Professional Documents
Culture Documents
Abstract
Leak before break is a fail–safe design concept for application in pressure vessels and piping of power and process plants.
A quantitative maximum allowable flaw size is required to establish to set acceptance/rejection limit to predict whether
the specific cracked pipe will leak or break. A new modification and its boundary based on Modified Two Parameter
Fracture Criterion is capable of separating the leak and break cases distinctly in order to predict the behavior of cracked
cylinders, pipelines and pressure vessels in advance for taking necessary precautions by the plant operator and also very
much handy for the designers. For the given operating pressure under the observed crack dimensions, whether the crack
will leak or break can be assessed from the boundary generated for the material concerned using Modified Two
Parameter fracture assessment procedure.
Keywords
Leak before break, modified two-parameter fracture criterion, cylindrical vessels
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
2 Proc IMechE Part L: J Materials: Design and Applications 0(0)
developed procedure is capable of separating the leak fracture toughness and hence it is a conservative cri-
and break cases distinctly in order to predict the terion, and it is only suitable under some conditions of
behavior of cracked cylinders in advance for taking loads and materials.
necessary precautions by the plant operator. Toughry8 has presented in the final report of The
Nondestructive Testing Information Analysis Center
(NTIAC) a sound technical basis for developing quan-
Literature survey titative maximum allowable flaw sizes and for setting
Zhou et al.5 highlighted that the concept of LBB was acceptance/rejection limit for the cylinder at the time of
first put forward by Irwin. Irwin and Hood developed retesting. Using traditional methods of re-testing, the
an analogous method, but both of these were very cylinders were rejected due to leaking, bursting, exces-
simple methods and gave conservative results. Irwin sive volumetric expansion or excessively large surface
used Linear Elastic Fracture Mechanics (LEFM) to flaws detected by visual examination.
show that the crack-driving force would be greater in Sharples et al.9 developed a structural integrity dia-
the radial direction than in the axial direction as long gram, a plot of crack depth against crack length that
as the axial crack length was less than twice the cylin- can be used to investigate a wide range of safety argu-
der thickness.6 Figure 1 shows the conceptual view of ments for flawed pressure vessels including LBB. It
a cylindrical pressure vessel with axial surface crack. enables clear margins to be shown for defects which
Accordingly, LBB was postulated to occur if the might exist in the vessels and indicates crack sizes and
length of the flaw was less than twice the thickness of loadings where the LBB case is valid. The use of this
the cylinder. diagram requires a model of crack shape development
as a crack grows through the wall of the vessel up to
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi the stage at which the deepest part of the crack breaks
KIC 5ys t þ rp ð1Þ through the wall.
Yun-Jae10 has proposed an approach for non-linear
2
a
where rp ¼ 2 2 is Irwin’s plasticity correction with fracture mechanics analyses (elastic-plastic and creep
ys
half crack length c ¼ t and ¼ ys . fracture mechanics) based on reference stress of cir-
Rana7 mentioned that the condition which satisfied cumferential and axial through-wall cracked pipes
LBB was that the surface crack whose length was under various loading conditions for LBB analysis.
equal to four times the thickness of cylinder must be Kim et al.11 presented a method to estimate elastic
stable under the highest circular stress, because this plastic J-integral for circumferential through-wall
kind of crack was most typical. Its criterion of stabil- cracked pipes for the LBB analysis of pressurized
ity is that the stress intensity factor is less than the piping. However, the aforesaid approaches involve
fracture toughness. The paper analyses LBB and plas- complicated computations for J-integral and require
tic fracture mode performance. However, their predic- stress–strain data of material concerned.
tion is stated to be 4–25% higher than the Both K-based and CTOD-based assessment routes
experimentally determined value for LBB perform- of BS7910 produced conservative predictions. API
ance. Moreover, analysis is based on plane strain 579 is mainly targeted at the assessment of ageing
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
Kannan et al. 3
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
4 Proc IMechE Part L: J Materials: Design and Applications 0(0)
fracture strength and corresponding stress intensity Here, a is crack depth, M and are correction
factor of two fractured specimens of relevant config- factors based on thickness t, crack length 2c and
uration, where fracture strength is nominal net section crack depth a. Detailed equations can be referred
stress corresponding to the failure load of specimen from the original literature. Using equations (2) and
containing crack and corresponding stress intensity (6), one can set up the following fracture strength
factor at failure is computed using well-established equation to determine the nominal failure stress (f )
solutions. for a specified crack size of a flawed cylinder under
The additional parameter p in equation (2) can be internal pressure.
