International Journal of Pressure Vessels and Piping: Sciencedirect

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

International Journal of Pressure Vessels and Piping 175 (2019) 103924

Contents lists available at ScienceDirect

International Journal of Pressure Vessels and Piping


journal homepage: www.elsevier.com/locate/ijpvp

Relevance of factor of safety based on number of cycles in the prediction of T


fatigue crack initiation as per A16 sigma-d approach
R. Suresh Kumara,∗, B.N. Raob, K. Velusamya
a
Indira Gandhi Center for Atomic Research, Homi Bhabha National Institute, Kalpakkam, Tamil Nadu, 603102, India
b
Indian Institute of Technology Madras, Chennai, India

A R T I C LE I N FO A B S T R A C T

Keywords: The focus of this paper is to bring out the conservatism present in the fast reactor design standard RCC MRx A16
Fatigue crack initiation in predicting fatigue crack initiation for a component that has a crack-like defect. Full-scale pipe bend test results
Sigma-d approach were used for comparing the fatigue life predicted based on the RCC MRx A16 approach. A representative pipe
Stress intensification factor bend of size 570 mm outer diameter and 15 mm thickness has been used for the purpose. The selected pipe bend
RCC MRx A16
is made up of austenitic stainless steel (SS 316 LN) and is provided with a surface notch 22.5 mm long and
Prototype testing
2.1 mm in depth to simulate the potential flaw, which could be present in the piping system. Cyclic testing has
been carried out to determine fatigue life. These tests were conducted in two stages. During the first stage, the
finite-sized notch was sharpened into a crack-like defect allowing it to develop into a natural crack (85,150
cycles). In the second stage, the testing was continued to estimate the number of cycles required to advance the
crack by > 50 μm and thereby compare prediction capability of the sigma-d approach given in the RCC MRx
A16. The number of cycles required for crack incubation and advancement was obtained by periodic mea-
surement of the measured readings obtained from the crack depth gauge. A comparison of the number of cycles
was predicted by the RCC MRx A16 method, and the experimental results show that the A16 prediction was very
conservative. Useful suggestions were proposed in this paper for improving the fatigue crack initiation prediction
capability.

1. Introduction SFR components, which have a crack-like defect, is dealt with using the
sigma-d approach given in the RCC MRx A16 standard [3]. For con-
Piping systems are the vital pressure boundaries in the case of so- firming the completeness of the sigma-d approach, one of the re-
dium-cooled fast reactors (SFRs). Secondary sodium main circuit presentative prototype specimens from the secondary sodium main
(SSMC) is one of the critical coolant carrying piping systems in a SFR. piping circuit of the typical SFR (Fig. 1) is identified, and the potential
The schematic of this piping system for a medium-sized pool type SFR defect is introduced and tested. The tested pipe bend has a size of
of 500 MWe capacity is given in Fig. 1 [1]. They are designed to operate 570 mm outer diameter, 15 mm thickness and bend radius of 820 mm.
at high temperatures (∼550 °C) in the creep regime. During the reactor A surface notch of size 22.5 mm long (2c), 2.1 mm in depth (a) and a
operation, these piping systems are subjected to various thermo-me- finite notch tip radius is introduced very close to the crown location [4].
chanical loadings including shutdown/startup of the reactor, where It is introduced on the external surface of the pipe bend along the
these piping temperatures vary from 200 °C to 525 °C along with the longitudinal direction to simulate the necessary conditions for com-
associated changes in its overall configuration. This, in turn, leads to parison.
cyclic loadings of the piping products, especially the pipe bend. Such SFR piping systems are generally designed as a thin-walled system
loads lead to low cycle fatigue, which can eventually lead to sodium due to low internal pressure [5]. The variation in pressure load is
leakage and the associated sodium fire. negligible during the reactor operating condition. Hence the pressure
Traditionally, RCC MRx RB [2] design code rules are used for the load is kept constant in this work. The pipe bend is subjected to cyclic
design of SFR class-1 systems. Large-sized, slender products of SFR load due to various plant operating conditions. This induces cyclic
piping system is made by welding the rolled plates. The probability of bending moment to the pipe elbow. During the in-plane bending, the
finding a crack-like defect is less in a class-1 system. Fatigue design of thin-walled 90° pipe bend is subjected to ovalisation of the circular


