The resistance of welded linepipes to sulfide stress cracking

The resistance of welded linepipes to sulfide stress cracking

Corrosion Science,Vol. 27, No. 10/11, pp. 1117-1135, 1987 Printed in Great Britain 0010-938X/87 $3.00 + 0.00 © 1987 PergamonJournals Ltd. THE RESIST...

2MB Sizes 1 Downloads 19 Views

Corrosion Science,Vol. 27, No. 10/11, pp. 1117-1135, 1987 Printed in Great Britain

0010-938X/87 $3.00 + 0.00 © 1987 PergamonJournals Ltd.

THE RESISTANCE OF WELDED LINEPIPES TO SULFIDE STRESS CRACKING Y. KOBAYASHI, K. UME, T. HYODO and T. TAIRA Nippon Kokan K.K., Fukuyama, Japan

Abstract--The SSC behavior of welded linepipes has been studied using tensile SSC tests (NACE TM-01-77), full thickness SSC tests, and full scale SSC tests. The results have indicated that homogenization of microstructures, which can be attained by reducing a carbon content to <0.05% or by quenching and tempering, improves the resistance to SSC. Most specimens from SAW welds fail at the heat affected zone (HAZ), regardless of differences in the microstructures of the parent materials, and they show nearly the same level of Oth/Cryratio. Detailed metallographic examinations have been performed in order to understand the influences of steel chemistries and heat inputs in welding on the SSC resistance of the HAZ. Relations between the results of the laboratory tests and full scale tests are also discussed. INTRODUCTION

SULFIDEstress cracking (SSC) resistance as well as hydrogen-induced cracking (HIC) resistance should be considered important to linepipes for sour oil/gas service because the pipes are highly stressed by internal pressure. In fact, a number of reports have indicated that some failures in sour gas pipelines could be caused by development of SSC near welds.l-3 The hardness dependency of SSC susceptibility could not explain cracking in such low or medium strength steels as welded linepipes. Since those failures were reported, metallographic investigation of welds has indicated that microstructural changes, including secondary precipitations during welding, as well as hardness and residual stress distributions, have a significant influence on the SSC resistance of the welds.4-8 In this paper, SSC behavior in welded linepipes is metallographically investigated focussing on the relationship between microstructures and SSC resistance of the welds. Full thickness specimens with weld reinforcements as well as NACE TM-01-77 specimens were used to determine the geometrical effects of specimens on SSC. Using results from a variety of laboratory tests and full scale SSC tests, a correlation between the small scale SSC test results and SSC occurrence in sour gas pipelines is also discussed. EXPERIMENTAL

METHOD

Test materials and specimen preparation The test materials used were API 5L X42-X80 steel plates and welded linepipes with various microstructures. These steel plates were produced mostly by the controlled-rolling (CR) process, but quenched and tempered tOT) plates were also tested for the sake of comparison. The chemistry range of all materials tested is shown in Table 1A. Welds were made in a laboratory using the steel plates in Table lB. The microstructure of Steel A, a typical HIC-resistant steel of Gr X52, consisted of fertile and This paper was presented at the Conference on 'Hydrogen Sulphide-lnduced Environment Sensitive Fraction of Steels' which was held in Amsterdam on 10-12 September 1986 and which was organized by the Working Party of the European Federation of Corrosion on Stress Corrosion Test Methods. Manuscript received 5 November 1986. 1117

Y. KOBAYASHIet al.

1118

TABrE 1. TEST MATERIALS (A) TEST STEELS

Grade

t (mm)

C

Si

Mn

P

X42-X70

16-25

0.02-0.17

0.19-0.30

0.89-1.91

0.004-0.021

S

Others

0.001-0.004 Cu,Ni,Cr, Nb V,Ti,B,Ca

(B) TEST STEELS

Steel

Microstructure

Grade

t (mm)

