Pd5500 Annex G
Pd5500 Annex G
Pd5500 Annex G
form
Suggested
working G1 I ~ o a case:
d
Clause 6.2.3 Circumferential moment on cylindrical shell
Sign convention: N,, N+ are tensile when +ve
Mx,M+ cause compression in the outer surface when+ve
with M, +ve in the direction shown;fx and &, are tensile when +ve
3hell mean radius r =
~~~
lined= I
Xfset from centre Moment M, =
note 1 1
For rectangular pad Cg = circumferential length =
C, = %(axial length) =
r W
”
t- p=
Le - 4c12
- L--= ”CX -
L r
W
2c ‘13
L - -
(PIT -
CX
Q B S I 11-1999 GI27
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
PD "5
E=
w -
I
r
"
-
t
Ci
n Cz
c,
O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
GD9
Licensed by Information Handling Services
S T D - E S 1 P D 5500-ENGL 2000 1624bh9 0832412 158 m
PD 5500.2O00 Annex G
E= L,--
r
For C+ > C, From figure G.20a c, = F'rom figure G.18 or 61 =
C1 -" C1 - "
figure G. 19 o='=36
CX r -&Er
= cz
E= L,-- W
r
NOTE 1. Position A corresponds to quadrants QI and Q4in figure G.24.
NOTE 2. To ensure correct summation in suggested working formG3, example A, letters have been inserted here forthe stress
components and theirsigns.
6.2.3.6 summation of maximum stresses due to local NOTE 4. Absolute values of shear stress are used in the table.
loads on a qylindricd SM This is because the actual shear stress pattern is complex and
because the formulae forshear stress due to shear force are
6.2.3.6.1 Gaeml approximate.
Although theexact location of the stsesses calculated NOTE 5. At the nOzzle 0.d. where a compensation pad is fitted,or
at the edge of a load on an attachmentor support, distribute N4,
in 6.2.3.1 to 6.2.3.4 is not known the stresses may be M+,Nx and M, as in 6.3.1.5. For a nozzle witkt a pad, an
considered to lie within the 180"sectors shown in additional hoop momentis to be added to M+ as in 6.2.7.
figure G.23. The sign of the stress in one sector is NOTE 6. In the calculation of total stress intensity (lines 27 to 29
known to be reversed in the oppmite sector. and 32 to 34 of the table) the pressure term has been omitted for
simplicity.
By dividing the loaded area into quadrants and
summing the maximum &esses in each quadrant, a
maximum combiied stress is obtained The method for
Stress o f one sign Stress of opposite sign
this is shown in suggested working form G3. The
stresses due to pressure are combined with those due
to the local loads. The combined stresses and stress
intensities are Bssessed against the allowable values
specified in A.3.3.
==+
The stress components should be insertedinto the
table accordrngto the correct convention. To define
this convention, each~ ~ E S Scalculated in suggested
worlung forms G1 and G2, includmg its algebraic sign,
has been assigned a letter. These numbers shouldbe
entered into suggested working form G3, example A, in
accordance with the convention shown.
NOTE 1. The signs of FR,Fc, ML,Mc and MT are positive when
they act in the direction shownin figure G.24.
NOTE 2. N , and N4 are positive for tensile membme stresses
and Mx and &i4 are positive when they cause compressivestresses
on the outer surface of the shell. Stresses& and& are positive
when tensile and negative when compressive.This is in Stress of opposite sign ;tress of one sign
accordance with 6.2.2.1.2. Longitudinal moment :ircumferential moment
NOTE 3. The letters A to D apply to the stresses resulting from a
radial load FR.When FR is positive,A and C represent positive Figure 6.23 Sector stresses
numbers in quadrant QI on the inside and B and D represent
negative numbers in quadrant QI on the inside.
0 BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
.. ... . .. - , ... . . .. ..
b) In the shell pluspad at the edge of a loaded area
I
.".
f,
F
1) circumferentially
MT (G.2.3-4)
f p = PDf%ìt I
2) Longitudinally
(G.2.3-6)
f p = PDIbat I
c) In the shell at the edge of pad, attachment or
support.