Corresponding author. Reactor Design Group, Indira Gandhi Center for Atomic Research, Department of Atomic Energy, Kalpakkam, 603102, Tamilnadu, India.
E-mail addresses: ersureshkr@gmail.com, suresh@igcar.gov.in (R. Suresh Kumar), bnrao@iitm.ac.in (B.N. Rao), kvelu@igcar.gov.in (K. Velusamy).

https://doi.org/10.1016/j.ijpvp.2019.103924
Received 11 April 2018; Received in revised form 24 May 2019; Accepted 25 June 2019
Available online 29 June 2019
0308-0161/ © 2019 Elsevier Ltd. All rights reserved.
R. Suresh Kumar, et al. International Journal of Pressure Vessels and Piping 175 (2019) 103924

Fig. 1. Schematic of Secondary Sodium Main Circuit (SSMC) piping system in a medium size pool type SFR.

cross-section. This ovalisation introduces high bending stress close to away from the notch tip is obtained using the Eq. (1). The value of Δσde
the crown with the associated risk of crack initiation/propagation [6,7]. will be lower than that predicted by Eq. (1) if the crack notch tip radius
The above discussed full-scale test results have been used to assess is taken into consideration. Due to the uncertainty in the measurement
the fatigue life predicted as per RCC MRx A16 [3] sigma-d approach. of crack tip radius in the experimental specimen, the influence of the
The salient features of the computational model, experimental results notch tip radius is not considered in this paper for comparison purposes:
on a full-scale pipe bend, and comparison of the test results with the
ΔKI
predictions by the sigma-d approach are presented in this paper. Details Δσde =
2πd (1)
of the recommendation made to RCC MRx in terms of following the
number of cycles based factor of safety (FOS) approach for a more 2 Δσ
realistic fatigue life prediction of prototype structures having a crack- Δεde = (1 + ν ) de
3 E (2)
like defect are also highlighted in this paper.
Δεtr = (k ν − 1) ∗ Δεde (3)

2. Fatigue crack prediction by the sigma-d approach The crack tip plasticity effect (Δεel+pl) is taken into account using
the Neuber's concept [13]. It is based on the strain energy invariance
The life of the SFR class-1 component without any defect is pre- principle using the elastically computed stress (Eq. (1)) and strain (Eq.
dicted using the fast reactor design code RCC MRx RB [2], and the life (2)). The crack tip triaxiality is introduced by the triaxiality factor [3]
of the same component with the defect is predicted by the sigma-d as given in Eq. (3). The material parameter kν and the cyclic stress-
approach in the RCC MRx A16 standard [3]. Many previous researchers strain curve required for this calculation are obtained from RCC MR A3
[8–11] discuss the convenience of the sigma-d approach in predicting [14] against the elastically computed maximum stress range (Δσde). The
the life of a component that has a crack-like defect under fatigue material properties associated with SS 316 LN are used for this purpose.
loading. Eq. (4) is used to obtain the total strain range, which is required for
The sigma-d approach predicts life, based on an elastically com- predicting the fatigue crack initiation:
puted stress at a characteristic distance of 50 μm away from the crack Δεtotal = Δεel + pl + Δεtr (4)
tip singular point (Δσde). This procedure is based on the concept of
linear elastic fracture mechanics (LEFM). The cyclic stress intensity Finally, the number of cycles required for the fatigue crack initiation
factor (ΔKI) of the pipe bend has a longitudinal surface defect and is is obtained from the design fatigue curve RCC MRx A3 [14] along with
computed as per the A16 compendium. The same is compared with the other material properties. The results are further compared with the
Newman-Raju correlation [12]. The value of Δσde at a distance, 50 μm, best-fit material fatigue curve.

