Asme CTOD Fracture [PDF]

Zitiervorschau

A critical view on the significance of HAZ toughness testing Andreas Liessem Marion Erdelen-Peppler

Europipe GmbH, Mülheim, Germany Salzgitter Mannesmann Forschung, Duisburg, Germany

International Pipeline Conference IPC 2004 October 4 - 8, 2004 Calgary, Alberta, Canada

TP63

EUROPIPE. The world trusts us.

Proceedings of IPC 2004 International Pipeline Conference October 4 - 8, 2004 Calgary, Alberta, Canada

IPC04-0315 A CRITICAL VIEW ON THE SIGNIFICANCE OF HAZ TOUGHNESS TESTING Andreas Liessem Technical Service Europipe GmbH Formerstrasse 49 45470 Ratingen, Germany [email protected] ABSTRACT Within the heat affected zone (HAZ) along the weld seam of LSAW linepipes discrete microstructural regions of reduced toughness can not be avoided and are commonly designated with the term Local Brittle Zones. The nature of these LBZ has been intensively investigated and the gathered knowledge is exploited in today’s steel technology, plate processing and pipe manufacturing. The HAZ toughness has been improved in general by reducing M-A constituents and by austenite grain refinement. Nevertheless local areas of low toughness within the CGHAZ can not be avoided completely. They are statistically distributed in every pipe. Furthermore it seems to be widely accepted that the structural reliability of LSAW linepipe produced and inspected with state-of-the-art technology is not influenced as these areas of low toughness have a limited size and distribution. This has been demonstrated by numerous investigations including small scale (CVN, CTOD), wide plate and burst tests. The essence of these investigations is that the failure behaviour of linepipe containing part wall defects in the HAZ is toughness independent. So far researcher’s world is clear and in good shape. Nevertheless many linepipe specifications tend to stipulate stringent test requirements with regard to acceptance criteria for the HAZ. In the occurrences of test failures a re-test procedure for test lot acceptance is carried out. As a matter of fact the LBZ are present along the weld seam over limited areas in each pipe. Therefore such a re-test procedure is regarded to be inappropriate in terms of quality inspection as it randomly sorts out pipes just by the statistical chance. With regard to HAZ toughness the pipes failed by this test do not differ from those pipes accepted and released for dispatch. As a final conclusion it can be stated that the existing test procedures for the HAZ toughness testing of the main standards and specifications do not reflect the current developments with regard to improved HAZ toughness achieved by the development of optimised steel composition and with regard to the enhanced defect detection probability along with modern NDT inspection methods. An amendment of the current test

Marion Erdelen-Peppler Salzgitter Mannesmann Forschung Ehinger Straße 200 47259 Duisburg, Germany [email protected]

procedures in this direction is proposed. Therefore proposals are made as start for a common discussion. Keywords: HAZ toughness, submerged arc welding, defect probability, CVN test, CTOD INTRODUCTION Within this paper the main influencing factors for the toughness in the heat affected zones within the production of LSAW pipes will be reviewed. Statistical evaluation of previous large scale production will be presented and reviewed with regard to the requirements of major standards and specifications. Beside the material properties the occurrence probability for defects and imperfections plays the second important role within the assessment of structural integrity. The use of most modern, automated welding and inspection technologies therefore represents a prerequisite to assure high quality with low defect probability and need to be considered in defining toughness requirements for CVN and CTOD testing. Fixed fracture toughness requirements as 0,20mm CTOD are more frequently specified, but do not adequately consider the intended application and service conditions and should be replaced by minimum values specific for the intended combination of applied stress, crack dimensions and material fracture toughness. NOMENCLATURE LBZ: Local brittle zone HAZ: Heat affected zone CTOD: Crack Tip Opening Displacement

HAZ TOUGHNESS IN LONGITUDINAL SUBMERGED ARC WELDED LINEPIPES High strength large diameter linepipes used for long distance or deep water pipeline are manufactured in the most economic way by submerged arc welding (SAW) in two passes. This high performance welding process is characterised by a

1

Copyright © 2004 by ASME

high heat input with cooling conditions that impacts on the toughness properties in the zone adjacent to the weld. The microstructure and herewith the toughness of this heat affected zone (HAZ) is mainly influenced by the cooling time t8/5 from 800°C to 500°C, the maximum reheating temperature and the chemical composition. The HAZ is characterised by a wide range of different microstructures, depending on the distance from the fusion line and the cooling conditions (Figure 1). Grain coarsened HAZ FL CV notch positions