1) Circumferentially
fP
PD (G.2.3-6) I
2) Longitudinally
fP =
PD
4e,p (G.2.3-7) I
where
F'igure 6.24 Notation for external loads at
a nozzle or attachment on a cylindrical
shell e- = shell
plate analysii thickness (see 1.6); I
eap = pad analysis thickness (see 1.6); I
%t =easp+ cap;
6.2.3.6.2 SEressform-
Q is as defined in 3.6.4.1;
pressure stxess formulae. e, is as defined in 3.6.4.1.
a) At nozzle 0.d
o BSI 042000
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI PD 5500-ENGL 2000 m Lb24bb9 0879333 T35 m
PD 6600.2000 Annex G
Suggested working form G3 Load case: Node o.d./pad o.&/ loaded area
dimensions')
Clause 6.2.3.6 Summation of Radial load FR I
I
Shell thickness/
maximum stresses due to local Shear forceFc Shell + pad
loads ona cylindrical shell ~clznessl)
Shear forceFL Shell id
Torsion momentMT Design pressure
Circumferential Design stress (J)
moment Mc
LodtUW Yield stress
moment ML I
Circumferential stresses Quadrant 81 82 83 o4
Mide
Outside mide
Outside nside
Outside mide Outside
- Surf'ace
Membrane component (N&) due to:
1 Radialload A A A A A A A A
2 Circumferential moment E E E E -E -E -E -E
3 Longitudinalmoment I -I I -I -I -I I I
4 Sub-total due to local loads
5 Pressure (fp from 6.2.35.2)
6 Sub40t.d (ef),
Bending component (W+@) due to:
7 Radialload B -B B -B B -B B -B
8 Circumferential moment F -F F -F -F F -F F
9 Longitudinal moment J -J -J J -J J J -J
10 sub-total(f+b)
11 Total circumferential stress cf&)
Longitudinal stresses
Membrane component (N&) due to:
12 Radial load C C C C C C C C
13 Circumferential moment G G G G -G -G -G -G
14 Longitudinalmoment K K -K -K -K -K K K
16 Sub-total due to local loads
16 Pressure cfp from 6.2.3.5.2)
17 Sub-total(f&
Bending component (W.#) due to:
18 Radial load -D D D -D D -D -D D
19 Circumferential moment H -H H -H -H H -H H
20 Longitudinalmoment L -L -L L -L L L -L
21 sub-total (f&)
22 Total longitudinalstress cfJ
shsar stresses (from 6.2.3.5.3) due to:
23 Torsion moment
24 Circumferential shear force
25 Longitudinalshear force
26 lbtal shear stress (T)
Check of total stress intensity
(membrane + bending) to k3.3.1
and A.3.3.2
27 fi = Gf+ +fx + du+-J,>~ + 43JJ/2
28 fi
I2
=v++fx- du+- f 3 + 431
29 -fi
')Delete as appropriate
+ dV+m -frnY'+ 4 3 1
IMaximum membrane stress intensity = maximum absolute value in rows 32.33 and 34 =
o BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
G/33
Licensed by Information Handling Services
STD*BSI P D 5500-ENGL 2000 m Lb24bb7 0832416 BT3 m
PD 5 6 W m Annex G
Suggested working form 63: Load case: D5 - N - 1 at nozzle 0.d. N o d e o.d.-/ 219 mm
example A l€dehre
1
-)
1
penetration weld,see 6.3.1.6 N~lllIIl
Circumferential 3630000 Design stress cf> 151.6 N/mm2
moment Mc N.Inln
Longitudinal 3630000 Yield stress 227.4 N/mm2
moment MT. N.IIWl
O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
Suggested working form 63: example A (continued)
Check of total stress intensity (membrane +
bendmg) to A.3.3.1 and A.3.3.2
fi -f1
34.5
1 207.3
8.2 188.6
124.3 149.3
95.7 111.9
-26.2 -18.7 -28.6 -37.4
49.5
-21.7
264.8
-29.9
224.9
126.6 167.3
85.3 27.7 134.9
-32.4 -41.3
Maximum total stress intensity = maximum absolute value in rows 27,28 and 29 = 254.8 (acceptable)
Allowable stress at nozzle = 2.25f = 341.1 , M - 2”. 2 *
Check of buckling stress to A.3.3.3
30 I
ROW 4 + row 10if row 4 is compressive I-113.0
81.2 I
24.9
-25.7 I I
31 ROW15 + row 21 if row 15 is compressive -90.5 68.5 -0.5 -9.9
Maximum compressive stress in rows 30 and 31 = -113
Allowable stress = -0.9 X yield stress = -204.7 (acceptable)
At edge of compensation pad, attachment or
support, check of membrane stress intensity
to A.3.3.1’)
34
fim -fim II
Maximum membrane smss intensity = maximum absolute value in rows 32,33 and 34 = 146.4
Allowable stress = 1.2f= 181.9
‘belete as appropriate
6.2.3.6.3 shear s t m x formulae Loads applied through square fittings of side 2Cx can
be treated approximately as distributed overa circle of
Due to: radius r, = C
,.
