes fixed in a horizontal position. The cold-drawn treatment) before use to improve ductility and toughness
and reduce welding stress [3]. Filler material composition
should be nearly the same as the base metal to ensure uni-
S. Adullah
Ramsis Engineering, P.O. Box 28967, East Riffa, Bahrain form strength and resistance to heat and corrosion [3, 4].
e-mail: [email protected] Stress corrosion cracking is a common failure mode in
tube-to-tubesheet joints. The failures are mainly related to
H. M. Ezuber (&)
improper tube expansion, abnormal clearance between tube
College of Engineering, University of Bahrain, P.O. Box 32038,
Isa Town, Bahrain and tubesheet [5], poorly executed tube-to-tubesheet welds
e-mail: [email protected] or to defects in the in-field PWHT [6].

123
 J Fail. Anal. and Preven.

The objective of this study is to determine the failure (Fig. 1). A close investigation of the leaked zone using PT
cause, identify the failure mechanism, and to give a solu- showed numerous cracks at tube-to-tubesheet welds. The
tion how the failure could be avoided. Failure analysis tests stress cracks are approximately straight and are of linear
including visual inspection and microscopic examination, and transverse types (Figs. 2, 3).
SEM, mechanical test (hardness), and nondestructive tests Samples extracted from the tube sheets with and without
(PT) were performed. Repair operation based on failure reported failures, containing two tube-to-tubesheet welds
analysis results and recommendations were conducted. The and the tubesheet ligaments between them, were sent for
repaired exchanger has been working effectively for sev- laboratory investigation. The outcome results are as
eral months without weld cracks. The failure analysis and follows.
repair procedure are outlined in this article.
Microscopic Examination

Inspection Examination on samples contained tube-to-tubesheet

revealed visible cracks on most of the specimens’ surfaces.
On Site Visual Inspection The cracks were through-wall and were circumferentially
oriented. Observations on the cross sections of the welds
On-site inspection of the failed heat exchangers showed are given in Fig. 3. The figure shows side-by-side images
that both exchangers leaked at tube-to-tubesheet welds of the vertical cross sections of a tube-to-tubesheet weld. It
is apparent from this figure that welds did not fully con-
sume the chamfered area of the weld prep. Cracks were
common in the areas where the weld preps were not con-
sumed in the welding process. The initiation of the
through-wall cracks took place in the area of the weld root
in the tube-to-tubesheet joint and propagated in the weld
metals. This side is submerged in the steam side of the
tubesheet. Parallel cracks (Fig. 3, right) appeared to run
side-by-side for much of the length of this through-wall
crack. No cracks were found at the tube or tubesheet
materials.

SEM–EDS Analysis

Figure 4 represents the EDS spectra for the chemical

compositions of the tube metal, tubesheet base metal, and
Fig. 1 Leaks appearance at the tube-to-tubesheet welds

Fig. 2 Crack appearance in the tube-to-tubesheet welds (left: transverse cracks; right: linear cracks)

123
J Fail. Anal. and Preven.

Fig. 3 Enlarged views on polished cross sections showing the through-wall crack

the deposited weld metal. The results presented in this effectiveness of the stress relieving treatment [4]. The high
figure clearly show that the chemical analysis of tube, hardness in the interface between weld and the tube HAZ
tubesheet, and the deposited weld metal are virtually are reported to be related to microstructural embrittlement
identical. The chemical compositions are consistent with [6].
what be expected for the SA 213-T11, SA 182-F11 CL2,
and ER 80S-B2 materials. Quantitative chemical analysis Possible Root Cause
of the tube, the tubesheet, and the weld material is given in
Table 1. The chemical composition of the materials com- The results of the chemical analysis show that there are no
plies with the manufacture’s specifications, and with apparent defects in the specification of the weld material,
ASTM specifications A387 Grade 11 (cover plate) and the tube material, or the tubesheet material, i.e., no dis-
ASTM A182 (pipes), which are correct for the type of crepancies with manufacture’s specifications were found in
service sought [3]. It is generally common to note that the chemical composition. The diameter of the tubesheet hole
carbon content of the filler metal is lower than that of the for tubes has uniform and proper diameter through the
base metal. The average values of tensile test, yield, and whole thickness of the tube (i.e., normal clearance between
elongations given in Table 2 are typical values of the steel the tube and the tube-sheet) and considered to be adequate
materials. and complies with the design specification. As a result,
Figure 5 is an overall SEM image of a polished cross fabrication error has been ruled out as the root cause of the
section of a dialed sample. A through-wall crack in a problem. Thermal calculation records confirm that the
deposited weld is observed. The crack propagation is quite exchanger is operated in nucleate boiling mode. A film
straight and there is no evidence of ductile behavior at and/or vapor blanketing of tubes is not possible. The
either end of the crack. tube wall temperature is very close to boiler feed water
(BFW)-saturated temperature or shell temperature. Based
Microhardness on these analyses, no excessive tube thermal expansion in
relation to shell is expected. The vapor blanketing or film
Tables 3 and 4 show approximate median hardness values boiling is unlikely considered as a cause of the tube weld
for various regions (tube-to-tubesheet: starting with the failures experiences. The root cause of the cracks is pos-
tube base material, passing through the HAZ, the weld root, sibly attributed to poor workshop in the tube-to-tubesheet
another segment of the HAZ, and finally in the deposited welds. Less than adequate weld throat thickness appeared
wed metal) along the maximum measured hardness. The to be the factor that governed whether or not a particular
weld metal and HAZ show much higher hardness than the tube-to-tubesheet weld remained intact. This is possibly
base metal. Some interesting differences are systematically caused by incorrect welding condition (incorrect torch/gun
observed in hardness values of HAZ of the tube and tube angle and incorrect edge penetration). The incorrect weld
sheet materials. Hardness of HAZ of the tube material is design and fabrication (incorrect weld size and quality) of
less than that of the tubesheet material. This denotes scant the weld joints thus results in an undetected lack of fusion

