Stress intensity factors for underclad and through clad defects in a reactor

pressure vessel submitted to a pressurised thermal shock

S. Marie
*
, Y. Menager, S. Chapuliot
CEA, DEN/DM2S/SEMT/LISN, Centre dEtudes de Saclay, 91191 Gif sur Yvette Cedex, France
Received 27 January 2005; revised 16 June 2005; accepted 16 June 2005
Abstract
CEA has launched important work on the development of a Stress Intensity Factors compendium for cracks in a Reactor Pressure Vessel
(RPV) taking into account the cladding.
The work is performed by Finite Element analysis with a parametric mesh for two types of defects (under clad defect and through clad
defect) and a wide range of geometrical and material parameters.
In addition, an analytical stress solution for Pressurised Thermal Shock (PTS) on the RPV is proposed to allow a complete analytical
estimation of the stress intensity factor K
I
for the PTS problem.
The results are validated by comparison with a complete 3D nite element calculation performed on a complex and realistic case study.
q 2005 Elsevier Ltd. All rights reserved.
Keywords: Thermal stresses; Internal pressure; Stress intensity factor KI; Thermal shock; PTS; Vessel; Cladding; RPV
1. Introduction
It is well known that the stress intensity factor is the
parameter which characterises the elastic stress eld at the
tip of a crack in a component. Thus, most methods of
assessment in structural analysis use this parameter as a
starting point with corrections for plasticity, creep,
triaxiality, etc. depending on the criteria and the material.
French documents such as the A16 appendix of RCC-MR
for defect assessment [1] permit users to calculate stress
intensity factors (K
I
) for a wide range of pipes, vessels and
defect geometries. However, no complete compendium or
precise formulation are available for cladded structures, and
more precisely for the austenitic steel cladding present on the
inner wall of the PWR Reactor Pressure Vessel (RPV). To
overcome this problem, CEA decided to provide a complete
and accurate formulation of K
I
solutions for such problems.
This work was performed by an analysis of stress
distributions through the thickness of the vessel for thermal
shock loading, and the development of the associated
inuence functions compendium by a series of Finite
Element (FE) calculations. In addition, to allow an
analytical estimation of K
I
under Pressurised Thermal
Shock (PTS), an analytical solution for the opening stress
under thermal shock is proposed. This solution is a
continuation of the thermal solution given by [2] and
shows that, associated with the proposed formula, an
accurate analytical K
I
can be obtained for PTS on a RPV.
2. Stress intensity factor for a cladded vessel
2.1. Defects under consideration equation
In a PWR, the vessel containing the core is made of a
ferritic steel cylinder (inner radius R
i
of about 2500 mm and
thickness hz250 mm) covered by an inner austenitic steel
cladding with a thickness r between 6 and 12 mm, to protect
the vessel against corrosion.
One of the scenarios considered for defect assessment of
this component is the pressurised thermal shock. Two kinds
of defect could be investigated:
the underclad defect conned in the ferritic thickness
(Fig. 1a),
International Journal of Pressure Vessels and Piping 82 (2005) 746760
www.elsevier.com/locate/ijpvp
0308-0161/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijpvp.2005.06.001
* Corresponding author. Tel.: C33 1 69 08 92 57; fax: C33 1 69 08 87 84
E-mail address: smarie@cea.fr (S. Marie).
the through clad defect (Fig. 1b) which is here
represented with a semi-elliptical crack in the ferritic
part and a constant width part in the cladding.
This study aims to propose tools for crack initiation
assessment. It does not treat crack growth or crack arrest
situations: The defects under consideration are small in
comparison with the vessel thickness and radius. In this
situation, the vessel curvature has a small effect and the
problem could be assimilated to the case of a cladded plate
with the same defects.
2.2. Opening stress representation
The expression which will be used for the stress intensity
factor is based on a polynomial relation of the nominal
opening stress (obtained in the defect-free structure). For a
circumferential defect this stress corresponds to the axial
stress, and for a longitudinal defect, this is the circumfer-
ential stress.
Under thermal shock, the continuity of the strains at the
interface between the cladding and the ferritic vessel, and
the different material properties (in particular, Youngs
modulus and expansion coefcient) lead to a discontinuous
stress distribution at the interface. Fig. 2 shows a typical
through thickness variation of the axial stress in this
situation.
The main difculty is then to take into account this stress
discontinuity in the stress intensity factor expression. To
overcome this difculty:
it is chosen to represent the through thickness stress
variation with the parameter u, equal to zero at the
cladding inner skin and normalised with the total
thickness hCr, which allows the same treatment for
the underclad defects and through clad defects.
Nomenclature
a Crack depth in the ferritic material
c Crack length
B Biot number
B* Equivalent Biot number for vessel with cladding
d Diffusivity (m
2
/s)
E
i
Youngs modulus of the material i (1Zausteni-
tic steel; 2Zferritic steel)
F
c
Fourrier number characteristic of the linear
shock
h Thickness of the ferritic vessel
H Heat transfer coefcient (W/m
2
/8C)
i
i
,i
ir
Axisymmetric shape functions
K
I
Stress intensity factor (MPa m
0.5
)
L
p
Plate width and length
P
i
Slope of curve giving temperature in uid as a
function of
~
t
r Cladding thickness
R
e
Outer radius of vessel
R
i
Inner radius of vessel
R
0
Radius of the interface
~
t Normalised time
T ~ x;
~
t Temperature in the vessel
T
1
~ x;
~
t Temperature in the cladding
T
2
~ x;
~
t Temperature in vessel with cladding
T
fluid

~
t Fluid temperature
T
1

~
t Uniform term in linear expression of cladding
temperature
T
i
Initial temperature of uid
T
f
Final temperature of uid for a thermal shock
u Radial distance to inner wall
x Radial position
~ x Normalised radial position ~ xZr
e
Kr=h
~ x
1
Normalised radial position of inner surface of
cladding
a
i
Thermal expansion coefcient of the material
i (8C
K1
)
DP
i
Variation of slope P
i
between two given instants
Dt
shock
Duration of linear shock
DT
1

~
t Linear term in expression of uid
temperature
ll
i
Thermal conductivity of the material i
n Poissons ratio
s
q
Tangential stress
s
i
, s
ir
Polynomial stress description
s
r
Radial stress
s
z
Axial stress
s
th
Thermal stress eld
s
p
Stress eld due to pressure
a
A
C
A
C
Under clad defect (b) (a) Through clad defect
2.c 2.c
h
r
h r
Outer skin
Inner skin
u
u
Fig. 1. The defects under consideration: (a) underclad defect, (b) through
clad defect.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 747
The stress variation s
f
in the ferritic part is represented
using a polynomial expression up to the 4th order:
s
f
Z

4
jZ0
s
j
u
h Cr
_ _
j
(1)
in the cladding, the additional stress variation s
r
is
represented using a linear expression as r is small in
comparison to h:
s
r
Zs
0r
Cs
1r
u
h Cr
if u!r (2)
Fig. 2 represents this stress decomposition.
To summarise, the opening stress in the cracked section
is:
s Zs
f
Cs
r
(3)
2.3. Stress intensity factor expression
The stress intensity factor is calculated here using shape
functions, linking this parameter to a particular component
s
j
or s
jr
of the opening stress representation. Usually, the
corresponding shape function is noted i
j
or i
jr
.
For a through clad defect, the stress intensity factor is
K
I
Z

4
jZ0
s
j
i
j
a Cr
h Cr
_ _
j
C

1
jZ0
s
jr
i
jr
a Cr
h Cr
_ _
j
_ _

pa Cr
_
(4)
For an underclad defect, the stress intensity factor is
obtained using the following relationship:
K
I
Z