a) tomion (MT) Loads applied through redangular brackets of sides
I T=T N T 2Cx and 2C+ can be treated approxhnately as
distributed overa circle of radius r, = m.
d o Tl
b)
circumferential
shear
force (Fc) The
followingforces
and
moments are set up
in
the
wall of the vessel by any local loador moment
I T=-
W C
doTl
((3.2.3-91 a) Meridionalmoment M,: acting per unit width on a
normal section, formed by the intersection of shell
longitudinal
c) shear
force (&J with a conesemi-vertex
of angle.
These forces and moments and the deflection of the 6.2.4.2 Stresses and &$ltxtions due to radid loads
shell dueto the load can be found interms of the F’igure G.26 shows aradial load appliedto a spherical
nondimensional parameters: shell through a branch of radius r,.
(G.2.311) The deflections, moments and membrane forces due to
the load W can be foundas follows fromSgUres G.27
and and G.28. For explanationof these curves see6.2.4.3.
For an example of their use see 6.2.4.4.
(G.2.312) a) Deflection from figure G27 and the relation
These two factors can be found quickly from the chart Wr
6=ordinateofcurveXz
in figure G.25, givenx, r, and theratio rlt. Et
The charts in 6.2.4.2 and 6.2.4.3 (figures G.27 to G.31) b) Meridional momentM, per unit width from
.- ._
...
give graphs of nondimensional functionsof these Sgure G.28 and the relatiox
deflections, forces and moments plotted against the M, = ordinate of M, curve X W
parameter S for given values ofu which have been c) Circumferential moment M+ per mit width from
derived from[31 and [91. figure G.28 and the relation:
The full curves in eachset of graphs give conditionsat M+ = ordinate of M+,curve X W
the edge of the loaded area whereu = s. The most d) Meridional membrane force N, per unit width
unfavourable combinationof bending and direct from figureG.28 and the relation:
stresses is usually found here.
The dotted curvesfor particular valuesof u give
N, = ordinate of N, curve X W/t
conditions at points inthe shell away form the edge of e) Circumferential momentN+ per unit width from
the loaded area wherex is greater than r, and u is figure G.28 and therelation:
therefore lessthan s. N+ = ordinate of N+ curve X Wlt
Since thecharts are nondimensional theycan be used
in any consistent systemof units.
The stresses and deflections found fromthese charls
w li be reduced by the effect of internal pressure but
this reduction is small and can usually be neglected in
practice. ( S e e [8] and [9].)
Values of r/t
.o1 .O2 .O3 .O4 .O5 .06.07.08 0.1 0.2 0.3 0.4 0.5 0.6 0.70.8
x / r or r. / r
Figure 6.25 Chart for finding S and u
o BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
GB9
Licensed by Information Handling Services
meridional
O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
~ ~
-1
-0.1
Nf
;e*
-0.01
-0.001
-0.0005
-0.0001 I
I
I I I I I 1 I I
0.0 0.5 1.o 1.5 2.0 2.5 3.0 3.5 4.0
s
Figure 6.27 Deflections of a spherical shell subjectedto a radial load W
o B S I 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
G/4 1
Licensed by Information Handling Services
-0.0001
-0.001
-0.01
-0.05
-0.1
0.0 0.5 1.o 1.5 2.0 2.5 3.0 3.5 4.0
Q B S I 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
G/#
Licensed by Information Handling Services
STD-BSI P D 5500-ENGL 2000 Lb24669 083242b 742 9
PD 55002(" Annex G
-1
2
2"
-0.1
-0.05
-0.01 1
0.0 0.5 1.o 1.5 2.0 2.5 3.0 3.5 4.0
5
O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
-1
S= 1.82~
\Irt
-0.1
-0.01
-0.001
-0.0001
0.0001
0.001
0.01
0.05
o. 1
I FigureG.28d)Circumferentialforce
load W
N,+in aspherical shell subjected to a radial
Q B S I 11-1999 G/&
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
6.2.4.3 Stresses, &$kctiom and slqpes due to an The slope of the branch dueto the external momentis
extemal moment found from
Figure G.29 shows an e x t e d moment appliedto a
spherical shellthrough a branch of radius r,.
