Вы находитесь на странице: 1из 10

Ramirez Supp May 2007corr:Layout 1 4/5/07 10:49 AM Page 125

WELDING RESEARCH
Weldability Evaluation of Supermartensitic
Stainless Pipe Steels
A look at the as-welded mechanical properties of supermartensitic pipe steel
welded joints, their susceptibility to hydrogen-induced stress corrosion,
and the influence of short postweld heat treatment
BY J. E. RAMIREZ

ABSTRACT. Significant interest exists in With the development of deep wells con- is effective to improve the resistance to
the use of supermartensitic materials for taining CO2 and H2S, many users began cold cracking or hydrogen assisted crack-
oilfield applications. However, the hard- using corrosion-resistant alloys (CRAs). A ing of the HAZ. Additionally, at this low
ness for both the weld deposit and the variety of CRA materials are now com- carbon content, the microstructure
heat-affected zone (HAZ) of different mercially available in a range of tubular changes from ferrite + martensite to sim-
material combinations may exceed the sizes. However, for reasons of cost, ple martensite with the addition of Ni
NACE requirement of 23 HRC (253 HV). martensitic-type stainless steels are pri- from 0 to 3%. Therefore, the high nickel
Therefore, further studies to quantify the marily used. content help limit the ferrite content or to
mechanical properties of welded joints Supermartensitic stainless steels have completely avoid its occurrence in the
and variables controlling the sulfide stress been developed with higher resistance HAZ and in the base material, which is the
cracking (SSC) resistance in the as-welded than conventional martensitic 13Cr to primary reason for any inadequate tough-
condition remain necessary both for eco- general and localized corrosion in CO2 en- ness or limited toughness in the HAZ of
nomical fabrication and to ensure reliable vironments at elevated temperatures martensitic 13Cr steels.
service operation. In this program, the as- (Ref. 1). They are also sulfide stress crack- The maximum hardness limit to avoid
welded mechanical properties of three dif- ing (SSC) resistant in environments con- the potential for sulfide stress cracking in
ferent supermartensitic pipe steels were taining a small amount of H2S (Refs. 1, standard martensitic 13Cr steels, as well as
compared using different welding con- 2–6). In a typical supermartensitic stain- for newer supermartensitic stainless steel,
sumables and welding procedures. The less steel, the C content is reduced to in CO2/H2S environments is 23 HRC per
susceptibility to hydrogen-induced stress below 0.03 mass percent in order to sup- the NACE MR0175. However, reliable at-
corrosion cracking of selected weldments press the reduction of Cr concentration in tainment of NACE MR0175 limits can be
under slightly sour conditions and under the matrix due to Cr carbide precipitation, extremely difficult, particularly in weld
cathodic protection was evaluated. Addi- about 5.5 mass percent Ni content is metals (Ref. 7). Therefore, a postweld
tionally, the influence of short postweld added to obtain the martensite single heat treatment (PWHT) operation proba-
heat treatment (PWHT) on the HAZ me- phase, and 2 mass percent Mo content is bly cannot be avoided.
chanical properties and on the mi- added to improve SSC and localized cor- However, the PWHT is more compli-
crostructure of three supermartensitic rosion resistance (Ref. 4). There are also cated with supermartensitic steels. The
stainless pipe steels was studied. The re- supermartensitic steels that contain presence of nickel depresses the Ac1 tem-
sults show that the filler metal and weld- around 4% Ni and 1% Mo. Thus, super- perature, so that tempering is necessarily
ing procedure combination affect the martensitic steels can play an intermedi- carried out at fairly low temperatures,
matching characteristics of the welded ate role between conventional martensitic 600°–620°C. Tempering reactions there-
joint and their toughness and ductility 13Cr and duplex stainless steels regarding fore tend to be sluggish while, even with
properties. Additionally, supermartensitic both the corrosion resistance and the ma- such a low temperature, the Ac1 may be ex-
steel welded joints with a maximum hard- terial cost. ceeded, leading to partial reformation of
ness ranging from 282 to 313 HV1 and In general, supermartensitic stainless austenite during heat treatment and for-
under an applied stress level equal to the steels have good weldability. At carbon mation of virgin martensite on cool out.
measured yield strength of the base mate- levels below 0.04% the hardness in the as- Maximum softening can normally be
rial did not crack under slightly sour con- welded heat-affected zone (HAZ) of achieved using two stage heat treatment
ditions. The result also showed that a short martensitic 13Cr base materials does not with a first intermediate cycle typically at
PWHT is effective in reducing the micro- exceed 350 HV 10, which is considered to 650–690°C and subsequent heat treatment
hardness of the HAZ to levels very close be the threshold of cold cracking. There- at 600–620°C.
to the hardness of the base metal in the as- fore, a decrease in C content up to 0.03% Therefore, it is desirable to obtain an
received condition. It is assumed that the estimate of the Ac1 temperature to set
main mechanism responsible for the PWHT conditions for different super-
changes of hardness with PWHT is the martensitic alloys. A fine balance must be
amount of reverted and stable austenite. KEYWORDS struck between achieving a temperature
high enough to give tempering, and low
Introduction Supermartensitic Pipe Steels enough to restrict austenite formation,
Heat-Affected Zone (HAZ) which will revert to virgin martensite on
Gradual depletion of easily obtainable Sulfide Stress Cracking (SSC) cooling. Some researchers (Ref. 8)
hydrocarbons has accelerated the produc- Postweld Heat Treatment pointed out that discernible hardening
tion of oil and gas from deep hot wells. (PWHT) after a heat treatment occurs only when
Welded Joints the Ac1 is exceeded by a temperature in-
J. E. RAMIREZ (jose_ramirez@ewi.org) is with terval sufficient to induce approximately
the Edison Welding Institute, Columbus, Ohio. 20% austenite. Unfortunately, a reliable

WELDING JOURNAL 125 -s


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:49 AM Page 126

WELDING RESEARCH
A B

Fig. 1 — Tensile properties of supermartensitic steel pipe S welded joints. A — UTS and yield strength, and B — percent elongation.

