Professional Documents
Culture Documents
WRC - The Welding Research Council brings together science and engineering specialists in
developing the solutions to problems in welding and pressure vessel technology. They exchange
knowledge, share perspectives, and execute R and D activities. As needed, the Council organizes
and manages cooperative programs.
MPC – A Council of the WRC, the Materials Properties Council is dedicated to providing
industry with the best technology and the best data that can be obtained on the properties of
materials to help meet today’s most advanced concepts in design and service, life assessment,
fitness-for-service, and reliability and safety.
PVRC – A Council of the WRC, the goal of the Pressure Vessel Research Council is to
encourage, promote and conduct research in the field of pressure vessels and related pressure
equipment technologies, including evaluation of materials, design, fabrication, inspection and
testing.
WRC Bulletins contain final reports from projects sponsored by the Welding Research Council, important
papers presented before engineering societies and other reports of current interest.
No warranty of any kind expressed or implied, respecting of data, analyses, graphs or any other
information provided in this publication is made by the Welding Research Council, and the use of any
such information is at the user’s sole risk.
All rights are reserved and no part of this publication may be reproduced, downloaded, disseminated, or
otherwise transferred in any form or by any means, including photocopying, without the express written
consent of WRC.
ISSN 0043-2326
Library of Congress Catalog Number: 85-647116
Library of Congress
Catalog Card Number: 85-647116
Copyright © 2009 by
Welding Research Council, Inc.
All Rights Reserved
Printed in U.S. A
FOREWORD
In recent years there has been increasing need to acquire technology for the safe handling
of hydrogen for petroleum and other energy related applications. The pace ofresearch into
metal/hydrogen interactions accelerated at the start of the space age due to the hydrogen's
efficiency per unit weight as a rocket fuel. Practical applications called for hydrogen to be
transported and used at high pressures thereby increasing its hazard potential. Exotic and
expensive alloys could be used to put astronauts into space because cost was of little concern
where safety was an issue. For most conventional applications steel is still the practical choice.
It has been long recognized that at ordinary and elevated temperatures hydrogen can be
destructive to steels, not just to high-strength steels, but also to steels of ordinary strength levels.
Hydrogen's capability to enter and then diffuse through the metallic lattice, accelerated by stress
gradients and seeking out points of weakness where it can concentrate or react, renders it capable
of destroying pressure retaining metals from the inside where damage defies detection and until
it is too late and the pressure containment as failed or is no longer safe.
This Welding Research Council (WRC) Bulletin is part of a series that captures the
essential studies of the interaction of steel-hydrogen interactions in recent years. Topics include
Modern Vanadium Steels for High Temperature Petroleum Reactors ( # 524), Fabrication and Repair of
Low Alloy Steel Pressure Equipment ( # 525), Performance of Steels in Hydrogen Charging
Environments ( # 526), Practical Aspects of Hydrogen Attack ( # 527), Test Methods for Hydrogen
Induced Cracking ( # 530), Metallurgical Studies of Steels for Sour Service Environments ( # 532),
Studies of Cladding and Overlayfor Pressure Vessel Service ( # 534), and Toughness, Fracture and
Fitness for Hydrogen Service ( # 535).
The papers included have been presented at international conferences sponsored by WRC's sister
organization the Materials Properties Council Inc. (MPC). The technology reported in this series provides
a comprehensive view of practical solutions to engineering problems and advances in the knowledge
about hydrogen and steel interactions.
Martin Prager
Executive Director, WRC
Table of Contents
The aim of this paper is to highlight a few aspects of The allowable design stress as given by the design code
engineering and operating high temperature hydrogen is generally related to either the hot tensile
A in ppm/.Jbar B in K
8.93 650
Ref. 2 1.0
7.6 630
Ref. 5
Ref. 6 14.8 1099
RDF T = 450°C
x
The Hydrogen Concentration during Operation and the )(
x
)<
)<)<~ )<
)(
)<
Effect of a cladding -
2 14Cr1 Mo Claddin~V "
~
[H] in cladding
t 200
[H] in CrMo-steel
~~~~~~~~~~~~~~~~-in_.Tppm
in ppm
- - prior to cooldown 5 .o
t - Disbanding of the Cladding -
• •
[H] in ppm [H] in ppm 'C) are low and that an incubation period is needed to
reach the equilibrium pressure inside a defect.
Calculation of the hydrogen concentration in the steel
: 6 which is needed at 200 'C to attain an internal
300
i pressure equal to the yield strength (i.e 4500 bar)
using the solubility data in the appendix gives a value
l
of 2.8 ppm. More realistic assumptions of a lower
hydrogen solubility and of significant diffusion rates
200 4 down to room temperature would result in a critical
hydrogen concentration well below 1 ppm for extension
:r----:;;.--------=
:I
of defects, but this does not seem realistic.
I
.,.....----.u:u······················································ ..• a) Relationship (1) is only valid in the engineering
range of hydrogen pressures. For very high pressures
it may be replaced by:
0 ..._~--~--...---+-------.-------.------..------0
6 4 2 0 10 20 30 C - A.Jp.e-(B-vp)/T (2)
cladding CrMo-steel distance
..,..._-------+I--------1•.,.. in mm with v - 0.071 K/bar (Ref. 13).
a) prior to cooldown ( 450°C, 150 bar) b) The assumption that above a certain temperature
sufficient hydrogen can diffuse into the defect to
after cooldown (after 50 min) build up the equilibrium pressure is an
after 200 hours over-simplification. Proper diffusion modelling is
required, so that factors like the size of the
defect, the finite amount of hydrogen available and
~[\
O>Q.
~.s 300
the temperature history can be taken into account.
Consideration should also be given to trapping of
part of the available hydrogen during cooldown. The
"O G>
0 (.) trapping ratio and the distribution and the size of
o.g the defects in the steel, the welds and the cladding
Q) Q) are strongly dependent on the type and the quality
200
:S£ of the material. Therefore, the tolerable hydrogen
.s Q)
,_...r::.
concentration during cooldown is expected to be
:I: .... strongly dependent on the material qualities as
...... 100
well.
T(°C) pH 2(bar) dT/dt(°C/hr)
c) The criterion that the internal pressure should
1: 450 150 675
equal the yield strength of the steel is
2: 450 150 150
inappropriate for defect growth. It should be
b) 0
0 50 100 150 200
replaced by a criterion based on fracture mechanics
principles. In a fracture mechanics analysis, the
time in hrs ... crack driving force which results from the internal
pressure inside the defect is compared with the
Fig.6. Degassing of a Disbonding Specimen fracture toughness of the surrounding steel. The
a) Hydrogen Concentration Profiles fracture toughness used in the analysis should
b) Hydrogen Concentration in the Cladding at account for any temper embrittlement and hydrogen
the interface during and after cooldown embrittlement of the steel. The analysis can be
extended to assess the risk of initiation of
Hydro9en solubility: CrMo steel Cladding unstable fracture when the vessel is under pressure
A (ppm/{ bar): 23.6 7.6 and the crack driving force resulting from the
B (K): 3257 630 membrane stresses is superimposed on the above crack
driving force resulting from hydrogen pressure in
the critical hydrogen concentration in CrMo steels for the defect.
hydrogen damage during shutdown is expected to be
largely dependent on the quality of the steel. In the API/MPC task group, a considerable amount of
hydrogen and temper embrittlement data have been
A well-known attempt to model the growth of defects as presented. An effort should be made to model the
a result of hydrogen pressure build-up inside the extension of defects during and after cooldown and to
defect in a steel which is over-saturated with hydrogen apply the generated data to derive guidelines for the
has been made by Geiger (Ref. 12). He postulates that avoidance of hydrogen damage during shutdown in terms
the pressure inside the defect can be calculated using of maximum hydrogen concentration during operation and
relationship (1) of the appendix: cooldown procedures.