4
jZ0
s
j
i
j
a Cr
h Cr
_ _
j
_ _

pa
p
(5)
where a is the depth of the defect, s
0
to s
4
, s
0r
and s
1r
are
the stress components dened previously, and i
0
to i
4
, i
0r
and
i
1r
are shape functions.
2.4. Shape functions determination
Each shape function is calculated for various representa-
tive geometries of vessel and defect using Finite Element
analysis.
2.4.1. Mesh description
The dimensional analysis of the problem shows that ve
non dimensional parameters are necessary to describe
precisely the problem: h/R
i
, a/h, a/r, a/c and E
1
/E
2
(n is
taken equal to 0.3 for both materials).
However, for small defects in comparison to wall thickness
(a/h/0.1), the curvature of the structure has no impact on the
inuence functions: the inuence function for h/R
i
%0.1 is
equivalent to that for cracks in a plate (h/R
i
Z0). Thus, calcu-
lations are made onplates, andinuence functions are assumed
to be the same for longitudinal or circumferential cracks.
As small defects are considered, the ratio a/h has a small
inuence on the shape function: the relevant dimensional
parameter to take into account the crack size is the cladding
thickness.
Finally, the dimension of the problem can then be
reduced to three parameters a/r, a/c and E
1
/E
2
.
The nite element models were made using both 2D and
3D isoparametric quadratic elements of the nite element
code CAST3M [3], developed by CEA. A representative
view of the meshes is shown in Fig. 3. The gure shows that
the crack front was accurately modelled, either for the
elliptical part of the crack front in the ferritic material or the
interface of the two materials (under clad defect). In our
calculation, the through clad defect is assumed to be semi
elliptical in the ferritic material. The crack shape is linear
inside the cladding (Fig. 1).
2.4.2. Boundary conditions
Only a quarter of the plate had to be modelled because of
symmetry conditions. No particular boundary condition was
imposed at the top of the plate. The width and length of the
plate were equal, and dened to avoid any signicant end-
effects using the formula:
L
p
Z4h (6)
2.4.3. Loading description
According to the principle of superposition, the shape
factors are then calculated one by one, applying the load
cases as corresponding pressures on the crack face:
Imposed pressure on the whole crack face for the
coefcients corresponding to the polynomial t in the
ferritic material.
0 1

0 1

f
0 1

r
= +
r + h r + h r + h
u u u
Fig. 2. Stress variation decomposition.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 748
For the particular case of a through thickness defect and
the linear t in the cladding, the load is only applied on
the cladding part of the face.
2.4.4. Calculation of the shape factors
The stress intensity factor K
I
was obtained from the
J-integral. The J-integral was calculated at nodes along the
crack front by the G(q) method [4], based on a domain
integral technique. K
I
was derived from the J-integral, under
the assumption that plane strain conditions prevailed along
the whole crack front, through the following formula:
K
I
Z

J
p
with E

Z
E
1Kn
2
(7)
Then, for each loading case, the shape factor is
determined by:
For a through clad defect:
i
j
Z
K
I
aCr
hCr
_ _
j

pa Cr
p
(8)
For an under clad defect:
i
j
Z
K
I
aCr
hCr
_ _
j

p$a
p
(9)
Such calculations were performed for a wide range of
these three parameters. Calculated results are given at the
deepest point of the crack (point A-Fig. 4), at the point at the
interface of the two materials (point C) and at the interface
point under clad defect (point B). This last point is only
dened for E
2
/E
1
Z1.
As quadratic elements are used, a mean value of the
shape factor is calculated for each element. For the deepest
point of the defect A, the three nodes of the element are
called point 1, point 2 and point 3 as represented in Fig. 4.
The second characteristic point under consideration is the
point C at the interface between the ferritic vessel and
the cladding (Fig. 4). The corresponding shape factor is the
mean value deduced for the adjacent quadratic element in
the ferritic part, represented by the nodes 4, 5 and 6 (Fig. 4).
But for this point, as the node directly at the interface is
affected by the cladding, the shape factor for point C is only
calculated using points 4 and 5.
Through clad defect
Underclad defect
Fig. 3. Meshes used for the stress intensity compendium development.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 749
For the through clad defect, the third point is point D
representative of the cladding (Fig. 4a): as clad thickness
is small in comparison to the other dimensions, the
corresponding value of the shape factor is the mean
value on the whole part of the crack front in the
cladding.
For the underclad defect, the third characteristic point
is point B, located at the middle of the intersection
between the defect and the interface (Fig. 4b). As for
point A, the value of the corresponding shape function is
the mean value from the adjacent element (points 7, 8,
and 9 in Fig. 4b).
For each case, the mean value is not a simple average
value on the element, but takes into account the inuence of
the quadratic interpolation function of the elements in the
G(q) method: the weight attributed to the node corresponds to
the integration of the interpolation function on the element,
i.e. 4/6 for the middle node and 1/6 for the edge nodes. Also,
as the edge nodes belong to two consecutive elements, the
weight is divided by 2. Finally, the mean shape functions are
deduced from the following relations [5]:
i
A
Z
1
6
i
point 1
C4$i
point 2
Ci
point 3