Table 1 — Supermartensitic Pipe Steels, Welding Process, and Filler Metals of softening, and the influence on tough-
ness need to be evaluated for different su-
S K N permartensitic steels.
Pipe inside diameter 228.6 mm (nominal) 5.2 in. (measured) 9.75 in. (measured) In this study, the as-welded mechanical
Pipe wall thickness 12.3 mm (nominal) 0.3 in. (measured) 0.50 in. (measured) properties of three different supermarten-
sitic pipe steels with a specified minimum
Pipe chemistry, wt-%
yield strength of 80 ksi were compared
Reported Analyzed Analyzed
Carbon 0.007 0.031 0.013 using different welding consumables and
Manganese 0.45 0.39 0.45 welding procedures. The susceptibility to
Phosphorus 0.018 0.012 0.017 hydrogen-induced stress corrosion crack-
Sulfur 0.0007 0.002 0.002 ing of selected weldments under slightly
Silicon 0.31 0.16 0.16 sour conditions and under cathodic pro-
Chromium 12.1 12.89 11.19 tection was evaluated. Additionally, the
Nickel 6.2 3.96 6.24
influence of short PWHT on mechanical
Molybdenum 2.53 1.01 2.58
Titanium 0.07 Not measured 0.015 properties and microstructure of Gleeble
Copper Not reported Not measured 0.57 simulated HAZ of the three supermarten-
Nitrogen 0.004 Not measured Not measured sitic stainless pipe steels was also studied.

Welding Process Filler Metals Experimental Procedures


SMAW E2553, ENiCrMo-3 E2209, ENiCrMo-3 E2209, ENiCrMo-3
(IN112) (IN112) (IN112)
Materials and Consumables: The gen-
GMAW MR-DP3W — —
GTAW DP-3WT — — eral dimensions and chemical composi-
tions of the three supermartensitic pipe
steels, and the welding process/filler metal
combinations that were used in this study
relationship between composition and Ac1 fully defined, but, for wall thicknesses above are listed in Table 1. The three different
does not seem to have been produced for some 10 mm, an external hardness of say pipe steels were identified as S, N, and K,
these alloys. Thus, further study of the 300 HV should be acceptable (Ref. 7). respectively. In general, the supermarten-
transformation behavior of commercial Therefore, widespread application of sitic steels S and N are of the type 13Cr-
supermartensitic steels in terms of both supermartensitic steels depends upon 5Ni-2Mo and the steel K is of type 13Cr-
the Ac1 and Ms/Mf temperatures, and of solving the challenge of girth welding 4Ni-1Mo. Additionally, these three steels
the tendency for supermartensitic steels to them in an economical way. At the same present some differences in the carbon
undergo reaustenization during temper- time, it should be recognized that blanket content as indicated in Table 1. The filler
ing heat-treatment cycles is necessary. impositions of hardness or other limits can metals DP-3WT and MR-DP3W (metal
On the other hand, in most situations, restrict welding procedures and increase cored), and the covered electrode E2553
the sour environment will be present only fabrication costs. Further studies to quan- are designated as superduplex. The cov-
on one side of the steel. In consequence, tify the mechanical properties of welded ered electrodes E2209 and ENiCrMo-3
there will be a gradient of hydrogen con- joints in the as-welded condition and vari- (IN112) are duplex and nickel-based elec-
centration through the thickness, from ables controlling the SSC resistance re- trodes, respectively.
high at the face in contact with the sour en- main necessary both for economical fabri- Welding: Welded sections of the three
vironment to very much lower at the free cation and to ensure reliable service supermartensitic steel pipes were pro-
surface. The risk of cracking will therefore operation. Short PWHT cycles are desir- duced with the welding process and con-
diminish through the thickness, so that able for productivity reasons. However, sumables combinations described in Table
higher hardnesses may be safely permitted recognizing the diffusional reactions in- 2. Each section had a gas tungsten arc
on the outside of the pipe or pressure ves- volved in tempering, the effect of short welding (GTAW) root pass made with DP-
sel. The relaxation possible has not been PWHT on partial reaustenitization, degree 3WT wire. The rest of each joint was com-

126 -s MAY 2007, VOL. 86


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:50 AM Page 127

WELDING RESEARCH

Fig. 2 — Charpy impact energy across welded joints of supermartensitic steel Fig. 3 — Microhardness profile near the cap of welded joints made with
S as a function of welding process – filler metal combination. supermartensitic steel S.