Fine Grain HAZ Intercritically heated zone

Intercritically reheated fine grain HAZ

Intercritically reheated grain coarsened HAZ

Measure

Figure 1: HAZ microstructural regions within two pass weld with typical Fusion Line CVN position (acc. [1]) Generally the lowest toughness values are expected in the grain-coarsened heat affected zone (CGHAZ), as the toughness decreases with increased heat input and an increase in grain size. With increasing cooling time the microstructure in the HAZ is transformed from martensite to upper bainite as shown in Figure 2. This transformation to upper bainite is shifted to longer cooling times by significant additions of nickel which leads to lower FATT even at higher carbon equivalent. 100% martensite (M) Upper bainite (UB) (with MA constituent) martensite + lower bainite 100 80

with positive influence on FATT

Upper shelf energy

Reduction of C and CE

l

l

Limitation of Nb

l

Controlled use of Ti

l

Addition of Ni

l

Low S, P, O

CEIIW =0.42

Remarks

Decreases hardenability and limits formation of M-A constituents Decreases hardenability and limits formation of M-A constituents

l

Restricts austenite grain coarsening Retards the formation of upper bainite to longer cooling times

l

Reduces precipitations and segregations

(0.11% C)

60 40 20

Table 1: Improvement of HAZ toughness by adapted steel composition

CEIIW =0.37

0 FATT (°C)

o f M-A constituents. Both microstructural features typically reveal low toughness properties. As the cooling conditions in two pass SAW welding remain virtually constant irrespective of wall thickness the grain growth and the formation of M-A constituents has to be limited by appropriate chemical composition. The main measures currently known to improve HAZ toughness by optimised chemical composition are summarised in Table 1. Reduction of both carbon and carbon equivalent play a predominant role in avoiding the formation of M-A constituents. Furthermore the controlled use of limited additions of microalloying elements is reported to inhibit the grain coarsening by formation of finely dispersed nitrides and/or oxides [3[5]. These favourable effects have been confirmed by weld cycle simulation tests for a variety of steel compositions. [2-[5].

CE IIW =0.55 (1% Ni)

-20 -40

CEIIW=0.72

-60

(3.3% Ni)

-80 -100

CEIIW =0.73

-120

(5% Ni)

-140

Range for pipeline girth welding

-160 -180

Range for two-pass SAW

-200 1

10 100 Cooling time from 800 to 500°C (s)

1000

10000

Figure 2: Correlation between FATT and the cooling time t8/5 for different steel compositions and microstructures occurring in the CGHAZ (acc. [2])

The review of the published results shows that there is a fair level of understanding and that this knowledge is continually exploited in the steel chemistry and processing routes for linepipes, by which the overall HAZ toughness has been considerably improved and the statistical probability of having low toughness values in the HAZ has been reduced. Nevertheless values below the minimum requirement are still found due to the presence of LBZ’s when the notch is favourably positioned in the CGHAZ. LBZ’s are discrete microstructural regions of low toughness within the CGHAZ, but surrounded by microstructures with higher toughness. As the toughness of the CGHAZ depends mainly on the chemical conditions and the welding procedure (heat input, cooling time), which are technically constant due to strict control within narrow ranges for a specific linepipe production, it becomes obvious that LBZ are existing in each pipe of production.

The upper bainite is coarse grained and exhibits fractions

2

Copyright © 2004 by ASME

HAZ TOUGHNESS TEST REQUIREMENTS AND PRODUCTION RESULTS In Table 2 the test and acceptance criteria of some major linepipe standards and specifications are compared. Differences can be found mainly with regard to notch position in the HAZ, the amount of tests and the acceptance criteria. The HAZ is generally tested in case of offshore application, whereas HAZ testing is not mandatory for most onshore standards. In most cases the specimen with notch position “Fusion Line” (FL) is defined to sample 50% weld metal and 50% HAZ. Other specification requires a position where the root of the notch is in the CGHAZ to the maximum extent possible.

location giving the lowest minimum average is set to be taken for production control tests.

