Anda di halaman 1dari 17

Proceedings of the ASME 2013 Pressure Vessels & Piping Division Conference

PVP2013
July 14-18, 2013, Paris, France

PVP2013-97622
STRESS ANALYSIS OF PIPE SUPPORT ATTACHMENTS: A COMPARISON OF
ANALYTICAL METHODS AND FINITE ELEMENT ANALYSIS FOR CIRCULAR AND
NON-CIRCULAR ATTACHMENTS
Anindya Bhattacharya
Technical Head, Stress Analysis
CB&I, 40 East Bourne Terrace, London, W2 6LG,
United Kingdom.
Phone: +442070535668

ABSTRACT
4. Comparison of FEA, WRC 107, WRC297 and Kellogg
methods w.r.t the following parameters:
Type of loading (Radial, Longitudinal, Circumferential)
applied in a stand-alone manner is absence of pressure
D , t and d ratios
T
T
D
Combined loading including pressure
Different element types

Despite the availability of special purpose FE codes with


post processing facilities as per rules of ASME SEC VIII
Division 2, use of simple analytical methods like ring loading
around a circumference or more complex methods like
Welding Research council bulletins 107 and 297, will continue
to be used in the industry for a significant period of time for
stress analysis of pipe support attachments. The reasons are
few: not all engineering companies have such custom made
FE codes, lack of trained personnel to work with general
purpose FE codes, ease of implementation of the available
methods and their successful design history, cost and time
issues with FE analysis etc. In this paper these available
methods will be reviewed based on their theoretical
background, their range of applicability w.r.t the typical design
parameters and their comparison with FE analysis. More
recent analytical methods based on mathematically accurate
space curves of intersections for circular attachments will also
be discussed. This study will include both circular as well as
non-circular attachments. This paper will highlight the
strengths and weaknesses of the conventionally used methods
especially with respect to their mathematical limitations to
make an analyst aware of the potential over conservatism and
under conservatism of these analytical methods. Finite
element analysis models will be discussed in detail
specifically in relation to elements used, element parameters,
boundary conditions and post processing.

NOMENCLATURE
, - components in the ( , ) coordinate of the main
shell
r,R - mid surface radii of the branch pipe and main shell
E, - Youngs modulus and Poisson ratio respectively
un - radial displacement

0 - diameter ratio = d
T

p
, ,z

D
thickness of main shell
Airy stress function
internal pressure
global cylindrical coordinates in 3D space

p , p - surface force components in the

directions

w - vertical displacement

ET 3
12( 1 2 )
loading in vertical direction
thickness of attached shell
ET
foundations stiffness = 2
R
direction of longitudinal axis of cylinder
section modulus of the trunnion pipe
area of cross section of trunnion pipe

H - flexural rigidity of shell =

INTRODUCTION

Z t -

In this paper, the subject matter has been structured in the


following manner:
1. Discussion of the available theoretical methods, from the
simplest to the advanced.
2. Brief overview of basic shell mathematical model.
3. Brief overview of available finite element options.

K x S A -

Copyright 2013 by ASME

2. AVAILABLE THEORETICAL APPROACHES TO


THE PROBLEM OF ANALYSING A CYLINDRICAL
SHELL WITH CYLINDRICAL OR NON-CYLINDRICAL
ATTACHMENTS.

M - resultant applied bending moment on trunnion


P - load per unit of circumference (applied as a ring
load)
rt - radius of trunnion
F - force on trunnion
1 , 2 - attachment parameters for rectangular attachments =
c1
c
& 2
R
R
c1 , c2 - half dimensions of the rectangle along
circumferential
and
meridional
directions
respectively
2 - Laplacian Operator, . ( )

2.1 Approach 1:
This approach is popularly known as Kellogg method in
the piping industry. This approach has been so named as it
appeared for the first time in [4] and is based on ring loading
around a circular cylinder.
Governing differential equation [1]:
For an axi-symmetric loading on a circular cylinder, the
governing differential equation is the well known beam on
elastic foundation equation:
d 4w
(1)
H 4 + Kw = Z
dx
d 4 w ETw Z
(2)
+
=
dx 4 HR 2 H
1.28
3( 1 2 )
Introducing 4 =
, i.e. =
considering
2 2
RT
RT
= 0.3 .

1. SHELL THEORIES:
There are various shell theories and each one has its own
protagonist. Any shell theory has to be evaluated within the
postulates of Sanders-Koiters approach [12, 21] which can be
summarised as follows:
1. The equations can be written in general tensor form.
2. The deformations are described by six strain measures,
three of which are components of the usual membrane strain
tensor and the other three deviate from the components of the
geometrical curvature change tensor only by terms that are
bilinear in the components of the curvature and membrane
strain tensor.
3. The stresses are described by six stress measures that satisfy
the equations of equilibrium without approximation.
4. The theory has a principle of virtual work that is exact for
displacements obeying the Kirchoff hypothesis; in conjunction
with approximate constitutive relations between the stress and
strain measures. Well-set boundary value problems can be
formulated, and the usual minimum and reciprocal relations of
structural mechanics hold good.
5. The theory contains an exact static-geometric analogy. This
analogy can be formulated by replacing the static quantities by
corresponding geometrical quantities in homogeneous
equations of equilibrium and the resulting equations become
identical with the compatibility conditions.
6. When applied to the symmetrical bending of shells of
revolution, the stress and strain measures agree with those
generally used. They are consistent with those of the most
simple curved beam theory.
For the present purpose, we will discuss the issue of
cylindrical pipes with circular (referred to as trunnion) as well
as non-circular (referred to as pipe shoes) attachments. Hence
there is no puncture in the header pipe. The mathematical
problem of the main shell with cut-out is a boundary value
problem of partial differential equation. It means that the
cylindrical shell equation, whose general solutions have many
unknown constants, is suitable on the shell surface with or
without cut-out. In order to determine the unknown constants
the boundary conditions have to be used.

we therefore get

d 4w
Z
+ 4 4 w =
4
H
dx

(3)

The solution of this differential equation and boundary


conditions are detailed in [1]
Extending the above analysis to a case of bending of a
cylindrical shell by a load uniformly distributed along a
circular section [1], we get:
P
Maximum Bending Moment =
, where P= load per unit
4
length of circumference.
Bending stress, bending =

1.17 P R
1.5

(4)

P, can be defined in terms of a local radial load, Pr and local


moment, M r . This is necessary because P is a line load
distributed around the circumference of the shell.
If a load Pr is divided by the attachment perimeter it becomes
Pr
for a nozzle of radius, rt . or a moment over section
2rt
Mr
modulus of the attachment becomes,
.
rt2
Flexural stresses are added to membrane longitudinal and
hoop stresses to get total stress = membrane stress in direction
i + flexural stresses in direction i computed by the expression
in eq-(4)

Copyright 2013 by ASME

To compute P, these were the steps followed:


Computation of loads in longitudinal and circumferential
directions by use of the following expressions:
longitudinal force = (longitudinal force x moment
arm)/ rt2
circumferential force = (circumferential force x moment
arm)/ rt2
radial force = radial load/ 2rt
equivalent circumferential force = 2 x circ. force + 1.5 x
radial force
equivalent longitudinal force = 1.5 x radial force +
longitudinal force
The above forces are used as P in eq-(4)
The reason behind the use of the factors 1.5 and 2.0 is
attributed to higher flexibilities in these directions.
The flexural stresses in longitudinal and circumferential
directions are then computed using these equivalent forces
and the membrane pressure stresses are then added to compute
the total stresses. Stress in the trunnion attachment is
F M
+
computed as
.
A S
2.1.1. The case of Pipe shoes:
Schematic arrangements for some pipe shoes are shown in
fig-(1). Dimension B stands for shoe/gusset width, G =
number of gussets, L = gusset spacing (this depends on the
design), S = number of spines, M = spine spacing, and A=
shoe length.
The approach taken for analysis of pipe shoes is similar to
that of trunnion type attachments. The computations of section
properties (few examples) are cited.
1. Pipe shoe with no gusset:
3
longitudinal moment of inertia = A

12
distance to centroid, longitudinally = A
moment of inertia, circumferential = A
distance to centroid = B

12

2. Pipe shoe with 4 gussets:


3
longitudinal moment of inertia = A

distance to centroid = A

12

+ 5 BL2
9

3
moment of inertia, circumferential = B

distance to centroid = B

Copyright 2013 by ASME

Fig-(2) [31, 25]


Bijlaard applied the radial force system, qn instead of
vertical force system, qz . When subjected to a radial force
system the resultants not only include moments, M xb or M yb
but also force, Fyb . This will be statically equivalent to
external load Z involving force, Fzb , transverse bending
moment, M xb and longitudinal bending moment, M yb . The

Fig-(1). Schematic design of some common pipe shoe types.


3. Saddle:
3
moment of inertia, longitudinal = A

distance to centroid = A

+ BL

12

3
moment of inertia, circumferential = B

distance to centroid = B

figure below shows the transverse bending moment, M xb case.


In fig-3(a), the linearly distributed force system, qz is
statically equivalent to the transverse bending moment, M xb
but in fig-3(b), the linearly distributed force system, qn is

+ AB

statically equivalent to the transverse bending moment, M xb


and force, Fy as in fig-3(d).

2.2 Approach 2: The WRC-107 approach based on


the work of Bijlaard [34, 26, 2]
Bijlaard derived a theoretical solution based on
Timoshenko equations [1, 26] for a cylindrical shell on end
supports under a force system, qn linearly distributed over a
square region defined by c , c , where c =

in
2
the developed surface. In deriving the equations it has been
assumed that 0 = 0 (circumferential strain).
The force/moment system is shown in fig-(2) below

Copyright 2013 by ASME

Fig-(3) [25]

qz = qn cos = qn 0
M x = 2c

cos , q y = qn sin = qn 0
sin
0
0

q z Rc sin d = 2c 2 R

qn 0

Circumferential Moment ( taken from A.3.3.2 and Table


A-4 of WRC 107)
For the thin walled model, the measured circumferential
and longitudinal stresses were both higher than the computed
values. Modifications were then done to Bijlaard's original
work for both longitudinal and circumferential stresses for the
bending components (and for the circumferential stress, for the
membrane component also) but for longitudinal stresses there
was minimal requirement for correction of the membrane
component. Correction factor used was around 2.7 for bending
component of circumferential stress + 20-25% for membrane
component and correction factor of 2.72 was used for bending
component of longitudinal stress (no correction for membrane
component). It is stated in WRC-107 that the modified curves
may be more conservative than the original work.
Longitudinal Moment ( taken from A.3.3.2 and Table A-4
of WRC 107)
For the thin walled model, the measured circumferential
and longitudinal stresses were both higher than the computed
values. Corrections were made to both the Membrane and
Bending components. For the Longitudinal stress, no
correction was required for the Membrane component. Higher
modification was required for membrane component (30%)
compared to bending component (18%) for circumferential
stress. Correction factor used was 6.75 in bending component
of longitudinal stress (no correction for membrane component)
Radial Load ( taken from A.3.3.4 and Table A-6 of WRC107)
Results agreed well on the transverse axis but the
theoretical results were conservative by factors as high as 2.0
on the longitudinal axis.

(5)

cos sin d
0

(6)

Fy = 2c

q y c d = 2c 2

qn 0

sin d
0

(7)

Bijlaard used Double Fourier Series to represent loads and


displacements.
The limitations of Bijlaards approach are the d 0.3
D
and the usage of Double Fourier series method which may not
show convergence for certain boundary conditions [32, 26].
The limitation of d is due to the use of radial force instead
D
of vertical force. This results in significant error outside
aforementioned d
limit. The stresses are computed in 8
D
specific locations around the intersection. The maximum
stresses need not be at these locations! Additionally stresses in
Trunnion/Shoes cannot be computed. For rectangular
attachments, the limitation is 1 & 2 < 0.5 .
It is to be noted that WRC-107 is not only based on
Bijlaards theoretical work but also experimental works by
Mehrson, Wichman and Hopper [34]. The bulletin shows a
comparison of the calculated and measured stresses for both
thick walled models and thin walled models. Following were
the main issues between the Experimental works and
Bijlaards work for thin shells.

Copyright 2013 by ASME

on the longitudinal axis. Results for transverse axis agreed


well for cases that are restricted by

dm
Dm

Based on this line of reasoning in-order to extend the


applicability of the thin shell theoretical solution for
cylindrical shells with cut-out, Xue et al adopted Morleys
equation [9], which has the same order of magnitude of
accuracy as the general thin shell theory, i.e. O T , instead
R
of Donnells. Morleys equation is expressed in complexvalued form by Simmonds [11] as follows:

Dm
2
T

( )

2.3 Approach 3: Post WRC-107 approaches WRC297 and works of Morley, Simmonds and Hwang et
al. [5, 33, 11, 16, 25]
Theory of thin elastic shells, in which T/R<<1 is
insignificant in magnitude is derived on the basis of LoveKirchhoff assumptions. A generally accepted fact is this
approach has an error of order of magnitude O T .
R
When a solution is derived by omitting some terms, which
has order of magnitude larger than O T
(such as
R

2
4
2
2
+ 4 i 2

( )

( )

(omitting terms of order of magnitude O

( )

= 0

R
2
4 = 12 (1 2 )
T
4 2
= un + i

ETR

( )

( )

) from
R
Flgges equation [13]. This equation is quite simple and can
be expressed in complex-valued displacement-stress function
form (Lekkerkerker [15] and Steele [3]) as follows:
2
4
2

(11)

where, and are the same as in eq-(10) and (9). The right
hand side of eq-(11) is a load function dependent on the
surface force components acting on the shell.
The cylindrical thin shell equations derived by
Goldenveizer, Morley, Simmonds and Timoshenko (which
was used by Bijlaard) have the same inherent error in order of
magnitude O T . The solution has the order of accuracy
R
O T . WRC-297, which is based on Steeles work on
R
shallow shell equations covers a range of only
r sin < 0.5 .
R
6
For detailed analysis of the approach taken by Xue et al
refer [17, 31].
In essence, the approach taken is to use compatibility
conditions enforced on the geometrically correct curve of
intersection as opposed to an assumed curve of intersection
and using theories which are of order O T
which may or
R
may not involve using different shell theories for intersecting
cylinders.
To summarize, different cylindrical shell equations are
suitable to different ranges of the developed surface. Fig-(4)
below shows the different ranges of developed surface [25].

for shallow shell equations), the accuracy of the


R
solution is bound to be lower. The detailed analysis of the
above-mentioned concept can be found in well-known
literature and textbooks of thin shell theory [6].
The exact equations for thin elastic cylindrical shells are
very complicated. For a problem of cylindrical shell with cutout [25], Donnell [8] presented an approximate equation
O

= P ( p , p , p )

( )

(8)
(9)
(10)

where, un , is radial displacement and , Airy stress function.