pleted using GTAW, shielded metal arc performed on all tested


welding (SMAW), or manual or mecha- specimens to observe
nized gas metal arc welding (GMAW) any cracks.
process. No preheat was used and 150°C Influence of Short
(300°F) maximum interpass temperature PWHT on Gleeble Sim-
was maintained. ulated HAZ: The trans-
Mechanical Testing: The mechanical formation tempera-
properties of the welded joints were char- tures, AC1 and
acterized by using tensile and bend tests at martensite start forma-
room temperature. Charpy V-notch im- tion, for each one of the
pact energy was determined at –30°C for steels were determined
the weld metal, fusion line, fusion line + 2 by dilatometric analy-
mm, and base metal locations. The weld sis. Heat-affected zone
metal, fusion line, and fusion line + 2 mm weld simulations were
Charpy samples were taken near the 3, 6, performed with a peak
and 12 o’clock positions of the weld. The temperature of roughly
weld metal, fusion line, and fusion line + 1350°C and a cooling
2 mm indicate the position of the middle rate between 800° and
of the sample relative to the welded joint. 500°C of 40°C/s simu-
Microstructural Characterization: lating normal GMAW
Microhardness profiles were determined heat inputs. A PWHT Fig. 4 — Microhardness of HAZs at different through-thickness locations
of welded joints made with supermartensitic steel S.
across some of the welded joints near the was performed in a
cap region. Additionally, the HAZ hard- Gleeble machine, at
ness of the joints W19 and W20 were de- temperatures equal to
termined at different through thickness Ac1, Ac1 + 40°C, and Ac1 – 40°C, for 5 and ksi). The percent elongation measured in
locations. The microstructure of the fu- 10 min. The resulting HAZ mechanical the tensile test samples of the super-
sion zone and HAZ of different welded properties (hardness and impact Charpy martensitic steel welded joints ranged
joints were characterized using light mi- V-notch energy) and microstructure of from about 6 to 33.
croscopy after proper sample preparation. three supermartensitic stainless pipe Normally, oil and gas companies re-
Corrosion Testing: Selected super- steels were evaluated. quire a yield strength overmatching in the
martensitic welded joints and base metals weld metal of circumferential joints in
were tested to evaluate their susceptibility Results and Discussions comparison with the pipe material. There-
to hydrogen-induced corrosion cracking fore, Ni-base electrodes (ENiCrMo-3) are
under cathodic protection (CP) and under Tensile Properties: The observed ten- not recommended to join X80 grade su-
slightly sour (H2S) service conditions. sile properties of the supermartensitic permartensitic steel pipes due to the re-
Table 3 shows a summary of the specimen welded joints are listed in Table 5 and sulting undermatching condition of the
identification, base material, loading some of the results are shown in Fig. 1. weld metal. The duplex and superduplex
method, specimen side in tension, and the The welded joint made with ENiCrMo-3 consumables provided an increasing de-
type of corrosion tests that were carried (IN112) or duplex filler metal failed in the gree of overmatching in the weld metal of
out. The test conditions under CP and weld metal. Welded joints made with su- X80 supermartensitic steel pipe joints.
slightly sour environment are reported in perduplex filler metal failed in the base The range of 6 to 33% elongation ob-
Table 4. After testing specimens were cut, metal. The yield strength of all the joints served in the supermartensitic steel joints
mounted, and polished. Visual inspection welded with either duplex or superduplex is influenced by the welding consumable,
with a magnification of up to 1000× was filler metal was higher than 550 MPa (80 matching characteristics of the joints, and

WELDING JOURNAL 127 -s


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:50 AM Page 128

WELDING RESEARCH
A B

Fig. 5 — Microstructures observed in the intermixing zone of the weld metals and in the HAZs of the supermartensitic steels. A — Cap; B — root.

Fig. 6 — Charpy V-notch absorbed energy at –30°C in simulated HAZ from Fig. 7 — Hardness of simulated HAZ from pipe steel S as a function of
supermartensitic pipe steel S as a function of thermal experience. thermal experience.

welding procedure used. Joints made with tride due to migration of nitrogen from dient across the fusion boundary of these
duplex and superduplex consumables the base metal into the niobium-rich weld welding joints with nonmatching consum-
using SMAW, GMAW, and GTAW metal. K steel has the highest reported ables. Additionally, the untempered con-
processes show an elongation between 20 level of nitrogen out of the three super- dition of the martensite and the present of
to 33%. The lower levels of ductility martensitic steels used in this study as strings of ferrite in the HAZ, as discussed
(14–15% elongation) measured in the shown in Table 1. in the following section, may have also af-
joints made with the nickel-base electrode Impact Absorbed Energy: It was ob- fected the absorbed impact energy.
(ENiCrMo-3) may be explained by the served, in general, that the absorbed en- The impact energy of the weld metal
concentration of plastic deformation in ergy in the welded joints at –30°C de- depends on the welding procedures and
the weld during the tensile test as a result creased from the base metal to the fusion filler metals used for welding the pipe
of the undermatching characteristics of line and to the weld metal. The average joints. Increasing average weld metal im-
the weld metal. Most of the plastic defor- impact energy of the supermartensitic pact energies in the range between 30 to
mation is localized in the weld metal in- steel pipes was about 200 joules. The de- 248 joules are achieved with the SMAW,
stead than in the complete gauge length. crease in impact energy in the fusion line GMAW, GMAW with heat input control,
However, the lowest level of ductility (6% and HAZ of the supermartensitic welded and GTAW process, respectively. De-
elongation) in the welded joints was ob- joints as compared to the base metals is creasing impact energies are obtained in
served in a K steel joint made with the considered to be a result of the mi- nickel-base, duplex, and superduplex weld
nickel-base electrode without heat input crostructural variation present near the metals deposited without heat input con-
control using a weave technique. The fur- fusion lines in these dissimilar metal trol. The minimum impact energy of the
ther decrease in ductility may have re- joints. The microstructural variation is a welded joints ranged from 25 to 181
sulted from the formation of niobium ni- result of the inevitable compositional gra- joules. The Charpy V-notch impact testing

128 -s MAY 2007, VOL. 86


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:52 AM Page 129

WELDING RESEARCH
A

Fig. 8 — Hardness of simulated HAZ from pipe steel N as a function of thermal


experience.