3 2 1

6 5 4

Standard

Application

Test

Acceptance criteria

DNV OSF101

Offshore

Charpy-V

38/45J at FL, FL+2, FL+5 (outside sub-surface and in addition mid thickness for t>20mm)

CTOD

Not required for HAZ 0.20mm min. for base material and weld (MPQT only)

Charpy-V

38/45J at the notch location with the lowest min. average during MPQT

CTOD

for information only from base material, HAZ, weld as an option during MPQT

Charpy-V

34/45J at FL, FL+2, FL+5 (outside subsurface)

CTOD

for information only from base material, HAZ and weld during MPQT

ISO 3183-3

Statoil RSP-230

Offshore

Offshore

ISO 3183-2

Onshore

Charpy-V CTOD

Both tests not required in HAZ

CSA Z245.1-02

On/Offshore

Charpy-V CTOD

Both tests not required in HAZ

DEP31.40.2 On0.30 /Offshore

Charpy-V

34/45J at FL, FL+2, FL+5 at mid thickness

CTOD

Not required

Table 2: Requirements of important technical standards with regard to HAZ toughness testing for grade X65 or equivalent The test and acceptance criteria according DNV OS-F101 turn out to be the most stringent ones. In total, 6 sets of specimen have to be positioned at different areas within the HAZ. In Figure 3 the notch positions of these 6 sets are marked in a macrograph of a weld with 34,9mm wall thickness. Compared to the STATOIL specification R-SP-230 and other standards for offshore application the amount of testing is doubled and consequently the probability of Charpy-V test failure. The current ISO 3183-3 requests only the notch locations no. 1 to 3 of Figure 3 to be performed during manufacturing procedure qualification test (MPQT). The notch

4a 5a 6a

Figure 3: Macrograph of weld cross section (34,9mm wall thickness) with possible HAZ notch locations acc. DNV OS-F101 With this macrograph it can be demonstrated that the weld seam geometry, which in practise is not perfectly symmetric, plays an important role with regard to HAZ Charpy-V notch testing. As the heat input increases with increasing wall thickness the width of the HAZ increases. Therefore it is the purpose of the linepipe manufacturer to limit this unavoidable energy increase as much as possible. The most effective measure to lower the heat input is increasing the welding speed. The welding lines of Europipe will therefore by equipped with new power sources allowing the use of up to 4 wires for inside welding, respectively up to 5 wires for outside welding. Hereby a heat input reduction of approximately 10% is enabled. Alternatively the heat input can be limited by reducing the weld seam cross section. Such a “narrow” weld cross section will have on the other hand a much steeper fusion line profile. A notch positioned at the fusion line boundary (50% weld/50% HAZ) of such a steep line lies with a higher percentage of its length in the CGHAZ as in case of a less steep fusion line as indicated in Figure 3 by notch position 4a. Of course the failure probability for a notch along such a steep fusion line is increased despite the lowered heat input. In Figure 4 the production results (SAWL 450, 34.9mm WT) for Charpy-V notch tests with notch location FL and FL+2mm according DNV OS-F101 are presented. It is evident that the great majority of values for FL are at a test temperature of –20°C significant above the minimum requirement of 36J. But some tests reveal values which are in the range of this lower limit with one value that failed. The distribution for specimens with notch location FL+2mm is shifted in general to higher values as the CGHAZ is sampled to a smaller amount. Failures are not reported for this notch location. This demonstrates the effectiveness of the measures with regard to optimised steel composition for HAZ toughness. As the results of FL+5mm are almost at the level of base material they are not shown here.

3

Copyright © 2004 by ASME

120

FL(root)+2mm

Min. Individual CVN toughness [J]

FL (root) 35 30 Percentage [%]

T= - 20°C 25 20 15

80 60 40 34J min. 20 0 20

5

FL

0 20 45 70 95 120 145 170 195 220 245 270 295 320 345 CVN toughness, individual Value [J]

Figure 4:

Influence of notch position on HAZ toughness

How the toughness values of HAZ depend on the heat input is revealed in Figure 5 and Figure 6 for different pipe productions of X65/X70 and –30°C test temperature. The test and acceptance criteria followed the requirements of the Statoil specification RSP-230. As the heat input increases with increasing wall thickness the minimum and average values for different HAZ notch locations are evaluated depending on wall thickness. As the width of the HAZ is growing with increased heat input/wall thickness the average toughness for the three different notch locations is reduced. For wall thickness below 25mm there is a more distinctive increase of the values reducing the risk of failures to virtually zero. With increasing wall thickness the probability for test failures increases to a maximum rate of 3% for FL notch position as the average values decreases. For FL+5mm test failures even for the thickest wall thickness have never been reported.