Eq-(8) can be decomposed into two second-order partial
differential equations and is easy to solve in polar coordinate
system for the problem of cylindrical shell with cut-out [25].
However, as pointed out by Koiter, eq-(8) can only be applied
for shallow shells. Koiter [7] had written, It has been noted
[9,14] that Donnells approximation is sometimes inaccurate
and the generalization of Donnells approximation is
applicable in the case of shallow shells in which the wave
length L of the deformation pattern on the middle surface is
always small compared with the minimum principal radius of
curvature R. Based on fig-6.14 in Donnells book [10], the
applicable range of shallow shell equation for the problem of
cylindrical shell with opening is only r sin = 0.5 .
R
6
The edge effects of general cylindrical shells and shallow
shells mathematically differ.

( )

Copyright 2013 by ASME

methods like FEM as currently an engineer in an industry does


not have an easy tool to compare the FE results against some
published benchmarks for d >0.5. In other words, as long as
D
we do not have analytical tools which are easily
implementable and which will address the problems to be
analyzed without having significant restrictions on geometry
and loading conditions, FE analysis should be the preferred
tool for analysis. The objective of this paper is to make an
analyst aware of the potential over conservatism and under
conservatism in the available and widely used methods if an
analyst is constrained to use them.
4.0 FINITE ELEMENT ANALYSIS APPROACH TO THE
PROBLEM
[33] provides an excellent discussion on the issues
involving conflicts between shell theory and finite element
analysis of shells. To briefly summarize them:

Fig-(4) [25]
Donnel [10] showed that his shallow shell equations could
be suitable to the range of < < i.e. 0 < 0.5 . In
6
6
[15] Lekkerkerker showed that the shallow shell equations
could be applied to the range of 0 0.25 . The different
applicable ranges adopted by different authors are dependant
on different allowable intrinsic errors.

Ill conditioning due to significantly different strain


energies between membrane and bending modes.
Use of low degree polynomial trial functions in the
displacement finite element method generally leads to
overstiffness in the response to bending actions.
Difficulty in deriving trial functions for in-extensional
bending.

Many authors [33, 18, 30] have recommended use of


hybrid elements. In this paper, however, we have used only
displacement based finite element method.
Finite elements available for shell analysis can be broadly
classified into the following groups:

3.0 DIFFICULTIES IN IMPLEMENTATION OF


ANALYTICAL SOLUTIONS:

1.
2.
3.

Analytical solutions (rather analytical solutions backed by


experimental findings like WRC 107/297 methods) are
extremely useful in addressing stress analysis issues of pipe
support attachments as they are available in almost all
commercial pipe stress codes and methods like Kellogg
method can easily be developed into spreadsheets. The
difficulty is of course the limited range of applicability of
these methods specially in relation to d ratio and for the
D
Kellog method , its main drawback is its mathematical
oversimplification of a problem, an issue which is not
negligible when the predominant form of the loading is
Radial. More advanced approaches as explained in section 2.3
of this paper have solved the problem up to d =0.8, but
D
these methods are yet not available in commercial pipe stress
codes or as WRC bulletins and it will be a while before they
will be available as handy tools for engineering applications.
Such methods can be used to validate numerical analysis

Degenerated solid elements.


Elements based on basic shell mathematical model.
Elements based on combination of plate and membrane
elements.

For a detailed discussion on type-1, refer [19]. The main


feature of these elements is the number and variety of adhoc
assumptions made to accommodate the standard procedures of
finite element formulation. The variation of strain through
thickness isnt linear. Assumptions regarding dependence of
determinant of Jacobian Matrix in the direction of thickness
can lead to violation of rigid body properties [19].
Type-2 elements are usually not available in commercial
FE codes. They suffer from rigid body motion problems [18].
The element S8R is ABAQUS is however close to these
elements as discussed in [18].
To explain the meaning of the term Basic Shell Mathematical
mode, we briefly describe the derivation of the governing shell
equations using the tensor approach which involves the
following steps [22]:

Copyright 2013 by ASME

Fundamental assumption of the shell theory based on LoveKirchoff hypothesis and zero strain in the through thickness
direction.
Expressing the base vectors of a surface located off middle
surface i.e. a general surface in terms of the base vectors of
the middle surface (both covariant and contra-variant
versions).
Expressing the metric tensor of a surface located off
middle surface in terms of the metric tensor (both
covariant and contra-variant versions) of the middle
surface.
Expressing the rotation vector.
Expressing the Cristoffel symbols and permutation
tensors (Levi-Cevita tensors) of the surface located off
the middle surface in terms of the corresponding tensors
of the middle surface.
Expressing the strain tensors of a surface located off
middle surface in terms of the strain tensor of the middle
surface (both covariant and contra-variant versions).
Strain tensors are expressed as the difference between
metric tensors and curvature tensors in the deformed and
un-deformed states
Writing expression for stress and moment resultants.
Using appropriate constitutive relations.