2
p - (C/A) .exp(2B/T) (1)
REFERENCES
1) API Publication 941
Steels for Hydrogen Service at Elevated Temperatures
and Pressures in Petroleum Refineries and
Petrochemical Plants.
Third edition, May 1983.
2) API publication 946
The effect of outgassing cycles on the hydrogen
content in petrochemical-reactor-vessel steels
July 1981.
3) MPC/API task group on materials for pressure vessel
service with hydrogen at high temperatures and
pressures.
HPV-43: Minutes of meeting of May 11 and 12, 1988.
4) MPC/API task group on materials for pressure vessel
service with hydrogen at high temperatures and
pressures.
HPV-41: Minutes of meeting of October 7 and 8, 1988.
5) Hydrogen distribution through pressure vessel wall.
Sakai,T. Asami K. and Takada H.
paper presented at the Pressure Vessel Research
Committee Meeting, Scottsdale, Jan 1981
6) The solubility of hydrogen at low pressure in iron,
nickel and certain steels at 400 to 600 °C.
Armbruster, M.H.
J. of Am. Chem. Soc. 65 (1943) 1043-1054
7) Influence of Alloy Additions on Hydrogen Diffusion
in Iron and Contribution to the System Iron
Hydrogen.
Geller, W. and Sun, T.
Arch. Eisenhuttenw., 21, 423-430 (1950)
8) Der Einfluss von Wasserstoff auf die
Gebrauchseigenschaften von unlegierten und
niedriglegierten Stahlen.
Haumann, W. e.a.
Stahl und Eisen 107 (1987), No. 12, 585-594
APPLICATION AT H2 ENVlRONMEN 1
(hydrogen reactor vessels)
400
I I
HYDROGEN INTRUSION IN CLAD STEEL .r.:
Ci 300
c:
....(!)
SHUTDOWN ti
...
C'O
JIS G3601 (;;.200MPa)
(!) 200
.r.:
HYDROGEN ACCUMULATION v;
ASTM A264 (;;.140MPa)
100~~~~~~~~-'-~~~--'
DISBONDING
Initial Reduction
30mm/pass 40mm/pass
{plate thickness : 100mm)
Figure 1 SCHEMATIC FLOW OF HYDROGEN- Figure 2 EFFECT OF INITIAL REDUCTION ON
-INDUCED DISBONDING SHEAR STRENGTH OF CLAD PLATES
gauge clad plates (80-200mm in thickness) has not been conventional overlaid clad steel and the rolled clad steel,
considered satisfactory. However, both the recent progress and, (2)the improvement of joining performance of rolled
in rolling technology and the use of mills with larger clad plate through proper selection for the base material
capacity attained a marked increase in shear strength at the and the insert metal. The commercial-scale production of
bonded interface, which was reasonably applicable to the heavy gauge y-stainless steel I Cr-Mo steel clad plate was
production of heavier-gauge clad plates. The example of also carried out using 9000 ton rolling mill in Fukuyama
the increase in bonding strength of austenitic stainless Works, NKK Corporation, and the properties were
steel clad plate (lOOmm in thickness) is shown in Figure discussed and compared with the conventional overlaid clad
2. The general feature of rolled clad steels and their plate.
production process will be discussed in the following
section. HOT ROLLED CLAD PLATES
The present study was performed to obtain the basic data
on (l)the difference of disbonding behaviour between the The typical process for producing overlay welded and hot
OVERLAID
~ T " /~~
I
I
overlay welding
\ rolling finishing
I
ROLLED
~ -- ~
slab composing rolling separationg
Table l COMPARISON OF ROLLED CLAD STEEL WITH OVERLAY WELDED CLAD STEEL
rolled clad plates was shown in Figure 3. In overlaying Therefore the production mill of 9000 ton-class was used
process, the banded weldment of cladding metal is in the present study. In the following sections the
performed, typically using submerged arc welding on base application of hot rolling process to production of the clad
steel. The surface finishing process usually follows. On plates for hydrogen environment was discussed.
the other hand, in the rolled clad production, the cladding
plates are composed with base steels by welding the EXPERIMENTAL PROCEDURE
circumference and degassing at the interface before hot
rolling. The insert metal such as Ni foil can be used in The chemical composition of the base steels and
oroer to help joint at the interface. Then the composed slab cladding metal used for the laboratory study is shown in
is heated and hot rolled for the interface bonding using the Table 2. The 21I4Cr- l Mo steels with lower C content and
plate mill. The clad plates used in the present study was a small addition of B+V were used as well as the standard
composed as the sandwich-type slab and the separation of 21; 4Cr-lMo steel. Type 347 austenitic stainless steel
two bonded clad plates was performed after rolling. The (18%Cr-10%Ni-Nb) with plate thickness of 20mm was
surface finishing is carried out at the final stage. always used as the cladding metal.
The advantageous feature of the rolled clad plate lies in The cladding conditions for rolling and overlaying were
the mass-productivity and capability in dimension summerized in Table 3. Slabs sized (200+20)mm x
enlargement of products. Moreover, the diffusion at the 240mm x 480mm were composed in advance of hot
interface related to the completion of the interface bonding rolling of the clads[5]. Ni foil of 100 µm in thickness was
is to be promoted by controlling the rolling parameter used as the insert metal.
relatively with ease, especially in the case a heavy-load The slabs were then heated at l 200°C and rolled to
reduction is available at rolling process. 55mm in thickness. The rolled plates were normalized at
c Si Mn p s Ni Cr Mo Nb v Ti B sol.Al total N
Std. 0.15 0.11 0.55 0.006 0.00< 0.03 2.43 0.96 - - - - 0.014 0.0044
Base metal 0.20V 0.08 0.19 0.55 0.004 0.001 0.50 2.42 1.04 - 0.20 0.007 0.0008 0.017 0.0033
0.34V 0.11 0.06 0.54 0.003 0.001 0.51 2.42 0.97 0.02 0.34 0.006 0.0009 0.016 0.0037
Claddinl!. meta tvoe347 0.04E 0.62 1.65 0.025 0.008 10.86 17.49 0.17 0.77 0.03 - - 0.005 0.0021
+ 690°C X 24h FC
900°C x 30 min., then tempered at 720°C x 30 min. overlaid clad plates. After reduction to 50mm in thickness,
followed by stress-relieving at 690°C x 24 hours. two layers of weld metal, type 309L and 347L as the first
Steel slabs normalized at 900°C x 30 min. and tempered and the second layer respectively, were overlaid using
at 720°C x 30 min. were used as the base steel of the submerged arc welding. Stress relieving heat treatment of
690°C x 24 hours followed the overlaying.
The plate specimen of (46+4)mm x 50mm x lOOmm in
dimension was taken from each clad plate and used for
hydrogen charging test. After verifying the bonding at the
150 interface using ultra-sonic testing(U.T.), they were
hydrogen charged in autoclave at 10-20MPa, 400-450°C
E
a. for 24hours. Then the plates were cooled to room
a.
c 100 temperature at the cooling rate 3.5-4.0°C I min., and kept
Q)
for 1 week before inspecting the disbonding. The scanned
c0
(.) data was picture-processed to measure disbonded area ratio.
c
Q) ti <t2 <h<t4
0
O> 50 RESULT AND DISCUSSION
.....
-0
>.