2
R. Suresh Kumar, et al. International Journal of Pressure Vessels and Piping 175 (2019) 103924

Fig. 2. Hoop stress distribution of (MPa) the pipe bend under in-plane bending moment and internal pressure.

3. Numerical simulation

The prototype pipe bend along with the associated straight portions
and the rigid loading portions are simulated using 3D solid elements for
numerical analysis. The analysis is carried out for the loading condition
of internal pressure of 0.6 MPa and an in-plane force of 130 kN along
the upward direction (Y-direction) as shown in Fig. 2.
The application of force with an offset (1860 mm) will create an in-
plane bending moment of 241.8 kN m in the pipe bend. It leads to a
closing mode of bending behaviour, which results in ovalisation of the
pipe bend. Numerical analysis helps in identifying the critical location
of the pipe bend from the fatigue damage point of view and also helps
to understand the structural behaviour of the pipe bend under the ap-
plication of the in-plane bending moment. This pipe bend has been
numerically analysed for the test loading conditions considering as well
as neglecting the presence of a surface defect. The results of both the
Fig. 4. Longitudinal stress distribution under the combined action of in-plane
simulations are presented below. bending moment and internal pressure.

3.1. Pipe bend without surface defect developed on the inner surface close to the crown location (78° from the
extrados along the circumferential direction). Hence, the preferred lo-
The hoop and the longitudinal stress distributions of the pipe bend cation and orientation to introduce the defect on the pipe bend is along
under the combined action of the in-plane bending moment and in- the longitudinal direction on the outer surface close to the crown lo-
ternal pressure are presented in Figs. 3 and 4. In both the figures, 0° cation. This defect will produce an early mode-1 crack initiation under
location represents the extrados, 90° represents the crown and 180° the above loading condition.
represents the intrados of the pipe bend. It shows that a maximum hoop
tensile stress of 508.4 MPa is developed on the outer surface close to the 3.2. Pipe bend with a surface defect
crown location (95° from the extrados along the circumferential di-
rection). A maximum longitudinal tensile stress of 362.0 MPa is Numerical analysis also has been carried out for the same pipe bend
in the presence of a semi-elliptical shaped crack-like defect on the outer
surface of the maximum hoop stress location, as identified in Section
3.1. Fine mesh refinement is deployed at the notch tip to predict the
peak stress variation. It is implemented using the standard crack block
element. Biased meshing pattern is followed at the area of singularity
such that the introduction of crack block helps in defining the fine mesh
size of the notch tip and coarse mesh in the far-field portion. The crack
block element will take care of the mesh connectivity with the far-field
geometry. The biased quadratic mesh is deployed such that at a dis-
tance of 50 μm from the crack tip, one element is included and over the
whole ligament thickness of 15 mm, ten quadratic elements are de-
ployed with a non-uniform biased mesh condition.
The computed hoop stress distribution in the pipe bend with the
surface crack-like defect is shown in Fig. 5. It is clear that the presence
of such a defect creates a very high-stress intensification. The maximum
Fig. 3. Hoop stress distribution under the combined action of in-plane bending elastically computed hoop stress at the defect tip is 8656 MPa against
moment and internal pressure. 508.4 MPa observed for the same pipe bend without any surface defect.

3
R. Suresh Kumar, et al. International Journal of Pressure Vessels and Piping 175 (2019) 103924

Fig. 5. Hoop stress distribution (MPa) along the pipe bend with a surface crack like defect under combined action of in-plane bending moment and internal pressure.

respective case. Accordingly, the obtained ΔKI is 33.3 MPa√m at the


crack depth location and 18.5 MPa√m at the crack surface location. The
ΔKI value predicted as per Newman-Raju correlation [12] was
34 MPa√m at the crack depth location and 18.5 MPa√m at the crack
surface location. For the subsequent calculation, ΔKI of 33.3 MPa√m has
been used. Thus, the obtained elastic stress (Δσde as per Eq. (1)) at a
distance of 50 μm from the crack tip is 1877 MPa.