STRI3 - Small Strain Triangular Element with 3 nodes and


quadratic variation of rotation (accurate representation of plate
bending because of linear curvature variation) and analytical
implementation of Kirchoff constraint at locations (DKT or
Discrete Kirchoff element).
STRI65 - Small Strain Triangular Element with 6 nodes and
Kirchoff constraint imposed numerically at points.
S8R- 8-node reduced integration element for small strain
formulation. This element has similarity with the Basic Shell
mathematical model as described in [18], although they are
not the same, the main difference being the use of Mindlin
hypothesis. This element is susceptible to element distortion.
The Hexagonal element used is a 20-node reduced integration
element. The method of analysis is Linear Elastic following
the Elastic Stress Classification Route of [28].
The issue of classification of the FE computed stresses on
the lines of [28] has been dealt with in numerous papers and
will not be repeated here. In a nutshell, local membrane
stresses are designated as Pl , primary + secondary stresses as
Pl + Pb + Q and peak stresses as Pl + Pb + Q + F in line with
[28]. Primary stresses develop to maintain equilibrium with
external loads, secondary stresses to maintain compatibility of
deformation (global) and peak stresses to maintain
compatibility of local deformation. Pl stands for local
primary stress, Pb for primary bending stress, Q for
secondary stress and F for peak stress. Peak stresses are
significant only from the standpoint of fatigue failure.
FE convergence theorems are in L2 or H 1 norms which are
difficult to implement when the exact solution is not shown
and in this presentation no attempt has been made to evaluate
the convergence using these norms. For checking the
convergence of an FE model percentage change in stress is
considered from a model with very fine mesh to gradually
becoming cruder. Stresses are checked at Gauss points for
accuracy and un-averaged. For convergence, monotonic
behavior is checked with a maximum permissible variation in
stress taken as 5%.The mesh size around the intersection is
taken as 0.3 rt with progressive mesh grading away from it.
For continuum four elements have been used through the
thickness at and close to intersections. The objective of the FE
analysis wasnt to catch the peak stresses which are used for
fatigue evaluation, because once the Pl + Pb + Q stresses are
computed, the fatigue stresses can easily be computed using
Fatigue Strength Reduction Factors (FSRF) [28]. The results
of the analysis can then be extended to compute
Pl + Pb + Q + F in a straightforward manner. [27] shows that
modeling of welds to properly simulate joint stiffness does not
have serious impact on the computed stresses and hence,
welds are not part of the models. FSRF can be avoided if
Dongs method [28] is used. However, this requires special
post processing ability of the FE Code. If welds are modeled,

In the Basic Shell Mathematical model version of Finite


Element implementation, the interpolation of the shell
geometry is accomplished using the Iso-parametric procedure.
Covariant and Contra-variant base vectors of the interpolated
surface are computed using the usual finite element
interpolation procedures and the First Fundamental form, the
Second Fundamental form and the Christoffel symbols are
then computed from these base vectors. In the Type-2
elements as described in [18], the normal vector is calculated
normal to the interpolated middle surface, although the normal
vectors at the nodal points are exactly normal to the middle
surface.
For a discussion on type-3 elements any standard text
book on FEM can be referred [19].
The FE code used for the analysis is ABAQUS ver. 6.9-1.
The ABAQUS element library [20] for shells is divided into
three categories consisting of general-purpose, thin, and thick
shell elements. Thin shell elements provide solutions to shell
problems that are adequately described by classical
(Kirchhoff) shell theory; thick shell elements yield solutions
for structures that are best modeled by shear flexible (Mindlin)
shell theory; and general purpose shell elements can provide
solutions to both thin and thick shell problems. All these
elements use bending strain measures that are approximations
of those of Koiter-Sanders version of shell theory [12].
For stress analysis, the following elements from
ABAQUS library have been used.

Copyright 2013 by ASME

Pl + Pb + Q can be evaluated at the weld toe directly by (even


though it is a singularity) linearization at the stress
classification line (SCL) as explained by Kalnins [29]. The
only issue with this procedure is the through thickness stress
component.
To avoid end effect, the location of the trunnion has been
taken as 5D [24] i.e. five times the Outside Diameter of the
Header Pipe with respect to the end of the header. The worst
aspect ratio around the intersection (HEX elements) was 6.0,
average aspect ratio 2.0. One end of the header was fixed in all
six DOFs and the other end is fixed in five DOFs. The DOF
along the longitudinal axis of the header was kept free to
generate longitudinal pressure stress (for models where
pressure was applied). Linear and full integration elements
were not selected in the quadrilateral and brick versions to
avoid shear locking. References [35, 24] provide excellent
guideline on modeling of Large Diameter Cylinder
intersections.

Table-1 (Contd)

5.0 RESULTS

36 inch header, 30 inch trunnion, wall thickness = 9.52 mm for


both. Magnitude of Force = 10KN, length of trunnion = 100
mm, d = 0.84 , t = 1
D
T

Radial
Force

Longitudinal
force

Circumferential
force

FEA shell element


(STRI3) Cylinder

FEA shell element


(STRI3) Trunnion

13

FEA shell element


(STRI65) Cylinder

FEA shell element


(STRI65) Trunnion

13

FEA continuum
element Cylinder

FEA continuum
element Trunnion

12

Loading Type

Table-2

The stresses shown in the tables below belong to the


Pb + Pl + Q category and are in N/mm2. Only maximum Von
Mises equivalent stress values are shown. For continuum
elements, stresses have been Linearized using [28] as a
guideline. For tables 1-5 the applied loadings are at the end of
the Trunnion which makes it a Shear Force + Bending
Moment at the Shell-Nozzle interface for the Longitudinal and
Circumferential Force applications.. Pressure is not a part of
the loadings in Tables 1-5. For WRC-107 and WRC-297
computations, code FE-107 has been used.

Radial
Force

Longitudinal
force

Circumferential
force

WRC 107 Cylinder

45

12

WRC 107 Trunnion

NA

NA

NA

WRC 297 Cylinder

51

16

WRC 297 Trunnion

56

16

Kellogg Cylinder

Kellogg Trunnion

0.5

0.2

0.2

Loading Type

Table-1
30 inch header, 24 inch trunnion, wall thickness = 9.52 mm for
both. Magnitude of Force = 10KN, length of trunnion = 100
mm, d = 0.8 , t = 1 :
D
T
Radial
Force

Longitudinal
force

Circumferential
force

FEA shell element


(S8R) Cylinder

21

WRC 107 Cylinder

45

16

FEA shell element


(S8R) Trunnion

15

WRC 107 Trunnion

NA

NA

NA

FEA shell element


(STRI3) Cylinder

17

WRC 297 Cylinder

50

22

FEA shell element


(STRI3) Trunnion

11

WRC 297 Trunnion

54

20

FEA shell element


(STRI65) Cylinder

20

Kellogg Cylinder

FEA shell element


(STRI65) Trunnion

14

Kellogg Trunnion

0.6

0.4

0.4

FEA continuum
element Cylinder

19

FEA shell element


(S8R) Cylinder

10

FEA continuum
element Trunnion

15

FEA shell element


(S8R) Trunnion

15

Loading Type

Copyright 2013 by ASME

Table 3

Table 4 (Contd.)