I
10 Disbonding condition
4 Hydrogen concentration in the clad steel of hydrogen
3 reactor vessels during operation and after shutdown was
2
I calculated evaluating the diffusivity of hydrogen in
05 0 50 100 austenitic stainless steel and Cr-Mo steel. The result was
Clodding Bose metal shown schematically in Figure 4. Hydrogen atoms entered
Distance from interface (rrrn) the clad steel at the inner surface of the vessel wall and
Inner surface Outer surface
diffused through the wan thickness to the outer, showing
the linear profile during the operation, but gathered
Figure 4 SCHEMATIC PROFILE OF HYDROGEN
preferentially at the interface showing the curved
DISTRIBUTION IN THE CLAD STEEL SIMULATING
distribution described in Figure 4 after shutdown [4]. The
THE HIGH PRESSURE REACTOR VESSEL AFTER
high concentration of hydrogen evaluated at the cladding I
SHUTDOWN[4]
base metal interface was due to a large difference in both
hours and cooled to room temperature at 4°C I min. U.T. 200 400 600
Max. hydrogen content at the
was carried out 5 days after completing the hydrogen
interface (ppm, calculated vo lue)
charging. By optical and scanning microscopy the crack
was observed in the cladding approximately 50 µm distant Figure 6 RELATIONSHIP BETWEEN CALCULATED
from the interface (Photo 1). The crack propagation was MAXIMUM HYDROGEN CONTENT AT THE
recognized on grain boundary and the feature of the crack INTERFACE AND MEASURED FRACTION OF
was similar to what was reported in disbonding of clad DISBONDED AREA
plates in high pressure hydrogen reactor plants[6].
According to these observation the hydrogen autoclave was using difference method. The calculation was carried out
verified to reproduce the interface disbonding of clad plates on the assumption that the distribution of hydrogen was
caused by shutdown of the reactors. proportional to the saturated solubility of hydrogen in the
Temperature and pressure dependence of hydrogen caldding and the base metal. Three combinations of
accumulation at the interface was simulated on the basis of temperature and pressure parameters, 450°C-20MPa,
Fick's second law, 450°C-1 OMPa and 400°C- l OMPa were assumed, and in
(1) each case the concentration showed the maximum value at
approximately 100 hours after cooling. Only this value is
Length (mm)
expressed in Figure 6. A linear relationship between the
0 110
O-t-~~~~~~~~~~~~~~~--1 calculated maximum hydrogen concentration at the
interface and fraction of disbonded area measured using
U.T. is shown both for overlaid and rolled clad plate.
E
E
~ <~~JY.. ~
Improvement of disbonding resistance in rolled
dad steel
60 The disbanding behavior of rolled and overlaid clad
Horizontal plates were tested and compared. Two hydrogen charging
conditions, 400°C-lOMPaH2 and 450°C-20MPaH2, were
E applied to specimen from both types of clad pates. The
5..c measured disbonded area fraction is shown in Figure 7. In
a.
Q) the case of lower temperature and lower hydrogen pressure
0
(400°C-10MPaH2), the rolled clad steel did not show
Figure 5 THE MAPPED ULTRA-SONIC RESULT disbanding, when both the standard Cr-Mo steel and
OF DISBONDING AFTER THERMAL HYDROGEN Cr-Mo-0.34% V were used as the base steel. The overlaid
CHARGING clad steel showed larger disbanded area fraction than any of
Table 4 COMPARISON OF CRACKING SENSITI- -lONi-Nb) was used for plates B and C. The standard
VITY IN OVERLAID AND ROLLED CLAD STEEL 2l/4Cr-lMo steel was used as the base metal for plates A
FACTORS AFFECTING CRACKING SENSITIVITY and B, while V-added 2 l; 4Cr- l Mo-0.2V steel was for plate
H2 accumulation site Harri phase formation C. The finish thickness of plates A, B, and C was
Overlaid C diffusion from Dilution of Cr+Ni in (143+5)mm, (75+4)mm, (100+4)mm respectively. After
Clad Steel base steel-->grain cladding rolling, dehydrogenation heat treatment was carried out,
boundary carbide -->martensitic phase and normalizing at 930°C, tempering at 730°C, and stress
Rolled Joint completing Solid-state bonding relieving at 690°C followed. Single Ni insert was used to
Clad Steel -->no microvoid -.;..less martensitic help bonding. In order to know the influence of thickness
at interface phase of the insert metal, l 00 µm-thick Ni foil was inserted as
Ni insertion single layer in the top-half of the slab, and two sheets of
-->suppressed C dif- the same foil were inserted in the bottom-half. For plate B
fusion and C, Ni foil insert of single layer was used.
-->less extra carbide The mechanical properties of the base metals of these
in cladding clad plates are shown in Table 6. The properties were
examined on the specimen taken at 500mm from the top
of the plate, at the center of the width and thickness. The
condition 450°C-20MPaH2. The comparison of overlaid
tensile tests were carried out for rolling(L), crossing(C),
and rolled clad steel on factors affecting cracking
and normal(Z) direction for plate A while specimen were
sensitivity is summarized in Table 4.
taken and tested only for crossing(C) direction for plate B
and C. The result contented ASTM A387 Grade 22, Class
PROPERTIES OF LARGE-SCALE
2 specification. The impact test was also performed for the
PRODUCED CLAD STEEL
specimen from the same position of plates, and the
transition temperature for each plate was measured to be
Procedure and mechanical properties
Based on the property evaluation for laboratory- lower than -20°C, as is shown in Table 6. Shear stress at
the cladding I base interface was larger than 280MPa for
-manufactured clad steels, three heavy gauge
both single- and double-insert portion of the clad plate.
disbonding-resistant clad plates were produced in trial in
For plate B and C shear stress was larger than 350MPa.
actual plate mills. The chemical composition of the
cladding and the base metals are shown as plates A, B, and
Effect of V addition in the base steel and
C in Table 5. The cladding metal of type 321(
microvoid at the interface
18Cr-10Ni-Ti) was used for plate A, and type 347(18Cr-
The resistance to disbonding of the trial product A is
Table 5 CHEMICAL COMPOSITION OF THE CLAD STEELS HOT ROLLED IN THE TRIAL MASS
PRODUCTION (MASS%)
c Si Mn p s Ni Cr Mo Nb v Ti B sol.Al total N
A 0.14 0.16 0.57 0.006 0.002 - 2.46 1.00 - - 0.008 - 0.018 0.0036
Base metal B 0.13 0.16 0.57 0.006 0.002 - 2.44 l.02 - - 0.012 - 0.010 0.0028
c 0.08 0.15 0.53 0.006 0.002 0.47 2.34 0.97 - 0.19 0.012 0.001( 0.013 0.0033
Cladding metal A 0.041 0.55 l.61 0.026 0.002 10.06 18.37 0.25 - - 0.427 - 0.055 0.0164
B, C 0.050 0.60 l.59 0.029 0.001 10.96 17.49 0.20 0.74 - - - 0.005 0.023 l
Table 6 MECHANICAL PROPERTIES OF THE BASE STEELS IN THE TRIAL MASS PRODUCTION
Location Di rec ti or Tensile test CXt) •> Impact test (Xt)b> Shear strength
YS(MPa) TS(MPa) El(%) vEs(J)c> vTs ('C)" > (MP a)
L 417 580 31.6 294 -50 -
A Top c 417 579 31.7 280 -48 293
z 417 576 25. 5 225 -29 -
Bottom - - - - - - 281
B Top c 446 592 30. 3 398 -77 364
Bottom c 460 572 31. 4 - - 353
c Top c 485 596 61. 2 340 -76 380
Bottom c 460 573 63. 2 325 -70 398
ASH! A387-22-2 ~310 515--690 ~18 - - -
a) A and B : 9~, GL=36111l11 , C : lH, GL=50mm
b) 2mmV-notch charpy
c) vEs : Upper she! f energy
d) vTs : 50% FATT
shown and compared with overlay welded clad plate in value than the laboratory value under the same condition.