T= -30°C Grade X65/X70

250

150 100 50

45J min. average average 20

FL+2mm

24

FL+5mm

FL+5mm

28

32

36

40

Wall thickness [mm]

Figure 6: Influence of wall thickness on minimum individual CVN toughness for different notch locations in the HAZ The current procedure applied in case of a test failure is to reject the initially failed pipe and to retest two other pipes from the same test lot. In case that all retests from the two pipes pass the minimum and the average requirement the lot is released except the initially failed pipes. Such a retest procedure is in contrast to the nature of LBZ’s in the CGHAZ along the weld seam. The LBZ’s are characterised by a limited size of a microstructural magnitude but occur with a statistical distribution in each pipe length. This has been proven by the results of additional tests from failed pipes. In all cases within Europipe’s experience this second test passed successfully the minimum requirements. As the statistical risk for failures is generally low the chance of a second failure of the same pipe is nearly zero. Concluding the above it can be stated that the pipes initially failed and consequently rejected do not have different properties as those pipes accepted and released. To increase the number of notch locations or a higher test frequency will therefore not lead to higher quality of the pipe delivery, but only to higher reject rates and costs.

The likelihood of the different known welding defects is depending on the type of welding process as the main welding parameters like heat input significantly differ. As described above the SAW process is characterised by a high heat input with the discussed detrimental effect on the HAZ toughness, but the likelihood for having any planar defects like lack of fusion is very limited. A weld macrograph with marked locations for possible defects in the HAZ is shown in Figure 7.

200

0

FL+2mm

24

LIKELIHOOD OF HAZ RELATED DEFECTS AND THEIR PROBABILITY OF DETECTION

300

Average CVN toughness [J]

100

10

FL

T= - 30°C Grade X65/X70

28

32

36

40

Wall thickness [mm]

Figure 5: Influence of wall thickness on average CVN toughness for different notch locations in the HAZ

4

Copyright © 2004 by ASME

1 2 4 3

Planar surface breaking defect

Type of defect

Defect position 1

Detectability

Undercut

1

Slag inclusion

2

Planar defects, safe detection

Lack of fusion

2

Lack of overlapping/ interpene tration

3 or 4

Virtually not existing for SAW and perfectly detectable by AUT Safe detection (AUT weld seam, X - Ray, macro at pipe ends)

Safe detection (Visual, expanding, AUT weld seam) Safe detection (Visual, AUT weld seam) Minor undercuts can be tolerated

Figure 7: HAZ related defects and detectability by inspection methods of Europipe It is obvious that the number of possible defects is small and their detectability is high if modern NDT techniques are used. The non-destructive inspection of the HAZ starts already in the plate mill where a 100% UT-coverage of the plate body can be offered by the plate mills associated with Europipe. The quality standard of Europipe includes visual, automated UT and X-Ray inspection [6]. As the pipes are cold expanded these inspection are performed before expansion according to a significant higher sensitivity than required by specification. Hereafter the pipe is plastically deformed by the expander. Although this process virtually constitutes only the last forming step, it fulfils also a function as quality control. It should be noted that the residual stress level of the pipe is herewith significantly reduced. The pipe is then loaded in the mill hydrotester, normally in the range of 95% yield stress. The second automatic UT inspection of the weld seam is performed after hydrotesting followed by X-Ray inspection of the pipe ends and possible weld repairs. By this extensive inspection procedure the probability for undetected defects is very low. In a risk analysis performed by DNV on behalf of Europipe, the probability of a linepipe delivered by Europipe with an undetected failure has been estimated to be 8.5 10-8 [7]. FRACTURE MECHANIC EV ALUATION OF HAZ TOUGHNESS A major application of the assessment routines is the investigation of the structural integrity of welds. Welds are thought to be generally defective, thereby increasing the risk of failure in comparison to non-welded components. An important difference between the two that has to be accounted for is the inhomogeneity in terms of mechanical properties of the weld. As shown before, low toughness is commonly associated with welds in pipes and structural steels. Although the alloying and welding techniques have been optimized to reduce the amount of such low toughness areas, commonly referred to as Local brittle zones (LBZ), they cannot be completely avoided. The important question that has been subject of discussion within