B C

Fig. 9 — Fraction of retained austenite at room temperature in a simulated HAZ from pipe steel S after PWHT for 10 min at different temperatures. A — 720°C;
B — 680°C; C — 640°C.

results of supermartensitic pipe steel S is the cooling rate and, therefore, increases –30°C required for different industrial ap-
shown in Fig. 2. the susceptibility to the formation of in- plications.
The increasing weld metal impact en- termetallic compounds. This results in de- Microstructural Characterization:
ergies achieved with the SMAW, GMAW creasing impact energies in welds made The maximum hardness of the HAZ cor-
using a weave technique, GMAW with with nickel-base (ENiCrMo-3), duplex, responding to different base metal/filler
heat input control, and GTAW process can and superduplex consumables and de- metal/welding process combinations is
be explained based on two main factors. posited without heat input control. There- listed in Table 6. The maximum hardness
First, a higher level of inclusions in the fore, the weld metals with the best impact of the HAZs near to the cap region range
weld metal is expected in joints made with absorbed energies are obtained with gas- from 301 to 341, 346 to 362, and 378 to 391
a flux-shielded process such as SMAW as shielded processes such as GMAW and HV1 in the supermartensitic steels S, N,
compared to joints made with gas- GTAW, and with control of heat input to and K, respectively. The hardness of the
shielded processes such as GMAW and avoid the formation of intermetallic com- three supermartensitic steels ranges from
GTAW. The higher the level of inclusions pounds in the duplex and superduplex mi- 281 to 301 HV1 and is included in Table 6
in the weld metal, the lower the expected crostructures. as a reference. Figure 3 shows the micro-
impact absorbed energy. Second, duplex However, most of the welding process- hardness profiles across different welded
and superduplex weld metals are very sus- filler metal combinations used to weld the joints corresponding to different filler
ceptible to the formation of intermetallic different supermartensitic steel pipes are metal/welding process combinations used
compounds that affect the impact ab- expected to provide average impact ener- to weld supermartensitic pipe steel S. Fig-
sorbed-energy of the weld metal. A high gies in the welded joints above the mini- ure 4 shows the hardness of the HAZs in
heat input or weave technique decreases mum absorbed energy of 27 to 40 joules at different through-thickness locations of

WELDING JOURNAL 129 -s


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:53 AM Page 130

WELDING RESEARCH
Table 2 — Characteristics of Different Welded Pipe Sections

Weld ID Base Metal Welding Process Filler Metal (a) Shielding Gas

W1 S Manual GTAW 1.2-mm DP-3WT Argon


W2 Manual GTAW 1.2-mm DP-3WT Argon
W3 SMAW 3
⁄32-in. ENiCrMo-3 None
W4 SMAW 3
⁄32-in. ENiCrMo-3 None
W5 SMAW 3
⁄32-in. ENiCrMo-3 None
W6 GMAW 1.2-mm MR-DP3W Argon
W7 GMAW 1.2-mm MR-DP3W Argon
W8 SMAW 3
⁄32-in. E2553 None
W9 SMAW 3
⁄32-in. E2553 None
W19 Mech. GMAW 1.2-mm DP-3WT 80%Ar-20%N2
W20 Mech. GMAW 1.2-mm MR-DP3W Argon
W10 N SMAW 3
⁄32-in. ENiCrMo-3 None
W11 SMAW 3
⁄32-in. E2209 None
W12 SMAW None (only root pass) None
W13 K SMAW 3
⁄32-in. ENiCrMo-3 None
W14 SMAW 3
⁄32-in. ENiCrMo-3 None
W15 SMAW 3
⁄32-in. ENiCrMo-3 None
W16 SMAW 3
⁄32-in. E2209 None
W17 SMAW 3
⁄32-in. E2209 None

(a) One to three root passes were deposited in each weld with GTAW using 1.2-mm DP-3WT wire.

two welded joints. The hardness of the


HAZ shows an increasing trend from the
root to the cap region of the joint.
Table 3 — Summary of Cathodic Protection and Slightly Sour Condition Testing The microstructure of the weld metals
ranges from fully austenitic in welds made
Sample Material Maximum Hardness, HV1 Loading Method Side in Tension Test with the Ni-base (ENiCrMo-3) electrode
to a ferritic-austenitic duplex microstruc-
W2 S 301 4-point Bend Cap CP ture with different levels of ferrite and
W19 S 320 4-point Bend Cap CP austenite in welds made with the duplex
W20 S 320 4-point Bend Cap CP and superduplex filler metals. The mi-
W8 S 315 4-point Bend Cap CP crostructure observed in the HAZ of the
W11 N 346 4-point Bend Cap CP
W17 K 378 4-point Bend Cap CP supermartensitic steel pipes corresponds
W2 S 301 4-point Bend Root H2S to untempered martensite as shown in Fig.
W19 S 313 4-point Bend Root H2S 5. Figure 5B shows some strings of re-
W20 S 296 4-point Bend Root H2S tained ferrite observed in the coarse-
W2 S 301 3-point Bend Root H2S grained HAZ of the supermartensitic steel
W19 S 313 3-point Bend Root H2S S. In the intermixing zone of the weld
W20 S 296 3-point Bend Root H2S metal near to the fusion line, a variety of
Base Metal S 282 3-point Bend Base metal H 2S
microstructures were present.
The increasing maximum hardness

Table 4 — Test Conditions under CP and Slightly Sour Environments


Test
Parameter Slightly Sour Environment
Cathodic Protection
Procedure A Procedure B
Solution NACE-seawater NACE-seawater 5%NaCl + Water +
Buffer solution(a)
pH 8.2 3.5–4.0 3.0–3.5
CO2 Pressure None 3 MPa (435 lb/in.2) 3 MPa (435 lb/in.2)
H2S Pressure None 0.001 MPa (0.15 lb/in.2) 0.001 MPa (0.15 lb/in.2)
Polarization (SCE) –800 mV Free corrosion potential Free corrosion potential
Applied Stress 100% yield strength 100% yield strength 100% yield strength
of base metal of base metal of base metal
Loading Mode Four-point bending Four-point bending Three-point bending
Specimen Geometry Flat-bar specimens Flat-bar specimens Flat-bar specimens
(in.) 5.5 × 0.675 × 0.275 5.5 × 0.675 × 0.275 1.7 × 0.675 × 0.125
Temperature, °C 25 ± 2 25 ± 2 25 ± 2
Exposure Time 30 days 30 days 30 days

(a) Solution containing CH3COONa-CH3COOH to control pH.