4.2. Fatigue crack initiation as per design code RCC MR A16 procedure

The number of cycles required for crack initiation has been pre-
dicted using the value of Δσde (1877 MPa) obtained from Section 4.1.
The Δσd value is calculated from the elastically computed peak stress
range (Δσde) using Neuber's concept. The total strain range (Δεtotal)
required for the prediction of the fatigue crack initiation is computed as
per Eq. (4). The computed total strain range (Δεtotal) is 2.0836% and the
same has been used for estimating the number of cycles required for
crack initiation. As per the procedure is given in the present RCC MRx
A16 [3], the number of cycles required for fatigue crack initiation is
Fig. 6. Cyclic test arrangements for the full scale representative pipe bend.
obtained from the design fatigue curve against Δεtotal/1.5. Accordingly,
the predicted number of cycles for fatigue crack initiation is 90 cycles.
The hoop tensile stress seen at a distance of 50 μm away from the crack The total strain range (Δεtotal) is divided by a factor of 1.5 to remove the
tip is 2200 MPa. In the vicinity of the crack, the stress variation is hy- FOS associated with the design fatigue curve.
perbolic, and the colour scale distribution provided by the software is
linear. Within the micron level of distance, the stress distribution be- 5. Experimental investigation
comes far-field stress. Thus in the global view, the stress distribution
looks blue, as shown in Fig. 5. Hence, other colours are not visible from The pipe bend cyclic test arrangement is shown in Fig. 6. It is
the engineering-sized pipe bend. A closer view of the elliptical defect subjected to a constant internal pressure of 0.6 MPa and an external
portion is included in Fig. 5 for further clarity. cyclic bending moment. The cyclic bending moment is applied with the
help of an actuator. The actuator's up and down movement creates a
4. Design code prediction for the fatigue crack initiation load variation of 130 kN, as shown in Fig. 6. This applied force acting at
an offset distance of 1860 mm away from the pipe bend mid portion
The number of cycles required to initiate fatigue crack has been induces an in-plane bending moment. An elliptically shaped surface
predicted as per the sigma-d approach given in RCC MRx A16 [3]. The notch of the size 22.5 mm (2c) long and 2.1 mm (a) in depth is provided
results are discussed below. along the longitudinal direction at the crown location, as shown in
Fig. 6.
4.1. Computation of Δσde The cyclic testing experiment is conducted in two stages. During the
first stage, the finite-sized notch is sharpened into a crack-like defect,
The computational procedure of Δσde is given in Section 2. ΔKI is enabling it to develop into a natural crack. The cyclic test is further
computed based on the method given in A16 compendium for the continued to measure the number of cycles required to advance the

4
R. Suresh Kumar, et al. International Journal of Pressure Vessels and Piping 175 (2019) 103924

crack by 50 μm is obtained by the extrapolation of the crack profile


measured at 85,150, 89,125 and 93,100 cycles. The initial crack depth
of 2.1 mm has grown to 2.2 mm at the end of 89,125 cycles and
2.4 mm at the end of 93,100 cycles. The crack initiation is considered as
the advancement of crack by 50 μm as per the design code practice of
RCC MR A16 [3]. Based on the measurements and a linear interpolation
between two consecutive measured values, the number of cycles re-
quired for the initial crack depth 2.10 mm to advance to 2.15 mm (i.e.,
50 μm growth) is 86,475.
To compare the above number of cycles (86,475) with the RCC MRx
A16 [3] prediction, it is necessary to compute the stress at the char-
acteristic distance (Δσde_θ) by considering the notch tip radius (ρ). It is
very sensitive to the notch tip radius (ρ) as given in Creager's equation
(Equation (5)).
2
Δσde _ θ =
ΔKI ⎡cos ⎛ θ ⎞ ⎛cos ⎛ θ ⎞ ⎞ + ρ cos ⎛ θ ⎞ ⎤
⎜ ⎟
⎢ ⎝2⎠ ⎝ 2 ⎠⎥
2πr ⎣ ⎝ ⎝ 2 ⎠⎠ 2r ⎦ (5)