C

Si

Mn

P

S

Others

A B

Ferrite-Bainite Ferrite-Bainite

)(65 X65

19 25

0.09 0.04

0.29 0.28

0.85 1.55

0.007 0.004

0.001 0.001

C

LowCarbon Bainite LowCarbon

X70

19

0.02

0.25

1.91

0.021

0.001

Nb, V, Ca Cu,Ni,Nb, V, Ca Nb,B,Ca

X80

19

0.05

0.19

1.90

0.011

0.001

D

Mo,Nb, B, Ca

YS MPa 404 502 566 612

pearlite. Steel B, consisting of ferrite and low carbon bainite (LCB), was a HIC-resistant and good weldability steel Gr X65. In contrast, Steels C and D were both low carbon steels having LCB microstructures and characterized by as high strength as Gr X70 and X80, respectively. All steels were hot-rolled to 19 mm or 25 mm thick plates, and the on-line accelerated cooling (OLAC) process was applied to Steel B following the CR processes. Welds of Steels A to D were made with the welding conditions shown in Table 2. The welds of Steel A were also made by multi-pass welding with a low heat input of 25 kJ cm -I, in order to compare SSC behaviour in the welds by the different heat inputs. N A C E TM-01-77 specimens were machined from the backing side of the weld, locating the heat affected zone (HAZ) at the middle of the gauge length. Bead-on-plate welding was also performed by tandem submerged arc welding (SAW) (50 kJ cm-t), single SAW (26 kJ cm -I) and GTAW (11 kJ cm -l) on Steels A-C. Four-point bend specimens were machined from these welds as shown in Table 2, and used for investigation of the hardness dependency of SSC resistance. Test conditions

Specimens and schematic drawings of the experimental arrangement are shown in Fig. 1. Tensile SSC tests were performed in accordance with NACE standard TM-01-77. For four-point bend SSC tests, 5 mm thick rectangular specimens were stressed by bending to either 80%, 100% or 120% of the yield strength TABLE 2.

WELDINGCONDITIONSAND SPECIMENPREPARATIONS

Specimen type

Thickness (ram)

Welding process

NACE TM-01-77, Full thick

19

2SAW

25

2SAW

Four-point bending

19 25

2SAW 1SAW GTAW

Pass Backing Final Backing Final Bead on Plate

Heat input (kJ cm- l)

Sampling

44

NACE TM 01-77

55 62

Resistance of welded linepipes to SSC

~-~ : ~-k20

j-

~..~u(6 0 }

1119

M,o '-i-~

,L2.5- L ~ (601

~20 k

200' (AI N A C E

SSC

Tes!

T hlckness t=$.O 10.2In}



,514 ~inl

(o) Specimen

t o

1 ° ~ 0 n

~ ' ~ J =100 (4In) ~ ~/2=30 --f--~,=40 ~ = 3 o

HE

{

--

Tt /

/

Tell Solution (NACE)

-~'

"~l'~-Thermostatlc Bath Gas

" l ~ = S

(b} Apparatus

(B) 4-point bend SSC Test

~

Specimen

~.

4-point bendhvl test Stress level : S : 12"D'E't/(3 ,t2-4~), D deflection

Cooler

/Test

--I

(C] Full Thickness SSC Test

Band Heater Cover I3

Tested Pipe

/

Length:2OOOmm |

IDio:609mm(Z4in.)/ 316~"

~

,

~1~

IPressurized

~rE-ilSolution Tank ~

111511Max. Pressure o

n

'

__

I

.~ i

~o

J

r- . . . . . . . . . . . . . . . . . . .

-5

[P,essure Ga~e]

I So'o"oo Tookl IO) Full Scale SSC Test

Fro,. l.

C02

N2

H2S

I~'o* Mo'erl I Gas C,,°Oerl

Specimens and experimental arrangement for SSC tests.

Y. KOBAYASHIet al.

1120

of the parent material and immersed in the NACE TM-01-77 test solution for 1 week. Full thickness specimens of the 38 mm wide and 50 mm long gauge lengths were also used for SSC testing to observe crack initiation sites and crack propagation paths in the welds. Hydrogen entry into a pipe wall was simulated in the full thickness coated specimens in which two or three surfaces were coated with a tar epoxy resin. In the full scale SSC testing, 2 m long pipes were stressed to either 80% or 85% of specified minimum yield strength (SMYS) by pressurizing with a test solution. The test solution was prepared in a tank and circulated to keep the solution chemistry constant in the pipe. 5% NaCI + 0.5% acetic acid saturated with 1 atm H2S gas, which is specified in the NACE TM-01-77 std, was used in all SSC tests. The pH value of this solution is about 3.5, which is lower than the NACE TM-02-82 solution, another H2S-containing solution commonly used in laboratory tests.