(G.2.41) I
In this case the deflections, moments and membrane where 61 is the maximum deflection at the edge of the
forces depend on the angle8 as well as on the branch for8 = O and u = S,ie.:
distance x from the axis of the branch. They can be
found as follows fromfigures G.30 and G.31. For
explanations of these curves see6.2.4.1. X ordinate of full curve in figure G.30
a) Deflections fromfigure G.30 and the relatiow for x = r, (G.2.42) I
MCos8$
d=ordinateofcurveX Et2
b) Meridional momentMxper unit width from
figure G.31 and therelation:
Mcos 8
M, = ordinate ofM, curve X -
fi
c) Circumferential momentM+ per unit width from
figure G.31 and therelatiom meridional
MCOS I3
M+ = ordinate ofM+ curve X -
fi
d) Meridional membrane forceN, per unit width
from figureG.31 and the relation:
M cos 8
N, = ordinate of N, curve X -
t f i
e) Circumferential membrane forceN+ per unit
width from figure G.31 and the relation:
Mcos I3
N+=ordinateofN+curveX-
t f i
Equal and oppositemaximum d u e s of all the above
quantities occurin the plane of the moment, ie. where
8 (seefigure G.29) = O' and 8 = 180'.
Y"
Figure 6.29 Spherical shell subjected to an
external moment
-1
-0.1
-0.05
-0.01
0.0 1 0.5 .o 1.5 2.0 2.5 3.0 3.5 4.0
S
0 B S I 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
GI47
Licensed by Information Handling Services
-1
-0.1
-0.01
-0.001
b
cn
O
o -0.0001
0.0001
S
0.001
0.01
0.05
o. 1
-1
-0.05
-0.01 I l I I I l l I I
I 0.0 0.5 1 .o 1.5 2.o 3.0 2.5
3.5
S
4.0
0.1
m
II
0.01
0.001
0.0005
0.0001
10
S = 1.82 x
Jrt
u = 1.82 r,,
1 fi
o. 1
sx
0.01
0.001
-0.001
-0.01
-0.05
-0.1
0.0 0.5 1.o 1.52.5 2.0 3.0 3.5 4.0
5
43 BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
G51
Licensed by Information Handling Services
STD-BSI PD 5500-ENGL 2000 Lb2Vbb9 0632434 B19
PD 5 ~ 2 0 0 0 Annex G
Wr
t - 12.5 r 1250-
a) Next to the bmnch When deflection = - 0.022
S = u = 1.09 (from figure G.25) Wr -0.022 X 4500 X 1250 = -o.oo425 mm
Ordinate of full curve in figure G.27 = -0.145. ËF = 1.86 X 105 X (12.5y
-
Interpolating similarly in figure G.28 gives:
:. Deflection = -0.145 X EtWr M
5
W = -0.01; 2 4 = +o.oos;
W
- -0.145 x 4500 x 1250 = o.o251
- 1.86 X 105 X (12.5)2 $
= -0.04;N t = +0.015
Ordinate of full Mxcurve in figure G.28 = + 0.067 W
:. Meridional M, = + 0.067W = 301 N ~ n n - h m Whence the:
ordinate of full M+ curve in figure G.28 = +0.02 meridional momentM, = -45 N a u n h q
:.Circumferential momentM+ = + 0.02W circumferential momentM+ = +22.5 Nmm/mm;
= 90 Nm-un/mm meridional membrane force N, = -14.4 N/-
Ordinate of full N, curve in figure G.28 = -0.11 circumferential membrane forceN+ = t625Nhtun
-0.11w The resulting meridional stresses are:
:. Meridional membrane forceN, =
-14.4 6 X 45
at the outsidef, = 12.5 + (12.5~
= -1.15 + 1.73 = +0.58 N/IIU$
ordinate of full N+ curve in figure G.28 = -0,034 at the insidef, = -1.15 - 1.73 = -2.88 N/mm2
The resulting circumferential s t r e s s are:
... N+ = -0.034w - 0.034 x 4500
t
"
6.2.6 Local loads on spherical shelfiozzle is the shear load applied at nozzle ( iN);
attachments
is the local wall analysis thickness of shell,
6.2.6.1 General adjacent to nozzle (in mm) (see 1.6);
6.2.6.1.1 Introduction is the wall analysis thickness of nozzle
The method of c a l c w local stress levels at a (in mm) (see 1.6);
