proceedings of the asme 2015 pressure vessels & piping ... · proceedings of the asme 2015...

Proceedings of the ASME 2015 Pressure Vessels & Piping Division Conference PVP2015

July 15-19, 2015, Boston, Massachusetts, USA

PVP2015-45757

Comparison of Methods for Evaluation of Crack Growth at Welds in Offshore Risers

Fengjie Yin, Mark Cerkovnik, Wasy Akhtar, Tracy Yang 2H Offshore Inc.

15990 North Barkers Landing, Suite 200 Houston, TX, USA

Email: [email protected]

ABSTRACT

In checking the fitness of fatigue critical welded

structure, the stress concentration at the weld due to the weld

geometry needs to be considered. Where fatigue is assessed

using crack growth methodology, two approaches are

commonly used. In the offshore industry in regions where BS

7910 [1] is followed, the effect of weld geometry is assessed

using the Mk factor approach. The Mk factor directly magnifies

the stress intensity. Mk factor solutions are available for T-butt

weld joints from the British Standard BS7910.

Alternatively, API579 [2] offers stress intensity solutions that

can account for the stress profile through the wall thickness of

the pipe. In using this method, the engineer will use an FEA

program to find the stress profile for use as an input for the

stress intensity factor computation. Since the goal is the

assessment of crack growth, the stress profile must represent the

cyclic changes in stress. Further, a histogram of such profiles is

required.

While the Mk factor approach of BS7910 offers the easier path

by supplying factors for pre-solved geometries, the API

approach offers an opportunity to refine the solution by

conducting relatively simple linear FEA of the un-cracked

component. This study compares the two approaches using an

example taken from offshore riser fatigue analysis.

NOMENCLATURE

CTOD crack tip opening displacement

ECA engineering critical assessment

FCG fatigue crack growth

FEA finite element analysis

FFS fitness for service

HAZ heat affected zone

K stress intensity factor

KDF know down factor

Mk stress intensity magnification factor

R reference stress

R Ratio stress ratio

SCF stress concentration factor

SCR steel catenary riser

SN stress life

TDP touch down point

VIM vortex induced motion

VIV vortex induced vibration

INTRODUCTION

Most pressure vessels and pipelines are largely static with

fatigue cycles coming only from start-up and shut down cycles.

However, there is an important class of lines that are dynamic,

where mechanical load cycles are a fundamental part of the

loading. One example of this class of structure is offshore risers

on floating production systems. The risers act as conduits from

the subsea wells to the platforms. The motions of the host vessel

and the hydrodynamic loads result in fatigue stress cycles;

primarily from bending and axial loading. Dynamic loads

include first and second order vessel motions, wave loadings,

vortex induced vibration (VIV), vortex induced vessel motion

(VIM), pressure variations and slugging. Offshore risers used

for deep water applications are often fabricated by the girth

welding method. One typical girth weld configuration is shown

in Figure 1. Some flaws will usually be introduced during

welding procedure.

While for most cases the SN approach is used in design of

risers, fatigue crack growth (FCG) methodology also has an

important place. ECA is used to define acceptable initial flaw

Learn more at www.2hoffshore.com

2

dimensions and develop corresponding non-destructive

inspection criteria. If defects are found during inspection, they

are assessed using fatigue crack growth methods in FFS

assessment. Also, when the riser operates in a sour or corrosive

environment, it is often more practical to characterize the crack

growth rates in the environment than to develop SN curves.

When the SN method is employed in the assessment of welds,

the local stress concentration factor (SCF) at the weld toe and

weld root is accounted for in the fatigue curve; i.e. the fatigue

test specimen used to develop the fatigue curve have the typical

local weld geometry. Hence, no additional correction is needed.

However, when using the fatigue crack growth method to assess

welds, the stress concentration factor at the weld due to mis-

alignment must be accounted for.

In this paper, two different approaches of accounting for stress

concentration factor in fatigue crack growth analysis are

compared and contrasted. The two approaches stem from the

two most commonly used codes used for fatigue crack growth

analysis, namely API RP 579 [2] and BS 7910 [1].

API RP 579 [2] provides solutions using weight functions that

allow the analyst to input the stress profile through the

thickness. The detailed weight function methodology is

described in Section C.14 of API RP 579 [2]. BS7910 [1] gives

adjustment factors Mkm and Mkb in tabular form that directly

adjusts the stress intensity factor.

Figure 1 – Typical Girth Weld Configuration

Fatigue crack growth assessment for a typical offshore riser

using these two approaches are compared in this paper.

