International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 1343

ISSN 2229-5518

Determination of Geometry Factor of Crack in Dented API 5L X65 Pipeline Using Finite Element Method

Arianta, Ricky L.T, Ahmad T., Rildova

Abstract— Oil and gas transmission pipelines during its construction, laying and operation period may undergo local plastic deformation due to mechanical impact, which some of them can be in the form of dent with crack. W ith the presence of crack, the acceptance criterion is no longer used alone where design or operational pressure are compared to SMYS. Therefore additional different acceptance criterion must be used based on fracture mechanics, where the stress intensity factor is compared to fracture toughness of material. Stress intensity factor depends on the geometry of pipe, size of crack, and the magnitude and direction of the loading system acting on the pipeline. The study was conducted based on the experimental test and numerical models using steel plate as one approach to large pipe diameter. The steel plate is made of a piece of material API-5L-X65 gas pipeline. Result of the research was the geometry factor of dented plate with crack. The geometry factor equation of dented plate with crack was then used to modify the geometry factor equation of crack pipe based on standard API RP 579. The value of geometry factor of crack in dented pipe was limited to ratio a/t = 0.5, a/c = 0.4 and 0 < Hr/t ≤ 2.

Index Terms— Pipeline, fracture mechanics, finite element, geometry factor, stress intensity factor, dent, crack, plate

1 INTRODUCTION

—————————— ——————————
Pipeline has a good safety factor and economics to delivery oil or gas from wells to production. However, during its construc- tion, laying and operation, pipelines sometimes subject to leaks or ruptures. Generally, the failure caused by external interference or mechanical damage. The type of mechanical damage which is common in pipeline is dents and cracks. The combination of dent and crack make pipeline is more vulnera- ble to failure. Therefore, it is important to evaluate the integri- ty of pipeline that contain a defect that contain both dent and crack.
Fitness for service (FFS) assessments are quantitative engineer- ing that are performed to evaluate the structural integrity of pipeline that may contain a flaw or damage and to make deci- sion run-repair-replace or down rate. API 579/ASME FFS-1 [4] is a standard method for assessment to evaluate the integrity of pipeline that has a size and a geometry of certain defects. The acceptance criterions of API 579 for dented pipe are i) the maximum allowable the dent depth, ii) Remaining Strength Factor (RSF) and iii) the maximum allowable stress of dented pipe. The recommendation from standard API 579 for dented pipeline case is reflected in the maximum allowable operating pressure (MAOP).
With the presence of crack in dented pipe, the design criterion is no longer used alone where the operational stress compare

————————————————

Arianta is currently pursuing doctoral degree program in civil engineering in Institute Technology of Bandung, Indonesia, PH-08122125915. E-mail: arian- ta.l@gmail.com

Ricky. L. T is professor in ocean engineering in Institute Technology of Ban-

dung, Indonesia,. E-mail: ricky@ocean.itb.ac.id

Ahmad. T is lecturer in ocean engineering in Institute Technology of Ban-

dung, Indonesia,. E-mail: a.taufik@ocean.itb.ac.id

Rildova is lecturer in ocean engineering in Institute Technology of Bandung, Indonesia,. E-mail: a.taufik@ocean.itb.ac.id

to SMYS (Specified Minimum Yield Strength) is set as an ac- ceptance criterion. Therefore additional different acceptance criterion must be used based on fracture mechanics, where the stress intensity factor is compared to fracture toughness of material. Stress intensity factor is the stress field around crack tip that depend on geometry of pipeline, size of crack and di- rection of the loading system acting on the structure. Geome- try factor is a function of crack depth, crack length, thickness of structure and the parametric angle of ellipse crack.
Yong Bai [6] has studied the case to determine the geometry factor of dented pipe with crack. He used theoretical approach from the stress intensity factor of crack plate by Newman and Raju [17], and the uniaxial tensile stress and bending moment of dented pipe by Shannon [6].
Experimental test and numerical analysis were carried out to determine the geometry factor (F) of dented pipe with crack. Experimental tests was validated the numerical model simula- tions. Steel plates was used as an approach to pipe geometry and it is made of a piece of material API-5L-X65 gas pipeline. XFEM (eXtended Finite Element Method) modeling as numer- ical simulation was used to determine the stress intensity fac- tor of dented plate with crack.