obtained from17–19 p
f
" ð1 m Þ
u
1 1 1
p ¼ 1
ln 1 pffiffiffi ð1 þ Þ u ð aÞ1=2 M f
ln 2 ð1 þ Þ ð1 m Þ 2 2 þ mþ 1¼0 ð8Þ
# KF u
1 pffiffiffi
þ 21 m ð4Þ Using the Newton–Raphson iterative method, the
non-linear equation (8) is solved for f . Failure pres-
sure of cylindrical vessel having an axial surface or
where through-thickness crack is obtained from equation (7).
4
¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi : ð5Þ Results and discussion
3 þ 9 8m
Rana and Selines23 conducted tests on cylinders fab-
The limiting values of KF and m and detailed pro- ricated with the material 4134 steel (see Table 1) and
cedure of evaluation can be found in Christopher 3AA 4130X (see Table 2). The cylinders were rated for
et al.17 The solution for stress intensity factor Kmax 31 MPa service pressure. Mean diameter and wall
in equations (2) and (3) for cylindrical vessels with thickness were 229 mm and 7.6 mm, respectively.
an axial surface crack as obtained from the finite Sharp, semi-elliptical flaws having 0.076 mm
element solutions by Newman22 is tip radiuses were introduced using the EDM
(electro-discharge-machining) process. The flaws
Kmax ¼ f ðaÞ1=2 M= ð6Þ were machined on the outside cylinder surface, and
the flaw’s major axis was parallel to the longitudinal
where axis of the cylinder. The flaw length varied from
25.4 mm to 76.2 mm, and the flaw depth to wall thick-
pbf Ri ness ratio (a/t) varied from 0.6 to 0.9. The cylinders
f ¼ ð7Þ
t were pressurized monotonically with water or nitro-
gen gas to failure. Using the procedure outlined in
Table 1. Test results23 of high-strength flawed cylinders manufactured from 4134 steel (sys ¼ 1097.1 MPa, sult ¼ 1179.9 MPa).
Failure mode
Test No. Do (mm) Thickness (mm) a (mm) 2c (mm) Pbf (MPa) Test23 Present analysis
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
Kannan et al. 5
Table 2. Test results23 of DOT 3AA 4130X flawed cylinders (sys ¼ 646.53 MPa; sult ¼ 783.15 MPa).
Failure mode
Test No. Do (mm) Thickness (mm) a (mm) 2c (mm) Pbf (MPa) Test23 Present analysis
Figure 2. Break–leak boundary for the test data23 of high-strength flawed cylinders manufactured from 4134 steel.
Section 3, the fracture parameters for the data in separates leak and break cases almost distinctly.
Table 1 were found as KF ¼ 405 MPaˇm, Figures 2 and 3 show the leak–break boundary gen-
m ¼ 0.7193 and; p ¼ 26.0635. The fracture parameters erated and the test results.23 It can also be seen in
for the data in Table 2 were found as Tables 1 and 2 that 100% of analyzed data comply
KF ¼ 246 MPaˇm, m ¼ 0.705 and p ¼ 25.15. Tables with the test results.
1 and 2 show the test results23 of high-strength Tables 3 and 4 show test results16 of Materials B
flawed cylinders manufactured from 4134 steel. It and C, respectively, referred as material having the
can be noted that few cylinders leaked where as strength of 724–860 MPa and 930–1100 MPa. Few
others ruptured. On observations, the leak cases are cylinders were tested with EDM flaws with semi-
mostly with a/t > 0.8. Hence, an attempt was made to elliptical shape and the remaining were tested with
plot both leak and break cases in a plane f vs 2C a. flaws made by CVN cutter. Flaws were made at exter-
Using the fracture parameters (KF, m and p) found nal surface in the longitudinal direction. The cylinders
as earlier and solving equation (8) for at ¼ 0:8, a failure were tested monotonically with water at room tem-
boundary has been developed. This boundary perature until LBB or fracture occurred. Fracture
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
6 Proc IMechE Part L: J Materials: Design and Applications 0(0)
Figure 3. Break–leak boundary for the test data23 of high-strength flawed cylinders manufactured from DOT 3AA 4130X steel.