FRACTURE MECHANICS BASED STRUCTURAL

INTEGRITY ASSESSMENT Fracture mechanics based structural integrity assessments,

which are commonly referred to as ECA or FFS assessment,

date back to the 1960s. In the offshore oil and gas industry, the

focus is more on assessing the structural integrity of welded

offshore structures. It is widely accepted that all welded

structures contain flaws. The flaws are assumed to be crack-like

flaws in the fatigue crack growth (FCG) analysis. In offshore

risers, girth welds, which include weld metal and heat affected

zone, are the weakest link from FCG point of view because of

the local stress concentration and the initial flaws that may be

introduced by the welding process.

GIRTH WELD Circumferential girth welds in offshore risers are fatigue

sensitive locations. Welding involves the heating of metal to its

melting temperature followed by rapid cooling. Without

dressing, the welding procedure usually leaves a weld toe both

at the external surface and the internal surface. Mk APPROACH Weld toe stress intensity magnification factor, or Mk factor, is

defined as the ratio of the stress intensity factor including the

stress concentration effect at the weld toe to the stress intensity

factor of the same crack without the stress concentration effect

from the weld toe.

Mk = (K with weld toe attachment) / (K without weld toe

attachment) (1)

The general form of the stress intensity factor solution is given

as:

KI = (Y)√(a) (2)

For fatigue assessment the following equation applies:

(Y) = Mfw{ktmMkmMmm+ktbMkbMb[b+(km-1)m]} (3)

It can be seen from these two equations (equations 1 and 2) [1]

above that Mk is applied by multiplying the stress intensity

factors from primary membrane stress and primary bending

stress.

Mk for surface cracks at weld toes of the T-butt welds (both 2D

solution and 3D) are presented in BS7910 [1]. These two types

of pre-solved solutions provide an easier way to apply the effect

of weld toe on stress intensity factor and eventually on crack

growth. As these solutions are readily available, Mk solutions

from BS7910 is widely used in the offshore industry in

conducting ECA or FFS assessment. The commercial software

CRACKWISE [3] also adopts this method. However, if the weld

toe geometry is very different from the T-butt weld, using the

pre-solved Mk solutions from the T-butt weld might result in in-

Weld Toe

Weld Root


3

accurate results. Therefore, , FEA is recommended to generate

the Mk solution for the exact geometries with the crack because

the numerical methods provide more realistic results consistent

with actual weld geometry. However, the determination of K for

weld toe with cracks in pipes generally requires complicated

modelling and therefore, not many solutions are currently

available.

STRESS PROFILE APPROACH The API579 [2] standard contains an extensive list of K

solutions. One of the K solutions allows the user to input the

arbitrary through-wall stress field, or in another words, the stress

profile through the wall thickness. Since the K solutions already

include the effect of the crack, the stress profile can be obtained

by simpler FEA to account for the effect of weld bead geometry

without the crack. This numerical analysis is simpler compared

to the FEA of the weld toe with a crack.

DEEPWATER RISER FATIGUE CRACK GROWTH APPLICATION In order to study how these two approaches influence the

predicted fatigue crack growth, an analysis is conducted with a

deep water steel catenary riser (SCR) in the Gulf of Mexico

(GOM). The SCR configuration is shown in Figure 2. One of

identified fatigue critical locations,,the touch down point (TDP),

is selected in the fatigue crack growth analysis.

Figure 2 – SCR Configuration and Fatigue Critical Locations

The SCR pipe data is given in Table 1. The girth weld geometry

is summarized in Table 2.

Table 1 SCR Pipe Data

Parameter Value

Material Steel

Steel Grade X65

Nominal OD (mm) 273

Nominal Wall Thickness (mm) 33

Yield Strength (MPa) 448

Ultimate Tensile Strength (MPa) 530

Elastic Modulus (MPa) 207,000

Poisson’s Ratio 0.3

Table 2 Girth Weld Geometry

Weld Toe Geometry Value

Attachment Length 6mm

Weld Bead Height 3mm

Global analysis in time domain is conducted for the SCR using

FLEXCOM [8]. The effective tension and bending time

histories are extracted from the simulation for the TDP. Stress

time histories are obtained and Rainflow counted to calculate

the stress histograms. Two sources are included in the fatigue

histograms: one from wave induced motion (stress histograms

are shown in Figure 3), and another one from vortex induced

motions (VIM) (stress histograms are shown in Figure 4). Wave

induced fatigue for the riser includes the first and second order

motion from the vessel, which transfers to the riser motion and

the wave loading directly on the riser. Vortex induced motions

of a vessel are the result of the exciting forces by vortex

shedding on the vessel, causing response near the resonant

period of the vessel. When the vortex induced motion of the

vessel is transferred to the SCR, this results in fatigue damage to

the riser.