2 STRESS INTENSITY FACTOR OF DENTED PIPES WITH

CRACKS

Yong Bai has determined the geometry factor equation of dented pipe with crack [6] theoretically. He is used the stress intensity factor by Newman-Raju [17]; the bending moment and uniaxial tensile stress in a dented pipe by Shannon [6] as an approach. The stress intensity factor (K) for wide plate un- der combined tension and bending is given below

IJSER © 2015 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 1344

ISSN 2229-5518

K = F σ

Q

πa + H

F 6M πa

Q t 2

(1)

tic asymptotic crack-tip functions (Fα (x)). The jump function and the enrichment function at the crack tips are given by
Where factor F, Q and bending correction H are given by Newman and Raju. σ is the nominal stress, t is the thickness of material and a is the crack length.

H (x, y ) = 

− 1

if (x x *).n ≥ 0

otherwise

(7)



θ θ θ

Shannon is given an approach solution for bending moment M
and uniaxial tensile stress σ in a dented pipe.

Fα (x) = 

r sin ,

2

r cos ,

2

θ

r sinθ sin ,

2 

(8)

σ = σ 1 −1.8

H r

(2)


r sinθ cos 

 2 

D

M = 0.85 σ H t Hr

(3)

4 DENT MODEL

Where D is the diameter of pipe, σH is the nominal hoop stress and Hr is the dent depth
Substituting equation 2 and 3 into equation 1,
Experimental tests are carried out to know the material behav- ior API-5L-X65 at the process of the formation of a dent. The experimental tests used plates as an approach to model pipe- line geometry with large diamater. The plate had the follow-
ing dimension length 70 mm, width 40 mm and thickness 4

  

  

K = F



1 − 1.8 H r  + 5.1H H r  σ

πa (4)

mm. Dent forming on steel the plate was made by applying

Q   D

t  

pressure on the indenter until it reached the depth of dent. The geometry of indenter was based on DNV RP F107, it had
Therefore, the geometry factor of dented pipe with crack can be expressed as
conical shape. Figure 1 shows the shape and dimensions of
indenter.

F   H r

H r  

FYongBai =

1 − 1.8

Q


 + 5.1H   

D t

(5)

  

  

3 THE EXTENDED FINITE ELEMENT METHOD

The extended finite element method (XFEM) is an extension of the conventional finite element method based on the concept of partition unit. This method was introduced by Ted Be- lytschko and his colleagues in 1999 [24]. XFEM models a crack as an enriched feature by adding degrees of freedom in ele- ments with special displacement functions. XFEM does not require the mesh to match the geometry of the discontinuities. It can be used to simulate initiation and propagation of a dis- crete crack along an arbitrary, solution-dependent path with- out the requirement of remeshing. XFEM can also be used to perform contour integral evaluation without the need to refine the mesh around the crack tip. The approximation for a dis- placement vector function u with the partition of unity en- richment is [1]
Figure 1. Indenter
The dent process produces both elastic and plastic responses in the material. When the indenter was removed, the elastic component of the deformation was recovered and the dent will move outward, therefore the depth of dent will decrease. This recovery is termed spring back. LVDT (Linear Vertical
Displacement Transducer) tool was installed below the plate

N

u = N

(x)

+ H (x)a

4

+ F

(x)bα

(6)

to measure the change of the dent depth.

i i

i =1

i α i

α =1

The change of the dent depth can be determined by the value of elastic recovery. Elastic recovery (ER) is the percentage
where Ni (x) is the shape functions on node I, ui is the nodal
displacement vector associated with the continuous part of the
finite element solution, ai is the nodal enriched degree of free-
change in the dent depth, during and after (spring back) the process of dent.
dom vector, and the associated discontinuous jump function