130 -s MAY 2007, VOL. 86


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:55 AM Page 131

WELDING RESEARCH
ranges of 301 to 341, 346 to 362, and 378
Table 5 — Tensile Test Results
to 391 HV1 observed in the HAZs near to
the cap in the supermartensitic steel S, N, Sample ID Tensile Strength Yield Strength Elongation Reduction Fracture
and K welded joints are in agreement with Area
the expected trend in untempered MPa (ksi) MPa (ksi) (%) in Area (%)
martensite microstructures with increas- Steel Pipe S
ing levels of carbon. The carbon content in
BM-T1 872 (126.5) 565 (82.0) 31.7 72.1 —
the steel controls the hardness level in the BM-T2 871 (126.3) 557 (80.8) 31.6 72.2 —
untempered martensitic microstructure. W2-T1 834 (121.0) 565 (81.9) 31.1 71.6 Base
As reported in Table 1, the carbon content W2-T2 836 (121.3) 586 (85.0) 30.5 71.0 Base
in S, N, and K steels is 0.007, 0.013, and W3-T1 834 (120.9) 488 (70.8) 14.9 28.5 Weld
0.031, respectively. The tempering effect W3-T2 812 (117.7) 496 (71.9) 18.4 32.6 Weld/HAZ
of subsequent passes induces some soft- W6-T1 842 (122.1) 579 (84.0) 25.8 71.4 Base
ening of the HAZ in the root and results W6-T2 829 (120.2) 592 (85.9) 21.2 74.3 Base
W8-T1 841 (121.9) 586 (85.0) 27.8 66.7 Base
in the increasing hardness levels observed W8-T2 827 (120.0) 587 (85.2) 28.6 72.6 Base
in the through-thickness microhardness W19-T1 838 (121.5) 564 (81.8) 29.4 70.3 Base
profile from the root to the cap zone in the W19-T2 848 (123.0) 570 (82.7) 31.5 70.2 Base
welded joints. W20-T1 834 (121.0) 574 (83.2) 32.3 67.1 Base
The hardness for the welding deposits W20-T2 823 (119.4) 598 (86.7) 27.5 72.4 Base
and the HAZ of all the material-welding Steel Pipe N
process-welding consumables combina- BM-T1 866 (125.6) 652 (94.5) 19.9 59.6 —
tions tested in this program exceeds the BM-T2 866 (125.6) 675 (97.9) 23.4 63.2 —
NACE requirements of 23 HRC (253 W10-T1 834 (121.0) 491 (71.2) 13.8 34.0 Weld
HV). Additionally, the hardness of the as- W10-T2 821 (119.0) 523 (75.8) 14.5 56.7 Weld
received supermartensitic steel pipes W11-T1 839 (121.7) 604 (87.6) 20.0 43.8 Weld
ranges from 282 to 301 HV1, therefore, is W11-T2 838 (121.5) 579 (84.0) 20.2 42.6 Weld
also above the maximum hardness limita- Steel Pipe K
tions established by NACE. BM-T1 849 (123.2) 698 (101.2) 23.9 70.7 —
However, there are several issues that BM-T2 847 (122.9) 703 (101.9) 23.7 69.5 —
need to be addressed regarding the maxi- W14-T1 754 (109.4) 537 (77.9) 5.7 23.0 Weld
mum hardness of 23 HRC allowed by W14-T2 777 (112.7) 561 (81.4) 6.4 21.3 Weld
W16-T1 824 (119.5) 623 (90.3) 18.8 72.5 Weld
NACE for use of supermartensitic steels W16-T2 845 (122.6) 627 (91.0) 21.3 71.9 Weld
in sour service. First, it has been indicated
that the ASTM E140 and BS 860:
Rockwell-Vickers correlations are not ap-
plicable to 13Cr-4Ni steels (Ref. 8), and if
the Hays-Patrick relationship is accepted, Table 6 — Maximum HAZ Hardness Near the Cap Region of the Welded Joints
the NACE limit of 23 HRC can be taken
Base Metal Welding Process-Filler Metal Maximum HAZ Hardness, HV1
as equivalent to 275 HV instead of 253
(Ref. 8). Second, in most situations the S Base Metal 282
sour environment will be present only on GTAW-SD filler 301
one side of the steel. Consequently, there SMAW-ENiCrMo-3 341
will be a gradient of hydrogen concentra- SMAW-E2553 315
tion through the thickness, from high at GMAW-SD filler 308
the face in contact with the sour environ- N Base Metal 301
SMAW-ENiCrMo-3 362
ment to very much lower at the free sur-
SMAW-E2209 346
face. The risk of cracking will therefore di- K Base Metal 282
minish through the thickness, so that SMAW-ENiCrMo-3 391
higher hardness may be safely permitted SMAW-E2209 378
on the outside of the pipe. The amount of
hardness relaxation in supermartensitic
steels has not been fully defined. For ex- Table 7 lists a summary of the results of the slightly sour conditions at hardness level
ample, in C-Mn steels with a wall thickness testing under CP and under slightly sour above the 23 HRC (253 HV) limit estab-
above 10 mm, an external hardness maxi- conditions. Even though the applied lished by NACE.
mum has been relaxed from 248 to 300 HV stresses used in this experimental work The observed resistance to SSC in
(Ref. 7). Additionally, the softening of the were equal to the measured yield strength slightly sour conditions of the tested su-
HAZ in the root region of the welded joint of the base metals, and the maximum permartensitic steels in spite of the high
induced by the tempering effect of subse- hardness of different specimens tested level of hardness may be explained based
quent passes will reduce the susceptibility under cathodic protection (–800 mV on the good corrosion resistance of these
to SSC in supermartensitic steel welded SCE) or under slightly sour conditions materials that results in low corrosion
joints. Third, there is a need to document (0.001 MPa, 0.15 lb/in.2 PH2S) ranges from rates. The SSC behavior of typical marten-
maximum hardness levels in supermarten- 301 to 378, and from 282 to 313 HV1, re- sitic stainless steel has been explained
sitic steel for sour service, as the existing spectively, cracking was not observed. based on the decrease of the hydrogen dif-
criteria are lower than the hardness of the These results indicate that under the fusion coefficient with an increase in con-
pipe materials as delivered. testing conditions used, the tested super- tent of alloying elements. Since hydrogen
Hydrogen-Induced Corrosion Crack- martensitic steels are not susceptible to content in steel is proportional to the in-
ing under CP or Slightly Sour Conditions: cracking under either proper CP or under verse of the hydrogen diffusion coeffi-