123
 J Fail. Anal. and Preven.

Fig. 4 EDS results showing typical chemical compositions for tube and tubesheet materials

in welded joints. Although the exchanger underwent suc- during the commissioning of the exchanger, (during the
cessful hydro test and further dye penetrant examinations start-up of the process, the boiling water coming out of the
before the start-up of the operation, the lack of fusion in cycle at the shell side is possibly caused by thermal
these routine testings is usually not detected. expansion of the tubes). This in turn results in mechanical
The initiation and propagation of cracks appear to occur stress on the welding joints leading to the initiation and
as a result of mechanical stress applied to the material propagation of cracks at defective weld joints. The storage

123
J Fail. Anal. and Preven.

Table 1 Average chemical composition (wt.%) for tube, tubesheet, and filler materials for the heat exchanger steels material
C Mn Si S P Cr Mo Ni Al Cu Fe

SA213-T11 0.10 0.452 0.65 0.018 0.009 1.23 0.49 0.07 0.022 0.17 Bal.
SA182-F11CL2 0.11 0.54 0.57 0.003 0.01 1.13 0.50 0.14 0.032 0.15 Bal.
ER80S-B2 0.09 0.55 0.48 0.006 0.012 1.35 0.55 … … 0.15 Bal.

Table 2 Mechanical properties for tube and tubesheet materials for Remove all the cracks by machining
the heat exchanger steels materials
Alloy Tensile test, psi 2% yield, psi Elongation, % PT examination

SA213-T11 65,950 54,500 57.8

SA182-F11CL2 71,600 40,200 34 Steam cleaning/
surface drying

Preheating
150-200oC

(1)
(3)
(2)

Preliminary welding 2nd pass welding

(root pass)

3rd welding pass

Heat Treatment
(350oC/2hr + air cooling 70oC/hr))
+ Rest for 48 hours

Postweld heat treatment

660oC/2hrs

PT examination

Fig. 5 SEM image of a polished cross section of the specimen Tube expansion
showing the through-wall crack in deposited weld metal
Final PT examination

Table 3 Approximate median and maximum hardness values for

tube material Hydro-test

Tube (left) Median, HV Maximum, HV

Fig. 6 Flowchart of repair procedure
Weld body 280 286
HAZ 1 200 203 of hydro test water (under a pressure) in the heat
Weld Root 280 203 exchangers for a period of more than 6 months may play a
HAZ 2 190 205 role in the crack mechanism failure.
Base material 145 153
Repair of the Tube-to-Tubesheet Weldments

Table 4 Approximate median and maximum hardness values for In order to eliminate the crack failures, the defective tube-
tubesheet material to-tubesheet welds was removed and replaced with prop-
erly designed and installed good quality welds.
Tubesheet (right) Median, HV Maximum, HV
The repair procedure is shown in Fig. 6 and is described
Weld body 280 288 as follows:
HAZ 260 280
(i) All the tube-to-tubesheet welds were removed by
Base material 185 221
machining.