Figure 8. The specimen for hydrogen charging was taken This was attributed to the difference in the soundness at
at the center of width in the thickness ( 46+4)mm, and lhe interface as rolled; the mill-produced plate A showed
hydrogen charged at 400°C- l OMPaH2, and at 450°C microvoids of approximately 4% in area fraction while no
-20MPaH2. Although disbonding was not observed at microvoid was observed in the clad plate used in
400°C-lOMPaH 2 for both single- and double-insert laboratory. This indicated that the micro voids also gave
portion, the trial product showed disbanding of 20-30% in the accumulation site for hydrogen and the lower reduction
area fraction at 450°C-20MPaH2, which was a larger applied in producing clad plate A led to cause a small
amount of microvoid resulting in a high sensitivity to
disbonding. However, still the disbonded area fraction in
plate A was approximately a half of that in the overlaid
~
~
50 l/2W 0 400°C, IOMP0H2
clad plate at the condition 450°C-20MPaH2, which was
0
<ll
6. 450°C, 20MP0H2
0 40 shown in Figure 9. The relationship between the fraction
-0 of microvoid at the interface and the fraction of disbonded
<ll Top
-0
c 30 area after hydrogen charging is shown in Figure 10.
0
.0
Vl /:::,.
'6 20 Hydrogen charging was carried out also on plate B and
'+-
0 C, and the detected disbanded area fraction is shown in
c 10
.Q Figure 9. Plate B and C did not show dis bonding after
t5 hydrogen charging at 400°C-lOMPa. At 450°C-20MPaH2,
0
.... 0 0 0 0 0
LL
C D Overlay welded a small area fraction (2%) of disbonding was observed in
A B
clod plate plate B, although no disbonding was observed in plate C
Single insert Double inserts
which was with V-added base steel. V addition to the base
Figure 8 RESISTANCE TO DISBONDING IN THE steel also improved resistance to disbonding in the large
TRIAL PRODUCT (steel A) scale trial production.
l::lad met:i-1 ~ 22 • ~ "'' ! j ____ j____ used as welding materials for the first layer and the second
CONCLUSION
5.0 o B
~
~ D. C head plate Heavy gauge, y-stainless-steel clad plate of Cr-Mo steel
0
Q)
...... 4.0 • Overlay 1Nelded clod plate by hot rolling was thermally hydrogen charged in an
0
"'O
Q)
"'O
.--..---· autoclave, and the resistance to the interface disbonding
was evaluated and compared with conventional overlaid
c
.8V1
3.0
.---- clad plate mainly in terms of disbonded area fraction at the
/6-06~
-
'6
0
c
2.0 interface. The result indicated superior resistance to
disbonding of rolled clad plate to that of overlaid clad plate
·n
0
0
I .0 both in single and in multiple hydrogen charging. The
'-
lJ._ formation of martensite phase caused by dilution of
0 cladding weld metal in base steel, and the carbide
2 3 4 5 precipitated at the grain boundary in weld metal by
Number of hydrogen charge diffusion of carbon from the base steel, were considered to
be the main factors which led to disbonding in overlaid
Figure 12 RESISTANCE TO DISBONDING UNDER clad steel. Both of these were suppressed in rolled clad
REPEATED HYDROGEN CHARGING steel showing the superior resistance to disbonding.
Further improvement in resistance to disbonding in
vessels and examine the resistance to disbonding properly, rolled clad steel was achieved by applying higher rolling
the evaluation after applying the repeated hydrogen reduction and/or using Cr-Mo-V steel as the base plate.
charging to clad steels is important, since the hydrogen Sufficient resistance was practically verified as received and
accumulation at the interface during repeated shutdowns after welding I head plate forming on the trial production
most possibly accelerates the disbonding. Specimens used of rolled clad plate.
for testing disbonding resistance after repeated hydrogen
charging were plate B, plate Cat the head plate flange, and
the overlaid clad plate. The cycle consisted of 48h heating REFERENCES
at 455°C-8.5MPaH2, cooling at 0.75°C/min down to
350°C, and continuing air cooling to the room [!]Naito et al.: Journal High Pressure Inst. Japan,
temperature. The cycle was repeated 5 times with l week's Vol.18, 1980, p.263.
interval between each, measuring the disbonded area by [2]Kikuta et al.: Quart. Journal Japan Welding Soc.,
U.T. at the end of each cycle. Vol. 2, 1984, p.463.
The result is shown in Figure 12. For the overlaid clad [3]Kikuta et al.: Quart. Journal Japan Welding Soc.,
plate disbonding was recognized after the first cycle, area Vol. 2, 1984, p.470.
fraction was increased to more than 2.5% after the second [4]Imanaka et al.: Kawasaki Steel Giho, Vol.17, 1985,
cycle, and further continuous enlargement of disbonded p.84.
area followed during the repetition. For plate B disbonding [5]Kawanami : Journal Japan Soc. Technol. Plasticity,
of less than 0.2% by area fraction was observed after 4 Vol.32, 1991,p.3.
cycles, and for the head plate of plate C no disbonding was [6]Naito et al.: Journal High Pressure Inst. Japan,
observed after 5 cycles. Vol.18, 1980, p.271.
From these results, the superior resistance to dis- [?]Sakai et al.: Journal Iron and Steel Inst. Japan,
bonding of the rolled clad plates to the overlaid clad plate Vol.72, 1986, p.1375.
was practically verified also in repeated hydrogen charging.
321) , the ferritic stainless steels • but also Typical results of shear and ram tests are also
more sophisticated alloys, high corrosion re- presnted on Figure 1. For the shear test , the
sistant products including duplex and super minimum ASTM 263-5 recommendation is
duplex stainless steels , super austenitic 140MPa . Both explosively bonded clads and
stainless steels • nickel base or copper alloys or hot rolled bonded clads present generally shear
more noble metal as titanium and zirconium . strength higher than an average value of 250
Other combinations are also availlable for spe- MPa.
cific applications [l] .
THEORIDCAL ASPECTS ABOUT
The two components , base metal and clad me- CLADDING AND HYDROGEN EM-
tal • are usually bonded together by explosion BRITTLEMENT
or hot rolling , or a combination explosion
and hot rolling . Table Il gives lists of hydrogen diffusivity
Mechanical properties of the bond ; Ultrasonic coefficients at 20 and 100°C for low alloyed
inspection is always used to check the bond normalized steels and austenitic stainless
quality of clad plates . In addition , to confirm steels. It is shown that the hydrogen diffusion
o-
coefficient is at least 1 5 times lower for aus-
Clad metal Process Plate th. Bend tests results tenitic stainless steels (fee structures). This
means that even with a high activity of
hydrogen on the inner wall of an austenitic
grade base clad comp. Tens. Side
clad plate reactor , only a little quantity of
hydrogen diffuses through the clad metal
UR 825 expl+roll 88 4 no crack no crack no crack towards the base material.
Moreover the time needed to obtain the steady
Monel400 hot roll 13,5 3 no crack no crack nocrad state (equilibrium conditions) through 1 mm
thick wall is of about 30 years at room
URB6 hot roll 25 3 no crack no crack no crack temperature and 115 days at 100°C.The use of
higher thicknesses of clad metal requires much
UR 625 hot roll 11 2 no crack no crack no crack
longer time to obtain the steady state since e
Titanium expl+rol 65 3 nocrad no crack no crad is proportional to square of the thickness (0 =
Table I : Examples of bend tests results T2/D)*. A two mm thick plate results thus in 4
x greater time of steady state.