Metallography Cross-sections of four-point bend specimens and full thickness specimens were microscopically examined for occurrence of SSC and crack propagation paths. In particular, Nital etching followed by electrolytic etching in a Picric acid plus NaOH solution was applied to delineate the microstructures of the HAZ and investigate initiation of SSC in association with the martensite-austenite (MA) constituent. EXPERIMENTAL

RESULTS

N A CE TM-01-77 test results In Fig. 2 are shown N A C E TM-01-77 SCC test results of Gr X42 to X70 welded

linepipes. All pipes were tested parallel to the circumferential direction and perpendicular to the straight seam welds. The ratios of threshold stress to yield strength (Oth/Oy) were in a wide range from 0.6 to 0.9 for parent materials but no strength dependency was seen in the steels of these grades. In contrast, the Oth/~yratios of the weld specimens were around 0.6, which was smaller by 0.3 as a maximum than those of the parent materials. Figure 3 illustrates the effects of carbon content and microstructure on the Oth/Oy ratios for the parent materials. Microstructural changes from ferrite/pearlite mixtures to the tempered martensite by quenching and tempering is effective for improving the SSC resistance; the improved level of the SSC resistance is also achieved by uniform low carbon bainite (LCB) or LCB/ferrite mixtures of low C-high Mn steels. It has also been noted that both lowering carbon content and sufficient alloying are essential for the LCB or LCB/ferrite microstructures and the improved SSC resistance. A low carbon steel containing a small amount of alloy

-9

g

1.0 03

"Q

0 CR (F-P) / k CR {Low C Balnite} r-I O & T

0.9

08

Open : Body Solid : Weld

A

m m~

0

~

O.7

N

0.6

o

0

• 0

00

0

t--

Orl CL•

t

300

[]





• ffi

0.5

FIG. 2.

T r a n s . Specimens

400

I

500 600 Yield S t r e n g t h MPa

700

Examples of NACE TM-01-77 SSC tests on welded linepipes.

1.0

"O~ 0 . 9 ,,e.-

o~ 0 . 8 .~_ ,,, 0.7 H S

rj') 0.6 0r '

0.5

_

to SSC at I HAZ

-7

I

I.0.4 0

,,,~- a

.

L

L

0.I0 wt%

Microstructt~nfl effects (m SSC resistance of parent materials.

1121

. . . . . ...a

50#m

0.05 Carbon Content F1G. 3.

t.

c~

o

N

c~

b

FIG. 7.

A typical example of SSC in weld ferrite/pearlite steel (Steel A).

1123

Ferrite-Peorite

0.70y

TTF=196hrs Hardness Distribution along Fusion Line

SSC Morphology

209-

Hv I Okg

218d

245 b" 221 !0 225

207b

232 j b 227 tt::) 145

2Jo~o

221 209

FIG. 8.

SSC morphology of multi-pass weld of ferrite/pearlite steel (Steel A).

1124

0165

-Sampling

-Etching I st:3% Nital 2nd:Picric Acid NaOH Wafer

5g 25g IOOml

Electro-etching 4A/cm 2 30~60sec

Ferrlte-Balnlte/X65 0.80y

Fl(;. 1(I. SSC and microstructures in intercritical H A Z .

1125

FI6. 11. Early stage of cracking in intercritical HAZ (Steel A).

1126

Gr.API 5 L X 5 2 28 in ODxO.750 In WT

or) 1.0

"0.8

0.6

O

kJ~!

0

0.4

o~

I

I0

I

50 Time

1

I00

to

Failure

500 (hours)

[

Appearance

L

,,

I000

i

Note

i

NACE

A

,

~*. . . . . ,

0

~ .............'.......... :~ * .~,,~ x . . . . . ::1

TM-OI-77 Round Bar

(6.35lH |

.'

|

~

C ~ (~ i

.

