Weld Improvement Example in Offshore Oil & Gas

Weld-impr-example_sect_1_5 1

LIFE TIME EXTENSION AND

INTEGRITY OF OFFSHORE STRUCTURES

5_WELD IMPROVEMENT METHODS

Why and How to grind

by use of a burr grinder?

• Crack growth due to fatigue loads – example

• Why do grinding improve the fatigue life?

• Corrosion and Crack growth

• Procedure for grinding

• Where to perform grinding by a burr grinder?

• Filings and protection

Crack growth in the braces

End of external doubling plate

Fatigue crack observed after 5 year

Seng

Typewritten Text

Seng

Typewritten Text

Seng

Typewritten Text

Weld Improvement Example in Offshore Oil & Gas

Seng

Typewritten Text

Seng

Typewritten Text

Seng

Typewritten Text

1


Class improvement *1.3,

Fatigue life *2.2

Ref- IIW, Commission XIII

Post Weld Treatment

Burr grinding TIG DressingRe-melting the material at the weld toe.

• Improved profile

• Remove weld toe flaws

• Plate thickness > 6mm


Fatigue life *2.2


Hammer PeeningCompressive residual stress at the weld

toe.


• Compress weld toe flaws



Fatigue life *2.2Needle PeeningCompressive residual stress at the weld

toe.


• Compress weld toe flaws



Fatigue life *2.2

Effect of grinding

• Remove smaller weld defects i.e. Longer crack initiation period

• Smother transitions i.e. reduce stress concentration that reduce the crack growth rate.

• A deep groove will increase the stress concentration i.e. increased crack growth rate

• Grinding increase the fatigue life with a factor of 2 to 6

• A ground grove of 1.5 – 2mm give capacity as a weld without grinding.

Corrosion and fatigue crack growth

• Corrosion results in a shorter crack initiation period

• Corrosion results in an increased crack growth rate

• Corrosion reduce the fatigue life estimate with a factor of 2-2,0

-1,0

0,0

1,0

2,0

3,0

4,0

5,0

20 220 420 620 820

Accep

tab

le G

roove D

ep

th

Initial Fatigue Life Estimate

Acceptable Groove Depth as function of initial

fatigue life estimate and condition at groove

No Shape

Haz

No gr.

Corrosion, No Shape

Haz

No gr.

Air, No Shape

Haz

No gr.

Air, Shape

Haz

Grind

Air, Shape

Haz

No gr.

Shape

Haz

No gr.

Area code

Pos. nr.

Clock Pos.

No Shape Haz No gr.

Shape Haz No gr.

Shape Haz Grind

PK3 3 0100 1,4 2,9 4,0

PK3 3 0200 0,7 1,9 2,8

PK3 3 0300 0,8 2,0 2,9

PK3 3 0400 0,5 1,6 2,4

PK3 3 0500 1,4 2,9 4,0

PK3 3 0600 3,0 6,8 8,0

PK3 3 0700 2,1 3,8 5,1

PK3 3 0800 1,2 2,6 3,6

PK3 3 0900 1,2 2,6 3,6

PK3 3 1000 1,5 3,0 4,2

PK3 3 1100 2,1 3,8 5,1

PK3 3 1200 3,0 5,8 7,5

Acceptable groove depth


Figure 1 A butt weld given damage due to corrosion. Weld material is grey, corrosion damage is red and groove shape is yellow.

Case: External corrosion resulted in a flaw at the brace in a butt weld. The flaw is ground by a burr grinder to an elliptic shaped groove in the direction of the stress. Wall thickness of the brace is 32mm and has a diameter of 2.5m. Max groove depth will as far as possible be located outside the weld and HAZ area.


Ok, However

only grinding of

weld toe was

required. (Also

weld toe between

the string of

welding.)

Not as wanted as

grind disk and flap

disk has been

used. Needle

peening was

required.

Figur 7.5 Slipt spor etter fjerning av Samgong-anode Normal quality, removal of anode.

FEM- model of weld detail of Ring stiffener

55 mm

20 mm

Grinding for weld improvement, has effect on

the Stress Intensity Magnification Factor, Mk

The Mk function estimated base on detailed analysis

da/dN = C⋅(Y·Mk · ∆σ·√πa )m

0,6

0,8

1,0

1,2

1,4

1,6

1,8

2,0

2,2

2,4

2,6

2,8

3,0

3,2

0,0 0,1 0,2 0,3 0,4 0,5

a/t

Mk

Membrane stress, rho = 1 mm

Membrane stress, rho = 10 mm

Bending stress, rho = 1 mm

Bending stress, rho = 10 mm

The Magnification Mk factor also for toe grinded welds Validation of the probabilistic fracture

mechanic model

• Deterministic validation. The FM model applied v.s. a

reference FM model.