* T = Thickness
the bond strength several destructive tests are
carried out . Table I present the typical results
Moreover the hydrogen activities resulting
of compression , tension and side bend tests .
from corrosion reaction on the surface of a
dad plate (stainless steel) or a C-Mn steel are
Figure 1 : Ram and Shear test results
MP a To Low alloyed Stainless
600 oc steel steel (304)
400
Diffusion 20 10- 1< D<lO-o tno- 11
coefficient 6
100 5.10- < D<5.10- 5 # s.10- 10
200 2
unit: cm /s
Steady state time 20 10-j< t<2.10-=> t # 10~ (30 years)
0 for a 1 mm thick 100 2.10-2< t<2.10- 3 t #2.10"/)
13 Cr 316L UR B6 UR625 C22 plate unit : s (115 days)
different since the corrosion rate of the carbon T bm =thickness of the base metal
manganese steel is much higher. Due to the T clad=thickness of the cladding
very low corrosion rates for stainless steels ,
very low hydrogen activity is expected . Then (for example with Sbm = 0.01 cm3 H /cm2/
2
is there risk of hydrogen embrittlement for the
base material ? First of all the hydrogen atm2 , Sclad = 103 Sbm• Tbm = 50 mm and
activity on the inner wall of the clad metal T clad = 6 mm we obtain at room temperature
already explained will be less than that of a ba-
se metal due to the limitation of the corrosion Ra=0,2).
reactions (cathodic part of the phenomena).
Then the hydrogen activity, as explained in For that reason, clad plates provide a safe
figure 2 is lowered through the thickness of solution in the most critical service conditions
the clad metal. At the interface due to the which can produce hydrogen embrittlement of
differences of diffusivity we observe a large <;-Mn base steels. To investigate the condi-
reduction of the hydrogen activity. This drop t1on.s and behaviour of material in hydrogen
of hydrogen activity can be obtained by the envrronment , C.L.I research center has deve-
following relationship : lopped a computerized code to determine the
hydrogen concentration in the pressure vessels
wall according to the thermal cycles . Such co-
de is very effective to determine the desirable
1 coo~ing conditions ( cooling rate , steps ... )
particularly for the petrochemical installations
1 + Sbm . Dlm:L . I..ct00
S clad D clad T bm
CLAD PLATES MATERIAL
with:
RaH =reduction of hydrogen activity at the
'I?e cladding technology included an optimisa-
bond tion of the chemical compositions and heat
Sbm =hydrogen solubility in the base metal treatment of both base metal and clad metal
Sclad=hydrogen solubility in the cladding for each specific application. Sevral high cor-
material rosion and high strength clad plates have been
designed for oil and gas as well as off-shore
Dbm =hydrogen diffusion coefficient in the
applications [2] .
base metal
Dclad=hydrogen diffusion coefficient in the
cladding
Grade Rp0,2 Rm A% Kv (T 0 C)
Th=50mm MPa MP a J
Solid rsss'!
30
Clad~ 0
Solid Clad Plate CP
20
Quenched Q+Temper
10
tentials results determined in 30 g/l NaCl solu-
tion at 60°C are presented on figure 3 and the
CPT in ferric chloride solution according
316L UR 825 UR UR UR UR UR ASTM G48A and B on the figure 4 and 5 .
45N B6 B6N 52N+ B26 625
UR B6N specificaly designed to improve the
Figure 7 :UR B26· test according SEP 1877 n for PROJECT APPLICATION GRADES
different state FIELD base clad
Sleipner gas separator SE500 URB6N
Sleipner gas separator SE500 URB6N
East Brae gas separator SE500 316L
East Brae gas separator SE500 316L
Ekofisk heat exchanger SE500 Ti
Embla gas separator SE500 URB6N
Troll gas gas separator SE500 316L
Solid Clad Plate CP
Quenched Q+Temper Froy gas separator SE500 URB6N
N'Kossa separators SE500 316L
phases stability presents an excellent corrosion
behaviour even after several heat treatment si- Table V : SE 500 clad plates Industrial applications
mulating fabrication steps . As shown in figure
3 , the pitting potential at 60°C appears to be
unaffected by several heat treatments which is
not the case for the nitrogen free UNS N08904 UK and Norway sectors) , and Table V pre-
( URB6) less stable after several heat treat- sents the main applications in off-shore . Seve-
ments . No detrimental effects of heat treat- ral projects using Superelso 500 + URB6N
ment are reported on the pitting potential of clad plates have shown the interest of this so-
UR B6N. This grade appears to be better adap- lution to improve the efficiency and reliability
ted to clad plates than standard N08904 . It is of the installation as well as the saving of
also important to point out that the high locali- weight and cost . Clad plates are used in many
sed corrosion resistance of UR B6N , particu- fields like ALWYN FIELD (TOTAL),
larly when it is compared to AISI 316L and Trenganu (PETRONAS - Malaysia), Lybian
Alloy 825 as show for ASTM G48A test on fi- (AGIP), BRAE B. (MARATHON),
gure 3. GULLFAKS A, B,C, OSEBERG 2,
SLEIPNER (STATOIL), FULMAR (SHELL),
SNORRE (SAGA) and BP MILLER (BP).
UR B26 -UNS 08926 - This alloy is a high Several refinery vessels are also built from
corrosion resistant material , particularly well clad plates like vacuum towers (ESSO -
adapted in high chloride media . Several type SHELL), Flexicokers (ESSO), heat
of tests have been carried out after cladding exchangers (ULTRAMAR - CND), Coke
operations , solution annealing and several ty- drums (CONOCO) or hydrotreater
pes of heat treatments in order to simulate (PETRONOR - Spain). Clad plates may also
temper and post weld heat treatments of the be used for carrying oil and gas from the field
carbon steel . The results presented on Figure to the continent in order to be treated on shore.
4 to 7 show the high corrosion properties of
this alloy . The tests in chloride media and
acidic environments show that the high corro- CONCLUSION
sion properties of UR B26 remain at level
equivalent to the solid plates after the heat This paper has presented :
treatments of the base material .
-Some theoritical considerations about
REFERENCES AND APPLICATIONS: cladding and hydrogen embrittlement. It has
been shown that austenitic stainless steel clad-
Figure 8 give typical applications of clad dings are very effective in reducing the
plates for offshore applications in North Sea ( hydrogen activity in the carbon manganese
J. G. Banker
Explosive Fabricators, Inc.
Louisville, CO
M. S. Cayard
CLI International, Inc.
Houston, TX
The resistance of the dad metal to disbonding is through the sample is restricted during cool-down by
dependent upon a variety of variables including the OD overlay, hence diffusion of hydrogen is
materials processing methods, clad manufacture tech- unidirectional, more accurately simulating service
niques and service conditions. conditions.
The second test used a newly proposed specimen Component Metal Variables
design which provides a more realistic simulation of
hydrogen diffusion during cool-down [1]. This The composition, manufacturing method, and
specimen, Figure 2, was 3" OD x 3" thick (75 mm OD heat treat condition of the component metals are
x 75 mm thick) with stainless steel on one flat face critical for hydrogen resistance. These variables are
and the OD. The stainless steel on the face was extensively presented elsewhere and will not be
applied by explosion bonding; the stainless on the addressed in detail here [1-3, 6]. The test specimens
OD was applied by weld overlay. Hydrogen diffusion were produced from Cr-Mo steel, SA-387,
manufactured using procedures incorporating both The stainless steel is then deposited onto the interior
vacuum degassing and sulfide shape control surface of the alloy steel components using one of
processes. The cladding metal was austenitic wrought several potential welding processes. In many cases,
plate of the niobium stabilized grade SA-240 Type the overlay metal is deposited in two sequential
347. layers to compensate for dilution in the initial layer
by the unalloyed base metal. The interface
The test specimens of this program were given characteristics are very dependent on the overlay
the heat treatments typically experienced in process and heat input parameters. Incorrect
manufacture and fabrication of refinery vessels. The selection or control of welding parameters can result
base steel was normalized and tempered prior to in formation of hard martensitic regions adjacent to
cladding. The stainless steel was solution heat treated the fusion line in the Cr-Mo steel which are
and water quenched prior to cladding. The clad susceptible to hydrogen cracking.