,,,~

"

. . . . .

,

iL

," C o a t i n g ~

,

~

~L,"

Full-thickness

Side surfaces coated

L

. ! ¸?"'¸,'111¸¸¸¸i¸.................................................:.......... Coatlng

O

'*

. Full-thlck~s| Outside weM

i=

& S~e surfaces coated

FIG. 13.

The effects of specimen geometries and hydrogen entry on occurrence of SSC.

1127

* Tested ,

Pipe

:

X65

711 OD. x 19.1 WT.

(mm)

Full-Scale SSC Test Condition : NACE soJ~./Stress, 80%SMYS/Durotion,2months

FIG. 16.

Examples of SSC developing near weld in full scale SSC test (pipe for ordinary use),

1128

Resistance of welded linepipes to SSC

OF-P

No Crack

~

~

"-" 1.2 0

0

I ~

~, 1.0

0

I~A(~A

o

, - 6 ~ AC~ ,~

Z~CI3~

nF-B

1129

ALCB

Crack •





IIII,





Di



m b \ b

0.8

i

ii

:'J/cm- 5 0 I

200

i

Jl I

I

250

300

Hv max FI~. 4.

i

26

L

350

10kg Load

Hardness dependency of SSC in welds (four-point bend SSC tests).

usually leads to a reduced pearlite/ferrite microstructure and results in a ath/Oyratio similar to those of the ferrite/pearlite steels. The OLAC process is another effective way of obtaining the LCB microstructures in this type of steel. In addition, the carbon content must be delicately controlled because thermal cycles in welding may cause grain boundary embrittlement in the H A Z of an extremely low-carbon steel. `) This indicates that fine and homogenized microstructures free from pearlite lead to a high resistance of the parent material to SSC, but the SSC resistance of the welds is influenced by other metallurgical factors rather than the microstructure of the parent materials. Four-point bend SSC test results The results of four-point bend SSC tests from the welds of Steels A -C are given in Fig. 4. SSC always developed in either the weld metal or the H A Z and the occurrence of SSC was therefore plotted against the respective maximum hardness in the figure. The critical hardness for SSC was around Hv (10 kg) 260 regardless of the steel chemistry or the levels of stress applied. The critical hardness obtained in the present study agrees closely with Hv 248 to 260 in the previous studies. L4,8These results indicate that, according to NACE MR-01-75, controlling the hardness of the weld as well as of the parent material below a critical value is important to pipelines for heavy sour services although the critical hardness also depends on the H2S concentration in the environment. Full thickness SSC test results Tensile SSC tests with full thickness weld specimens were performed to investigate geometrical effects of the specimen on SSC occurrence. The weld of Steel D, in which the maximum hardness was > H v 260, developed SSC across the weld metal and the HAZ. 5 In contrast, the welds whose hardness was well below Hv 260 showed SSC only in their H A Z as shown in Fig. 5. The configurations of these cracks were

1130

Y. KOBAYASIIIet al.

Type

Morphology

Note

Ferrlte-Peorllte , Ferrite ,

Hlqh HIC

after PWHT

Ultra Low Carbon

sensitivity

QT (Norma)

Balnlte [Ultra Low C -Nb)

Bainite (Plate QT)

Pipe SR Ft(;. 5.

SSC morphologies of welds.