• Probabilistic FM model v.s. probabilistic SN-model

• Sensitivity sign check

• Standard checks in structural reliability i.e.

- absolute sensitivity value check

- importance factor check

- simulation v.s. FORM/SORM


The Challenge of Quality - Workmanship

• Case Veslefrikk B

– Detailed engineering and analyses

– Detailed follow up activity at site

– The Integrity engineers at site

• Case Drilling Unit– Detailed engineering and analyses

– Normal follow up activity at site

– No Integrity Engineer, Site and Inspection Engineers

Case Drilling Unit – North Sea,

Detailed FEM , FLS-analyses in 1999 and 2002:

• Reconstructed in year 2000

•Detailed As-Is and probabilistic fatigue assessment in 2004

•RS2009 included extended inspection scope, twice of normal

scope

•IS2011 NDT scope closely to 150% of normal renewal scope

Drilling Units – after RS2004

• The experience shows that the critical fatigue failure are mainly due to gross

errors in design, fabrication or inspection scheduling.

– The FLS analysis may be insufficient w.r.t. extent or quality

– gross error during design or fabrication

– knowledge not developed based on in-service observations i.e. validate the

prediction model (SRA-model)

– Re-analysis based on incomplete as-is description or re-assessment

process have not included knowledge from fabrication and maintenance

• Successful implementation of SRA for fatigue assessment fulfil the NORSOK

requirements:

To predict and to determine with a reasonable level of confidence the

existence, extent and consequences of deterioration, damages and defects of

the structural components

• We recommend to include an inspection program for verification of the

predictions of the SRA model in addition to the inspections scheduled by the

SRA model

Every inspection of a defined inspection point is given one row in

the spreadsheet. The spreadsheet is based on “positive” reporting.

i.e. both “finding” and “no-finding” are recorded.

Number of items (rows in spreadsheet): 1124 + 282 Number of findings: 204 + 83 (43 hot spot)

Number of cracks: 100 + 83 (43 hot spot)

Number of cracks not finally repaired: 47 (24 hot spot)

All 0-30%

prob fat

30-70%

prob fat

70-100%

prob fat

# of cracks: 83(43) 42(17) 31(18) 10(8)

# not finally repaired:

47(24) 24(12) 20(10) 3(2)

Table 4.1 Number of cracks detected up to now in year 2009. The

numbers in brackets are number of different hot spot given detection

of crack.

Drilling Units – offshore RS2009

Year detected Potentially

fabrication defect

Potentially propagating

cracks

Most likely

propagating cracks

Per cent belief of

crack to be propagating

< 30%

Between

30-70%

>70%

1980 4 6 3

1981 1 1

1982 2 1

1990 2 1

1994 1

1998 2 15 10

2003 2 1

2004 22 19 7

2009 42(17) 31(18) 10(8)

Total 72(47) 76(63) 35(33)

Classification of cracks. The number in brackets represents number of hot spot given

detection of crack and not the actual number of cracks.


# of Hot Spots

within group

Range of Fatigue life in

years

Cumulative list

# of Hot Spots

0 1-5 0

2 5-10 2

13 10-15 15

21 15-20 36

36 20-30 72

25 30-40 97

38 40-60 135

62 60-100 197

91 100-300 288

121 300-∞ 409

Summary of the calculated fatigue life estimates. * - high SCF included due to lack of information for hot spots that

potentially have low fatigue lives.



Inspection scheduling – Principle

The theory predicts approximately 3-15 per cent of the scheduled inspection to detect small fatigue cracks. The probability for detection of a through thickness crack is less than 0.5 per cent. The answer is given by the future inspections. Significant deviation between prediction and observation motivate for a re-assessment.

•For RS2009, internally a total 200 NDT inspections done•Findings are 10 or 42 fatigue cracks detected•Percent detection of crack; 10/200 = 5%, 42/200 = 21%

•Inspection result indicate the fatigue crack potential to be higher than assumed in the inspection scheduling.•Inspection result indicate very poor welding quality, re-welding and improvement will be proposed.

Present fatigue crack growth potential

According reporting in April 2009 the following cracks were identified as the most critical:

Item Description Fatigue life in years

Used SCF

1 SC1-10-06, end of stiffener, ground out by burr grinding to a depth of 15mm, wt=24mm

0,18, (52) 4,3 (*1,1)

2 SK3-5-0600 – weld repair given initial large root failure, latest repair in June 2009 reports root

failure of 3-4mm.

0,21 (32,5) 3,7 (*1,1)

3 SK3-5-1200 – weld repair given initial large root failure, latest repair in June 2009 reports root

failure of 3-4mm.

0,34 (52,2) 3,7 (*1,1)

4 SK3-5-0300- end of stiffener, ground out by burr grinding to a depth of 9mm, wt=23mm

0,39 (39,9) 3,35 (*1,1)

Table 1 The critical item and fatigue life estimates for present condition. Fatigue life and

SCF in brackets represent the values without any crack or groove present.