Figure 3 – Wave Fatigue Stress Histograms at TDP


4

Figure 4 – VIM Fatigue Stress Histograms at TDP

Crack growth analysis is conducted with an internal surface flaw

in the circumferential orientation with initial flaw size of 4mm

length (2c) by 2mm depth (a). A simplified illustration of the

flaw is shown in Figure 5.

Figure 5 – Circumferential Flaw in the Internal Surface of the

Pipe

The crack growth curve (Paris curve) used is shown as

“Modified in Air Curve with Knock Down Factor of 10” in

Figure 6. This curve is modified from the BS7910 in air curve

(R>=0.5) by removing the threshold and adding a knock down

factor (KDF) of 10 in the higher K region. The BS7910 in air

curve (R>=0.5) is shown in Figure 6 for comparison purpose.

This is an estimated Paris curve to consider that the internal

surface of the riser might be exposed to sour service. Both the

BS7910 in air curve (mean + 2 standard deviation with R

ratio>=0.5) and the modified in air curve with a knock down

factor of 10 are given in Table 3. R ratio, or stress ratio is

defined under fatigue loading conditions as the minimum

applied stress divided by the maximum applied stress for any

given loading cycle. The crack growth threshold value K0 is

defined in Table 4 for both the BS7910 in air curve and the

modified in air curve. For K less than K0, da/dN is assumed

to be zero.

Figure 6 – Paris Curves: Modified BS7910 In Air Curve with Knock Down Factor (KDF) of 10 and BS7910 In Air Curve

Table 3

Paris Curves: Modified BS7910 In Air Curve with Knock Down Factor (KDF) of 10 and BS7910 In Air Curve

R

Ratio

Stage A Stage B Transition

point ΔK

(N/mm3/2

) A m A m

Modified In Air Curve with Knock Down Factor of 10

≥ 0.5 2.10 x

10-17

5.10 1.29 x

10-11

2.88 404.9

BS7910 Steels In Air Curve [1]

≥ 0.5 2.10 x

10-17

5.10 1.29 x

10-12

2.88 144

Table 4

Fatigue Crack Growth Threshold Stress Intensity Factor Range Values

Material Environment ΔK0, N/mm3/2

SCR Steel X65 Potential Sour

Service 0

Steels, including

austenitic Air up to 100°C 63 [1]


5

Crack growth analysis is conducted with BS7910 K solutions

(internal surface flaw in cylinder) with and without Mk effect

applied using an in house software 2HFLAW [4]. 2HFLAW is a

computer program that performs unstable fracture and crack

growth analysis. 2HFlaw includes both BS7910 and API579 K

and reference stress (R) solutions, and has been validated

against Crackwise [3] and SignalFFS [5]. Crackwise adopts

BS7910 K and reference stress (R) solutions, while SignalFFS

mainly applies API579 K and R solutions. The 3D Mk solution

in BS7910 [1] developed from the T-butt weld is adopted in the

analysis. Crack growth analysis is also conducted with API579

K solutions (internal surface flaw in cylinder) with and without

stress profile applied to the stress histogram input. The stress

profile is obtained from the numerical model of the girth weld.

A 2-D finite element model of the internal girth weld bead

geometry (weld toe length: 6mm and weld toe height: 3mm) is

constructed in ANSYS v. 15.0 [6] to solve for the stress profile

across the pipe wall thickness when subjected to an end cap load

of 10,000psi with the other end of the pipe as fixed support. The

finite element model along with a zoom in of the weld bead

region is shown in Figure 7. Solid element type PLANE 182 is

used in the model. Weld mismatch is not included in this model.

Figure 7 – Finite Element Model of Weld Bead Geometry

Figure 8 – Finite Element Model with Mesh

For the majority part of the model (other than the weld bead and

vicinity), mesh size of 2mm is used. Mesh sensitivity is

conducted for the weld bead and vicinity. The finite element

model with the mesh details is shown in Figure 8. The

normalized stress profile as a function of distance from the weld

toe obtained using the three mesh sizes is shown in Figure 9.

The solution at the finest mesh size of 0.05mm is considered to

be essentially converged given the singular behavior of the

stress field at the weld toe. The stress profile using this mesh

size is used in accordance with the guidance of API-RP-579 [2]

to conduct fatigue crack growth analysis.