α

H ER = 0

H r

× 100%

(9)

(H(x)) across the crack surfaces; bi
is the product of the nod- H 0
al enriched degree of freedom vector, and the associated elas-

IJSER © 2015 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 1345

ISSN 2229-5518


Where H0 is the maximum of dent depth and Hr is the dent depth after spring back. Experiments have been repeated three times to ensure the repeatability of the result. Table 1 shows the result of measurement the dent depth. The deeper of the dent on plate, the value of elastic recovery will decrease. This means the material will become plastic and there was no sig- nificant changes in the dent depth.

Table 1. Result Dent Depth Measurement

Elastic Recovery

19.00

14.00

9.00

Figure 2. Relation Dent Depth

Specimen H0 Hr

4.00

(mm) (mm)

% Average

4 5 6 7 8 9

H0 (mm)

Figure 3. Elastic Recovery
Finite element modeling was conducted to simulate the pro- cess of dent and spring back. An eight-node hexahedral solid element, C3D8 of ABAQUS, was used for modeling dent and spring back on plate specimen. Figure 4 shows the modeling of indenter and plate specimen.

Figure 2 shows a relationship values between maximum dent depth (H0) and dent depth after spring back (Hr). It has a line- ar trend with the line is increase. Elastic recovery has similar trend but the line will decrease towards deepening dent (fig- ure 3).

8

7

6

5

4

3

2

4 5 6 7 8 9

H0 (mm)


Figure 4. Indenter and Plate Models
Combined hardening between kinematic and isotropic hard- ening was used to model the behavior of material API-5L-X65 in dent and spring back modeling. The material properties of API-5L-X65 was referred to Akid & Fatoba [3]. The parameters H0 and Hr as results from the experiment were used to vali- date the numerical modeling. The value of young’s modulus should be changed accordance to the dent depth, therefore the parameter H0 and Hr of model will be close to H0 and Hr from experiment. With trial and error to determine the young’s modulus, figure 5 shows the young’s modulus degra- dation accordance to the dent depth.

IJSER © 2015 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 1346

ISSN 2229-5518


K = σ F a , a , φ  ⋅ πa

(10)

100000

80000

 


I t c Q

60000

40000

20000

0

y = -7970,9x + 113216

R² = 0.9712

3 4 5 6 7 8 9

Hr (mm)

Where σ is the applied stress, a is the crack depth, Q is the shape factore for an ellipse and F is the geometry facator as function of crack depth, crack length, plate thickness and the parametric angle of the ellipse. The purpose of benchmark XFEM modeling was to determine the meshing technique in crack area. Partition of mesh was more refined near the crack area. The element size for near crack area is 0.25 mm of hexa- hedron element. Figure 7 shows the mesh configuration near the crack area.

Figure 5. Young’s Modulus Degradation
The process of dent caused young’s modulus degradation. The deeper of dent depth, the value of young’s modulus will de- crease (fig. 5). Empiric equation to predict the young’s modu- lus for API-5L-X65 material as presented in figure 5, linear equation as follow

E ' = −7970,9( H r ) + 113216

Where: E’ is the young’s modulus degradation
(7)

5 XFEM MODEL

Experimental test in the form of dent resulted the initial crack on the side of the arch dents (fig. 6). Strain gage was attached on crack tip zone. It was used to record the strain on crack tip from propagate until fracture. Tensile tests have been con- ducted three times to ensure repeatability of the results. Figure
7 shows the average of strain on crack tip from three times test. The maximum strain on crack tip and far field stress on plate were used to validate the finite element modeling of ten- sile test.