i
C
Z
1
5
i
point 4
C4$i
point 5

i
B
Z
1
6
i
point 7
C4$i
point 8
Ci
point 9

_
for an underclad defect 10
i
A
Z
1
6
i
point 1
C4$i
point 2
Ci
point 3

i
C
Z
1
5
i
point 4
C4$i
point 5

i
D
Z

n
k
n
i
M
n

n
k
n
_

_
for a through clad defect 11
where n is the number of nodes in the cladding, i
Mn
is the
shape function of the node n of the cladding and k
n
the weight
for this node (2 for an edge node, 4 for a middle node and 0 for
the interface node to avoid any inuence of the ferritic part).
Tables 14present the calculated shape function for the
two kinds of defect.
2.5. Complementary solutions for the innite defects
The compendium for an innite defect (a/c)Z0
(Tables 57) was completed by existing solutions for the
deep conguration for through-clad defects [5]. In this case,
the ratio E
1
/E
2
equals 1.
2.6. Validation
The consistency of the present results with existing
solutions in the literature was investigated for specic cases:
the solutions for the through clad defects when the
cladding has same material properties than the vessel has
been compared to the existing solutions for a plate
containing a semi-elliptical defect [1] (with same
dimensions a and c). As the defect considered here is
not a real ellipse (see Fig. 1), the agreement is not perfect
but becomes better and better when the ration a/r
increase, i.e. the when the rectangular part of the defect
becomes small.
The solution for under clad defects which a/cZ0 has
been compared to the existing solution of the dissym-
metric innitely long defect in a plate [6]. A good
agreement is found. For example, for a vessel thickness
of 200 mm, a cladding size of 6 mm and a ratio a/rZ2, a
difference of 6% is obtained between the two compendia
for the deepest point of the defect.
3. Nominal stress
When dening a complete analytical method for a given
geometry, a compendium for the nominal stress is also
needed. Here, two types of loading are considered: internal
pressure and thermal shock. For each, the corresponding
through thickness stress variation is derived analytically and
linearised for small values of a/h.
4
5
6
A
C
D
1
2
3
4
5
6
1
2
3
A
C
7
8
9
B
Fig. 4. particular points of the defects for the K
I
calculation.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 750
Table 1
Shape factors for an under clad defect, with E
1
/E
2
Z1
a/c a/r i at the deepest point i at the interface point Point B
i
0
i
1
i
2
i
3
i
4
i
0
i
1
i
2
i
3
i
4
i
0
i
1
i
2
i
3
i
4
1 0 0.550 0.550 0.550 0.550 0.550 0.263 0.263 0.263 0.263 0.263 0.580 0.580 0.580 0.580 0.580
1 1/8 0.550 0.537 0.524 0.512 0.500 0.264 0.238 0.214 0.193 0.174 0.582 0.530 0.484 0.442 0.404
1 1/4 0.551 0.526 0.504 0.484 0.466 0.264 0.217 0.178 0.147 0.121 0.582 0.490 0.414 0.350 0.297
1 1/2 0.551 0.511 0.477 0.448 0.423 0.265 0.186 0.131 9.27!10
K2
6.61!10
K2
0.584 0.430 0.320 0.240 0.182
1 1 0.554 0.493 0.445 0.409 0.379 0.267 0.148 8.32!10
K2
4.80!10
K2
2.85!10
K2
0.593 0.359 0.225 0.145 9.74!10
K2
1 3/2 0.557 0.482 0.429 0.389 0.357 0.270 0.125 6.05!10
K2
3.08!10
K2
1.67!10
K2
0.602 0.318 0.178 0.106 6.76!10
K2
1 2 0.560 0.476 0.418 0.376 0.345 0.272 0.111 4.79!10
K2
2.26!10
K2
1.19!10
K2
0.611 0.292 0.152 8.70!10
K2
5.43!10
K2
1/2 0 0.642 0.642 0.642 0.642 0.642 0.244 0.244 0.244 0.244 0.244 0.635 0.635 0.635 0.635 0.635
1/2 1/8 0.643 0.626 0.610 0.594 0.579 0.245 0.220 0.197 0.177 0.160 0.639 0.584 0.535 0.490 0.450
1/2 1/4 0.645 0.613 0.585 0.559 0.536 0.245 0.200 0.164 0.134 0.110 0.640 0.542 0.461 0.393 0.336
1/2 1/2 0.647 0.594 0.550 0.513 0.481 0.246 0.171 0.119 8.36!10
K2
5.89!10
K2
0.644 0.480 0.361 0.276 0.213
1/2 1 0.653 0.572 0.511 0.464 0.426 0.248 0.135 7.47!10
K2
4.20!10
K2
2.42!10
K2
0.658 0.408 0.262 0.174 0.121
1/2 3/2 0.660 0.560 0.490 0.439 0.400 0.251 0.114 5.34!10
K2
2.61!10
K2
1.35!10
K2
0.673 0.367 0.214 0.133 8.84!10
K2
1/2 2 0.666 0.553 0.478 0.425 0.385 0.253 0.100 4.16!10
K2
1.86!10
K2
9.16!10
K3
0.688 0.342 0.187 0.113 7.39!10
K2
1/4 0 0.677 0.677 0.677 0.677 0.677 0.215 0.215 0.215 0.215 0.215 0.657 0.657 0.657 0.657 0.657
1/4 1/8 0.680 0.661 0.643 0.626 0.610 0.216 0.193 0.173 0.155 0.139 0.661 0.606 0.556 0.510 0.469
1/4 1/4 0.682 0.648 0.617 0.590 0.564 0.216 0.176 0.143 0.116 9.45!10
K2
0.664 0.564 0.481 0.411 0.353
1/4 1/2 0.687 0.629 0.581 0.540 0.505 0.217 0.149 0.103 7.11!10
K2
4.92!10
K2
0.670 0.502 0.381 0.293 0.228
1/4 1 0.696 0.607 0.540 0.488 0.448 0.219 0.117 6.26!10
K2
3.40!10
K2
1.87!10
K2
0.689 0.432 0.281 0.190 0.134
1/4 3/2 0.706 0.595 0.518 0.462 0.420 0.220 9.70!10
K2
4.34!10
K2
2.00!10
K2
9.50!10
K3
0.707 0.392 0.233 0.148 0.101
1/4 2 0.714 0.588 0.505 0.447 0.404 0.222 8.40!10
K2
3.28!10
K2
1.34!10
K2
5.87!10
K3
0.726 0.369 0.207 0.128 8.60!10
K2
1/8 0 0.689 0.689 0.689 0.689 0.689 0.185 0.185 0.185 0.185 0.185 0.665 0.665 0.665 0.665 0.665
1/8 1/8 0.692 0.673 0.654 0.637 0.621 0.185 0.165 0.147 0.132 0.118 0.670 0.614 0.564 0.518 0.476
1/8 1/4 0.696 0.660 0.629 0.600 0.574 0.185 0.150 0.121 9.79!10
K2
7.92!10
K2
0.674 0.573 0.489 0.419 0.360
1/8 1/2 0.701 0.642 0.592 0.550 0.515 0.185 0.126 8.61!10
K2
5.88!10
K2
4.02!10
K2
0.681 0.512 0.389 0.300 0.234
1/8 1 0.713 0.620 0.551 0.498 0.456 0.186 9.74!10
K2
5.11!10
K2
2.70!10
K2
1.43!10
K2
0.702 0.443 0.290 0.197 0.140
1/8 3/2 0.723 0.609 0.529 0.471 0.427 0.187 8.01!10
K2
3.46!10
K2
1.51!10
K2
6.70!10
K3
0.722 0.404 0.242 0.155 0.107
1/8 2 0.732 0.602 0.516 0.456 0.411 0.188 6.86!10
K2
2.54!10
K2
9.63!10
K3
3.78!10
K3
0.742 0.381 0.216 0.135 9.19!10
K2
1/16 0 0.693 0.693 0.693 0.693 0.693 0.156 0.156 0.156 0.156 0.156 0.667 0.667 0.667 0.667 0.667
1/16 1/8 0.696 0.677 0.658 0.641 0.624 0.156 0.139 0.124 0.111 9.91!10
K2
0.672 0.616 0.566 0.520 0.478
1/16 1/4 0.700 0.665 0.633 0.604 0.578 0.156 0.126 0.102 8.20!10
K2
6.62!10
K2
0.676 0.576 0.492 0.421 0.362
1/16 1/2 0.706 0.646 0.596 0.554 0.518 0.156 0.106 7.19!10
K2
4.87!10
K2
3.31!10
K2
0.684 0.515 0.392 0.302 0.236
1/16 1 0.718 0.625 0.555 0.501 0.459 0.157 8.11!10
K2
4.20!10
K2
2.18!10
K2
1.13!10
K2
0.706 0.446 0.292 0.200 0.142
1/16 3/2 0.729 0.613 0.533 0.475 0.430 0.158 6.62!10
K2
2.79!10
K2
1.18!10
K2
5.05!10
K3
0.726 0.407 0.245 0.158 0.109
1/16 2 0.739 0.607 0.520 0.459 0.414 0.158 5.63!10
K2
2.02!10
K2
7.31!10
K3
2.70!10
K3
0.747 0.385 0.219 0.138 9.43!10
K2
0 0 0.680 0.680 0.680 0.680 0.680 0.000 0.000 0.000 0.000 0.000 0.680 0.680 0.680 0.680 0.680
0 1/8 0.681 0.661 0.643 0.625 0.608 0.000 0.000 0.000 0.000 0.000 0.681 0.625 0.574 0.528 0.486
0 1/4 0.683 0.647 0.615 0.586 0.559 0.000 0.000 0.000 0.000 0.000 0.683 0.582 0.498 0.427 0.368
0 1/2 0.687 0.627 0.576 0.534 0.498 0.000 0.000 0.000 0.000 0.000 0.689 0.520 0.397 0.307 0.241
0 1 0.699 0.606 0.536 0.482 0.439 0.000 0.000 0.000 0.000 0.000 0.707 0.448 0.295 0.203 0.145
0 3/2 0.710 0.594 0.513 0.455 0.410 0.000 0.000 0.000 0.000 0.000 0.726 0.410 0.248 0.161 0.112
0 2 0.720 0.588 0.501 0.439 0.394 0.000 0.000 0.000 0.000 0.000 0.748 0.388 0.223 0.141 9.73!10
K2
S
.
M
a
r
i
e
e
t
a
l
.
/
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
P
r
e
s
s
u
r
e
V
e
s
s
e
l
s
a
n
d
P
i
p
i
n
g
8
2
(
2
0
0
5
)
7
4
6