Figure 9 – Normalized Stress Profile

The crack depth (a) growth curves calculated with BS7910 K

solutions with and without Mk effect applied are shown in

Figure 10 together with the crack depth growth curves

calculated with API579 K solutions with and without weld toe

stress field profile. Similarly, the crack length (2a) growth

curves are shown in Figure 11. The difference between the crack

growth curves of BS7910 with Mk = 1 and API579 without

stress profile mainly comes from the different K solutions from

BS7910 and API579, which is very small in this case. The

difference between the crack growth curves of BS7910 with Mk

= 1 and BS7910 with 3D Mk shows the effect of Mk, and it can

be seen that the necessity to include the stress concentration

resulting from the discontinuity of the weld toe geometry.

However the Mk factor solutions adopted here are from weld toe

magnification factors calculated for T-butt weld [7]. Using the

Mk solutions from T-butt weld is shown to be overly

conservative for the girth weld. It is thus recommended to

calculate the Mk factors with geometry similar to that expected

in a girthweld rather than the more extreme T-butt weld

geometry. It is noted that generating and solving the 3D crack

models required do present difficulties

Using the stress profile approach with API579 K solutions, the

crack growth curve predicts slower crack growth compared to

the Mk solutions derived from T-butt weld [7]. This is

considered as a more practical and accurate approach in

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0 5 10 15 20 25 30 35

No

rmalized

Str

ess

Distance from Weld Toe (mm)

STRESS PROFILE BENEATH WELD TOEMesh Convergence Check

Pipe Wall Thickness = 33.1mm, 3mm by 6mm Weld Bead

Mesh size - 0.2mm Mesh size - 0.1mm Mesh size - 0.05 mm


6

comparison to using Mk solutions from BS7910.

Figure 10 – Crack Depth Growth Curves

Figure 11 – Crack Length Growth Curves

ECA during design phase of offshore riser is usually conducted

to define acceptable initial flaw dimensions and develop

corresponding non-destructive inspection criteria. Additional

ECA input parameters are listed in Table 5. Allowable initial

flaw dimensions calculated with BS7910 K and reference stress

(R) solutions with and without Mk effect applied are shown

together with those calculated with API579 K and R solutions

with and without weld toe stress field profile in

Figure 12. The difference between the allowable initial flaw

dimensions of BS7910 with Mk = 1 and API579 without stress

profile mainly comes from the different K and reference stress

(R) solutions from BS7910 and API579. The difference between

the allowable initial flaw dimensions of BS7910 with Mk = 1

and BS7910 with 3D Mk shows the effect of Mk, and it can be

seen that ignoring the effect of the stress concentration from the

weld toe could result in under-conservative allowable initial

curves, which leads to under-conservative non-destructive

inspection criteria. Some weld joints, which should have been

rejected, could accidentally be accepted because of ignoring the

weld toe geometry stress concentration.

Using the stress profile approach with API579 K and reference

stress solutions, the predicted allowable initial flaw sizes are

larger than those with BS7910 K and reference stress solutions

with Mk solutions derived from T-butt weld [7]. This approach

releases the over conservatism in applying T-butt weld stress

concentration for girth weld.

Table 5

ECA Input Parameters

Parameter Value

Residual Stress 100% Yield

Extreme Loading Occur at End of Life 228 MPa

CTOD 0.38 mm

Design Life 20 Years

Factor of Safety 5

Figure 12 – Allowable Initial Flaw Sizes

DISCUSSION

Sample calculations are conducted with both crack growth

analysis in the operation phase and ECA in the design phase for

a typical deep water SCR. The Mk combining with BS7910

approach is easier in application if using existing solved

solutions; however it requires more intensive FEA work to

derive the accurate Mk solution with the corresponding weld

geometry including the crack and pipe data. The stress profile

combining with API 579 weight function approach is a more

practical and accurate approach in comparison to using Mk

solutions from BS7910, but requires the user to generate a stress

profile appropriate to the expected girth weld geometry.


7

REFERENCES

1. BS7910: “Guide to Methods for Assessing the Acceptability

of Flaws in Metallic Structures”, Third Edition, December

2013.

2. API 579 -1 /ASME FFS-1 2007 Fitness-For-Service.

3. TWI, CRACKWIDE 5, 2014.

4. 2H Offshore Inc., “2HFLAW – Unstable Fracture and Crack

Growth Analysis – User Manual”, 9406-MAN-1041,

Revision 1.2.0.1, July 2013.

5. Quest Integrity Group, SignalFFS.

6. ANSYS Inc. – “ANSYS Workbench”, Version 15.0.

7. Bowness D. and Lee M.M.K, “Prediction of Weld Toe

Magnification Factors for Semi-Elliptical Cracks in T-Butt

Joints”, International Journal of Fatigue 22, 2000.

8. MCS, FLEXCOM 8.0.


proceedings of the asme 2015 pressure vessels & piping ... · proceedings of the asme 2015...

Documents