A precise measurement of the notch tip radius prior to the com-


mencement of the experiment plays an important role in predicting
fatigue life. By considering the uncertainty in the measurement of the
tip radius, an alternate method has been adopted in this paper, wherein
the fatigue life has been compared for a sharp crack-like defect (i.e. ρ→
0). For this purpose, the number of cycles, responsible for forming a
Fig. 7. Crack profile measurement. crack-like defect from a finite-sized notch (crack incubation), is sub-
tracted from the observed total number of cycles. So the influence of the
crack by 50 μm or more. It is used for comparing the RCC MRx A16 [3] notch tip radius can be eliminated. The remaining number of cycles
sigma-d prediction. The number of cycles required for the crack in- responsible for the advancement of fatigue crack by 50 μm is 1325
cubation and advancement is estimated with the periodic measure- cycles. This has been further compared with the RCC MRx A16 [3]
ments taken with the crack depth gauge. prediction corresponding to the σde computed as per Eq. (5) by con-
The working principle behind the crack depth gauge is based on the sidering (ρ = 0).
alternating current potential difference (ACPD) technique. Crack depth
gauge houses a display unit and four spring-loaded contact pins, placed
6. Comparison of the results
in a square matrix. These four pins are positioned across the surface
notch, whose profile is to be measured. Constant alternating current is
6.1. Hoop stress distribution
passed through two of these pins, which are on one side of the notch,
and the voltage drop is measured through the other two pins, kept on
The hoop stress distribution in the pipe bend under the application
the other side of the notch. Based on the pre-calibrated material data,
of a global in-plane bending moment of 241.8 kN m is predicted as per
these voltages can be interpreted as notch dimensions. By repeating this
the empirical formula, provided in the design code RCC MRx RB [2].
measurement along the visible direction, the crack advancement can be
Accordingly obtained hoop stress variations at the inner and outer
captured. The crack profiles measured during this experiment are pre-
surfaces of the pipe bend are compared with the 3D finite element si-
sented in Fig. 7.
mulation results. It is shown in Fig. 8.
Fig. 7 depicts that, up to 85,150 cycles, there is no advancement in
This comparison shows that the 3D finite element simulation pre-
the initial notch profile, i.e., at least 85,150 cycles are required for the
dicts more realistic stress pattern than the design code. The RCC MRx
incubation of crack. The pipe bend is a ductile material. It may be
RB [2] empirical formula predicts symmetric behaviour of stress in the
highlighted that a relatively large number of cycles required for crack
inner and outer surfaces of the pipe bend. However, this is incorrect
incubation can be attributed to crack tip plasticity and notch tip
because the material available at the outer surface is more than that in
bluntness. The number of cycles required for the advancement of the
the inner surface due to the bend curvature. Hence, the outer surface

Fig. 8. Comparison of 3D FEM predicted hoop stress distribution with RCC MRx RB values.