WELDING JOURNAL 131 -s


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:56 AM Page 132

WELDING RESEARCH
than the toughness of the base metal in the
Table 7 — Summary of Test Results (Cracking) for Corrosion Testing under CP and Sour as-received conditions. The results of the
Conditions Charpy V-notch tests for the HAZ of su-
permartensitic pipe steel S is shown in Fig.
Specimen Steel Test Hardness, HV1 Results 6. On the other hand, the toughness of the
simulated HAZ of pipe steel K after a sin-
W2 S CP-Cap tension 301 No cracking
gle- and double-thermal cycle, and after
W19 S CP-Cap tension 320 No cracking
W20 S CP-Cap tension 320 No cracking PWHT at 785°C was lower than the base
W8 S CP-Cap tension 315 No cracking metal in the as-received condition. How-
W11 N CP-Cap tension 346 No cracking ever, metallographic evaluation of sam-
W17 K CP-Cap tension 378 No cracking ples from the simulated HAZ of pipe K
W2 S H2S-Root Tension-Procedure A 282–301 No cracking material indicated a nonuniform mi-
W19 S H2S-Root Tension-Procedure A 313 No cracking crostructure as the result of a nonuniform
W20 S H2S-Root Tension-Procedure A 296 No cracking
thermal experience across the section of
W2 S H2S-Root Tension-Procedure B 282–301 No cracking
W19 S H2S-Root Tension-Procedure B 313 No cracking the sample. This may have resulted from
W20 S H2S-Root Tension-Procedure B 296 No cracking an improper contact between the samples
Base metal S H2S-BM Tension-Procedure B 282 No cracking and the grips in the Gleeble machine. Due
to the size of the pipe, specimens from
pipe steel K were subsize. Therefore, the
results obtained from the evaluation of the
simulated HAZ from pipe steel K are con-
Table 8 — Transformation Temperatures of Supermartensitic Pipe Steels S, N, and K sidered to be invalid, and no further analy-
sis was attempted from the samples pre-
Pipe Steel Ac1 (°C) Ac1(a) Ac3 (°C) Ms (°C)
pared from pipe steel K.
S 680 630 780 201 The microhardness of the simulated
N 695 602 765 224 HAZs of supermartensitic pipe steels S
K 745 680 820 277 and N after different welding thermal cy-
cles and PWH treatments are listed in
(a) Determined based on chemical composition.
Table 9. Figures 7 and 8 show how the av-
Ac1 (°C) = 850 – 1500 (C + N) – 50Ni – 25Mn + 25Si + 25Mo + 20 (Cr–10) (Ref. 9) erage microhardness of the simulated
HAZs of pipe steels S and N changed with
thermal cycles and with temperature and
cient, the hydrogen content in high- properly controlled between the protec- holding time during tempering. The mi-
chromium steel may become quite large. tion potential and the hydrogen evolution crohardness of the HAZs from pipe S in-
As result, the high-Cr steel may have a potential. In practice, the production of creases from 285 to 316 HVN (1 kgf) after
high susceptibility to hydrogen embrittle- hydrogen is expected to be very limited at a single- or double-thermal cycle. Temper-
ment due to the smaller hydrogen diffu- a potential slightly lower than the hydro- ing at 720°C does not affect the hardness
sion coefficient. Therefore, the best way to gen equilibrium potential and, in the ab- of the HAZ. However, as the tempering
control the susceptibility to hydrogen em- sence of H2S, potentials 100 mV lower temperature decreases from 720° to
brittlement of high-Cr steels is by de- than the hydrogen equilibrium potential 640°C, the microhardness decreases. A
creasing the hydrogen permeation rates. may be acceptable in practice. microhardness of 289 HVN is obtained in
The low corrosion rate of supermarten- Influence of Short PWHT on Gleeble the HAZ after a PWHT at 640°C for 10
sitic steel decreases the amount of hydro- Simulated HAZ: The transformation tem- min — Fig. 7. This result shows that a short
gen generated on the surface, which may peratures of pipe steels S, N, and K, as de- PWHT is effective to reduce the micro-
result in lower hydrogen permeation rates termined by dilatometer analysis, are hardness of the HAZ to levels very close
in the steel a better SCC resistance. shown in Table 8. Additionally, the tem- to the hardness of the base metal in the as-
Materials may also be susceptible to perature at which the austenite starts to received condition. As shown in Fig. 8, sin-
hydrogen embrittlement caused by ca- form during heating, Ac1, was determined gle- and double-thermal cycles increase
thodic protection dependent on the envi- based on chemical composition, and it is the hardness of the HAZ of pipe steel N
ronmental conditions. The lower limit of also included in Table 8 as a comparison. from 291 to 320 and 330 HVN, respec-
the cathodic protection is in principle lim- The Ac1 temperature, based on chemical tively. Postweld heat treatment of the
ited by the hydrogen equilibrium poten- composition, was determined using an HAZ at 735° and 695°C for 5 and 10 min
tial. The hydrogen evolution potential is a equation that was developed empirically further increases the hardness to about
function of the pH and is equal to from data for 13Cr steels with carbon con- 350 HVN. However, PWHT at 655°C for
(–0.24–0.059 pH) VSCE. For a pH of 8.2, tent below about 0.05 wt-% (Ref. 9). The 5 and 10 min softens the HAZ to a level of
the hydrogen evolution potential is equal difference between the transformation about 314 HVN.
to –0.72 VSCE. Therefore, the potential of temperatures of the different super- A fully martensitic microstructure was
–0.8 VSCE impressed on the sample during martensitic pipe steels can be explained observed in the simulated HAZ of both
the CP testing was more negative than the based on the effect of the alloying ele- pipes steel S and N after a single-thermal
hydrogen evolution potential and demon- ments, especially Ni, on the stability of cycle. The main characteristic observed in
strates the hydrogen-induced cracking re- austenite. pipe steel S was the increase of retained
sistance observed in the tested super- The impact Charpy V-notch toughness austenite, at the grain boundaries, as the
martensitic steels. of the HAZs of pipes S and N in every tempering temperature was decreased
Therefore, supermartensitic 13Cr steel thermal condition, single-/double-thermal from 720° to 640°C as shown in Fig. 9. The
can be perfectly protected against corro- cycle, and tempered at temperatures fraction of retained austenite in the grain
sion without any risk of hydrogen embrit- equal to Ac1, Ac1 + 40°C, and Ac1 – 40°C, boundaries increases with holding time
tlement when the cathodic protection is for 5 and 10 min, were equal, or higher during PWHT. On the other hand, etching