7
6
0
7
5
1
3.1. Thermal stress eld
The stress eld for a thermal shock in a cladded vessel
has been derived in [2]. This solution is an extension of the
well-known solution for a linear thermal shock, taking into
account the cladding and a complex variation (even cyclic)
of the uid temperature.
The geometry considered is represented in Fig. 5. d is the
thermal diffusivity, l the thermal conductivity and H
the coefcient for heat exchange between the uid and the
vessel. To solve the problem, the non-dimensional distance
~ xZR
e
Kx=h and the non-dimensional time
~
tZt$d=h
2
are
introduced. The Biot number is also given as BZHh/l.
Material 1 is the cladding and material 2 is the ferritic
vessel.
To treat a complex uid temperature variation, the
approach consists of breaking down the uid temperature
into a succession of linear shocks (Fig. 6). The uid
temperature is represented by a series of temperatures T
uid,
iZ0,1,.
. The variation in temperature between instants i and
iC1 is considered to be linear.
~
t
iZ0;1;.
designates the
instants associated with temperatures T
uid,iZ0,1,.
. The
slope corresponding to the variation between instants
~
t
i
and
~
t
iC1
is designated P
i
. This slope is then broken down as
follows:
P
iC1
ZP
i
CDP
iC1
(12)
where DP
i
represents the variation of the slope from instant
~
t
i
. The relationship (12) is then written:
Table 2
Shape factors for an under clad defect, with E
1
/E
2
Z0.7
a/c a/r i at the deepest point i at the interface point
i
0
i
1
i
2
i
3
i
4
i
0
i
1
i
2
i
3
i
4
1 0 0.558 0.558 0.558 0.558 0.558 0.284 0.284 0.284 0.284 0.284
1 1/8 0.558 0.544 0.531 0.518 0.506 0.284 0.256 0.231 0.208 0.187
1 1/4 0.559 0.533 0.511 0.490 0.471 0.284 0.233 0.192 0.158 0.131
1 1/2 0.560 0.517 0.482 0.452 0.426 0.285 0.200 0.141 0.101 7.20!10
K2
1 1 0.562 0.498 0.449 0.411 0.381 0.287 0.160 9.04!10
K2
5.25!10
K2
3.14!10
K2
1 3/2 0.565 0.487 0.432 0.391 0.359 0.290 0.136 6.61!10
K2
3.39!10
K2
1.86!10
K2
1 2 0.568 0.480 0.421 0.378 0.346 0.293 0.120 5.25!10
K2
2.51!10
K2
1.33!10
K2
1/2 0 0.657 0.657 0.657 0.657 0.657 0.262 0.262 0.262 0.262 0.262
1/2 1/8 0.658 0.640 0.622 0.606 0.590 0.262 0.236 0.212 0.191 0.171
1/2 1/4 0.660 0.626 0.596 0.569 0.545 0.263 0.215 0.176 0.144 0.118
1/2 1/2 0.662 0.606 0.560 0.521 0.487 0.264 0.184 0.129 9.04!10
K2
6.39!10
K2
1/2 1 0.669 0.583 0.519 0.470 0.431 0.266 0.146 8.09!10
K2
4.58!10
K2
2.66!10
K2
1/2 3/2 0.675 0.570 0.497 0.444 0.403 0.269 0.123 5.82!10
K2
2.87!10
K2
1.50!10
K2
1/2 2 0.681 0.562 0.484 0.428 0.388 0.271 0.108 4.55!10
K2
2.06!10
K2
1.03!10
K2
1/4 0 0.695 0.695 0.695 0.695 0.695 0.229 0.229 0.229 0.229 0.229
1/4 1/8 0.698 0.678 0.659 0.642 0.625 0.230 0.206 0.184 0.165 0.148
1/4 1/4 0.701 0.665 0.632 0.603 0.576 0.230 0.187 0.152 0.124 0.101
1/4 1/2 0.706 0.645 0.593 0.550 0.514 0.231 0.159 0.110 7.61!10
K2
5.28!10
K2
1/4 1 0.716 0.621 0.550 0.496 0.453 0.233 0.125 6.72!10
K2
3.66!10
K2
2.03!10
K2
1/4 3/2 0.726 0.608 0.527 0.469 0.424 0.235 1.04!10
K1
4.68!10
K2
2.17!10
K2
1.04!10
K2
1/4 2 0.734 0.601 0.513 0.452 0.408 0.236 9.00!10
K2
3.55!10
K2
1.47!10
K2
6.51!10
K3
1/8 0 0.709 0.709 0.709 0.709 0.709 0.195 0.195 0.195 0.195 0.195
1/8 1/8 0.712 0.691 0.672 0.654 0.636 0.195 0.174 0.155 0.139 0.124
1/8 1/4 0.716 0.679 0.645 0.615 0.587 0.195 0.158 0.128 0.103 8.37!10
K2
1/8 1/2 0.723 0.659 0.606 0.562 0.524 0.195 0.133 9.10!10
K2
6.22!10
K2
4.26!10
K2
1/8 1 0.734 0.636 0.562 0.506 0.462 0.196 0.103 5.41!10
K2
2.86!10
K2
1.52!10
K2
1/8 3/2 0.745 0.623 0.539 0.478 0.433 0.197 8.47!10
K2
3.67!10
K2
1.61!10
K2
7.19!10
K3
1/8 2 0.755 0.616 0.525 0.462 0.416 0.198 7.26!10
K2
2.70!10
K2
1.03!10
K2
4.09!10
K3
1/16 0 0.713 0.713 0.713 0.713 0.713 0.163 0.163 0.163 0.163 0.163
1/16 1/8 0.717 0.696 0.676 0.658 0.640 0.163 0.146 0.130 0.116 0.104
1/16 1/4 0.721 0.684 0.650 0.619 0.591 0.163 0.132 0.106 8.57!10
K2
6.92!10
K2
1/16 1/2 0.728 0.664 0.611 0.566 0.528 0.163 0.111 7.51!10
K2
5.10!10
K2
3.46!10
K2
1/16 1 0.741 0.641 0.567 0.510 0.466 0.164 8.48!10
K2
4.39!10
K2
2.28!10
K2
1.19!10
K2
1/16 3/2 0.752 0.628 0.543 0.482 0.436 0.165 6.93!10
K2
2.93!10
K2
1.24!10
K2
5.32!10
K3
1/16 2 0.762 0.621 0.529 0.466 0.419 0.165 5.89!10
K2
2.12!10
K2
7.70!10
K3
2.85!10
K3
0 0 0.702 0.702 0.702 0.702 0.702 0. 0. 0. 0. 0.
0 1/8 0.703 0.682 0.662 0.643 0.625 0. 0. 0. 0. 0.
0 1/4 0.705 0.666 0.632 0.601 0.573 0. 0. 0. 0. 0.
0 1/2 0.709 0.645 0.591 0.546 0.508 0. 0. 0. 0. 0.
0 1 0.722 0.622 0.548 0.491 0.446 0. 0. 0. 0. 0.
0 3/2 0.733 0.609 0.524 0.463 0.416 0. 0. 0. 0. 0.
0 2 0.744 0.603 0.510 0.446 0.399 0. 0. 0. 0. 0.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 752
Table 3
Shape factors for a through clad defect, with E
1
/E
2
Z1
a/c a/r i at the deepest point i at the interface point
i
0
i
1
i
2
i
3
i
4
i
0r
i
1r
i
0
i
1
i
2
i
3
i
4
i
0r
i
1r
1 1/8 0.229 0.215 0.205 0.196 0.189 0.035 0.027 0.262 0.226 0.197 0.173 0.153 0.139 0.115
1 1/4 0.308 0.277 0.256 0.239 0.226 0.048 0.030 0.353 0.268 0.209 0.166 0.133 0.179 0.123
1 1/2 0.399 0.337 0.300 0.275 0.255 0.062 0.027 0.458 0.282 0.186 0.127 0.089 0.224 0.113
1 1 0.488 0.385 0.333 0.299 0.274 0.069 0.020 0.565 0.263 0.139 0.079 0.047 0.262 0.087
1 3/2 0.532 0.407 0.347 0.309 0.282 0.068 0.015 0.619 0.241 0.112 0.058 0.033 0.272 0.068
1 2 0.558 0.419 0.355 0.314 0.286 0.063 0.011 0.652 0.225 0.096 0.048 0.027 0.270 0.054
1/2 1/8 0.339 0.302 0.278 0.260 0.245 0.100 0.072 0.292 0.242 0.207 0.179 0.156 0.175 0.136
1/2 1/4 0.464 0.382 0.336 0.305 0.281 0.141 0.078 0.397 0.284 0.216 0.168 0.133 0.230 0.145
1/2 1/2 0.602 0.451 0.381 0.337 0.306 0.179 0.070 0.515 0.297 0.190 0.128 0.088 0.286 0.133
1/2 1 0.721 0.502 0.411 0.357 0.320 0.185 0.049 0.617 0.274 0.142 0.080 0.047 0.315 0.099
1/2 3/2 0.771 0.522 0.422 0.364 0.325 0.168 0.035 0.661 0.249 0.114 0.058 0.033 0.312 0.075
1/2 2 0.798 0.533 0.428 0.368 0.328 0.150 0.026 0.685 0.230 0.097 0.048 0.027 0.301 0.059
1/4 1/8 0.466 0.381 0.337 0.307 0.284 0.203 0.127 0.304 0.238 0.198 0.169 0.146 0.203 0.147
1/4 1/4 0.649 0.477 0.400 0.353 0.320 0.289 0.140 0.409 0.275 0.203 0.156 0.122 0.264 0.155
1/4 1/2 0.819 0.550 0.443 0.382 0.340 0.338 0.122 0.509 0.280 0.175 0.115 0.078 0.308 0.137
1/4 1 0.927 0.592 0.466 0.396 0.350 0.307 0.079 0.581 0.249 0.125 0.067 0.039 0.317 0.099
1/4 3/2 0.962 0.606 0.474 0.401 0.353 0.261 0.053 0.608 0.220 0.096 0.046 0.025 0.305 0.074
1/4 2 0.979 0.612 0.477 0.403 0.355 0.224 0.038 0.620 0.198 0.078 0.036 0.019 0.290 0.057
1/8 1/8 0.634 0.460 0.388 0.344 0.313 0.358 0.194 0.311 0.228 0.186 0.156 0.133 0.225 0.151
1/8 1/4 0.845 0.563 0.452 0.389 0.347 0.462 0.204 0.391 0.253 0.184 0.139 0.108 0.268 0.152
1/8 1/2 0.984 0.621 0.485 0.411 0.363 0.465 0.161 0.458 0.247 0.152 0.098 0.065 0.289 0.129
1/8 1 1.047 0.645 0.499 0.419 0.368 0.377 0.095 0.505 0.210 0.101 0.052 0.028 0.287 0.091
1/8 3/2 1.064 0.651 0.502 0.421 0.369 0.309 0.062 0.520 0.180 0.074 0.033 0.016 0.274 0.067
1/8 2 1.073 0.655 0.504 0.422 0.370 0.261 0.044 0.526 0.158 0.057 0.023 0.011 0.261 0.052
1/16 1/8 0.829 0.544 0.437 0.378 0.338 0.543 0.269 0.295 0.211 0.170 0.142 0.120 0.230 0.152
1/16 1/4 0.983 0.620 0.486 0.413 0.364 0.584 0.248 0.349 0.224 0.162 0.121 0.093 0.255 0.148
1/16 1/2 1.061 0.654 0.506 0.426 0.374 0.523 0.179 0.396 0.211 0.128 0.081 0.053 0.269 0.124
1/16 1 1.095 0.667 0.513 0.430 0.376 0.404 0.102 0.429 0.174 0.081 0.040 0.020 0.266 0.088
1/16 3/2 1.104 0.671 0.515 0.431 0.377 0.326 0.065 0.439 0.146 0.056 0.023 0.010 0.255 0.065
1/16 2 1.110 0.673 0.515 0.431 0.377 0.274 0.046 0.442 0.125 0.042 0.015 0.006 0.243 0.051
0 See innitely long longitudinal and circumferential defects (Tables 57)
S
.
M
a
r
i
e
e
t
a
l
.
/
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
P
r
e
s
s
u
r
e
V
e
s
s
e
l
s
a
n
d
P
i
p
i
n
g
8
2
(
2
0
0
5
)
7
4
6