classified into four different morphologies, depending on steel chemistry, levels of HIC resistance, and type of post-weld heat treatment. 5 SAW welds in linepipes usually showed cracks of Type I in the full thickness SSC tests. The crack was initiated and propagated along the intercriticai HAZ, which was heated to temperatures just above the Ac], but no crack developed near the fusion line, the hardest portion of the weld. As shown in Fig. 6, a number of planar-oriented fine cracks was also found in the intercritical HAZ. Supercritical tensile stress is considered responsible for the generation of these cracks because the weld showed no cracks in the same region when tested without applied stress. Figure 7 compares the crack path and the hardness distribution in the weld of Steel A. The same crack morphology, as shown in Fig. 8, was obtained for the multi-pass weld by the relatively low heat input of about 25 kJ cm -l. The crack developed along the intercritical HAZ but not along the fusion line, though the multi-pass weld consisted of a region a little harder than the example in Fig. 7. These observations indicated that SSC was generated and propagated along the softened region of the HAZ. Stress relief heat treatment on the welds caused no change in the crack morphology.5 SSC in the intercritical HAZ were also common in the NACE TM-01-77 SSC tests of these welds. SSC test results of simulated H A Z Different microstructures in the HAZ of Steel A were simulated by use of the Gleeble Tester and examined for SSC resistance by the NACE test. In Fig. 9a, the same stress, 70% of the yield strength of the original material, was loaded to all specimens, and SSC resistance was compared in terms of time-to-failure. In Fig. 9b, the yield strengths of the specimens heated in different thermal cycles were determined, and each specimen was loaded to 80% of its yield strength in the SSC test. Both results showed that heating to about 800°C, which was similar to the thermal cycle for the intercritical HAZ, resulted in the highest susceptibility to SSC. It is also seen in these figures that the highest SSC susceptibility of the intercritical H A Z is not only because of the lowest strength (hardness) but because of its microstructural characteristics.

Rcsistancc of welded lincpipcs to SSC

0.70y

1131

of Bose Metol

I000 500

I00 I--

g

l+

IO

I ntercrltlcal Region

la_ I 1400

I 1200

I I000

I 800

I~, I 600 As Roll

Max. Temp, (*C) (a) 0.80y (actual YS)

I000 500

I-/

I00 It. II--

~'-.

Infercrlflcal Region

I0

"" " ' " ,Ik.~, . 0 . . -

I 1 1400 1200

500

A$ :o1 . 400 { Warent] Matedel 300 )-

1

1

I

I000

800

600

Max. Temp. (*C} (b)

FtG+ t),

The effects of thcrmul cycles on SS(" rcsistmacc (Steel A).

DISCUSSION

Laboratory SSC testing of welds indicates that SSC is generated in the hardest regions whose hardness is above Hv 248-260; in the welds with maximum hardness well below the critical value, on the other hand, SSC was generated and propagated along the softened intercrital HAZ. Factors influencing the latter SSC behavior are now discussed from a metailographical standpoint. The meaning of laboratory SSC test results is discussed through comparisons of the SSC occurrence between the laboratory tests and full scale tests.

Metallurgical study Hardness is a primary factor controlling the SSC resistance of steels, and HRC22 (Hv 248), according to the NACE MR-01-75 std, is a well-known safe criteria for valves or oil country tubular goods (OCTG). In the last decade, a number of researchers have been concerned with whether this criteria applies to welds consisting of localized hard regions. Taira et al. ,4 TWI 8 and the present study have showed that, in general, controlling a maximum hardness below about Hv 250 is important for

1132

Y. KOBAYASflIet al.

welds of many kinds. As seen in Figs 7, 9a and 9b, however, most weld specimens were ruptured in the intercritical H A Z though their hardness was below the critical value. Cracks of this type are associated with microstructural changes and reduction of yield strength, both of which are caused by heating to just above the Ac]. The thermal cycles in welding decreased the yield strength of the intercritical H A Z by about 10% of the original yield strength and this is significant because the O'th/O'yratios for the welds are usually smaller by about 10% of the yield strength than those of the parent materials with ferrite/pearlite microstructures. The softened region undergoes macroscopically no localized plastic deformation in a tensile test because of mechanical restraint by the surrounding portion. Microscopically, however, the softer region can be plastically deformed to a larger extent when the weld is stressed to near yielding, which may lead to a high possibility of developing SSC in the intercritical HAZ. The intercritical H A Z can be characterized by a high tendency towards formation of the MA constituent. Figure 10 shows cracks and surrounding microstructures in the intercritical HAZ. In Fig. 10D, a typical crack in the parent material with the ferrite/pearlite microstructure is also shown for the sake of comparison. SSC is usually found in pearlite in the parent material of pearlite and ferrite microstructures. In particular, these cracks appear to be generated at the cementite/ferrite boundaries. The cracking behaviour of the intercritical H A Z depended on the steel chemistry of the parent materials. In the case of Steel A (Figs 10A and I1), the Gr X52 ferrite/pearlite steel, the MA constituent was rarely found, and cracking appeared at the retained pearlite rather than at the MA. Increase in the alloy and carbon content increases the tendency towards formation of the MA in the HAZ. For this reason, cracking at the MA is more likely in the H A Z of Gr X60 or higher strength steels. The H A Z in the Steel B weld (Fig. 10B), Gr X65 ferrite/LCB steel, showed a considerable amount of the MA formed in welding. However, cracks here were developing in the bainitic microstructure in association with carbide precipitates rather than at the MA. The largest amount of the MA was observed in the Steel D weld (Fig. 10C), Gr X80 LCB steel, as compared with the others. Some of the cracks were developing inside the MA or along the MA/matrix boundaries, and therefore the presence of the MA can explain cracking in this region. These observations suggest that the presence of the MA is not always a primary factor controlling the SSC susceptibility of the intercritical HAZ, but it is significant only in such high strength welds as Gr X80. A decrease in the yield strength combined with the metallographical features described above can be responsible for the rupture of the intercritical H A Z in the SSC tests, with slightly smaller ath/O'y ratios for the weld than for their parent materials.