132 -s MAY 2007, VOL. 86


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:56 AM Page 133

WELDING RESEARCH
Table 9 — Results of Hardness Measurements [HVN (1 kgf)] in Simulated HAZ from Pipe Steels S and N after Different Thermal Cycles and PWHT

Specimen HVN (1 kgf) Average HVN Specimen HVN (1 kgf) Average HVN

Base Metal 285 Base Metal 291


S1 318 316.5 N1 318 319.2
319.5 315
315.5 318
313 326
S2 315.5 316.3 N2 326 329.2
321 329
314.5 326
317.5 338
313 327
S1T3-5 314 315.6 N1T3-5 350 353.7
311 356
319.5 354
318 355
S1T3-10 319 311 N1T3-10 335 352
309.5 340
316.5 357
299.5 357
355
351
S1T2-5 305 309 N1T2-5 345 344.4
303 346
306 345
316 314
315 340
346
S1T2-10 300 304 N1T2-10 350 349.8
310.5 354
298.5 350
307 345
S1T1-5 302 294.8 N1T1-5 306 311.8
293 314
291 318
293 309
S1T1-10 279 289 N1T1-10 309 316
295 316
291 318
291 321

of pipe steel N only revealed very small tent is greatly increased, the relative en- form to virgin martensite on cooling from
amounts of retained austenite at the grain richment is reduced, and on cooling to the PWHT temperature; the reaction se-
boundaries when PWH treated at 695°C. room-temperature transformation to vir- quence depends on the C content of the
The increasing fraction of austenite ob- gin martensite will occur, thus decreasing steel, and the effect will be of less signifi-
served in pipe steel S as the PWHT tem- the amount of austenite retained at room cance with very-low-C alloys. The result-
perature is decreased from 720° to 640°C temperature. ing higher Ms temperature of pipe steel N
is the result of the reversion and stability Based on the microstructural analysis may be responsible for the absence of any
of austenite as a function of time and tem- and as expected, the Ac1 determined by major fraction of austenite at room
perature. At PWHT temperature, refor- dilatometry is higher than the temperature.
mation of austenite is promoted. Since the equilibrium-transformation temperature. An important additional consideration
reaction is diffusion controlled, the The heating rate affects the transforma- is the effect of retained austenite on the
austenite formed after PWHT will differ tion temperature determined by dilatom- corrosion performance of supermarten-
compositionally from austenite retained etry; the higher the heating rate, the sitic stainless steel pipe. It has been ob-
after the welding operation. In the latter higher the transformation temperature. served that there is no harmful influence
case, the martensitic matrix and residual According to the microstructural analysis, of retained austenite on the corrosion re-
austenite will be of identical composition, the equilibrium-transformation tempera- sistance of supermartensitic steels (Ref.
whereas the reverted austenite formed ture for pipe steel S is below 640°C. 10). Additionally, a higher content of re-
during heat treatment will be enriched in To some degree, the effect of PWHT tained austenite reduces the diffusible hy-
Ni, C, and N. The degree of enrichment depends on the Mf of the alloy. If this is drogen and the SSC susceptibility of the
determines the stability of the austenite low, substantial austenite will be retained steel. Due to the high solubility of hydro-
formed during the PWHT operation. If at room temperature after PWHT simply gen in austenite, austenite act as a hydro-
the Ac1 is exceeded only slightly, by say because the material remains above the gen trap and reduces the effective diffu-
40°C, then the enriched austenite formed Mf point. However, PWHT may induce sion coefficient of hydrogen.
will be stable on cooling to room temper- sufficient carbide precipitation to deplete The changes in the mechanical proper-
ature. At temperatures further above the the matrix of C and raise the Ms. The ties on tempering can now be discussed on
Ac1 point, the equilibrium austenite con- austenite is destabilized and will trans- the basis of the microstructural changes. It