7
6
0
7
5
3
Table 4
Shape factors for through clad defect, with E
1
/E
2
Z0.7
a/c a/r i at the deepest point i at the interface point
i
0
i
1
i
2
i
3
i
4
I
0r
i
1r
i
0
i
1
i
2
i
3
i
4
i
0r
i
1r
1 1/8 0.238 0.223 0.211 0.202 0.193 0.041 0.032 0.308 0.265 0.231 0.202 0.178 0.173 0.142
1 1/4 0.321 0.286 0.262 0.244 0.230 0.056 0.035 0.412 0.310 0.241 0.191 0.152 0.220 0.150
1 1/2 0.416 0.345 0.306 0.278 0.258 0.072 0.032 0.532 0.323 0.212 0.144 0.100 0.275 0.138
1 1 0.506 0.393 0.337 0.302 0.276 0.081 0.023 0.651 0.298 0.156 0.088 0.053 0.320 0.105
1 3/2 0.549 0.413 0.350 0.311 0.283 0.077 0.017 0.707 0.271 0.125 0.064 0.037 0.328 0.081
1 2 0.574 0.424 0.357 0.316 0.287 0.071 0.012 0.738 0.250 0.106 0.052 0.029 0.323 0.064
1/2 1/8 0.362 0.320 0.293 0.272 0.255 0.117 0.083 0.344 0.284 0.242 0.209 0.182 0.215 0.167
1/2 1/4 0.497 0.402 0.351 0.316 0.290 0.166 0.091 0.466 0.329 0.249 0.193 0.153 0.282 0.177
1/2 1/2 0.644 0.471 0.393 0.345 0.312 0.209 0.081 0.599 0.340 0.216 0.144 0.099 0.349 0.160
1/2 1 0.761 0.518 0.420 0.363 0.324 0.211 0.056 0.706 0.309 0.159 0.089 0.052 0.377 0.118
1/2 3/2 0.805 0.535 0.429 0.369 0.328 0.188 0.039 0.747 0.277 0.126 0.064 0.036 0.369 0.089
1/2 2 0.827 0.543 0.433 0.372 0.330 0.165 0.028 0.767 0.253 0.106 0.052 0.029 0.353 0.069
1/4 1/8 0.506 0.408 0.358 0.324 0.299 0.235 0.147 0.356 0.277 0.230 0.196 0.168 0.246 0.178
1/4 1/4 0.706 0.508 0.421 0.368 0.331 0.334 0.161 0.476 0.316 0.233 0.178 0.138 0.318 0.186
1/4 1/2 0.881 0.578 0.460 0.393 0.348 0.385 0.137 0.584 0.317 0.197 0.129 0.087 0.366 0.163
1/4 1 0.977 0.612 0.477 0.403 0.355 0.339 0.086 0.656 0.277 0.138 0.074 0.042 0.371 0.116
1/4 3/2 1.002 0.621 0.482 0.406 0.357 0.283 0.057 0.679 0.241 0.104 0.050 0.026 0.355 0.085
1/4 2 1.012 0.624 0.484 0.407 0.358 0.240 0.040 0.688 0.215 0.084 0.038 0.019 0.336 0.066
1/8 1/8 0.690 0.494 0.412 0.363 0.329 0.406 0.219 0.358 0.262 0.213 0.179 0.153 0.266 0.180
1/8 1/4 0.914 0.597 0.475 0.405 0.360 0.518 0.228 0.445 0.287 0.208 0.157 0.121 0.313 0.179
1/8 1/2 1.047 0.649 0.502 0.423 0.371 0.512 0.177 0.516 0.276 0.169 0.108 0.071 0.337 0.151
1/8 1 1.093 0.664 0.509 0.426 0.373 0.407 0.103 0.562 0.231 0.111 0.057 0.030 0.332 0.105
1/8 3/2 1.100 0.666 0.510 0.426 0.373 0.329 0.066 0.575 0.196 0.079 0.035 0.017 0.316 0.078
1/8 2 1.103 0.666 0.510 0.426 0.373 0.275 0.046 0.579 0.171 0.061 0.024 0.011 0.300 0.060
1/16 1/8 0.888 0.577 0.461 0.396 0.353 0.594 0.294 0.337 0.242 0.195 0.163 0.139 0.267 0.179
1/16 1/4 1.044 0.652 0.507 0.428 0.376 0.634 0.270 0.396 0.254 0.183 0.138 0.106 0.296 0.173
1/16 1/2 1.117 0.680 0.522 0.436 0.382 0.564 0.193 0.447 0.238 0.144 0.091 0.059 0.312 0.145
1/16 1 1.138 0.685 0.523 0.436 0.381 0.431 0.108 0.480 0.193 0.090 0.044 0.022 0.307 0.102
1/16 3/2 1.138 0.684 0.522 0.435 0.380 0.345 0.069 0.488 0.160 0.061 0.025 0.011 0.294 0.076
1/16 2 1.138 0.683 0.521 0.435 0.379 0.288 0.048 0.489 0.136 0.045 0.016 0.006 0.281 0.059
0 See innitely long longitudinal and circumferential defects (Tables 57)
S
.
M
a
r
i
e
e
t
a
l
.
/
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
P
r
e
s
s
u
r
e
V
e
s
s
e
l
s
a
n
d
P
i
p
i
n
g
8
2
(
2
0
0
5
)
7
4
6