Figure 6. Strain on Crack Tip
To evaluate XFEM result in Abaqus, a stress intensity factor comparison was made against benchmark case. The bench- mark case was a wide range of semi-elliptical surface cracks in finite-thickness plates [18]. The stress intensity factor is de- fined as
Figure 7. Mesh refinement control
Equation 10 was used to determine the geometry factor from XFEM results. Table 2 shows the comparison of geometry fac- tor results from XFEM model with the reference solution ob- tained by Newman and Raju in terms of the % relative error.
Table 2. Comparison of Geometry Factor

IJSER © 2015 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 1347


ISSN 2229-5518

1.500

1.000

0.500

0.000

0 20 40 60 80 100

Crack Angle ()

Figure 10. Angle of Crack Path

Newman-Raju Model

Figure 8. Comparison of Geometry Factor
The highest % relative error was 8.296% and still below 10%. Therefore, the comparisons shows good agremeent with the results published by Newman and Raju, for case a/c = 0,4 and a/t =0,5. The mesh configuration with element size 0.25 mm will be applied to the model dented plate with crack. Figure 9 shows the mesh configuration for XFEM model of dented plate with crack. Deformation of dent shape from spring back analysis is used for XFEM model. On XFEM model, the resid- ual stress from spring back analysis is not applied to model. Therefore, the stress intensity factor from the XFEM modeling is due to external load only. The dimensions of crack are length (2c) 10 mm and the depth of crack (a) 2 mm, with semi- elliptical shape.

Figure 9. Mesh Configuration on Dented Plate with Crack
Model
The result of XFEM modeling was the stress intensity factor (KI ) along the cracked path. Stress intensity factor equation given by Newman and Raju is used to determine the geometry factor of dented plate with crack. The value of geometry factor dented plate with crack is function of angle of the crack path. Figure 10 shows the angle of crack path.
Figure 11 shows the result of geometry factor along crack path as function of parametric angle (φ) and the dent depth. The maximum geometry factor occurs at φ = 00 and the minimum geometry factor occurs at φ = 900. The deeper of the dent depth, geometry factor become smaller.

Figure 11. Geometry Factor along Crack Path
Geometry factors for dented plate with semi-elliptical surface crack (a/c = 0.4, a/t = 0.5) as a function of the parametric an- gle (φ) and the dent depth to plate thickness ratio (0 < Hr/t ≤
2), are given in table 3.
Table 3. Geometry Factor of Dented Plate with Crack

Geometry factor for dented plate could be obtained by equa- tion 11

FDentedPlate =

FTotal

FNewmanRaju

(11)

Where FTotal is geometry factor for dented plate with cracks (table 3) and FNewmanRaju is geometry factor for finite-thickness plate that obtained by Newman and Raju [18]. Table 4 shows the geometry factor for dented plate (a/c = 0.4, a/t = 0.5) as a

IJSER © 2015 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 1348

ISSN 2229-5518

function of the parametric angle (φ) and the dent depth to plate thickness ratio (0 < Hr /t ≤ 2).
Table 4. Geometry Factor of Dented Plate

6 GEOMETRY FACTOR OF DENTED PIPE WITH CRACK

Stress intensity factor equation of surface crack pipe was used to determined geometry factor of dented pipe with crack. The
tio (0 < Hr /t ≤ 2), are given in table 5
Table 5. Geometry Factor of Dented Pipe with Crack

Pipeline is a pressure vessel that has thin wall thickness. Equa- tion 12 is for thick wall pressure vessel, therefore the stress intensity equation based on API RP 579 need to be modified. Modified equation 12 for thin wall pressure vessel case as fol- low.
   
   
equation is based on API RP 579 [4]. Figure 12 shows crack
  2
+ 2  a  + 3
a   
location on pipe.

pD

G0

G1 R


G2 R    πa

(14)

K I =


FDentedPlate
o
o   
2t
  a

4

a
  Q
  + 4G3  
+ 5G4    

   

o


   

o

   
   
Comparison of the stress intensity factor of dented pipe with crack between Yong Bai and Arianta is presented below. The bending moment from Yong Bai equation is neglected, com- parison only due to hoop stress that acting on pipeline.

