Rudi M. DENYS Universiteit GENT, · PDF fileLabo Soete - UGent, Belgium 6 Girth weld defect/flaw acceptance ... Manual versus mechanized welding Type of flaws: Non-planar (SMAW/FCAW)

Rudi M. DENYS

Universiteit GENT, Belgium

Maaats

–M

ar

2012

Labo Soete - UGent, Belgium 2

Scope of Presentation

�Girth weld (pipeline) integrity

• Stress based design (“Stable soil”)

• Strain based design (Earthquakes, soil settlement, land slides, ..)

�Defect acceptance

• Workmanship

• Fitness-for-purpose (ECA)

• Major codes

• Comparison

� Issues

• Toughness testing

• Weld strength mismatch

• Non-destructive inspection

� Final observations

Maaats

–M

ar

2012


Stress vs strain based design – 1/2

REMOTE Straining capacity

Threshold

toughness

Strain based design

Ap

pli

ed

str

es

s

0.5 %

NSY

Remote stress < YS

Remote yielding

Remote stress > YS

Str

ess-b

ased

desig

n

Defect(3 % max)

X %. SMYS

Maaats

–M

ar

2012


Stress vs strain based design – 2/2

0 2 4 6 8 10

Remote strain (%)

Re

mo

te s

tre

ss

Onset of

Remote yieldingY0.5t

uEL

Ultimate strain

Capacity

Safety margin

Allowable

stress ( < YS)

STRESS

0 2 4 6 8 10

Remote strain (%)

Re

mo

te s

tre

ss

Onset of

Remote yieldingY0.5tuEL

Ultimate strain

Capacity

Safety margin

Strain

Demand

STRAIN“Very safe”

Today: SMYS and SMTS

Future: Actual properties

Maaats

–M

ar

2012


Girth weld (pipeline) integrity – Stress based designGirth weld (pipeline) integrity – Stress based design

� Girth welds failure is prevented by using:

�Design codes (Max. hoop stress – x % of SMYS)

�Material selection codes• Material properties (Yield strength, Y/T ratio*, )

• Adequate toughness (Charpy V or CTOD requirements)

• Matching welding consumables (transverse tensile test)

• Good weldability (prevention of cold cracking)

• Hardness restriction (sour service)

�Non-destructive inspection• X-ray (film) and / or ultrasonic testing

� However, … girth welds are a possible weak link in a pipeline string when flaws/defects occur.

Str

ess

Strain

UTS

SMYSx*SMYS

Hoop direction

* Axial tensile properties are “neglected” (Longitudinal vs Spiral welded pipe)

Maaats

–M

ar

2012


Girth weld defect/flaw acceptance - 1/2Girth weld defect/flaw acceptance - 1/2

� Girth welds might contain non-planar and / or planar defects

� To avoid unnecessary repairs*, welding standards allow some defects

� Acceptance criteria:

� Workmanship levels

• Are based on experience

� Engineering Critical Assessment (ECA) methods• Based on Fracture Mechanics and Plastic Collapse theories (stress based

design)• Under development (strain based design)

� UGent approach• Curved wide plate test (stress based design)• Full scale tensile test

* Repair of rejected (“small”) defects can be harmful (undermatch weld/new anomalies)

Maaats

–M

ar

2012


Girth weld defect acceptance - 2/2

Flaw/defect acceptance criteria

Workmanship criteria

Compliance with known (API, CEN, ..)or new criteria

Engineering CriticalAssessment (ECA)

ACCEPTABLE

Compliance with"ad-hoc“ criteria

YES YesYes No

REPAIRor

CUT OUT

No

Curved Wide Plate Testing

Maaats

–M

ar

2012


CWP and FST tensile testing

Full Scale Pipe Testing(Effect of internal pressure)

Curved Wide Plate Testing

Cooling boxes

1200 mm

300 CWP

Maaats

–M

ar

2012


1200 mm

300 CWP

Bending

Comparison of full-scale bend / CWP test results

CWP test provides conservative estimates

0,470,69

3,23

> 1,78

0

1

2

3

4

5

7t 5t

Defect length (mm)

Fa

ilu

res

tra

in(%

)

Curved wide plate tests

Full scale bend tests

X70

-30

" x

9.9

mm

Maaats

–M

ar

2012


Tolerable flaw size for stress/strain based design

Predicted (tolerable) dimensions (ECA)(EPRG or other methodologies)

Acceptable dimensions (sizing error)

Height and length!!