36 inch header, 12 inch trunnion, and wall thickness = 9.52


mm for header; and 6.35 mm for trunnion. Magnitude of Force
= 10KN, length of trunnion =100 mm, d = 0.34 ,
D
t = 0.67 :
T

Radial
Force

Longitudinal
force

Circumferential
force

WRC 297
Trunnion

90

13

44

Kellogg Cylinder

Loading Type

Radial
Force

Longitudinal
force

Circumferential
force

Kellogg Trunnion

WRC 107 Cylinder

48

10

31

FEA shell element


(S8R) Cylinder

19

10

WRC 107 Trunnion

NA

NA

NA

FEA shell element


(S8R) Trunnion

20

WRC 297 Cylinder

54

30

41

FEA shell element


(STRI3) Cylinder

17

WRC 297 Trunnion

103

30

75

FEA shell element


(STRI3) Trunnion

19

Kellogg Cylinder

15

11

22

FEA shell element


(STRI65) Cylinder

19

10

Kellogg Trunnion

FEA shell element


(STRI65) Trunnion

20

FEA shell element


(S8R) Cylinder

46

16

29

FEA continuum
element Cylinder

17

11

FEA shell element


(S8R) Trunnion

48

16

31

FEA continuum
element Trunnion

19

FEA shell element


(STRI3) Cylinder

42

14

26

FEA shell element


(STRI3) Trunnion

43

15

27

FEA shell element


(STRI65) Cylinder

45

16

28

FEA shell element


(STRI65) Trunnion

47

15

30

FEA continuum
element Cylinder

44

13

27

Loading Type

FEA continuum
element Trunnion

46

14

29

Loading Type

Table 5
24 inch header, 8 inch trunnion, and wall thickness = 9.52 mm
for header and 8.18 mm for trunnion. Magnitude of Force =
10KN, length of trunnion = 100 mm, d = 0.36 , t = 0.86 :
D
T

Table 4
24 inch header, 20 inch trunnion, and wall thickness = 9.52
mm for header and 6.35 mm for trunnion. Magnitude of Force
= 10KN, length of trunnion = 100 mm, d = 0.84 ,
D
t = 0.67 :
T

Radial
Force

Longitudinal
force

Circumferential
force

WRC 107 Cylinder

47

21

53

WRC 107 Trunnion

NA

NA

NA

WRC 297 Cylinder

69

31

77

WRC 297 Trunnion

74

34

78

Kellogg Cylinder

16

20

40

Kellogg Trunnion

Radial
Force

Longitudinal
force

Circumferential
force

FEA shell element


(S8R) Cylinder

48

26

46

WRC 107 Cylinder

44

20

FEA shell element


(S8R) Trunnion

43

21

43

WRC 107
Trunnion

NA

NA

NA

FEA shell element


(STRI3) Cylinder

44

22

40

WRC 297 Cylinder

44

23

FEA shell element


(STRI3) Trunnion

38

19

36

Loading Type

10

Copyright 2013 by ASME

44

22

40

FEA shell element


(STRI65) Trunnion

39

19

37

FEA continuum
element Cylinder

46

24

44

FEA continuum
element Trunnion

41

19

43

Table 6 is to reflect the effect of applying the Forces and


moments at the Shell-Nozzle Interface as opposed to at the
end of the Trunnion in Tables 1-5. Pressure is not a part of the
Loading.
Table 6

Circumferential
moment

FEA shell element


(STRI65) Cylinder

Longitudinal
moment

Circumferential
force

Shear Force
(Circumferential)

Longitudinal
force

Shear Force
(Longitudinal)

Radial
Force

Loading Type

Radial Force

24 inch header, 8 inch trunnion, and wall thickness = 9.52 mm


for header and 8.18 mm for trunnion. Magnitude of Force =
10KN, length of trunnion = 100 mm, d = 0.36 , t = 0.86 :
D
T

Torsional Moment

Table 6 (Contd)

Loading Type

Table 5 (Contd)

FEA shell
element
(STRI65)
Cylinder

47

10

20

106

361

FEA shell
element
(STRI65)
Trunnion

45

23

104

402

FEA
continuum
element
Cylinder

44

18

106

360

FEA
continuum
element
Trunnion

47

20

104

398

For tables, 7-9, applied load in longitudinal, circumferential


and radial directions = 10KN (applied together), pressure =
18.9Barg.

36 inch header, 12 inch trunnion, wall thicknesses 9.52mm


and 6.35mm for header and trunnion respectively. Loads
applied at shell nozzle interface, Moment=10KN-m and
Force=10KN. d = 0.34 , t = 0.67
D
T

For the WRC-107 analysis, pressure loading has NOT been


added as a radial load at the trunnion attachment.

WRC 107
Cylinder

48

13

99

310

WRC 107
Trunnion

NA

NA

NA

NA

NA

NA

WRC 297
Cylinder

54

13

153

413

WRC 297
Trunnion
FEA shell
element (S8R)
Cylinder

30 inch header, 24 inch trunnion, wall thickness =9.52 mm for


both (results shown for maximum Pb + Pl + Q in MPa)

Circumferential
moment

Longitudinal
moment

Torsional Moment

Shear Force
(Circumferential)

Shear Force
(Longitudinal)

Radial Force

Loading Type

Table 7

WRC 107 Cylinder

258

WRC 107 Trunnion

NA

Kellogg Cylinder

87

Kellogg Trunnion

FEA shell element (S8R) Cylinder

103

46

10

19

21

295

108

752

363

FEA shell
element (S8R)
Trunnion

48

FEA shell
element
(STRI3)
Cylinder

46

21

106

359

FEA shell
element
(STRI3)
Trunnion

46

24

103

403

23

105

401

11

121

FEA shell element (S8R) Trunnion

63

FEA shell element (STRI3) Cylinder

120

FEA shell element (STRI3) Trunnion

59

FEA shell element (STRI65) Cylinder

125

FEA shell element (STRI65) Trunnion

66

FEA continuum element (Shell) Cylinder

126

FEA continuum element Trunnion

70

Copyright 2013 by ASME

Table 8

Table 10

36 inch header, 30 inch trunnion, wall thickness =9.52 mm for


both (results shown for maximum Pb + Pl + Q in MPa)
WRC 107 Cylinder

307

WRC 107 Trunnion

NA

Kellogg Cylinder

100

Kellogg Trunnion

0.8

FEA shell element (S8R) Cylinder

146

FEA shell element (S8R) Trunnion

75

FEA shell element (STRI3) Cylinder

143

36 pipe, wall thickness = 9.52 mm. Shoe design corresponds


to 3-gusset, A=450, B=500, shoe plate thickness = 10 mm, L =
350 mm (refer fig-1); magnitude of load = 40KN. Pressure is
not applied. 1 = 0.56 , 2 = 0.49
Radial
Force

Longitudinal
Force

Circumferential
Force

WRC 107 Cylinder

181

18

61

Kellogg Cylinder

30

18

40

Loading Type

Kellogg Shoe

75

20

36

FEA shell element (STRI3) Trunnion

76

FEA Shell element (S8R)


Cylinder

FEA shell element (STRI65) Cylinder

145

FEA Shell element (S8R) Shoe

77

30

32

FEA shell element (STRI65) Trunnion

76

82

18

35

FEA continuum element (Shell) Cylinder

148

FEA Shell element (STRI3)


Cylinder

FEA continuum element Trunnion

74

FEA Shell element (STRI3)


Shoe

82

18

35

Table 9

FEA Shell element (STRI65)


Cylinder

75

22

37

36 inch header, 12 inch trunnion, wall thickness = 9.52 mm for


header and 6.35 mm for Trunnion (results shown for
maximum Pb + Pl + Q in Mpa)

FEA Shell element (STRI65)