product was given a simulated Postweld Heat
Treatment (PWHT) at 1,275°F (690°C) for 12 hours Explosion Clad. Explosion cladding uses the
prior to hydrogen testing. energy of an explosive detonation to create metal-
lurgical bonding between the stainless steel and base
Clad Manufacture Variables metal [7]. Prior to bonding, the component metals
are heat treated to achieve optimum performance
Clad for hydrogen service refinery vessels is characteristics. Typically, the stainless steel is solution
primarily manufactured using one of three different heat treated and water quenched; the base steel is
manufacturing technologies: roll bond clad, weld normalized and tempered or quenched and tempered.
overlay, and explosion clad. Each of these processes During the bonding process the faying surfaces are
results in distinctly different interface metallurgical dynamically cleaned by jetting away the surface
characteristics and potentially significant differences layers. The nascent surfaces are then brought
in hydrogen disbonding characteristics. Although not together under pressures of several million psi. There
addressed herein, there are significant variations in is no bulk heating of the metals, subsequently the
the cladding processes of individual manufacturers heat treat conditions are not altered by the bonding
which may result in further variance in the hydrogen operation and no measurable diffusion occurs. The
service performance of the specific products. clad plates are then formed into cylinders and heads
for subsequent vessel fabrication. If hot forming is
Roll Bond Clad. Roll bond clad is produced by required for fabrication some diffusion will occur in
hot rolling a multilayer pack, consisting of cladding the interface region; however, since the combined
metal plates and base metal plates appropriately time and temperature exposure of these operations is
prepared for bonding. Subsequent to the hot rolling typically short, interface carbide concentrations are
operation the composite plates are heat treated to much lower.
achieve the required metallurgical conditions for each
component metal. Typically the clad plate is first The three different processes result in clad
heated to the stainless steel solution annealing metals that differ significantly in cladding metal
temperature followed by air cooling. The plate is then chemistry uniformity and corresponding corrosion
heated to the normalizing range for the Cr-Mo steel resistance, in bond strength, in bond uniformity, and
followed by air cooling or quenching. Lastly, the plate in component metal mechanical properties. The effect
is tempered at the appropriate temperature for the of these variables upon hydrogen disbanding is
base steel; a temperature which is normally above the currently not well known. Service experience of
vessel PWHT. The clad plates are formed into vessels constructed by methods other than weld
cylinders and heads for subsequent vessel fabrication. overlay has not been well documented, consequently
The manufacturing and heat treating sequence may laboratory testing is being used to evaluate clad
induce significant diffusion between the component disbonding characteristics.
metals resulting in a concentration of carbides in the
interface region. Service Operation Parameters
Weld Overlay. Weld overlay is typically produced Exposure Conditions. The in-service exposure
in a very different sequence of operations. The base conditions influence the probability of hydrogen
steel is manufactured to the required condition and disbanding. The potential for disbonding increases
fabricated into and heads prior to overlay. with increasing temperature and/or hydrogen partial
pressure. The higher the service temperature and/or was produced in the center of the explosion bonded
hydrogen partial pressure, the higher the hydrogen dad face, Figure 3. Ultrasonic inspection indicated no
concentration available at the base metal/dad metal other nonbonds in the clad face or the overlayed OD.
interface during shutdown conditions. Consequently
the pressures that can be generated by internal
accumulation of molecular hydrogen are much
greater. The pressure and temperature conditions of
this test program were chosen to represent a
significant percentage of refinery vessels as
/
documented elsewhere [8]. I
/
'
Cool-Down Rate. The cool-down rate used during
shutdown has a dramatic effect on the susceptibility
of hydrogen disbonding, with increasing cooling rates
increasing the potential for disbanding. Under higher
cooling rates the diffusion rate of hydrogen in the Cr-
Mo steel decreases dramatically, potentially creating
conditions in which the hydrogen cannot outgas to
the exterior surface. The result is a high
concentration of supersaturated atomic hydrogen in
the steel which can coalesce at bond anomalies
creating disbanding. At sufficiently high cooling rates,
it is expected that any of the stainless steel/Cr-Mo
clad products can be forced to incur hydrogen
cracking of the steel and/ or separation of the dad.
The 300°F (167°C)/hr rate used in this test program
is considered to be more extreme than can possibly
be achieved in realistic refinery operation. Intentional Defect 1~
1'347 ~losive Clad
Base Material
047 Weld Overiay
TEST PROGRAM
:l
g
Specimens A and B exhibit pretest properties ~ 240
which are typical of production clad materials. Table w
1 presents the comparative data of Specimens A and
B with that of 59 clad plates produced during 1992
z
0
~
I 34 7 STAINLESS STEEL ---b-; - - Cr Mo STEEL
and 1993.
160 ,---..........,--,.----r--,---T- -
___ - - - ··-- ···· ---------·· - - I
,,_,,_,.
--.------.~-,---,-
0 030 0 025 0 020 0 015 0 010 0 005 0 005 0 010 0 015 0 020 0 025 0 030
Specimen A were examined for hardness and diffu- BONOZONE
sion in the interface region. The specimens were
DISTANCE FROM BONDZONE
given the same heat treatments as Specimen A prior
to examination. Figure 6 presents the microhardness
scan across the bond zone. Figure 7 presents the scan Figure 6. Microhardness scan across explosion
for Carbon, Chromium, and Nickel across the bond bonded interface of Specimen A material after all
zone. The data indicate no hardness increase in the heat treatments of this program
REFERENCES
~
2 ° Current Solutions to Hydrogen Problems in Steels, p.
--:::x:-7~S~
01
12
I
() 11- +"" 340, (Metals Park, OH: American Society for Metals.
000
1982).
---,--~,-----,..------,--,------,-----r
10 -·- ...--,--------,---- - . - - 006
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14
t
!101i020NE
DISTANCE FROM BONDZONE (INCHES)
TABLE I
NOTE:Production shear data is taken from 59 clad plates produced in 1992 and 1993 and is exclusively
from clad plates of stainless steels types 321 and 347 clad to Cr-Mo steels SA-387, Grades 11 and 22,
Classes 1 and 2. Production UT data is from all clads produced in a 2 month period of 1993, yielding
18,000 sq-ft (1,700 sq-m).
C.CASCIARO, P. MARINELLI
Belleli Ricerche Scpa
TARANTO, ITALY.
475 •c
:ai:: 0.4
0
..0
rJl
0.3
Q
150 bar 0.2
0.1
Typicai cycie of 0 p
G H L M N 0
hydrogenation
Welding Procedure specifications
12
.....=
-e
u
u
~
~
50 1111 M (AW)
8 G(AW)
500 M (PWHTreated)
G (PWHTreated)
40
400
2001----......
10•,.._--'-~..l..--I...~'---'---'~-'---'----'
~ 400
>
::r::
'-'
VJ
VJ
(].)
:::::
] 300
..c
c
I-<
u
......
~
200
1001...--'---'-~....__.___,~_.__._~.._~
1.0
Fig.8 Sample G: Cr profile across fusion line (750x)
0.8
0.6
t\2
ro
~on
.....,
i::: 0.4
"O
i:::
0
,.0
.,...,
VJ
Cl
0.2
0.0..._'-l~""-1~-'-...._....__._...._....__,__,__,__.___.
tested in this work: all peak hardness values are of about JO. C. Casciaro, M. Olive, A. Solina, R. Valentini
400 HV after PWHT and no meaningful relation has "Hydrogen diffusion and distribution in weld-clad
been found between peak hardness values and disbonding fabricated products with special reference to dis bonding",
susceptibility. Weld. Int., Vol.7, 1993, p.544.