(l = 25/50 mm x h = variable)

Workmanship

Flaw size limits (via CWP / FST)

Fla

w h

eig

ht

Flaw length

WM vs ECA limits:

significant safety

margin

Strain based design

No height limit

Maaats

–M

ar

2012


Workmanship codes – Facts 1/2 Workmanship codes – Facts 1/2

� Based on what can be seen on the film (length)

� Are in use since 1938 (Pressure Vessels)

� Go / no go acceptance levels (API 1st Ed. - 1953)

� Various requirements (API, BS, CSA, CEN...)

� Toughness requirements ?

� Strength mismatch (transverse / all-weld metal tension tests)

� Specify/limit allowable length (planar flaws)

� High level of safety which is not known

� Ensure safe (girth) welds

� WMS are conservative

� Developed for manual welding

WMS criteria for GMAW welds ?

Maaats

–M

ar

2012


UGent - CWP data base (X60/X70 results)

Data base has been used to develop the EPRG guidelines

0

2

4

6

8

0,00 0,05 0,10 0,15 0,20 0,25 0,30 0,35 0,40

Defect area ratio dr (= lh/st)

Rem

ote

str

ain

at

failu

re, e (

%)

Remote strain = 0.50 %

Overmatched welds

Undermatched welds

Plain pipe

Y/Tpipe ≤≤≤≤ 0.90

CVN ≥≥≥≥ 30/40 J

and / or

CTOD ≥≥≥≥ 0/10/0.15 mm

50 x 3 mm2

Scatter ?

Maaats

–M

ar

2012


Workmanship codes – Facts 2/2Workmanship codes – Facts 2/2

� Issues

� WMS have no relationship to applied stress (Max ? – Hydro test / MAOP)

� WMS don’t refer to pipe grade

� WMS do not incorporate effect of the thickness effect

� WMS assume matching welds

� Progress in welding technology

� Manual versus mechanized welding

� Type of flaws: Non-planar (SMAW/FCAW) versus planar flaws (GMAW/SAW)

� Progress in NDE

� Past: X- ray inspection (film)

� Today: Automated Ultrasonic Inspection (AUT)

Maaats

–M

ar

2012


Fitness-for-purpose (ECA) codes (stress based design)Fitness-for-purpose (ECA) codes (stress based design)

� Require information on:

� Applied strain (stress)

� Strength, weld strength mismatch and toughness

� Defect length, height and depth (location)

� Can NDE procedure detect the significant flaws?

� Assessment methodology (failure conditions)

� Brittle fracture (low toughness)

� Plastic collapse

� Different approaches are available

� Determine tolerable (≠≠≠≠ critical) defect dimensions

Maaats

–M

ar

2012


Input parameters - Stress vs Strain Based Design

STANDARD INPUT FACTORS

- Applied (remote) stress

- Toughness

- Defect size (length, height, location)

- Tensile properties

STRESS

based design

Variability of input parameters

+

Supplementary input parameters

STRAIN based design

- Applied / remote strain (STRAIN DEMAND)

- Actual Pipe and Weld Metal Tensile Properties

- Full stress-strain curves (Uniform elongation capacity)

- Y/T ratio pipe and weld metal

- Level of weld strength mismatch

- Tearing resistance

- Internal pressure

- Wall thickness variations, etc.,

Specified valuesSpecified values

Maaats

–M

ar

2012


Flaw acceptance limit for strain based design ?

Use existing

(API / CEN) criteriaMethodology ?

Validate theirapplicability

forStrain-based designs

Girth weld integrity(Flaw acceptance criteria for high plastic strains)

Flaw acceptance limitfor high plastic strains

Remote yielding

Flaw

TIER 1

Workmanship Criteria

TIER 2

Fitness-for-Purpose(ECA)

Under development

Options:

Numerical / FAD

Experimental

Traditional pipe/weld qualification procedures provide insufficient

information.