5
R. Suresh Kumar, et al. International Journal of Pressure Vessels and Piping 175 (2019) 103924

will experience relatively less bending stress than the inner surface. based on strain range for the accurate fatigue life estimate. The above
This is not reflected in the design code correlations, whereas the 3D hypothesis can be further validated with multiple prototype specimen
finite element simulation results are on the expected lines. However, testing towards confirming the FOS value like 20 or different.
the overall hoop stress distributions and the maximum stress values
predicted by both these methods are in good agreement.
8. Conclusion
6.2. Fatigue crack initiation
Numerical analysis has been performed before prior to the full-scale
experiment. The 3D finite element simulations predicted a more rea-
The number of cycles for fatigue crack initiation, as predicted by the
listic stress distribution than that predicted as per the design code
sigma-d approach, is 90. The corresponding value as per the experi-
empirical relations. Based on the findings of the analysis, a finite-sized
mental data is 1325. Thus, the experimentally observed number of cycle
notch is introduced at the critical location to enhance mode-1 crack
is very high in comparison with the design code prediction.
initiation as early as possible. The cyclic testing experiment was con-
ducted in two stages. During the first stage, the finite-sized notch was
7. Results and discussion sharpened into a crack-like defect, enabling it to develop into a natural
crack (85,150 cycles). In the second stage, the test was further con-
Usually, a FOS value of 2 is applied to the total strain range, and tinued to estimate the number of cycles required to advance the crack
FOS of 20 is applied to the number of cycles from the test data to derive by 50 μm and the results were compared with the predictions of sigma-d
the design fatigue curve [15]. The same philosophy is adopted for de- approach, as given in RCC MRxA16 [3].
riving the design fatigue curve as per RCC MR A3 Gen [14]. Hence, to The number of cycles required for crack incubation and advance-
get a more realistic prediction for the number of cycles required to ment was obtained by a periodic measurement and extrapolation of the
initiate the crack as per RCC MRx A16 [3], the respective FOS has to be measured readings. The results indicate that the present design code
removed from the design fatigue curve. The influence of these factors predictions for fatigue crack initiation behaviour as per the sigma-d
on fatigue life prediction and experimental observation are presented in approach have a high degree of conservatism. It predicts the number of
Fig. 9. cycles from the design fatigue curve by considering a FOS of 1.5 over
Since sigma-d approach predicts very high stress at a distance d the total strain range. As per this procedure, the number of cycles
from the crack tip, the respective total strain range predicted will also predicted is 90, whereas the number of cycles predicted by considering
be very high. In case of higher strain range (i.e. Δεtotal > 0.5% at room FOS 20 on the number of cycles is 690. The experimental observation is
temperature), the design fatigue curve matches the curve derived by 1325 cycles.
dividing a factor 20 on the number of cycles than the curve derived by The sigma-d approach predicted very high stress at the character-
dividing a factor of 2 on total strain range (Δεtotal) as indicated in Fig. 9. istic distance d from the crack tip. In case of higher strain range (i.e.
If we consider this aspect, then the FOS to be considered for predicting Δεtotal > 0.5%), the design fatigue curve would be closer to the curve
the fatigue crack initiation as per design approach will be 20 against the derived by dividing a factor based on the number of cycles rather than
number of cycles. Accordingly, the predicted number of cycles required that obtained by dividing a factor based on total strain range (Δεtotal).
for fatigue crack initiation is 690 cycles, as shown in Fig. 9. Hence, it appears that it is more appropriate to consider a FOS 20 over
In summary, the number of cycles required for fatigue crack in- the number of the cycles than considering FOS 1.5 over the total strain
itiation predicted by RCC MRx A16 [3] considering an FOS of 1.5 on the range (Δεtotal). Based on this, it is suggested to apply a FOS based on the
total strain range gives 90 cycles, and the same predicted by adopting number of cycles rather than the FOS based on strain range for a rea-
FOS 20 on the number of cycles gives 690 cycles. Among these values, listic estimate of fatigue life as per sigma-d approach. The above
the number of cycles predicted by considering FOS 20 on the number of method can be further validated with multiple prototype testings.
cycles is closer to the experimental results of 1325 cycles. Still, the
difference persists because the design fatigue curve is derived from the
average best-fit material curve and is also partially due to the mea- Acknowledgements
surement sensitivity error. Based on this, it can be summarised that in
the present design code RCC MRx A16 [3], the prediction for the fatigue The authors acknowledge the support rendered by the technical
crack initiation, as per the sigma-d approach, has a high margin of staff of the Structural Mechanics Laboratory of Reactor Design Group
conservatism. Hence, it is recommended to apply a suitable multi- for the completion of the experimental work.
plication factor based on the number of cycles rather than the FOS

Fig. 9. Fatigue curves of SS 316 LN material at room temperature as per RCC MRx A3 [14].