Shoe

80

30

33

FEA Continuum element


Cylinder

81

24

33

FEA Continuum element Shoe

78

35

32

WRC 107 Cylinder

321

WRC 107 Trunnion

NA

Kellogg Cylinder

100

Kellogg Trunnion

0.8

FEA shell element (S8R) Cylinder

157

Table 11
30 Pipe, wall thickness 9.52 mm, Shoe design corresponds to
3 Gusset, A=450, B=500, Shoe plate thickness=10 mm, L=350
mm (refer fig-1), Magnitude of load=40KN. Pressure is not
applied. 1 = 0.67 , 2 = 0.60

FEA shell element (S8R) Trunnion

108

FEA shell element (STRI3) Cylinder

155

FEA shell element (STRI3) Trunnion

102

Loading Type

FEA shell element (STRI65) Cylinder

159

FEA shell element (STRI65) Trunnion

98

FEA continuum element (Shell) Cylinder

154

FEA continuum element Trunnion

103

Results for Pipe Shoes: (Stresses at locations of


singularities have not been considered)
Note: WRC-107 method has been used even though in most
cases 1 , 2 are above the allowable limit. So far Pipe Shoes
are concerned, the typically used dimensions render them
unsuitable for use of WRC-107. Despite this fact, the author in
his experience has seen its usage for computation of local
stresses at Shoe Attachments and its use is mostly due to
availability of this module in most common pipe stress
programmes. For the WRC 107 computation of Pipe Shoes,
the geometry of the attachment has been considered as
Rectangular solid. Pipe Stress Program CAESAR II Version
5.2 has been used for this purpose. For Tables 10, 11 and 12,
2c1=500 mm and 2c2=450 mm.

12

Radial
Force

Longitudinal
Force

Circumferential
Force

WRC 107 Cylinder

173

21

63

Kellogg Cylinder

27

17

38

Kellogg Shoe

FEA Shell element (S8R)


Cylinder

60

15

18

FEA Shell element (S8R) Shoe

80

22

12

FEA Shell element (STRI3)


Cylinder

62

14

18

FEA Shell element (STRI3)


Shoe

75

22

13

FEA Shell element (STRI65)


Cylinder

60

14

20

FEA Shell element (STRI65)


Shoe

82

22

14

FEA Continuum element


Cylinder

63

18

25

FEA Continuum element Shoe

82

20

16

Copyright 2013 by ASME

Table 12

Table 14 (results shown for maximum Pb + Pl + Q in MPa)

24 pipe, wall thickness = 9.52 mm. Shoe design corresponds


to 3-gusset, A=450, B=500, shoe plate thickness = 10 mm, L =
350 mm (refer fig-1); magnitude of load = 40KN. Pressure is
not applied. 1 = 0.84 , 2 = 0.75

30 header, wall thickness = 9.52 mm


WRC 107 Cylinder

312

WRC 107 Shoe

NA

Kellogg Cylinder

161

Radial
Force

Longitudinal
Force

Circumferential
Force

WRC 107 Cylinder

174

23

65

Kellogg Cylinder

24

15

34

FEA shell element (S8R) Shoe

115

Kellogg Shoe

FEA shell element (STRI3) Cylinder

131

FEA Shell element (S8R)


Cylinder

35

12

22

Loading Type

FEA Shell element (S8R) Shoe

73

22

10

FEA Shell element (STRI3)


Cylinder

34

12

22

Kellogg Shoe

12

FEA shell element (S8R) Cylinder

126

FEA shell element (STRI3) Shoe

118

FEA shell element (STRI65) Cylinder

128

FEA shell element (STRI65) Shoe

113

FEA continuum element (Shell) Cylinder

132

FEA continuum element Shoe

119

FEA Shell element (STRI3)


Shoe

50

22

10

FEA Shell element (STRI65)


Cylinder

35

13

22

FEA Shell element (STRI65)


Shoe

53

17

10

WRC 107 Cylinder

298

FEA Continuum element


Cylinder

39

15

24

WRC 107 Shoe

NA

FEA Continuum element Shoe

57

21

13

Kellogg Cylinder

136

Kellogg Shoe

12

FEA shell element (S8R) Cylinder

80

FEA shell element (S8R) Shoe

85

FEA shell element (STRI3) Cylinder

84

FEA shell element (STRI3) Shoe

89

FEA shell element (STRI65) Cylinder

82

FEA shell element (STRI65) Shoe

83

Table 15 (results shown for maximum Pb + Pl + Q in MPa)


24 header, wall thickness = 9.52 mm

For Tables 13-15, applied load in longitudinal, circumferential


and radial directions = 40KN( applied together), pressure =
18.9 barg. Pressure has been applied but not as radial thrust
load.
Table 13 (results shown for maximum Pb + Pl + Q in MPa)
36 header, wall thickness = 9.52 mm
WRC 107 Cylinder

330

FEA continuum element (Shell) Cylinder

88

WRC 107 Shoe

NA

FEA continuum element Shoe

92

Kellogg Cylinder

186

Kellogg Shoe

12

FEA shell element (S8R) Cylinder

180

FEA shell element (S8R) Shoe

155

FEA shell element (STRI3) Cylinder

184

FEA shell element (STRI3) Shoe

156

FEA shell element (STRI65) Cylinder

182

FEA shell element (STRI65) Shoe

153

FEA continuum element (Shell) Cylinder

188

FEA continuum element Shoe

159

13

Copyright 2013 by ASME

6.0 DISCUSSION OF RESULTS AND SCOPE FOR


FUTURE WORK

the pipe shoe can be seen as analogous to this parameter.


Significant differences exist for the Circumferential loading
case also. For both Trunnion and Pipe Shoes, for some cases ,
specially for the Radial Load scenario, stresses in the
shoes/Trunnion elements exceed stresses in the cylinder
which clearly shows the risk of using the Kellogg method for
computing stresses in the Pipe support attachments. A point to
note is that, the method of computing stresses in the Pipe
supports cannot be technically stated as Kellogg Method as
[4] only discusses computation of local stresses in the
Cylinder. The context of using the term Kellogg method for
the method of computing stresses in attachments is due the
fact that this computation based on elementary beam theory is
an essential feature ( in authors experience) of the
spreadsheets which use the Kellogg method to compute Local
stresses in the Cylinder. Hence the caution is using elementary
beam theory analysis for computation of local stresses in
Attachments. Significant differences in results have not been
seen in Finite element approach using different element types.
This however should not be taken as a blanket statement as the
models had proper mesh grading with adequately small
element size and the element distortion control was well
within the recommended limits of the FE code. For improper
mesh grading, element size, significantly distorted elements
and improper integration methods, significant differences in
results can be seen between the elements, especially for
Triangular elements which suffer from geometric anisotropy.
The stress analyst should carefully study the theory manual of
the FE code which he/she should be using with respect to
applicability, element distortion and integration rules.