Experimental tests and theoretical considerations about 11. J.W. Elmer, S.M. Allen, T.W. Eagar
the thermal cycle of the weld, lead to conclude that "Microstructural development during solidification of
different strip widths don't induce different disbonding stainless steel alloys", Metall. Trans.,20A, 1989, 2124
behaviours, when the appropriate sets of welding
parameters are used. The only difference is the number of
overlap zones that seem to be more critical, and therefore
a better behaviour is achievable with 90mm wide strips.
A similar research has been carried out, also, on
Electro Slag (ES) cladding procedure; it will be the
subjiect of a separate report. Results obtained by ES
cladding procedure were in many cases better than those
obtained by SAW procedure.
ACKNOWLEDGEMENTS
P. Bocquet
Creusot-Loire lndustrie
71202 Le Creusot, France
The study carried out covers the hydrogen induced The cooling down of pressure vessels from high
cold cracking phenomenon, so called disbanding, at the temperature and hitjh hydrogen pressure induces a hydro-
interface of 2.25 Cr - 1 Mo steel plates, cladded with gen saturation at room temperature at the overlay/base
stainless steel. metal interface region [1]. The disbanding phenomenon
The main factors responsible for the disbanding phenome- can immediately occur if this region contains sensitive
non are the presence of austenitic coarse grains and the microstructures such as :
thin martensitic layer near the interface. It seems that - a thin martensitic layer carburized by the postweld
the thin martensitic layer has the more detrimental heat treatment (PWHT),
effect. The solutions using a low heat input to limit - carburized austenitic coarse grain boundaries near
only the amount of austenitic coarse grains are partly this thin martensitic layer.
efficient under the worst conditions. We have thus in- The disbanding phenomenon can propagate along these two
vestigated a very efficient solution consisting in the regions.
performance of a duplex heat treatment, which mainly Any parameter which generates these microstructu-
acts on the thin martensitic layer and prevents any res promotes the sensitivity to the disbanding phenome-
risk of disbanding. non [2] [4]. For instance, a PWHT at a higher tempera-
The Maglay process can be applied, using wide ture or the use of the electroslag welding process
strips, even in the presence of a large amount of aus- increases the risk of disbanding [3].
tenitic coarse grains and a strongly carburized thin
martensitic layer. CURRENT SOLUTIONS
TRIAL PROCEDURE
.
690'f - 30.,.H
r-"-1
J Tu_o~~- 1s H
- strip dimensions : 60 x 0.5 mm - 75 x 0.4 mm - 150 x I . ... +
0.4 mm, . \
- welding speed and welding current : conventional and I . I \
higher values,
- first layer microstructure : normal delta ferrite
. \ ... \
(approx. 6 %) and high delta ferrite (> 10 %) and low R.PWHT TIME S.PWHT TIME
silicon using a special 309 L strip,
- first layer chemical composition : 309 L or 347 type,
- one or two layers. TEMPERATURE ('Cl TEMPERATURE ( 0 0
The table 1 shows the welding conditions applied
and the chemical composition obtained in the first 690°C - 30 H
layer. 1--------.,
I I 650°C - 30 H
I I :---·--·--·-\
POSTWELD HEAT TREATMENT CONDITIONS (PWHT) I I 600°C . I
I \ {i"nT\ I
!
I
i
I I I I
Four types of PWHT were used, according to Fig. 1: I I I I I
I
I
.
I I I I
- a single heat treatment called regular PWHT (R.PWHT) - I I I I I
. I
I
I 11 I
690°C - 30 hours, I II I ! i
- a special heat treatment (S.PWHT) - 690°C - 30 hours + II I I \
TABLE 1. WELDING CONDITIONS - CHEMICAL COMPOSITION AND MICROSTRUCTURE OF THE FIRST LAYER
'O
'VHS ES\I 75 x O.'I MM/
30 ,Lf'.a. 00
I
'O'•
/' 6
20
I
I
I
I
I
-A ...........
,, .... .,..,,,,,. ..... ------')<
10
I
I 0
• ...
,. ,, ...
;.~
,_,.,,,, , ... "
A/
100 200 300 •100 500 600 700 COOLING RATE ·uu
These results lead to two important observations : Influence of the Type of Weld Overlay
- for comparing the results, it is necessary to use the
same cooling rates. In the opposite case, the risk of - for the slow cooling rate (150°C/h), no difference is
errors may be important, observed between all the procedures (Fig. 5). The criti-
- a cooling rate capable of discriminating weld overlay cal disbanding threshold was not reached.
procedures can be applied.
TABLE 2. DISBONDING RESULTS
EFFECT OF THE WELDING PROCEDURE - for the medium cooling rate (250°C/h), a significant
difference was observed between the welding procedures
In order to understand the effect of the welding (Fig. 6). The use of wide strips (150 mm) increases the
procedure, we have analysed the interface behaviour under disbanding sensitivity.
the worst conditions (R.PWHT 690°C - 30 hours).
D!SBONDING m DISBONDING (%)
50
"t
40 j 40
WllJE TRIPS
30
~
30
NllRR~ STR f PS
i
20
20
W!IJE STRIPS
= = cs:=
10 10
= =
SPECll\l
30~~
0
0 PROCESS SllW1 Sf\112 ES\11 ES\12 ES\14 ES\15 ES\16 ES\17
~_09L+
308L ~giht
3091.+
o~~rf 308t ;:;r
µvoe>
308L •>34]"' 308l 347
60. 5 • 75 • ZS • 150. 150 • J'.>U • 150 •
•Hw n .....
''HH h 4HH tJ 4HH 0.4191 kJ.•1m l,411'1 ~J.4HH
Fig. 5 Effect of the welding procedure Fig. 6 Effect of the welding procedure
(specimen cooling rate = 150°C/h) (specimen cooling rate = 250°C/h)
. for narrow strips (less than or equal to 75 mm), the In fact, the cooling rate threshold to discriminate
disbanding level is the same in SAW ans ESW processes, the effect of the welding procedure varies with the strip
even using various parameters (SAW2 and ESW2 tests). size. If the cooling rate is slow, any welding procedure
Micrographic examinations show small amounts of marten- cannot be discriminated. The medium cooling rate only
si te in the weld overlay near the carburized thin mar- discriminates the welding procedures using wide strips.
tensitic layer (Fig. 7). The presence of martensite does The high cooling rate only discriminates the welding
not decrease the disbanding level. procedures using narrow strips.
DISBONDJNG CU
H/t~ROW STRIPS WIDE STRIPS 0
50 100 150 200 250
50
~j
Fig. 9 Effect of the welding heat input on
QO
= the austenitic coarse grain ratio
30
Furthermore, strips with special chemical composi-
tions give a smaller percentage of planar grains because
20
= the high level of delta ferrite in the first layer pre-
vents the generation of austenitic coarse grains near
SPECIAL the thin martensite layer (Fig. 10).
309 L STRIP
10 ESW4 test
(X 200)
The use of a narrow strip enables to apply a low On these figures, the points representative of
heat input, thus to obtain a better austenitic micro- strips with special chemistry, are outside the scatter
structure. After R.PWHT, the effect on the disbanding band, containing all the other points. In fact, the dis-
level is depicted on Fig. 11 thru 13. banding phenomenon depends on the coarse grain ratio
(Fig. 14).