Correlation of laboratory SSC tests with small specimens and full scale SSC tests The NACE TM-01-77 SSC test is intended to screen materials for use in the sour environments, but not to assess the adaptability of a material to a specific sour environment use, because of the geometrical effects on the hydrogen entry into steel, residual stresses, surface conditions, and other causes. Although it has been reported that HIC behaviour in actual pipes can be simulated by single surface exposure HIC tests, because the hydrogen concentration is almost the same, 1°'11 degrees of HIC

Rcsislance of welded linepipes to SSC

/>

I00 (I)

80

0

6o

NACE" Solution I5~NoCI+O.5~,CHaCOOH*H2SI

IZE ...J (..)

zo

o

0-0"0 0 20

40

0 ~"

q:)

60

I" C L R ] std.

Fro. 12.

1133

,o

80

I00

%

HIC behaviour in standard H I C test and single surface exposure HIC test.

occurrence are much greater in the standard HIC tests than in the single surface HIC tests (Fig. 12). For the same reason, the occurrence of SSC was compared in bare specimens and coated specimens and the specimens and stress vs time-to-failure curves are shown in Fig. 13. The ratios of ath to Oy were similar in the NACE TM-01-77 SSC test and the full thickness SSC test with the bare specimens. Larger Oth/Oy ratios were obtained with the full thickness coated specimens, as compared with the former two, and the ratio was between 0.8 and 0.9, about 40% larger than those of the bare specimens, when only one surface was exposed to the environment. Figure 14 compares calculations of hydrogen concentration patterns in these four small specimens with that in an actual pipe. The calculations indicate that the situation in the actual pipe in the sour environment can be best simulated in the specimen with only one surface allowed to take up hydrogen. Use of the bare specimens, in which the hydrogen concentration is uniform, results in the development of SSC at a lower stress than that present in a pipeline. Comparisons of the SSC threshold stress (Oth) in the NACE test and SSC behavior in the full scale SSC test for a wide range of welded linepipes is shown in Fig. 15. All pipes were stressed to 80% or 85% of SMYS in the full scale tests. The tests were continued for 2 months without any leakage through cracking. The solid marks (O II) mean SSC developing in the full scale tests, while the open marks (O [2) mean no SSC. One pipe for ordinary use showed SSC in the H A Z in association with HIC (Fig. 16) when stressed 1.27 times the O~h of the NACE SSC test but others for ordinary use showed different degrees of HIC development, depending on their HIC resistance, but no SSC. Another SSC was found in a pipe for sour service when the pipe was tested at 1.36 times the ~rth. These observations suggest that the occurrence of SSC in the pipe differs from that in the NACE SSC test because of the difference in the hydrogen concentration pattern, and the threshold stress for SSC development in the pipeline is higher by about 30% than that of the NACE SSC test. Further systematic studies are needed for understanding of environmental effects on SSC in welded linepipes.