70


Figure 12. Outside Crack in Pipe

60

50

 2G

2

+ 2G1

a   a   40



 + 3G2   + 30

Yong Bai

K I =

pR

2 2

Ro

3

Ro

4

πa

(12)

20

Arianta

Ro Ri

a

a   Q

 4G3  

+ 5G4    10


     

o o

     

0

0 10 20 30

Where p is the pressure, Ri is the inside radius, Ro is the out- side radius and G0 , G1 , G2 , G3 , G4 are the geometry coeffi- cient. Geometry factor for dented pipe with crack is obtained by multiplied the geometry factor of dented plate with geome- try factor of crack pipe. The equation of geometry factor for dented pipe with crack as follow.

Dent Depth (mm)

Figure 13. Comparison of Stress Intensity Factor
The stress intensity factors of dented pipe with crack found by
Yong Bai and Arianta have a similar tendency to decrease
with increasing the depth of dent. Different approach in de-

  a

a  

veloping the KI equation of dented pipe with crack, will gen-

 2G + 2G


 + 3G2   + 

erate different value between Yong Bai and Arianta. Differrent

FArianta = FDentedPlate

3

a

Ro

Ro  

a  

(13)

approach used was; i) the location of the hoop stress and, ii)
determination of the geometry factor of dented pipe with

 4G3  

+ 5G4   

crack.

  Ro

Ro  

Where FArianta is the geometry factor of dented pipe with crack as result of this study. The geometry factor for dented pipe with crack plate (a/c = 0.4, a/t = 0.5) as a function of the parametric angle (φ) and the dent depth to plate thickness ra-
Different approach used to develop the stress intensity factor of dented pipe with crack will produce different result. Differ- ent parameters that used to develop the stress intensity factor equation of dented pipe with crack determines the location of

IJSER © 2015 http://www.ijser.org

International Journal of Scientific & Engineering Research, Volume 6, Issue 1, January-2015 1349

ISSN 2229-5518

hoop stress of dented pipe and the geometry factor of dented pipe with crack.

7 CONCLUSION

A three dimensional model with XFEM model was used to determine the stress intensity factor of dented plate with crack. The material behavior in model based on experimental test. Dent depth, strain at crack tip and far field stress as result of experimental tests, were used for validation to numerical modeling of dented plate with crack.
Based on XFEM model result due to tensile stress, the presence of a dent on plate the geometry factor become smaller at the deepest point of crack. The geometry factors of dented plate with crack have tendency to be smaller when the dent depth become deeper.
Stress intensity factor equation of outside crack pipe based on API RP 579 was used to determine the geometry factor equa- tion of dented pipe with crack. Comparison of the stress inten- sity factors of dented plate with crack with Yong Bai have sim- ilar trend. It is become smaller when the dent depth deeper. Different approach resulted different slope of the stress inten- sity factor graph when it was plotted as a function of the dent depth. The geometry factors of dented pipe with crack was limited to a/c = 0.4, a/t = 0.5 and the dent depth to plate thickness ratio (0 < Hr /t ≤ 2).

XFEM for Integrity Assessment of Subsea Equipment. Master Thesis, Univer- sitas Chalmers, 14-16.

[16] Marko Vrh, Miroslav Halilovic, Boris Stok (2008). Impact of Young’s Modulus Degradation on Springback Calculation in Steel Sheet Drawing. Jurnal Me- chanical Engineering vol 54, 288-296, Strojniski vestnik.

[17] Newman, C.J., Raju, S.I. (1981), An Empirical Stress-Intensity Factor Equation

For The Surface Crack, Jurnal Engineering Fracture Mechanics, Pergamon, 15,

,1-2. 185-192.

[18] Newman, C.J., Raju, S.I. (1979), Stress Intensity Factor for A Wide Range of

Semi Elliptical Surface Crack, Jurnal Engineering Fracture Mechanics, Per- gamon, 11, 817-829.

[19] Oliveira, Z.F. (2013), Crack Modeling with Extended Finite Element Method, Master Thesis, Instituto Superior Tecnico, 35-38.

[20] Oh, K.C., Kim, Y.J., Baek, J.H., Kim, Y.P., Kim, W. (2007), A phenomenological

model of ductile fracture for API X65 steel, International Journal of Mechanics

Sciences, Elsevier, 49, 1399-1412.