Tables 1,2,4 show that WRC 107 and WRC 297 results
show significant differences with respect to FE results for the
radial load case. This is because of the high d ratio and
D
radial as opposed to vertical load representation of the same in
WRC 107 as explained in section 2.2 of this paper. Tables 3, 5
and 6 show that the results are comparable (even for the
Radial load case) indicating the criticality of the d factor in
D
WRC 107/297 approaches. For the Kellogg Method, the
significant difference is for the radial load case. This is
because of the basis of the method being axi-symmetrical ring
loading which significantly deviates from the actual
mathematical model in the radial load situation. The Kellogg
method also underestimates the stresses in the Trunnion. This
is due to the use of simple beam theory as opposed to shell
theory and the non-consideration of the compatibility
requirement between the header pipe and the Trunnion in this
method. Kellogg method also in most (but not all) cases
predicts lower magnitude of stresses in the Longitudinal and
Circumferential Force applications. However the allowable
stresses in the Kellogg method as long as they are specified as
the [28] allowable for local primary stress, the error will not in
general make the analysis non-conservative except for the
Radial Load scenario. For Tables 7, 8 and 9 which are for the
combined load scenario, WRC 107 results show significantly
higher magnitudes of Pb + Pl + Q with respect to FEA. Even
though the Pressure loading has not been modeled as a Radial
loading for these Tables, which would have resulted in even
higher magnitudes of Pb + Pl + Q if the direction of this load
would have been in the same direction as the additive radial
load, but the simplistic way of computing pressure stresses
also (as in Tables 7, 8 and 9) induces higher stresses in the
WRC 107 type of analysis. Pressure induced loading at a
cylinder to cylinder interface with or without other external
loadings is complicated and WRC 107 analysis which
considers the loading on the cylindrical surface as a
rectangular loading cannot predict the stresses correctly and
will err on the conservative side for most cases. WRC 107
/297 analysis has shown lower magnitudes of Stress for Shear
Forces and Torsion moments (Table 6 where the loadings have
been applied at the Shell-Nozzle Interface) with respect to
FEA. However, these loadings, in general are not the
governing loads in piping applications. When using WRC107/297 modules of a Pipe Stress Program, an analyst should
review the program document to see how pressure is modeled
in these modules.

The present analysis has to be extended for different load


combinations with varying magnitudes of the individual load
vectors to quantify the degree of over or under conservatism
of the available analytical methods. The present analysis
mainly focuses on the stand alone effect of individual load
vectors (although Tables 7-9 as well as Tables 13-15 does
address combinations but more tests need to be done with
varying magnitudes of the individual load vectors) . Effect of
variance in mesh grading and element size should be checked
to assist an analyst in selection of the best element for these
applications, if an analyst so desires. In the present scope of
work, the use of proper mesh grading, element size and
integration rules have ironed out significant differences
between the individual elements. Hence, the take away
message for an analyst with respect to individual element
types is, as long as mesh grading , element size, distortion
control and integration rules are properly used, there are no
preferred elements , although the analyst should carefully read
the Theory manual of the FE code which he/she intends to use.
7.0 CONCLUSIONS:

Results in tables 10, 11 and 12 again show that the pattern


of variance in results between FEA and WRC 107/297 is most
significant with respect to radial loads. The reason can still be
attributed to the d even though in case of the shoe
D
dimension d is strictly not applicable but the dimension of

1. Use of a particular shell theory requires an understanding of


the order of magnitude of error inherent in that theory and its

14

Copyright 2013 by ASME

applicability vis--vis the problem to be analyzed specially


with reference to d and D ratios.
D
T

107/297 based analysis is also not correct as the branch is not


pressurized for Pipe support applications where by branch
we mean the Trunnion.

2. Shell theories should be evaluated on the basis of SanderKoiter postulates.

8. Finite elements for shell analysis have different approaches


based on the theoretical considerations that form the basis of
their developments, with elements based on basic shell
mathematical model being least popular because of the
problem of addressing rigid body motion. Commercial FE
codes should be having Hybrid elements in the element library
for shell applications.

3. The use of an axi-symmetric loading model (which in this


paper has been referenced to as Kellogg method) has been
historically the most popular method for analyzing both
cylindrical and non-cylindrical attachments.
4. WRC-107 method which is based on Timoshenko equations
has the same error O T
as Morley, Simmonds and
R
Goldenveizer equations. WRC-107 results may be more or
less conservative than FE results. Results are generally overly
conservative for d > 0.5 . The analysis results show that in
D
some cases but not all (generally computation as per Kellogg
method has shown lower magnitude of stresses with respect to
WRC 107 or FE analysis), Kellogg method significantly
underestimates the stresses in Trunnion and Pipe Shoes. Hence
it is recommended that this method should not be used and
hence should not be used for evaluating stresses in Pipe
supports. A point to note is, the method as at appears in [4]
addresses only the local stresses at the cylinder, so the
evaluation of stresses in attachments cannot technically be
addressed as Kellogg method, rather calculation based on
elementary beam theory. It is against this later which, the
author in his experience has seen as widely used in the
Industry as part of the spreadsheets based on Kellogg
method is what this caution is directed at.

( )

9. Degenerated solid elements have used adhoc assumptions


on shell theory to work within the constraint of finite element
formulation. Assumptions regarding the mathematical form of
dependence of the determinant of the Jacobian Matrix on the
thickness direction coordinate can lead to violation of rigid
body properties.
10. Not much difference has been found in results using 8node reduced integration shell element developed on the line
of Mindlin hypothesis, triangular elements based on discrete
Kirchoff constraints (imposed analytically or numerically) and
use of solid elements for circular attachments. Stresses at
locations of singularity have to be carefully addressed [29].
The pattern of results i.e. relative invariance with respect to
element types need not be always correct depending on the
D
ratio, element distortion, element size and use of
T
alternate numerical integration rules. In general, as long as
thin shell theory is valid and reduced integration rule is used
for shear flexible elements, with proper mesh grading and
keeping the element size at the intersection region
significantly less than rt , type of element is usually not a
significant parameter. Stress Analyst should carefully review
the Technical Manual(s) of FE Code for the capabilities and
limitations of the available elements from the element library.

5. If an analyst is constrained to use Kellogg method for


analysis of local stresses on the pipe at support locations, the
allowable stress should not be exceeded beyond the allowable
for local primary stresses as per [28].
6. WRC-297 method is based on shallow shell theory and the
order of magnitude in error is due to omission of some terms
which are of the order O

( T R)

11. Analytical methods with d

as high as 1.0 with ease of


D
implementation is required not only because the available
methods like WRC107/297 etc are inadequate for such
applications but also as a tool to properly benchmark the FE
results. Till such time, FE models will continue to be
benchmarked against WRC 107 type of analysis for similar
loading within the limits of the applicable geometry.

and has shown overly

conservative behavior, specially for the Trunnion stresses for


most of the cases analyzed. Use of WRC 297 for Pipe support
attachments is not recommended.
7. When comparing results between an analytical and FE
approach, it is best to check the model on a component by
component basis i.e. the model is loaded with only one
force/moment component in the absence of pressure. This
check will show stresses because of which components are
over/under represented in the final results. Since WRC107/297 does not have a provision for checking pressure
loading, simulating the same by a modified radial load (=
applied radial load + pressure times area) or superposing the
results with the usual membrane stresses in the header pipe
due to pressure generally makes the analysis overconservative. The modification of the radial load in a WRC-

12. Additional tests need to be done for Pipe Shoes for varying
effects of D and combined loadings. In authors opinion it
T
is futile to expect usability of WRC-107 for shoe attachments,
as based on typical dimensions of Pipe Shoes, these geometric
parameters will in most cases be not satisfied.
13. WRC 107 /297 analysis has shown lower magnitudes of
Stress for Shear Forces and Torsion moments (Table 6 where
the loadings have been applied at the Shell-Nozzle Interface)
with respect to FEA. However, these loadings, in general are
not the governing factors in piping applications.