D!SBOND!NG m
DISBONDING (%)
50
50
110
SPEClf'EN COOLING RATE = 675°C/H
110
30
30
~~~~~ ~g:g:~: SPEW'EN COOLING RATE 250"C/H
20 DESW1 75•0,qMH
llESW 2 75 • 0.4 MM
)( ESW 3 75 • O,q MH 20
6ESW q 150 x 0,q MH
10 AESW 5 150 • O,q MH
~HSll 6 150 • o,q MH
TESll 7 150 • 0.4 MH 10
?HS ESW 75 X 0.4 MH
SPECil'EN COOLING RATE = 150°C/H
50 100 150 200 250 300 HEAT l NPUT
(KJ/CH)
10 20 30 110 50 AUSTEN ITI C COARSE
GRAIN RATIO (%)
Fig. 11 Effect of the welding heat input
(specimen cooling rate = 675°C/h) Fig. 14 Effect of the austenitic coarse grain ratio
on disbanding
D!SBOND!NG (%)
Some examinations were performed with a transmis-
e SAW1 60 • 0.5
0 SA\12 75 • O,q MH
MM sion electron microscope using accurately located thin
o ESW1 75 • O,q MH foils in order to observe the carburized martensitic
50 111 ES\12 75 • 0.11 MH layer. Fig. 15 shows Maglay interfaces using strips with
6 ES\ill 150 • o,q MM
A ESW5 150 • o.q MM standard and special chemistries. No difference can be
® ESW6 150 x 0.11 MH observed. In both cases, the precipitates M23 C6 are
110 "ES\17 150 • o.q Ml1
similar. Thus, the sensitivity of the carburized layer
is probably similar because the same PWHT was performed.
Therefore, it seems normal to only observe the effect of
the austenitic coarse grain microstructure.
30
20
10
.. 0
Ill 0
Fig. 13 Effect of the welding heat input Fig. 15 Transmission electron microscope examinations
(specimen cooling rate = 150°C/h) (martensitic thin layer microstructure)
.7~~
- a duplex PWHT (D.PWHT). This type of hIBat treatment
consists in performing a first PWHT at normal tempera-
ture (690°C approximately) to obtain the best mechanical 20
properties for the 2.25 Cr - 1 Mo material. Then after
~·_.,?s.PWHT
cooling down to room temperature a second heat treatment
is carried out at a lower temperature to de-sensitize
the interface.
- a single PWHT at a lower temperature (LT PWHT), for
10
0
.
/
/6
/
e
//
("
\
.. ~ /
- _,/
/
./·
/
D.P~
upgraded 2.25 Cr - 1 Mo steels (e.g. ASME Code Case 100. 200 300 HEAT INPUT
(KJ/tH)
1960). Decreasing the temperature leads to an interface 2 LAYER OVERLAY WITH NORMAL 309 l STR IP
microstructure less sensitive to the hydrogen induced
cracking and to less stresses in the interface region.
FRAMATOME has studied the effect of these solu- Fig. 17 Effect of the PWHT type
tions. (specimen cooling rate = 250°C/h)
Effect of a Duplex Postweld Heat Treatment Electron examinations using thin foils were carried
out (Fig. 18). After RJPWHT (690°C - 30 hours), the
As the disbanding level depends on the welding fresh martensite was not revealed, probably due to the
heat input, we have analysed the effect of the PWHT type low amount of the thin martensitic layer which is auste-
for various welding procedures and each cooling rate nitized and to the difficulties met to accurately posi-
applied to disbanding specimens. tion the thin foils.
2 layer-overlay with standard strip. For various
cooling rates applied to disbanding specimens, the fol-
lowing behaviour was observed (Fig. 16 and 17). M23 C6
DISBOND!NG (%)
0. 22 )-Im
100
\ /--:/
200 300 HEAT INPUT
Disbanding tests were carried out after a PWHT at
6505C - 30 hours (called LT PWHT) for three types of
(KJ/CH) weld overlay :
- 2 layer ESW process using special 309 L strips ( 75 x
Fig. 19 Effect of the S.PWHT for various 0.4 mm and 150 x 0.4 mm)
welding procedures - 1 layer ESW process using standard 347 strip (150 x
(specimen cooling rate = 675°C/h) 0.4 mm).
The Fig. 22 compares the disbanding level for
each overlay, after cooling at 675°C/h. The efficiency
DISBONDING CXl
of the LT PWHT is obvious, even for the one layer weld
x 2 LAYER OVERLAY WITll SPECIAL 309 L STRIP <D.PWllT CONDITION) overlay which contains a lot of austenitic coarse
grains (ESW7 test). The disbanding level is less than
50 or equal to 10 %.
DISBONDING <Xl
30 50
20 40
D.PWllT SCATTER BAND OF 2 LAYER
10
Fig. 20 Effect of the D.PWHT for various
welding procedures
(specimen cooling rate = 675°C/h) 0
PROCESS SW7 C150•0.4MHl
/.
PWltT
30 H 30 H SPWllT DPWflT 30 H 30 H DPWHT 30 H 30 H SPWllT
40 TYPE RPWllT Tf'lffr l«'lfrr TNlT TPlfJT
30
~--
- the amount of austenite at 650°C is less than that
10
obtained at 690°C. Thus the fresh martensite is less
-~----01 e S.PWllT important after cooling down to room temperature. There-
I o D.PllHT fore, the sensitivity to disbanding generation after
:./0 I
I
LT PWHT is probably less important than after R.PWHT at
10 20 30 40 50 AUSTEN ITl C COARSE 690°C but more important than after D.PWHT.
GRAIN RATIO (%)
- the carburization of the interface (thin martensitic
layer and austenitic coarse grain boundaries) is less
Fig. 21 Effect of the austenitic coarse grain ratio important than after R.PWHT, S.PWHT and D.PWHT. Thus
for S.PWHT and D.PWHT types the sensitivity to disbanding propagation is less im-
(specimen cooling rate = 675°C/h) portant in the coarse grain boundaries.
As a conclusion, we can assume that the LT PWHT is [3] SAKAI T., ASAMI K., KATSUMATA M., TAKADA H. and
an other solution to significantly decrease the risk of TANAKA 0. : "Disbanding of Weld Overlay in Pressure
disbanding. But the best solution would be the following Vessels and its Prevention". Proceedings of the
duplex PWHT : 650°C with cooling down to room tempera- First International Conference, American Society
ture+ an otherPWHT at 600°C. In this PWHT, the genera- for Metals, WASHINGTON DC, November 1982, pp. 340-
tion and propagation of the disbanding phenomenon would 348.
be very low.
[4] COUDREUSE L., GERARDIN D., CHARLES J., BLONDEAU R.
SOLUTIONS PROPOSED TO PREVENT THE DISBONDING PHENOMENON and PALENGAT R. : "Etude du Phenomene de Disbanding
a l'Interface entre Divers Revetements en Aciers
We want not to apply solutions that would lead to Inoxydables Austenitiques et l'Acier 2.25 Cr - 1 Mo".
obtain any martensite in the first layer of weld overlay Cinquiemes Journees d'Etude sur les Appareils a
or high welding speed, or high delta ferrite in the first Pression, AFIAP, Vol. 3, 1986, p. 179-194.
layer. For one or two layer overlays, using 75 x 0.4 mm
strips, we propose the following solutions to avoid any [ 5] GI TOS MF,, ROBINSON JL. and GOOCH TG. : "Disbanding
risk of disbanding : of Stainless Steel Cladding in High Temperature
- for the pressure vessels as stress relieved at 690°C, Hydrogen Service", internal report.
performance of a duplex heat treatment 690°C + 600°C -
5 hours,
- for the pressure vessels as stress relieved at a low
temperature, performance of the duplex PWHT : a first
PWHT at low temperature, then a second PWHT at 600°C -
5 hours.
CONCLUSIONS
REFERENCES