7
6
0
7
5
4
P
iC1
Z

iC1
jZ0
DP
j
(13)
Using the solution corresponding to a linear shock, it is
then possible to propose a simple analytical solution for an
homogenous vessel, using the same constants as for the
linear thermal shock (Biot number B, etc.):
T ~ x;
~
t Z
T
i
c
~
t%0
T
fluid

~
t C

N
nZ0
T
n
cosg
n
~ xf
n

~
t c
~
tO0
_

_
(14)
with
f
n

~
t
i
ZK

i
jZ1
DP
jK1
g
2
n
1Ke
Kg
2
n
~ t
i
K~ t
jK1

(15)
with g
n
dened by the solution of
g
n
tan g
n
ZB; g
n
20; p=2 Cnp (16)
T
n
Z
2sin g
n
g
n
Csin g
n
$cos g
n
(17)
However, this solution assumes independence of the
thermal parameters (conduction, diffusivity and coefcient
of heat exchange) of temperature in the component.
To take into account the cladding, a linear temperature
distribution in the cladding is assumed. This hypothesis is
acceptable if the thickness of the cladding is small compared
to that of the vessel itself. An equivalent Biot B
0
number was
determined making it possible to apply the solution for a
uniform vessel to a vessel with cladding.
in the vessel:
T
2
~ x;
~
t Z
T
i
c
~
t%0
T
fluid

~
t C

N
nZ0
T
n
cosg
n
~ xf
n

~
t c
~
tO0
_

_
(18)
in the cladding:
T
1
~ x;
~
t
Z
T
i
c~ t%0
T
fluid

~
t C

N
nZ0
T
n
cos g
n
K
l
2
l
1
~ xK1g
n
sing
n
_
_
_
_
f
n

~
t c
~
t%0
_

_
(19)
with
f
n

~
t
i
ZK

i
jZ1
DP
jK1
g
2
n
1Ke
Kg
2
n
~ t
i
K~ t
jK1

(20)
g
n
tan g
n
ZB

(21)
B

Z
B
1CB
l
2
l
1
~ x
1
K1
(22)
Table 5
Shape factors (i
0
to i
4
) for innitely circumferential through clad defect
(E
1
/E
2
Z0.7 and 1)
aCr=hCr i
0
i
1
i
2
i
3
i
4
0.00 1.122 0.683 0.526 0.441 0.378
0.10 1.155 0.694 0.530 0.443 0.387
0.20 1.255 0.735 0.554 0.459 0.399
0.30 1.400 0.791 0.586 0.480 0.415
0.40 1.583 0.861 0.625 0.506 0.433
0.50 1.804 0.946 0.672 0.537 0.455
0.60 2.060 1.046 0.726 0.572 0.481
0.70 2.352 1.158 0.791 0.615 0.512
0.80 2.700 1.307 0.880 0.676 0.557
Table 6
Shape factors (i
0
to i
4
) for innitely long longitudinal through clad defect
(E
1
/E
2
Z0.7 and 1)
aCr=hCr i
0
i
1
i
2
i
3
i
4
0.00 1.122 0.683 0.526 0.441 0.387
0.10 1.176 0.702 0.535 0.446 0.390
0.20 1.338 0.767 0.572 0.471 0.408
0.30 1.592 0.865 0.627 0.507 0.434
0.40 1.959 1.004 0.704 0.558 0.470
0.50 2.481 1.197 0.810 0.626 0.519
0.60 3.222 1.467 0.955 0.719 0.584
0.70 4.253 1.837 1.152 0.844 0.672
0.80 5.535 2.297 1.397 0.999 0.781
Table 7
Shape factors (i
0r
and i
1r
) for innitely long longitudinal and circumfer-
ential through clad defect (E
1
/E
2
Z0.7 and 1)
r=aCr i
0r
i
1r
0.000 0.000 0.000
0.050 0.056 0.002
0.100 0.107 0.006
0.150 0.154 0.012
0.200 0.197 0.021
0.250 0.238 0.031
0.300 0.276 0.042
0.350 0.313 0.055
0.400 0.350 0.070
0.450 0.387 0.087
0.500 0.425 0.106
0.550 0.464 0.128
0.600 0.505 0.153
0.650 0.550 0.181
0.700 0.598 0.214
0.750 0.651 0.252
0.800 0.710 0.296
0.850 0.774 0.348
0.900 0.845 0.409
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 755
The thermal stress eld through the thickness [2] related to the
temperature variation is equal to:
s
th
r
x Z
Ex
1Kn
1
x
2
x
2
KR
2
i
R
2
e
KR
2
i
_
R
e
R
i
axTxx$dxK
_
x
R
i
ax$Tx$x$dx
_