WELDING JOURNAL 133 -s


Ramirez Supp May 2007corr:Layout 1 4/5/07 10:57 AM Page 134

WELDING RESEARCH
is assumed that three mechanisms are re- ness of the weld metal and HAZ result- containing CO2 environments. Corrosion/96,
sponsible for the changes of hardness with ing from all the material-welding Paper 58, NACE International, Houston, Tex.
PWHT, i.e., reversion of austenite, reduc- process-welding consumables combina- 4. Amaya, H., Kondo, K., and Hirata, H.
tion in dislocation density, and precipita- tions exceeds the NACE requirements 1998. Effect of chromium and molybdenum on
corrosion resistance of super 13Cr martensitic
tion. In pipe steel S, the hardness de- of 23 HRC (253 HV).
stainless steel in CO2 environment. Corro-
creases with a decrease in PWHT sion/98, Paper 113, NACE International, Hous-
temperature from 720° to 640°C and an in- • The tested supermartensitic steels are ton, Tex.
crease in holding time from 5 to 10 min not susceptible to cracking under either 5. Miyata, Y., Kimura, M., Toyooka, T., and
due to the increase in the fraction of sta- CP or slightly sour conditions at hard- Murase, F. Weldable martensitic stainless steel
ble austenite. Additionally, a reduction in ness levels above the 23 HRC (253 HV) seamless pipe for linepipe application.
dislocation density and decrease of C level limit established by NACE, under the Kawasaki report: 59–71.
in the untransformed martensite may also testing conditions used including an ap- 6. Asahi, H., Hara, T., and Sugiyama, M.
influence the hardness level. On the other plied stress level equal to the measured 1996. Corrosion performance of modified 13Cr
OCTG. Corrosion/96, Paper 61, NACE Inter-
hand, the hardness of pipe steel N in- yield strength of the base material.
national, Houston, Tex.
creases with PWHT at the two highest 7. Pargeter, R. J., and Gooch, T. G. 1994.
temperatures, 735° and 695°C. Higher lev- • The transformation temperature at Welding for sour service. Update on sour ser-
els of C and lower levels of Ti in pipe steel which the austenite starts to form dur- vice: materials, maintenance and inspection in
N as compared to pipe steel S may in- ing heating, Ac1, of pipes steels S, N, and the oil and gas industry. Conference, London:
crease the susceptibility to secondary K, as determined by dilatometric analy- 17 pp.
hardening during PWHT due to the for- sis, are 680, 695, and 745°C, respec- 8. Gooch, T. G. 1995. Heat treatment of
mation of carbides based on V, Cr, or Mo. tively. The martensitic start tempera- welded 13%Cr-4%Ni martensitic stainless steel
Additionally, the formation of Cu precip- tures, Ms, are 201°, 224°, and 277°C for for sour service. Welding Journal 74(7): 213-s to
223-s.
itates may also play a factor in the increase pipes S, N, and K steel, respectively. Ac-
9. Gooch, G. T., Woollin, P., and Haynes, A.
of hardness during PWHT at the two high- cording to the microstructural analysis, G. 1999. Welding metallurgy of low carbon 13%
est temperatures. Softening of the HAZ of the equilibrium-transformation tem- chromium martensitic steels. Proceedings of Su-
pipe steel N during PWHT at 655°C may perature, Ac1, for pipe steel S is below permartensitic Stainless Steels ‘99 Congress,
result from a reduction dislocation density 640°C. 188–195. Belgium.
and from the fact that some C has been 10. Kimura, M., Miyata, Y., and Kitahaba,
taken out of solid solution and has formed • The fraction of retained austenite in the Y. 2001. Effect of retained austenite on corro-
some carbides. simulated HAZ in pipe steel S at room sion performance for modified 13% Cr steel
temperature increases as the tempering pipe. Corrosion 57(5): 433–439.
temperature was decreased from 720°
Conclusions to 640°C and increases with hold time
during PWHT.
The following conclusions can be
drawn from this work: • A microhardness of 289 HVN is ob-
tained in the HAZ of supermartensitic
• Nickel-base ENiCrMo-3 electrodes are pipe steel S after a PWHT at 640°C for
not recommended to join X80 grade su- 10 min. This result shows that short
permartensitic steel pipes due to the re- PWHT are effective for reducing the REPRINTS REPRINTS
sulting undermatching condition of the microhardness of the HAZ to levels
weld metal and the low levels of ductil- very close to the hardness of the base To order custom reprints
ity (6 to 15% elongation) of the welded metal in the as-received condition. of 100 or more of articles in
joints.
• The toughness of the simulated HAZ of Welding Journal,
• Duplex and superduplex consumables pipe steels S and N in every thermal call FosteReprints at
provided an increasing degree of over- condition, single-/double-thermal cycle, (219) 879-8366 or
matching in the weld metal of X80 su- and after PWHT is equal to or higher
than the toughness of the base metal in
(800) 382-0808 or.
permartensitic steel pipe joints result-
ing in a yield strength of the joint above the as-received conditions. In general, Request for quotes can be
550 MPa (80 ksi) and an elongation be- the change in toughness reflected the faxed to (219) 874-2849.
tween 20 to 33% when using SMAW, change in hardness; i.e., a reduction in You can e-mail
GMAW, and GTAW processes. hardness was accompanied by an in-
crease in toughness.
FosteReprints at
• All of the welding process-filler metal sales@fostereprints.com
combinations used to weld the different References
supermartensitic steel pipes provided
averaged impact energies in the welded 1. Duisberg, J., Niederhoff, K. A., and
joints above the minimum absorbed en- Popperling, R. K. 1996. 13%Cr steels for flow- Want to be a
ergy of 27 to 40 joules at –30°C required line and pipeline applications. Stainless Steel Welding Journal
for different industrial applications. World 8(6): 29–33. Advertiser?
2. Biagiotti, Jr., S. F., and Reichman, J. S.
• The maximum hardness observed in the
1995. Justifying the use of 13Cr for corrosive For information, contact
HAZ near to the cap in the S, N, and K
CO2 operations. Corrosion/95, Paper 81, Rob Saltzstein at
NACE International, Houston, Tex. (800) 443-9353, ext. 243,
supermartensitic steel welded joints 3. Ueda, M., Amaya, H., Kondo, K., Ogawa,
ranges from 301 to 341, 346 to 362, and K., and Mori, T. 1996. Corrosion resistance of or via e-mail at
378 to 391 HV1, respectively. The hard- weldable super 13%Cr stainless steel in H2S salty@aws.org.

134 -s MAY 2007, VOL. 86

Вам также может понравиться