1134

Y. KOBAYASHIer al. c/co. 1.0

//////////////////// C/Co • 0.99

/

C/Co • 0.99

0

,O__ /

"//I////////"

~////////.£'C{///I/////;

/

3&l

39.1

,,q

o. NACE-SSC Specimen

I I

I//I////

C. Full-thlckne. SSC q)eclmen ITwo surfaces coaled) c/co=

b. Full-thickness SSC specimen ( Non Coated) c/co=

"/I////.I

.0//17/I/.

/.LL/-~-~-~.~-/-~

0.8 0.6 0.4 02

,lL 0 38,1

_

d. Full-thickness SSC specimen (Three surfaces coated)

e. Actual pipe Imm)

Note) C :Hydrogen content Co :Hydrogen content at corroded surface / / / / : Corroded surface

FI6. 14.

Calculations of hydrogen concentration patterns in specimens and pipe.

500 "~

0

CR PLATE

I-

[]

O t.T PLATE

i

o

03 03

~;

~ go _~

F[t~. 15.

:1

0"[-1=l.360"th(For Sour Gas ervl

400

<

I

,

/ /

X'.t /'O'.=l.2T~th /-i,= /~/ ~ /I o/o "~'I/D l

oo/

/

----ISolid :SSC op,.

:

.0ssc

200 200

500 300 400 O"th MPo Threshold Stress of NACE SSC Test

AcomparisonofSSCoccurrenceinNACETM-01-77testandfullscaleSSCtest.

o

Resistance of welded linepipes to SSC

1135

SUMMARY

SSC behaviour has been studied with various laboratory SSC tests and full scale SSC tests. The results obtained are summarized as follows. (1) For the parent materials of welded linepipes, homogenized fine microstructures, such as tempered martensite and low carbon bainite, generally show a better SSC resistance than the usual ferrite/pearlite microstructures. They can be obtained by reducing the carbon content below 0.05% as well as by quenching and tempering. (2) Welds in low or medium strength steels fail mostly at the intercritical H A Z in the SSC tests. A significant decrease in the yield strength and microstructural changes, which are caused by the thermal cycles during welding, are both responsible for SSC developing along the relevant region of the weld. (3) Comparison of hydrogen concentration patterns in specimens with that in a pipe shows that a coated full thickness specimen allowed to take up hydrogen from one surface can best simulate the situation of the pipe exposed to a sour environment, and shows about 40% higher O'th/Oy ratio than bare specimens. (4) Comparison of the NACE TM-01-77 test results with occurrence of SSC in the full scale SSC tests suggests that the threshold stress for SSC development in the pipeline is higher by about 30% than that of the NACE TM-01-77 test because of differences in the hydrogen concentration pattern.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

REFERENCES A. A. OMAR,R. D. KANEand W. K. BOYD, Corrosion/81. Toronto, Canada, Paper No. 186 (1981). G. J. BIEFER, Mat. Perf. 21, 19 (1982). G.J. BIEFERand M. FICHERA, CANMET Rep. MRP/PMRL 82-3, March (1982). T. TAIRA, K. TSUKADA,Y. KOBAYASHI,H. [NAGAKIand T. WATANABE,Corrosion 37, 5 (1981). K. UME, T. TAIRA, T. HYODO and Y. KOBAYASHI,Corrosion/85, Boston, Paper No. 240 (1985). C. CHRISTENSENand R. T. HILL, Corrosion/85, Boston, Paper No. 241 (1985), H. J. CIALONE and D. N. WILIJAMS, Proceedings o f A G A NG-18/HLP Joint Tech. Meeting on Linepipe Research, Sept. (1984). TWI Group Sponsored Project Report 5537/28/85. T. TAIRA, K. MATSUMOTO,Y. KOBAVASm,K. TAKESm6Eand I. KOZASU,Int. Conf. on Technology and Applications of HSLA Steels, Philadelphia (1983). T. TAIRAand Y. KOBAYASm,Corrosion/81, Toronto, Paper No. 183 ( 1981 ). T. TAmA, Y. KOBAYASm, K. UME, T. HYODO, N. SE~:I and T. NAKAZAWA,Corrosion/83, Anaheim, Paper No. 156 (1983).