Maaats

–M

ar

2012


Failure conditions (stress based design)

� Brittle fracture� Toughness = defect size x applied stress

• Brittle fracture is possible if CVN < 27 J

• EPRG excludes BF if CVN > 30/40 J (min./avg.)

• CTOD requirements ? – Can vary from 0.05 up to 0.25 mm

� Plastic collapse

� Applied stress = Flow stress x Defect size• Toughness threshold: 30 J / 40 J at design temperature (EPRG)

• Flow stress = f(Y/T) becomes a key variable

• Overmatching is a beneficial factor !!!!

� FAD diagram (combines BF and PC assessments)

� API, CSA, BS, …

A A

Cross Section AA

Plastic Deformation

ContainedYielding

Global (NSY)Collaps

e

Maaats

–M

ar

2012


Major ECA (pipeline specific) codesMajor ECA (pipeline specific) codes

� Semi-empirical (numerical + limited validation)

� API 1104 Appendix A

� CSA Z 184 Appendix K (and J)

� Others: (BS7910, DnV RP F101, …..)

� FAD diagram / various options (Level 1, Level 2, … )

� Empirical

� EPRG Tier 2 guidelines – 2010 / AS 2805 / CEN 12732

� Each of these codes have different requirements as to

� Applicability

� Weld procedure qualification (property requirements)

� Mechanical and toughness properties of weld and pipe material

� Inspection (sizing error)

� Predictions are not consistent (different approach / provisions)

Maaats

–M

ar

2012


ECA vs workmanship

� EPRG philosophy towards ECA based defect (anomalies) acceptance limits:

� Unless pipe and weld materia;s are fully documented, defect acceptance should be based on workmanship standards such as API 1104 or EPRG-Tier 1.

� ECA defect size limits should only be used when defects exceeding the workmanship limit are identified during a post construction audit.

� ECA based defect size limits are not intended to be used to justify poor workmanship.

� The presence of a larger defect, or many defects in a single weld, indicates poor quality control of the welding, and remedial action to maintain good workmanship is required.

� However, EPRG:

� Accepts that defects exceeding workmanship levels do not necessarily affect the fitness-for-purpose of a girth weld.

Maaats

–M

ar

2012


API vs EPRG

Plastic collapse solution needs to be revised !

PIPE PROPERTIES ● Pipe grade Not mentioned up to L555 (X80)● Y/T ratio 0.95 / unlimited 0.90

PIPE DIMENSIONS

● Pipe diameter D ≥ 30 inch● Wall thickness t 7.5 < t ≤ 25.4mm

User input (max. 0.5 %) 0.5 % (default value)

Stress or strain Strain

EXPERIMENTAL VALIDATION

● Full-scale bend tests 69 18

● Curved wide plate tests 31 (Grade X60 pipe) 485 (upto Grade X80)

TOUGHNESS REQUIREMENTS

● CVN (Min/Ave.) at minimum design temp. 30/40 J 30/40 J

Sampling position Close to weld crown Root region

Shear area 50% No requirement

● CTOD - Bx2B (Min) at minimum design temp. 0.10 or 0.25 mm Not required

WELD METAL REQUIREMENTS

● Cross tensile Yes Welding proc requirement

Reinforcement in place Yes No

Test requirement TS > SMTS Pipe TS > SMTS Pipe

● All weld metal tensile Not specified Yes

● YS matching requirement No gross undermatching SMYS + 100 MPa

APPLIED STRESS /STRAIN

API 1104 Appendix A

Option 1 (2007)

EPRG - Tier 2

(2009)

D/t ≥ 10

Maaats

–M

ar

2012


Toughness testing

� API 1104 and CSA Z186� Toughness obtained from CTOD testing

� DnV RP F101 � Toughness obtained from CTOD / SENT testing

� EPRG� Toughness obtained from Charpy V testing

CVN test can be used to ensure failure by plastic collapse

Strainbaseddesign

Maaats

–M

ar

2012


Toughness

� CTOD testing can produce disquieting low results, suggesting that only very small flaws would be acceptable

� API / CSA assumes that CTOD is the key variable in fitness-for-purpose based girth weld defect acceptance;

� If, the EPRG (30/40 J) toughness requirement is met, EPRG concludes that:

� Y/T ratio (strain hardening capacity) and,

� weld strength mismatch

have a greater effect on the allowable defect size than CTOD toughness has.