_
_

_
(23)
s
th
q
x Z
Ex
1Kn
1
x
2
x
2
CR
2
i
R
2
e
KR
2
i
_
R
e
R
i
ax$Tx$x$dx
_

_
C
_
x
R
i
ax$Tx$x$dxKax$Tx$x
2
_

_
(24)
s
th
z
x Z
Ex
1Kn
2
R
2
e
KR
2
i
$
_
R
e
R
i
axTxxdxKaxTx
_

_
_

_ (25)
where x is the radial position (xZR
i
Cu), a(x) the thermal
expansion coefcient eld, and T(x,t) the evolution of the
temperature eld during the transient: this evolution is given in
detail in [2] as a function of the geometrical and thermal
parameters of the problem.
Since the defect is small aCr=hCr%0:1, the term
aCr=hCr
j
rapidly becomes negligible in the K
I
formulae, and it is only necessary to know the stress
distribution in the neighbourhood of the interface of the two
materials (only the stress distribution on the crack length is
needed). In addition, only the longitudinal stress is
considered s
th
z
: Due to the large radius of curvature of the
vessel, the circumferential stress is very near to
s
th
z
zs
th
q
Zs
th
. As the defect is small in comparison
with the vessel thickness, the Taylor series of the function
sth(x,t) about the interface (xZR
i
Cr) up to order 1 is
accurate enough, and will be suitable to represent the
opening stress for circumferential and longitudinal defects:
in the ferritic material, the development gives the
following result:
Fluid Tfluid
H
Outer surface
heat insulated
Vessel 2
Conductivity
Diffusivity d2
Thickness h2
Young Modulus E2
Temp. ) t
~
, x
~
( T
2
Cladding
Conductivity 1 2
Diffusivity d1
Thickness h1
Young Modulus E1
Temp. ) t
~
, x
~
( T
1
1 x
~
= 0 x
~
=
1
x
~
x
~
=
r
z
1
Fig. 5. Diagram of problem covered with allowance for cladding.
T
0
T
0
t
0
t
1
t
2
t
0 t
1
t
2
t
0 t
1
t
2
t
0
t
1
t
2
t
0
t
1
t
2
T
2
T
1
Fluid temperature
Time
=
Time
Time
Time
Time
P
0
P
1
P
2
+
+
+
Fig. 6. Breakdown of uid temperature into a succession of linear shocks.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 756
s
th
0
t Z
E
2
$a
2
1Kn
1K
h
2$R
m
_ _
S
1
KS
2
C
h
R
m
S
3
_
C
r
h
a
1
a
2
K1
_ _
T
f
tKT
i
KS
1
C
a
1
a
2
S
2
CS
4
_ __
(26)
s
th
1
t ZK
E
2
a
2
1Kn
S
4
(27)
Then
s
th
u; t Zs
th
0
t Cs
th
1
t
u
h Cr
: (28)
in the austenitic cladding:
s
th
0r
t Z
E
1
E
2
K1
_ _
s
th
0
t C
E
1
a
1
1Kn
!
a
2
a
1
K1
_ _
T
f
tKT
i
CS
2
C
r
h
l
2
l
1
K
a
2
a
1
_ _
S
4
_ _
(29)
s
th
1r
t Z
E
1
E
2
a
1
l
2
a
2
l
1
K1
_ _
s
th
1
t (30)
s
th
u; t Zs
th
0
t Cs
th
0r
t Cs
th
1
t Cs
th
1r
t
u
h Cr
(31)
with
S
1
Z

N
nZ0
T
n
$f
n
sing
n

g
n
(32)
S
2
Z

N
nZ0
T
n
$f
n
cosg
n
(33)
S
3
Z

N
nZ0
T
n
$f
n
1Kcosg
n

g
2
n
(34)
S
4
Z

N
nZ0
T
n
$f
n
$g
n
sing
n
(35)
where T
n
, f
n
and g
n
are time dependent functions dened for
the temperature distribution (Eqs. (17), (20) and (21)). The
additional linear stress component in the cladding becomes
signicant as soon as the cladding and the vessel show
different material properties (E, l and a), as it is the case for
the cladded RPV. For the study case presented in Section
3.2, which is very representative of the real conguration,
the additional constant stress s
th
0r
is about 30% of the
constant stress in the vessel s
th
0
when the elastic stress
intensity factor at the deepest point of the defect reaches its
maximum value.
3.2. Internal pressure stresses eld
For internal pressure, an analytical solution is developed
for the cladded vessel: the problem is axisymmetric
(geometry and loading) and is solved using the displacement
method. Considering the different symmetries of the
problem, the displacement eld could be represented as
follow:
x
j
ZxF
j
x

e
r
CG
j
z

e
z
(36)
The resolution of the equilibrium equation leads to:
F
j
x Za
j
C
b
j
x
2
G
j
z Zc
j
z
_

_
(37)
which corresponds to a stress eld given by:
s
1
r
ZA
1
K
B
1
x
2
; s
1
q
ZA
1
C
B
1
x
2
; s
1
z
ZC
1
s
2
r
ZA
2
K
B
2
x
2
; s
2
q
ZA
2
C
B
2
x
2
; s
2
z
ZC
2
_

_
(38)
where A
1
, A
2
, B
1
, B
2
, C
1
and C
2
are constants to be
determined. To do it, following boundary conditions are
considered:
inner surface is subjected to internal pressure p:
s
1
r
x ZR
i
ZKp5A
1
K
B
1
R
2
i
ZKp (39)
outer surface is free:
s
2
r
x ZR
e
Z05A
2
K
B
2
R
2
e
Z0 (40)
radial stress continuity at the interface between the cladding
and the ferritic vessel:
s
1
r
x ZR
0
Zs
2
r
x ZR
0
5A
1
K
B
1
R
2
0
ZA
2
K
B
2
R
2
0
(41)
radial strain continuity at the interface between the cladding
and the ferritic vessel:
3
1
r
x ZR
0
Z3
2
r
x ZR
0
5
1Kn
E
1
A
1
C
1Kn
E
1
B
1
R
2
0
Z
1Kn
E
2
A
2
C
1Kn
E
2
B
2
R
2
0
(42)
axial strain is constant:
3
1
z
Z3
2
z
5
C
1
E
1
Z
C
2
E
2
(43)
average axial stress is related to the pressure:
s
moy
z
Z
r$s
1
z
Ch$s
2
z
h Cr
Z
p$R
2
i
R
2
e
KR
2
i
5
r$C
1
Ch$C
2
h Cr
Z
p$R
2
i
R
2
e
KR
2
i
(44)
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 757
Solving this system of equations leads to the denition of
two coefcients k
L
and k
C
:
These two parameters allow a simple description of the
solution: k
L
and k
c
are respectively related to the
circumferential stress s
q
and the longitudinal stress s
z
. For
a RPV case with small cladding thickness, k
L
and k
C
can
reasonably be approximated to 1. As for the thermal stress
eld, the Taylor series of the function s
p
(x,t) about the
interface (xZR
i
Cr) up to order 1 is accurate enough.
It gives:
in the ferritic material:
s
p
u; t Zs
p
0
t Cs
p
1
t
u
h Cr
(46)
in the austenitic cladding:
s
p
u; t Zs
p
0
t Cs
p
0r
t Cs
p
1
t Cs
p
1r
t
u
h Cr
(47)
Table 8 gives expression for s
p
0
, s
p
1
, s
p
0r
, and s
p
1r
, for each
type and orientation of the defect. For the through clad
defect, the pressure on the crack faces is taken into account
in the s
p
0
stress expression.
4. Application to a case study
We give in this section an example using the proposed
compendium to illustrate the use of such development: the
solution obtained with the complete analytical method is
compared to a complete elastic 3D FE calculation with
CAST3M.
4.1. Presentation of the case study
The RPV loading is a small break loss of coolant accident
(SBLOCA) during normal operation. This kind of accident
is related to the fracture of the primary loop. The loss of
coolant leads to an increase of the temperature of the vessel
and then possibly core fusion. To avoid this scenario,
emergency cooling of the core is performed by injecting
cold water, so the temperature and the internal pressure fall
quickly. The aim of this study is to check the integrity of the
vessel during this thermal shock, considering a defect at the
inner skin of the vessel. For simplicity, the thermal shock is
assumed to be axisymmetric. All the data given in this
example were proposed in the SMILE project [7] on the
warm pre-stressing effect, for benchmarking on a realistic
conguration.
4.1.1. Flaw type and geometry
In this study, an axial through clad defect is assumed to
be present in the wall of a PWR RPV. The cylinder
geometry of the RPV and the dimensions of the postulated
crack are given in Table 9.
4.1.2. Material properties
All relevant temperature dependent material properties
are given in Table 10 for the base material (ferritic steel),
Table 8
Normalised stresses of the linear distribution for pressure
Defect Longitudinal defect Circumferential defect
ThroughKclad defect underKclad defect throughKclad defect underKclad defect
s
p
0
t=pt 2R
2
e
=R
2
e
KR
2
i
R
2
e
CR
2
i
=R
2
e
KR
2
i
R
2
e
=R
2
e
KR
2
i
R
2
i
=R
2
e
KR
2
i