15

Copyright 2013 by ASME

ACKNOWLEDGMENTS

[17]

The author wishes to acknowledge Professor M.D.Xue of


Tsinghua University, Department of Engineering Mechanics
for providing some valuable suggestions and document
references and for answering some questions on her paper. The
author also wants to thank Dr.Subrata Saha of Reliance
Industries Ltd India, Mr. Suraj Kunder of Costain UK and excolleague and friend Mr.Arijit Chatterjee for providing
valuable guideline and suggestions.

[18]

[19]

REFERENCES
[1]
[2]

[3]

[4]
[5]

[6]
[7]

[8]
[9]

[10]
[11]

[12]

[13]
[14]
[15]

[16]

[20]

Timoshenko, S: 1940, Theory of plates and shells, MCGraw Hill, New York
Bijlaard, P.P., 1955, Stresses from Radial loads and
External moments in Cylindrical pressure vessels,
Welding Journal, Miami, FL, US Vol. 34
Steele, C.R and Steele, M.L.M 1983, Stress analysis of
Nozzle in Cylindrical vessels with external Load,
ASME Journal of Pressure Vessel Technology, Vol. 105
M.W.Kellogg Company,2011,Design of Piping
Systems
Mehrson, J.L, Mokhtarian, K., Ranjan, G.V and
Rodabaugh, E.C., 1984, Local Stresses in Cylindrical
Shells due to External Loadings on Nozzle Supplement
to WRC bulletin 107WRC Bulletin No. 297
Goldenveizer A.L., 1961, Theory of Elastic Thin shells,
Pergamon, Oxford,
Koiter, W.T, 1959. A Consistent First Approximation
in the General Theory of Elastic Shells, Proceedings of
the Symposium on the Theory of Thin Elastic Shells,
Delft, the Netherlands, W.T.Koiter ed., North-Holland,
Amsterdam.
Donnell, L.H., 1933, Stability of Thin Walled Tubes
under Torsion, NACA Report No. 479
Morley, L.S.D., 1959, An Improvement on Donnells
Approximations for Thin walled Circular Cylinders,
Q.J.Mech., Appl. Math., Vol. 12
Donnell, L.H, 1976, Beams, Plates and Shells, McGraw -Hill, New York, Chapter 6
Simmonds , J.G., 1966, A Set of Accurate Equations for
Circular Cylindrical Elastic Shells, ,Int. J. Solids
Structure., Vol. 2,
Budiansky,B and Sanders , J.L, 1963, On the best first
First Order Linear Shell Theory, Progress in Applied
Mechanics (The Prager Anniversary Volume),
Macmillan, London.
Flugge, W.1967, Stresses in Shells, Springer, Berlin.
Hoff, N.J., 1955, The accuracy of Donnells equations
J.Appl. Mech Vol. 22
Lekkerkerker, J.G., 1972. The Determination of Elastic
Stresses near Cylinder-to-Cylinder Intersection. Nuclear
Eng. Des., Volume 20.
Xue, M.D., Du, Q.H., Li, D.F. and Hwang, K.C., 2006.
Theoretical Stress Analysis of Intersecting Cylindrical

[21]
[22]
[23]

[24]

[25]
[26]

[27]

[28]
[29]

[30]
[31]

[32]

[33]

[34]

16

Shells Subjected to External Forces on Nozzle, ASME


Journal of Pressure Vessel Technology, Vol. 128.
Ming-De Xue, Qing-Hai Du, Keh-Chih Hwang, ZhiHai Xiang, 2010, An analytical method for cylindrical
shells with nozzles due to internal pressure and external
loads Part ITheoretical foundation; ASME Journal
of Pressure Vessel Technology Vol 132
Bathe, K.J.,Lee, P.S., 2005, Insight into finite element
shell discretizations by use of the basic shell
mathematical model Computers and Structures Vol. 83
Mac Neal, R.H., Finite Elements their design and
performance, Marcel-Dekker, NY.
Theory Manual ABAQUS, a product of Dassault
Systems, SA.
Mollman, M, 1981, Introduction to the Theory of Thin
Shells, Wiley-Blackwell.
Flugge, W., Tensor analysis and Continuum Mechanics,
Springer, Berlin.
Chen., M.W.L.Y., Li, J.G., 2000, A two step approach of
Stress Classification and Primary Structure Method.
Trans ASME Vol. 122
Xue,L.,Widera, G.E.O.,Sang, Zhifu., Feb. 2006,
Flexibility Factors for Branch Pipe Connections
subjected to in-plane and out of plane moments ,Journal
of Pressure Vessel Technology
Private communication, Professor M.D.Xue of
Tsinghua University, Peoples Republic of China.
Bhattacharya A, 2012, A Comparison of Simple
Analytical Methods for evaluating local Stresses at Pipe
Supports with Finite Element Analysis results
NAFEMS UK Conference 2012.
Bhattacharya A, 2011, A Finite Element based Study on
Stress Intensification Factors (SIF) for reinforced
Fabricated Tees NAFEMS World Congress Boston
2011
ASME Boiler and Pressure Vessel code, Sec VIII Div2
(2007 Edition), ASME Publication
Kalnins, A, 2008, Stress Classification Lines Straight
through Singularities ASME PVP Conference, Chicago,
ILL, USA
Bathe, K.J., Chapelle, D., 2010, Finite Element
Analysis of Shells Fundamentals, Springer.
Xue, M.D., Li, D.F., Hwang, K.C, 2005, A Thin Shell
Theoretical Solution for Two Intersecting Cylindrical
Shells Due to External Branch Pipe Moments ASME
Journal of PVT Vol. 127
Wong, F.M.G, 1984, Stresses and Flexibilities for
Pressure Vessel Attachments. Dissertation thesis for
Master of Science in Nuclear Engineering.
Morley, L.S.D, Morris, A.J.1978, Conflict between
Finite Elements and Shell Theory, Paper presented at
Second World Congress on Finite Element Methods,
Bournemouth, UK.
Wichman, K.R., Hooper, A.G., Mehrson, J.L., Stresses
in Spherical Shells due to External Loadings, Welding
Research Council Bulletin No. 107.

Copyright 2013 by ASME

[35]

Koves, W., Mokhtarian, K., Rodabaugh, E., Widera, W.,


2006, Large Diameter Shell Intersections, Welding
Research Council Bulletin No. 497.

17

Copyright 2013 by ASME