Failure mode
Do (mm) Thickness (mm) a (mm) 2C (mm) Pbf (MPa) sys (MPa) sult (MPa) Test13 Present analysis
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
Kannan et al. 7
Table 3. Continued
Failure mode
Do (mm) Thickness (mm) a (mm) 2C (mm) Pbf (MPa) sys (MPa) sult (MPa) Test13 Present analysis
parameters evaluated using fractured data as reheated to obtain yield strengths in the neighborhood
KF ¼ 289 MPaˇ m, m ¼ 0.8792, and p ¼ 50.1227 for of 1102.4 MPa to conduct test. Since it was not pos-
material B and KF ¼ 283 MPaˇ m; m ¼ 0.5486; sible to produce high strength cylinders with the
p ¼ 18.8425 for material C. Figures 4 and 5 show required low toughness range of 414–483 MPaˇm
the break–leak boundary and the test data.16 It can at 1102.4 MPa yield strength level, cylinders having
also be seen in Tables 3 and 4 that 85% of analyzed identical physical size and tensile properties of high
data comply with the test results.16 However, scatter strength cylinder were used for testing purpose.
in test results can be dealt only with statistical Sharp, semi-elliptical shaped, EDM flaws of varying
analysis. lengths of 25–76 mm and varying depths (a/t of
Rana7 conducted pressure test on DOT 3AA 0.5–0.9) were machined on the outside surface of the
(Manufactured from modified AISI 4130 steel) high reheat treated test cylinders. Four flawed cylinders
pressure gas cylinders with nominal wall thickness of were subjected to fatigue loading by pressurization
7.6 mm and a diameter of 229 mm. The cylinders were with water to failure. These cylinders were subjected
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
8 Proc IMechE Part L: J Materials: Design and Applications 0(0)
Failure mode
Present
Do (mm) Thickness (mm) a (mm) 2c (mm) Pbf (MPa) sys (MPa) sult (MPa) Test13 analysis
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
Kannan et al. 9
Table 4. Continued
Failure mode
Present
Do (mm) Thickness (mm) a (mm) 2c (mm) Pbf (MPa) sys (MPa) sult (MPa) Test13 analysis
Figure 4. Break–leak boundary for the test data16 of DOT 3AA 4130X – Material B flawed cylinders.
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
10 Proc IMechE Part L: J Materials: Design and Applications 0(0)
Table 5. Test results7 of DOT3AA (manufactured from modified AISI 4130 steel) material (sys 1104 MPa and sult ¼ 1207.5 MPa).
Failure mode
Test No. Do (mm) Thickness (mm) a (mm) 2c (mm) Pbf (MPa) Test7 Present analysis
Figure 5. Break–leak boundary for the test data16 of DOT 3AA 4130X – Material C flawed cylinders.
to stresses that were lower by approximately 6% com- able to reach the outer surface. Modified Two
pared to the stress level of 537.4 MPa of a high Parameter Fracture Criterion that has been
strength cylinder. In addition, six flawed cylinders validated successfully for fracture assessment is now
were monotonically pressurized with water to failure. validated here for making a boundary between leak
The failure mode is defined as either LBB or fracture. and break.
All ruptured cylinders exhibited flat fracture with
fairly large size shear lips, indicating plane stress
(elastic-plastic) fracture mode. Table 5 shows these
Conclusions
test results.7 Fracture parameters evaluated using The leak before break (LBB) concept could be applied
the fractured cases are KF ¼ 210 MPaˇ m, for the protective design against pipe break. A simple
m ¼ 0.8911 and p ¼ 54.74. Figure 6 shows the break– and reliable modified two parameter fracture criterion
leak boundary and the test data.7 It can also be seen in (MTPFC) has been found to clearly differentiate leak
Table 5 that 90% of analyzed data comply with the and break cases. This new approach does not require
test results.7 It is observed that mostly the leak cases crack growth analysis. The fracture parameters neces-
are of lower toughness as compared to fracture cases, sary for the fracture strength equation of the criterion
in which crack grow gradually and leak. Leak cases can be evaluated from the test data on simple fracture
demand initiation stresses that permit a little crack specimens. For the given operating pressure under the
growth leading to a through crack. Fractured cases observed crack dimensions, whether the crack will
are of higher toughness and after reaching instability leak or break can be assessed from the boundary gen-
load, the fracture occurs before the crack growth is erated for the material concerned using Modified Two
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
Kannan et al. 11
Figure 6. Break–leak boundary for the test data7 of DOT3AA (manufactured from modified AISI 4130 steel) material.