Maaats

–M

ar

2012


EPRG Guidelines 2001/2010 - Y/T ≤ 0,90

Wall thickness range*: 7 mm ≤ t ≤ 30 mm

= 3 t*

= 5 t*

= 7 t

Not applicable

h

(mm)

Flaw

height

> X80

3.2 t≤ 5

3.9 t≤ 4

5.3 t

l

(mm)

≤ 3

≤ X80+ 120 MPa

Predicted

lengths

Allowable flaw length

per 300 mm (12”) length of weld

Pipe Grade

(SMYS)

* Wall thickness, t > 8 mm for h = 4 mm and t > 10 mm for h = 5 mm

R1

R1

h

t300l

+

−=

Remote yielding

Remote stress >YS

Plastic collapse

Remote stress < YS

“7/5

%”

Maaats

–M

ar

2012


API vs EPRGAPI vs EPRG

0

3

6

9

0 25 50 75 100 125 150

Tolerable length (mm)

To

lera

ble

he

igh

t (m

m)

Without length

correction

D t = 40" x 0,7"

D/t = 57

Y/T = 0.90 (Pf = 0.95)

WMS

API 1104

collapse solution level 1

(CTOD > 0.10 mm)

"Denys" collapse model

(W = 300)

EPRG

limits

CVN = 30/40 JAPI 1104

collapse solution level 1

(CTOD > 0.25 mm)

Maaats

–M

ar

2012


EPRG vs API 1104 (with and without length correction)

0

3

6

9

0 25 50 75 100 125

Allowable defect length (mm)

All

ow

ab

le D

efe

ct

he

igh

t (m

m)

Denys - 300 mm collapse

EPRG 2009

API 2008 - CTOD = 0,10 mm - NC

With length correction

API 2008 - CTOD = 0.25 mm - NC With length correction

Dt = 40" x 0,62" (15.7 mm)

D/t = 57

Y/T = 0.90

WMS

EPRG

limits

CVN = 30/40 J

API 1104

Y/T = 0,90

(CTOD > 0.25 mm)

API 1104

Y/T = 0,90

(CTOD > 0.10 mm)

Maaats

–M

ar

2012


Tensile properties (“X80”) – Specimen geometry effects

Pipe Wall:

18 mm

5 o'clock

2 o'clock

WELD

3 o'clock

YSmax = 617

YSmin = 579

500

550

600

650

700

0 2 4 6 8 10

Engineering strain (%)

En

gin

ee

rin

g s

tre

ss

(M

Pa

)

Round bar - 3 o'oclock Round - 5

Round - 2 Flat full thickness - 2

Flat full thickness - 3 Flat full thickness - 5

Max. YS Min. YS

Luders plateau

uEL

Representative stress-strain curve ???

Maaats

–M

ar

2012


Effect strain hardening (Y/T ratio) – Can be an issue !!

Y/T = 0,92

Net Section Yielding

Y/T = 0,86

Gross Section Yielding

Surface breaking flaw: 75 mm x 3 mm

Low Y/T ratio steels perform better

Maaats

–M

ar

2012


Weld Strength Requirements – Strain based design

Undermatched weld

Applied strain: 1,2 %

Overmatched weld

Applied strain: 2,0 %

Weld reinforcement

can be effective

Flushed weld

Matching / Over-matching is a minimum requirement!

Maaats

–M

ar

2012


Factors affecting level of mismatch

VARIABILITY: Cannot be avoided

Incorporate effect of:

Specimen size

Sampling position

Sampling direction

(trans vs axial)Level of

Weld Metal Yield Strength

Mismatch

COMPAREWeld

reinforcement

"Heterogeneous

weld

Plate

PropertiesRolling practice

Alloy content/design

Wall thickness

Pipe forming

BASE Metal

Yield Strength

Pipe

Rolling

Welding

procedure

Groove/Bevel

Design

Consumables

Wire/rod

composition

Flux/Shielding gas

composition

WELD Metal

Yield Strength

Maaats

–M

ar

2012


Weld strength mismatch (Stress BD)

465 485 505 525 545 565 585 605 625 645 665 685 705

Yield Strength (MPa)

Pro

ba

bili

ty

SM

YS

(X

70

)

SMYS (X80) + 120 MPa

Tail Overlap

EPRG Tier 2

(Undermatching)

1SD

(20 MPa)

WM

Yield strength

distribution

(EPRG TIER 2 req.)