6
R. Suresh Kumar, et al. International Journal of Pressure Vessels and Piping 175 (2019) 103924

Nomenclature G. Raghava, Fatigue crack growth behavior in pipes and elbows of carbon steel and
stainless steel materials, Procedia Eng. 55 (2013) 703–709.
[5] S.C. Chetal, V. Balasubramaniyan, P. Chellapandi, P. Mohanakrishnan,
a Depth of the surface defect, mm P. Puthiyavinayagam, C.P. Pillai, S. Raghupathy, T.K. Shanmugham, C.S. Pillai, The
2c Length of the surface defect, mm design of the prototype fast breeder reactor, Nucl. Eng. Des. 236 (2006) 852–860.
d Characteristic distance from the crack tip, mm [6] S. Bhandari, M. Fortmann, L. Grueter, J. Heliot, P. Meyer, B. Percie Du Sert,
A. Prado, H. Zeibig, Crack propagation in a LMFBR elbow, Nucl. Eng. Des. 91
E Young's modulus of elasticity, MPa (1986) 107–119.
kv Triaxiality factor [7] A. Sauter, K. Turan, K. Kerkhof, G. Wagemann, Fracture Mechanics analysis of a
ΔKI Cyclic Stress Intensity Factor in Mode-1, MPa√m pipe elbow, Nucl. Eng. Des. 112 (1989) 211–220.
[8] P. Chellapandi, S.C. Chetal, B. Raj, Assessment of Sigma-d approach for creep da-
Δσd Elasto-plastic stress at a distance d from the crack tip, MPa mage estimation of FBR components with crack like defects at welds, Int. J. Press.
Δσde Elastically computed stress at a distance d from the crack tip, Vessel. Pip. 82 (2005) 739–745.
MPa [9] D.G. Hooton, et al., Application of the sigma-d method for the estimation of creep
crack incubation at austenitic weld boundaries, Proc 2nd Int. Conf. Integr. High
v Poison's ratio of the pipe bend material
Temp. Welds, 2003, pp. 425–434.
[10] J. Chattopadhyay, B. Dutta, H. Kushwaha, 'Leak-before-break qualification of pri-
References mary heat transport piping of 500 MWe Tarapur atomic power plant, Int. J. Press.
Vessel. Pip. 76 (4) (1999) 221–243.
[11] B. Drubay, S. Marie, S. Chapuliot, M.H. Lacire, B. Michel, H. Deschanels, A16: guide
[1] Rangasamy, et al., Challenges during Construction of Sodium Piping Systems for for defect assessment at elevated temperature, Int. J. Press. Vessel. Pip. 80 (2003)
500MWe PFBR', International Conference on Fast Reactors and Related Fuel Cycles: 499–516.
Next Generation Nuclear Systems for Sustainable Development (FR17), [12] J.C. Newman, I.S. Raju, NASA Technical Memorandum 85793, (1984).
Yekaterinburg, Russian Federation, 2017, pp. 26–29 June. [13] T.H. Topper, R.M. Wetzel, J. Morrow, Neuber's Rule Appplied to Fatigue of Notched
[2] RCC MRx RB, Components of Nuclear Installations CLASS N1Rx Reactor Specimens, 1967 NAEC-ASL 1114, June 1976.
Components, (2012). [14] RCC MR A3, General Contents of Appendix - A3, (2012).
[3] RCCMRx. A16, Section III – Tome 1 – Subsection Z - Appendix A16 : Guide for [15] C.E. Jaske, W.J. O'Donnell, Fatigue design criteria for pressure vessel alloys, J.
Prevention of Fast Fracture, Leak before Break Analysis and Defect Assessment, RCC Press. Vessel Technol. 99 (4) (1977) 584–592.
MRx, 2012.
[4] P. Arora, P.K. Singh, V. Bhasin, K.K. Vaze, D.M. Pukazhendhi, P. Gandhi,

You might also like