s
p
1
t=pt K2R
2
e
=R
2
e
KR
2
i
hCr=R
i
K2R
2
e
=R
2
e
KR
2
i
hCr=R
i
0 0
s
p
0r
t=pt 0 0 R
2
i
=R
2
e
KR
2
i
E
1
=E
2
K1 R
2
i
=R
2
e
KR
2
i
E
1
=E
2
K1
s
p
1r
t=pt 0 0 0 0
Table 9
Geometry of the PWR RPV and of the postulated crack
Internal diameter of the vessel 2.R
i
5 000 mm
Wall thickness h 250 mm
Cladding thickness r 6 mm
Crack type Through clad surface crack
(surface breaking aw)
Crack depth aCr 16 mm
Crack length 2.c 96 mm
Table 10
Physical material properties of base material (ferritic steel)
Temperature T
in 8C
20 100 200 300 400
Modulus of
elasticity E
in MPa
211000 206000 199000 192000 184000
Poissons ratio v 0.3 0.3 0.3 0.3 0.3
Thermal
conductivity l in
W m
K1
K
K1
44 44 43 41 39
True heat capacity
C
p
in J.kg
K1
.K
K1
460 490 520 560 610
Thermal
expansion
coefcient
between 20 8C and
T in 10
K6
K
K1
11.7 12.7 13.2 13.6 14
k
L
Z
2R
2
0
R
2
e
KR
2
i

R
2
e
KR
2
0
1KnR
2
0
C1 CnR
2
i
C
E
1
E
2
R
2
0
KR
2
i
1KnR
2
0
C1 CnR
2
e

; k
C
Z
1 C
r
h
1 C
r
h
E
1
E
2
(45)
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 758
and in Table 11 for the austenitic cladding: the properties
are representative of an RPV at beginning of life.
4.1.3. Loading conditions
The evolution during the transient of the uid tempera-
ture T
f
(t), the heat transfer coefcient H(t) and the absolute
internal pressure p(t) are given in Table 12. For the
analytical solution an average value of H(t) is considered:
H
average
Z3000 W m
K2
K
K1
. For the 3D FE calculation, the
variation of H with temperature is taken into account.
4.2. Interpolation in the compendium
As recommended, to use this compendium in this
particular conguration, a linear interpolation in the a/r,
a/c and E
1
/E
2
dimensions was used:
at the deepest point (A): i
0
Z1,029; i
1
Z0,634; i
0r
Z0,
279; i
1r
Z0,053
at the interface point (C): i
0
Z0,633; i
1
Z0,214; i
0r
Z0,
326; i
1r
Z0,074
Only linear terms are determined because stress
developments were made up to order 1.
4.3. K
I
calculation
Since we only used descriptions of the elds up to order
1, the formula to determine K
I
becomes:
K
I
Z s
0
i
0
Cs
0r
i
0r
Cs
1
i
1
Cs
1r
i
1r

a Cr
h Cr
_ _

pa Cr
_
(48)
where s
j
Zs
th
j
tCs
p
j
t for each j.
4.4. Comparison with the complete 3D FE calculation
To validate the results obtained with the simplied
method, a complete 3D FE thermo mechanical calculation
was performed on a RPV mesh similar to the one presented
in Fig. 1. K
I
was derived from the J-integral.
The K
I
evolutions versus time given by the two methods
for the two characteristic points of the defect are presented
in Fig. 7. The maximum of the stress intensity factor
evolution is the most important value in terms of component
safety: the maximum is reached at the deepest point. The
simplied method gives a value only 1. 0% higher than
the complete 3D FE calculation. At the interface point, the
difference is greater: it reaches 2.5%, but this value is less
important for the safety of the component. Also, the
beginning of the thermal shock, up to the maximum values
of K
I
, is well represented by the analytical method.
These differences are very small in comparison to the
difculty of the problem, and are mainly due to the
consideration of an average heat transfer coefcient and to
the interpolation of the shape factors in the compendium.
This result shows that it is now possible to obtain a
reasonably accurate estimation of K
I
by an analytical
analysis of the Pressurised Thermal Shock on the RPV.
Table 11
Physical material properties of the austenitic cladding
Temperature T
in 8C
20 100 200 300 400
Modulus of
elasticity E in MPa
200000 194000 186000 179000 172000
Poissons ratio n 0.3 0.3 0.3 0.3 0.3
Thermal
conductivity l in
W.m
K1
.K
K1
15 16 17 19 20
True heat capacity
C
p
in J kg
K1
.K
K1
500 500 520 540 590
Thermal
expansion
coefcient
between 20 8C and
T in 10
K6
K
K1
15 16 17 17 18
Table 12
Loading conditions of the transient
Time t (s) Pressure p(t)
(bar)
Temperature T
f
(t)
(8C)
H (W m
K2
K
K1
)
0 158 298 20000
250 78 272 10000
500 73 140 8000
800 65 92 5000
1200 53 79 3500
1450 48 70 3300
1800 41 30 2200
1950 22 45 2700
2750 15 15 1500
7000 15 15 900
0
20
40
60
80
100
120
140
4000 2000 0 6000 8000
Time (s)
K
I

(
M
P
a
.
m
0
.
5
)
KI(A) - F.E. calculation
KI(C) - F.E. calculation
KI(A) - analytical solution
KI(C) - analytical solution
Fig. 7. Application of the analytical method to predict the K
I
variation of a
through clad defect during a LOCA thermal shock(A) is the deepest point
and (C) is the surface point.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 759
This constitutes a major step for a sensitivity study or
probabilistic approach to the problem.
5. Conclusions
In this paper, we have presented a stress formulation and
the associated compendium of stress intensity factors for
defects in a RPV wall in the presence of cladding. This work
was performed because existing bibliographical sources did
not cover these crack congurations.
A complete set of nite element calculations was
conducted by CEA. A parametric mesh generator was
used to analyse 72 2D and 3D congurations for each type
of defect. The shape factors i
0
to i
4
for under clad defects are
given in Tables 1 and 2, and i
0
to i
4
, i
0r
and i
1r
for through
clad defects are given in Tables 3 and 4 for a wide range of
geometrical and material parameters:
a=c21; 1=2; 1=4; 1=8; 1=16; 0
a=r 21=8; 1=4; 1=2; 1; 3=2; 2
E
1
=E
2
20:7; 1
To complement this numerical work, an analytical
solution of the stress distribution in the wall thickness
under PTS is proposed. This solution is a continuation of the
thermal solution proposed in [2], which allows a complete
analytical determination of K
I
for the PTS problem.
The use of this solution associated to the compendium
was illustrated on a complex case study by comparison with
a complete 3D FE calculation. The difference for the
maximum of K
I
reached during the thermal transient is only
1.0%, which constitutes a very good result for an analytical
application to this complex problem.
References
[1] RCC-MR 2002-Annexe A16, 2002,Tome I: vol. Z, AFCEN, Paris.
[2] Marie S. Analytical expression of the thermal stresses in a vessel or a
pipe with cladding submitted to any thermal transient. Int J Press Vess
Piping 2004;81:30312.
[3] F.E. software cast3M, 2004, www.cast3m.cea.fr.
[4] Suo XZ, Combescure A. On the application of the G(q) method and its
comparison with De Lorenzis approach. Nucl Eng Des 1992;135:
20724.
[5] Chapuliot S, Formulaire de K
I
pour les tubes comportant un defaut de
surface semi-elliptique longitudinal ou circonferentiel, interne ou
externe. CEA report. CEA-R-5900, ISSN N80429-3460.
[6] Le Delliou P, Barthelet B. Inuence coefcients to calculate stress
intensity factors for an elliptical crack in a plate. Proc Press Vess Piping
Conf, Vancouver 2002.
[7] Moinereau D, Bedzdikian G. SMILE: a european R&D program for the
validation of the warm pre-stress effect in a reactor pressure vessel
structural integrity assessment. Proc 11th Int Conf Nucl Eng, Tokyo;
2003.
S. Marie et al. / International Journal of Pressure Vessels and Piping 82 (2005) 746760 760