[21] Paik, J.K., Lee, M.J., Lee, D.H. (2003), Ultimate strength of dented steel plates under axial compressive loads, International Journal of Mechanical Science, Pergamon, 45, 433-448.

[22] Sarangi, H., Murthy, K.S.R.K., Chakraborty, D. (2013), Experimental Verifica-

tion of Optimal Strain Gage Locations for the Accurate Determination of

Stress Intensity Factor, 13th International Conference on Fracture, Beijing, China.

[23] Shivakumar, K.N., Newman, C.J. (1991), Stress intensity factors for large as- pect ratio surface and corner cracks at a semi-circular notch in a tension spec- imen, Journal Engineering Fracture Mechanics, Pergamon, 38, 6, 467-473.

[24] Sukumar, N., Moes, N., Moran, B., Belytschko. (2000), Extended Finite Ele-

ment Method for Three-Dimensional Crack Modeling, International Journal

for Numerical Method in Engineering, 48, 1549-1570.

[25] Seifi, R. (2011), Effect of Residual Stress on Fracture Parameters of Through

Cracks in Welded Plates, Journal Procedia Engineering, 10, 1895-1900.

[26] Yu, H.Y. (2009), Variation of elastic modulus during plastic deformation and

its influence on springback, Journal Materials and Design, 30, 846-850.

8 REFERENCES

[1] Abaqus 6.12 (2012), Online Documentation, Dassault Systemes, Rhode Island. [2] Abu Naim Md Rafi. (2011), Structural Behaviour of Dented Pipelines, Master

Thesis, Universitas Windsor, Canada, 32.

[3] Akid, R dan Fatoba, O. (2014), Low Cycle Fatigue Behaviour of API 5L X65

Pipeline Steel at Room Temperature, Jurnal Procedia Engineering, 74, 279-286. [4] API Recommended Practice 579. (2000), Fitness For Service, American Petro-

leum Institute, C-29.

[5] Arun Shukla. (2005), Practical Fracture Mechanics in Design, Second Edition,

Marcel Dekker, Kingston, Rhode Island, U.S.A.

[6] Bai. Y dan Song. R. (1997), Fracture Assessment of Dented Pipes with Cracks and Reliability-Based Calibration of Safety Factor, Int. Jour. Pressure Vessel and Piping, 74, 221-229.

[7] Blacut, J., Iflefel, B.I. (2007), Collapse of pipes with plain or gouged dents by

bending moment, Int. Jour. Pressure Vessel and Piping, Elsevier,84, 560-571.

[8] Christopher R. A. dan John F. Kiefner. (1997), Effect of Smooth and Rock

Dents on Liquid Petroleum Pipelines. American Petroleum Instiute.

[9] Cleveland, R.M., Ghosh, A.K. (2002), Inelastic effects on springback in metals, Journal of Plasticity, Pergamon, 18, 769-785.

[10] Cosham, A., Kirkwood, M. (2000), Best Practice in Pipeline Defect Assess- ment, Proceedings of International Pipeline Conference, Calgary, Alberta, Canada.

[11] DNV RP F107. (2010), Risk Assessment of Pipeline Protection, Det Norske

Veritas.

[12] Hearn, E.J. (2000), Mechanics of Materials 1, Butterworth-Heinman, 215-224. [13] James W. Dally dan Robert J. Stranford. (1987). On Measuring the Instantane-

ous Stress Intensity Factor for Propagating Cracks. Departemen Teknik Me- sin, Universitas Maryland, U.S.A. 3223-3230.

[14] Jovicic, G., Zivkovic, M., Jovicic, N. (2009), Numerical simulation of crack

modeling using extended finite element method, Journal of Mechanical Engi- neering, 55, 1-6.

[15] Leven, M dan Rickert, D. (2012). Stationary 3D Crack Analysis with Abaqus

IJSER © 2015 http://www.ijser.org