Remaining fatigue capacity of brace after leakage

Use of BS7910 for flaw assessment

Procedure applied:

- Use of quality class

- Model of through thickness crack in a plate - A half pipe model given the crack at the edge

The analyses done is not a detailed assessment including estimation of crack size given

final cracking. However, the analyses give an indication of time from leakage to loss of

brace. Based on the result in table 1 and 2 we select the SCF = 2.5 and SCF = 3.5 as upper and lower boundary for a reasonable time estimate. The fatigue life estimate is

than between 0,26 and 1,07 year.

NB!! We have limited the assessment to one crack even if section SK3-5 has critical hot

spot at three clock positions.

Conclusions – 10.09.2009

• Ref. conclusion after RS2004

• Additions:

– The internal inspection for RS2004 was not managed by skilled

personnel – most likely the poor welding quality would have been

focused in 2004 if the inspection team of 2009 had been used

– Modification of cracked area and other high stress area , will require very

god weld and detailed design quality. Ongoing activity and proposal to be

delivered DNV the following weeks.

.

Case Drilling Unit, North Sea, Status March 2013

• Extended and frequent NDT inspection

• No findings compromise the safety against larges disaster

• Number of findings larger than acceptable level of 3-15 % finding and 0.5%

fatigue failure (TTC - Through Thickness Crack)

• 21% and 5% during offshore survey RS2009

• 10% and 1% totally for RS2009, (1633 NDT inspections )

• 2% and 1% for IS2011. ( 107 NDT inspections )

• Des.2012: Ca. 40 NDT for SC4 og PC4 => 50% og 22% byAS2012

• Mars2013: 153 NDT, 7 (14) find => 5% (9%) og 0% by AS2013


Compensating measures:

•Install stress gouges

•New structural analyses delivered from 2 independent vendors

•Inspection of critical area after storm periods

•Start of design for re-construction

Case Drilling Unit, North Sea, Status March 2013 Preliminary results from the stress gouge measurement after 1

winter season

• Measurements indicate 20% longer fatigue life estimate than the analyses

from vendor 1

• The fatigue load effect in the braces are high but within acceptable limits

• The fatigue load effect is locally within the column base tank too high.

Reconstruction is required.

Centre line

Clockposition

MeanRatioS

95% Confidence, Upper limitRatioAWSG

95% Confidence,Lower limitRatioAWAA

0600 0,97 <0,95, 2,03> <0,72, 0,8>

0300 1,16 <1,04, 2,64> <0,71, 1,11>

0900 0,56 <0,65, 1,05> <0,43, 0,5>

1200 0,44 <0,34, 1,04> <0,24, 0,44>

SCF=1,53, Traditional design

R=2000mm,

SCF=1,08

SCF=1,2, behind the TIP,

Factor for fatigue life = 2,3

Weld improvement factor = 4

Total factor = 9.2

End of external doubling plate, (#72)

Fatigue crack observed after 5 year


Alternative Follow-Up Regime for Structure:

• Level 0 : Standard “DNV” Follow-Up Scheme - IIP

• Level 1 : Based on detailed in-service FLS analyses

• Level 2 : RBI analyses including inspection history

• Level 3 : Extended by As-Is description and improved

follow-up regime for inspection, repair, modifications and

weld improvement

Validation Studies – The Benefit of Refinements

• Studies for Philips Petroleum, bp, TFE, PSA-N and HSE-UK in the period 1995-1999,

• 3 published reports by HSE-UK and more than 10 papers

• Detailed assessment of more than 500 fatigue cracks

Main Conclusion• Traditional follow-up regime – level 0 and level 1

– 20-40% of the detected cracks are in accordance with the initial fatigue analyses

– Crack growth does mainly occur at details given poor fabrication quality

– All units given comparable amount of inspections

• AIMS based approach – level 3 and partially level 2– The calculations are based on As-Is description

– 70-80% of the fatigue crack growth are according the calculated crack growth potential

– The scope of inspection varies from 30 to 200% of a traditional follow up scope

Competance at Site

Unit 1 – Level 3

• Extensive and detailed follow up regime at site by competent

personnel

• Upon 10 years operation

– 73 fatigue cracks of 1 at weld improved area i.e. hot spot given ultra high

fatigue loads

Unit 2 – Level 2

• Traditional site team for inspection, repair and modification

• Upon 5-years operation

– 180 fatigue cracks at 143 different locations

– All cracks at hot spot given high fatigue loads

– Main reason: Poor workmanship for weld improvement and local modification

Page 52

Integrity Management

vs.

Inspection Management

Inspection

Integrity Engineering and Inspection

working together on-site, on-shore and off-shore.

Integrity Eng.

Condition

Assessment –

The process of

What to be done

Perform

inspection –

The process of

How to do