Parameter fracture assessment procedure. The fact 9. Sharples JK and Clayton AM. A leak-before-break
that whether it would be safe or break/leak will be assessment method for pressure vessels and some cur-
easily evident. rent unresolved issues. Int J Pres Ves Piping 1990; 43:
317–327.
10. Kim Y-J, Huh NS and Kim YJ. Engineering J-estima-
Funding tion methods for LBB analysis of nuclear piping. JSME
This research received no specific grant from any funding Int J Series A 2005; 48: No.1.
agency in the public, commercial, or not-for-profit sectors. 11. Kim Y-J. Non-linear fracture mechanics analysis of
through-wall cracked pipes for leak-before-break ana-
lysis. J Pressure Equipment Syst 2004; 2: 71–78.
References 12. Andrewsd RM, Harrisona JD, Wiesnera CS, et al.
1. Pacholkova S and Taylor H. Theoretical background of Engineering critical analysis to BS 7910-the UK guide
‘‘leak-before-break’’ as a concept in pressure vessels on methods for assessing the acceptability of flaws in
design. Metal 2002; 14–16.5: 1–8. metallic structures. Int J Pressure Vessels Piping 2000;
2. Nam K-W and Abn S-H. Crack opening behavior of 77: 883–893.
penetrated crack under fatigue load. KSME Int J 2002; 13. Structural Integrity Assessment Procedure for
16: 24–31. European Industry – SINTAP, Final Report, Swinden
3. Kawaguchi S, Hagiwara N, Masuda T, et al. Evaluation Technology Centre, Rotherham, UK, November 1999.
of leak-before-break (LBB) behavior for axially notched 14. Zerbst U, Hamann R and Wohlschlegel A.
X65 and X80 line pipes. J Offshore Mech Arct Eng 2004; Application of the European flaw assessment procedure
126: 350–357. SINTAP to pipes. Int J Pressure Vessels Piping 2000;
4. Drubaya B, Mariea S, Chapuliota S, et al. A16: guide for 77: 697–702.
defect assessment at elevated temperature. Int J Pres Ves 15. Darlaston BJL. Some aspects of leak before break; their
Piping 2003; 80: 499–516. quantification and application. Nucl Eng Des 1985; 84:
5. Zhou J-Q and Shen S-M. Development of the leak- 225–232.
before-break assessment method to reduce some short- 16. Rana MD, Smith JH and Tribolet RO. Technical basis
comings. Int J Pres Ves Piping 1996; 69: 75–77. for flawed cylinder test specification to assure adequate
6. Wilkowski G. Leak-before-break: what does it really fracture resistance of ISO high-strength steel cylinder.
mean? J Pressure Vessel Technol 2000; 122: 267–272. J Pressure. Vessel Technol 1997; 119: 475–480.
7. Rana MD. Experimental Verification of Fracture 17. Christopher T, Sankaranarayanasamy K and
Toughness requirement for LBB performance of 150– Nageswara Rao B. Correlating cryogenic fracture
175 Ksi strength level gas cylinder. J Pressure Vessel strength using a modified two-parameter method.
Technol 1987; 109: 435–439. J Engng Fract Mech 2005; 72: 475–490.
8. Toughry M. Development of accept reject criteria for 18. Christopher T, Sankarnarayanasamy K and Nageswara
requalification of high pressure steel and aluminium Rao B. Fracture strength of flawed cylindrical pressure
cylinders. Non-destructive Testing Information vessels under cryogenic temperatures. Cryogenics 2002;
Analysis Center (NTIAC), Austin, Final Report 2002. 42: 661–673.
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016
12 Proc IMechE Part L: J Materials: Design and Applications 0(0)
Appendix
Notation
a crack depth
c half the crack length
Do outer diameter of cylindrical vessel
Downloaded from pil.sagepub.com at Middle East Technical Univ on February 10, 2016