123
 J Fail. Anal. and Preven.

Fig. 7 Schematic diagram for preheating, welding range, and post- Fig. 9 Schematic diagram for post weld heat treatment temperature–
heating temperature–time profile time profile

(v) Welding process (multipass welding). This process

started by carrying out a preliminary welding (root
pass). A filler material GTAW wire ER80S-B2 was
used to obtain similar properties to the substrate.
After conducting the preliminary welding process,
the surface was subjected to heat treatment (HT) at
350 C for 2 h followed by gradual air cooling with
a rate of 70 C/h (max). The surface was then left
for 48 h to insure that no hydrogen damage is
developed. Second- and third-pass welding proce-
dure were followed as similar to the preliminary step
(heating at 150–200 C, welding, HT at 350 C, air
cooling with 70 C/h, followed by 48 h between
each welding pass). Preheating along with heat
treatment is required to minimize the detrimental
Fig. 8 Schematic diagram for preliminary heating process effect of high temperature and severe thermal
gradients in welding [7].
(vi) Postweld heat treatment (PWHT) at 660 C, soaking
(ii) PT examination was conducted to insure that no the metal at this temperature for 2 h, followed by air
cracks were developed during machining. slow cooling with a rate of 70 C/h as per applicable
(iii) Cleaning the surface using pressurized steam: This codes and clients specific request (Fig. 9). The post
process was aimed to clean weld grooves area heat treatment to a temperature of 600–700 C is
around the tubes and to remove all the traces of dye- required to soften the HAZ, remove the residual
penetrant residue. The cleaning process, however, hydrogen, restore the toughness, stabilize the micro-
was followed by drying process using compressed structure of the weld metal, and reduce the welding
air. stresses [3]. In Cr–Mo alloy steels, the weld
(iv) Preheating the surface at a temperature range of 150 transforms uniformly to martensite or bearlite. It
(min)–200 C (max) as per ASME code B31.3, will then normally be subjected to a PWHT which
Fig. 7. The heating process was conducted using has the effect of tempering and stress relieving the
heating elements (ceramic fiber and insulation joint as a whole [8].
attached to mesh with stainless steel wire, Fig. 8). (vii) Tube expansion was then applied as a method of
The preheat treatment is needed to prevent harden- sealing tube-to-tube sheet joints.
ing and weld cracking. Preheating reduces stresses, (viii) Final PT examination was conducted to ensure no
limit, or temper martinstics areas, and reduce the cracks were developed during the welding or
amount of hydrogen retained in the weld [3]. expansion procedure.

123
J Fail. Anal. and Preven.

Finally, a shell test was performed to insure that no without any problems, indicating thus that the repair was
leak occurs on the tube-to-tubesheet joints. The heat successful.
exchanger was then pressed into operational service. The
exchanger has been in use for several weeks after the
repair. No leaks have been reported during this period. References
These results proved that the repair performed was a
1. Harison, J. (ed.): Standards of the Tubular Exchanger Manufacture
successful one.
Association, p. 280. TEMA, Tubular Exchanger Manufacture
Association, INC, New York (1999)
2. 2007 ASME Boiler and Pressure Vessels Code, ASME, Section
Conclusion VIII, Division 1, Rules for Construction of Pressure Vessels,
American Society of Mechanical Engineering, p. 306 (2007)
3. Kuppan, T.: Heat Exchanger Design Handbook. Taylor and
The cracks found at tube-to-tubesheet welds during the Francis Group, New York (2000)
start-up operation were identified to be due to stress cor- 4. Oates, W.R., Saitta, A.M. (eds.): ASW Welding Handbook.
rosion cracking. Failure analysis revealed no apparent Materials and Application, vol. 4, part 4, 8th edn., p. 45. American
Welding Society, Miami (1998)
defects in the weld material, the tube material, or the tube
5. El-Batahy, A., Amin, M., Fathy, G.: Failure of a column reboiler in
sheet material. The root cause of the cracks is possibly a MEG recovery unit. Mater. Perform. J. 47(8), 66–69 (2008)
attributed to poor workshop in the tube-to-tubesheet welds. 6. Otegui, J.L., Fazzini, P.G.: Failure analysis of tube–tubesheet welds
Repair procedure including the removal of the defective in cracked gas heat exchanger. Eng. Fail. Anal. 11, 903–913 (2004)
7. Rao, K.R. (ed.): Companion Guide to the ASME, Boiler, Pressure
tube-to-tubesheet welds and replacing with properly
Vessels Code, vol. 1. ASME Press, New York (2002)
designed and installed welds was satisfactory. In fact, the 8. Griffing, L. (ed.): Metal and their weldability. In: Welding
repaired heat exchanger has been used for several weeks Handbook, Section 4, 6th edn. AWS, Miami (1972)

123

View publication stats