the last years is the significance of such LBZ to the structural integrity of welded components. A considerable amount of research work was directed towards fracture toughness tests, wide plate tests and the different behaviour of those tests when encountering areas of low toughness in the HAZ. The influence of the size of LBZs [8[9], the sampling methods [10], notch location, weld geometry mismatch and loading mode [11[12] were investigated. It was shown that the CTOD test is very sensitive of boundary conditions, therefore it was even considered questionable it the critical CTOD is a material constant [13]. Wide plate tests were conducted to correlate the results of fracture toughness tests to the structural behaviour of welded components [14-[18], the discussions were controversial. Whereas some authors did not see the safety of the structures endangered [16[18], others found low CTOD values to coincide with fractures in wide plate tests [14,[15]. The following explanations for the discrepancy between fracture toughness tests, wide place tests and actual structures in service were found: • Fracture Toughness tests like CTOD are very “critical” in terms of placing pre-cracked defects so that they sample a large quantity of low toughness microstructure • Constraint is much higher in the deep notched SENB specimens than in the actual structures, therefore instability predictions based on CTOD results may be overly conservative and not realistic [11] • The probability for a combination of large defects, low toughness and high loads is very low in the real structure • In a large structure, the areas of low toughness are surrounded and supported by tougher material, CTOD tests aim at excluding this effect • If the above named mechanism does appear in a CTOD test in form of a pop-in, the assessment is very stringent although the appearance of pop-ins is also attributed towards the test itself (high constraint-high load) It should be noted, that the research work summarized above was mainly conducted on welds on structural plates for offshore industry. A relatively small amount of experimental work has been conducted on welded pipes exposed to internal pressure. In all cases, the pipes failed at stress levels above the yield strength, regardless of the low toughness values found in the HAZ. In [19], CTOD values down to 0.022 mm were obtained at -10°C. Ring expansion and burst tests were conducted where the test pipes contained defects in a range of depths up to 50% of the wall thickness and lengths up to 450 mm. The failures initiated from defects successfully positioned into the CGHAZ, yet the failure pressure level was well above the levels predicted by fracture mechanics assessment according to PD6493 [20], in fact, the test results suggested that the failure behaviour was toughness -independent. In [21], the defects were exposed to fatigue loading to sharpen the notch tip radius, the defect depth equalled 50% of the wall thickness. Regardless of the sharpened notch tip which is in this respect comparable to the notch tip in fracture mechanics tests, the pipes failed at stress levels above yield strength. The study was aimed at evaluating fracture mechanics concepts, it is noteworthy in this context that the Battelle concept [22] yielded

5

Copyright © 2004 by ASME

tool to estimate the critical crack sizes, the dimension of the assumed defects are so large that it is virtually impossible for them to be completely embedded in low toughness areas. 8 7 6 5 acrit [mm]

the most accurate predictions. A recent study [23] on seam weld defect tolerance applied a constraint based fracture mechanics analysis on a surface breaking defect. Despite the low toughness values that were found in the CTOD tests, the ring tension tests which were conducted exhibited plastic collapse failure mode. Standard fracture mechanics assessment routines implied the defective ring to be unsafe, whereas constraint based fracture mechanics assessments were able to predict the plastic collapse failure mode in a conservative manner. The question whether or not a brittle failure may take place is always related to the triangle of load-defect-toughness, the focus in the past being clearly held on the estimation of the toughness. Especially the estimation of a realistic defect size is of large importance when assessing the impact of low toughness. Modern NDT systems allow a safe detection especially of the more detrimental planar defects, a fact that should be considered in an ECA. A calculation based on BS 7919 [24] utilizing the computer program Crackwise [25] was conducted to demonstrate these effects. The load was assumed to be 70% of SMYS of grade X80 which is realistic for an internal pressure load in a pipeline. A calculation of the critical flaw height for a surface flaw and an embedded flaw positioned in different ligament heights varying the assumed toughness from CTOD = 0.01 mm to 0.1 mm was conducted. CTOD values below 0.1mm represent the lower bound HAZ results in critical situations, e.g. heavy wall pipes, as can be seen in Figure 8.

4

Detection level of NDT system

3 2 1 0 0

0,01

0,02

0,03

0,04

0,05

0,06

0,07

0,08

0,09

0,1

CTOD [mm]

Figure 9: Critical flaw height versus toughness for a surface flaw in an X80 pipe subjected to a load equal to 70% SMYS

14

12

60 T= - 10°C X65, 31,9mm WT

10 2a [mm]

50

Percentage

40

8 12-14

6

10-12 8-10

30

4

20

2

6-8 4-6 2-4 0-2

0

10

9

0,01

0,02

0,03

Ligament height [mm] 0,04

0,05

CTOD [mm]

0