P(Axial direction)

WM

DnVPipe - SMYS+120/150 MPa

EPRG - MLYS+100 MPa

Pipe metal

Yield strength

distribution

(Hoop direction)

MW

YS

MLYS

Maaats

–M

ar

2012


Weld metal yield strength requirement (Strain BD)

0,00

0,01

0,02

0,03

0,04

0,05

0,06

555 575 595 615 635 655 675 695 715 735 755 775 795 815 835

Yield Strength (MPa)

Pro

babili

tyPlate/pipe and weld metal yield strength distributions

SM

YS

Pipe

1,18 SMYS

Strain

based

design

10 % on Ymax

34 % on SMYS

EPRG

P

(CEN pipe)

Tentative

Maaats

–M

ar

2012


All weld metal testing

UM

4

15

Prismatic

(rectangular)

specimen

6 m

m

d = 5.5 mm

d = 6.0 mm

7 9

.8

5.5

8.0

4

CAP

ROOT

MT

� Specimen geometry

� Cross section

� Sampling position

� Low/high strength root

� High/low strength fill

Maaats

–M

ar

2012


Effect of sampling on YS and YS/TS ratio

UM

UM

Verify sampling

position !

Maaats

–M

ar

2012


Ju

ly 2

6,

201

0

Variation of weld metal tensile properties - 1

� SMAW weld

� Pipe: X70

General average : 536 MPa5

39

50

8

50

8

49

1

51

3

49

2 49

8

53

3

50

8

58

0

55

4 55

9

56

0

49

7

51

3

55

3 55

7

56

5

56

2

52

8

537

494

557

542 543

527

513

523

545

547

485

505

525

545

565

585

1 2 3 4 5 6 7 8 9 12

Sampling location (o'clock position)

Weldmetal yieldstrength(MPa) min max ave

Maaats

–M

ar

2012


Flaw treatment

WELD FLAW(S)

VERIFY

FLAW

dimensions

Type

Surface breaking

Embedded

Depth

Length

Height

Interaction

(Multiple flaws)

Re-categorisation

(Near surface

breaking flaws)

Flaw Treatment

Maaats

–M

ar

2012


Surface breaking vs embedded defects

Inspection accuracy

� Focus inspection on root flaws

� Height and LENGTH inspection errors

� The inspection error for embedded defects is a minor problem

� Flaw treatment (re-categorization)

Remote strain at failure = 0.40 %



Girth welds contain flaws of comparable

size

Suggestion

� Eliminate ALL workmanship root flaws

� This measure is very effective if plastic straining capacity of flawed welds is a design requirement

Maaats

–M

ar

2012


Defect sizing – defect geometry effect

Surface breaking defect

Weld defects

"Salami" Low temperature tensile test

Buried defect of variable height

Verification of defect dimensions:

Macro sectioning (salami technique) vs low temperature tensile testing

Maaats

–M

ar

2012


Final UGent recommendations

� Level 1 : Workmanship criteria

� Level 2 : Pipe yielding (< 0,5 % strain)

� EPRG criteria

� Level 3 : Full ECA

� Are input parameters available ?

� Which methodology (API, BS, CSA,…) is best?

� Level 4 : Model (wide) plate testing

� Sub-standard welds

� Non-elastic design

EPRG requires that the

application of the ECA

defect limits can only be

justified when higher

weld quality is assured

Pipeline industry does not have yet generally accepted guidelines for

strain based designs

Documents

Rudi M. DENYS Universiteit GENT, · PDF fileLabo Soete - UGent, Belgium 6 Girth weld defect/flaw acceptance ... Manual versus mechanized welding Type of flaws: Non-planar (SMAW/FCAW)