nozzle junctionis based on data given in[25]. Using
thisdataitispossibletoestimahthemaximumstress is the nondimensional parameter= -
which can occur at a spheremozZle attachment dueto
the application of internal pressure,thrust, external is the maximum stress due to local loading
moment and shear force.The method covers both (inN/mm2) (see equations (G.2.51)
flush and protruding nozzles.In the original work the to (G.2.54));
nozzle length is treated as semi-infinite withoutany is the circumferentialstress (in~/-z);
restriction on its length It is, however, considered
necessary to stipulate a lower limit on the internal is the meridional stress (longitudinal in a
cylindrical shell) (inNhnm2);
protrusion equal to I/%?.Nodes with internal
protrusion lessthan I,&? should be treatedas flush is the yield stress in simple tension
nozzles. In this way some additional conservatismwill (in N/mm2);
be introduced for those protruding nozzles where the is the external moment shakedown factor
internal projection does notsatisfy this (see equation (G.2.H));
recommendation. is the overall external moment shakedown
All the stress concentration factors given in factor (see6.2.6.6);
figures G.32 to G.39 inclusive are based on the
maximum principal stress theow. is the internal pressure shakedownfactor
(see equation(G.2.6-1));
The stress concentration factors given in 6.2.6.2
to 6.2.6.7 are based on data obtained fora sphere of is the overallinternalpressure shakedown
constant thicknessT,whereas in practice T is looked factor (see6.2.6.6);
upon as the local shell thicknessd a c e n t to the is the radial thrust shakedown factor (see
nozzle, themain vessel being of a smaller thicknessT equation (G.2.6-2));
For thesecurves to be valid the thickness ofthe shell is the overall radial thrust shakedown factor
should not be reduced to Twithin a distance H as (see 6.2.6.5).
defined in3.6.4.3.4.
Work in progress showsthat h e n the vessel ~ c k n e s s 6.2.6.2 Maximum stress at a spherdmzzle junction
is reduced fromT to T at a distance H h m the nozzle, due to applhtion of intemal pressure
higher stresses than those given infigures G.32 to G.39 Figure G.32 gives plots of stress concentration
inclusive may occur forsmall values of p and high factors(s.c.f.s) against the nondimensional parameterp
values of v?". Further guidancecannot be given at the for various nozzldshell wall tfT d os for flush
present stage. nozzles. The maximum st;ress, ,,u is then calculated
This procedure providesa method of computing by multiplying the s.c.f.thus obtained by the nominal
maximumstresseswhichoccurintheshellratherthan press~res t r ' e s ~given by- PR .
in the nozzle. In some instancesd c ~ &eses d 2T' Le*:
may be higher in the nozzle wall than in the vessel
shell, especially for very thin nodes. These are not
(G.2.51)I
considered for the reasons stated in [31]. F'igure G.33 gives similar plots for protriidhg nozzles.
6.2.6.1.2 Notacion Before using figure G.33 a check should be made to
For the purposesof 6.2.6 and 6.2.6, which are ensure that the internal nozzle protrusionis equal to or
applicable to radial nozzles only, the following symbols greater than if it is not, Sgure G.32 shouldbe
apply All dimensions are in the corroded condition used as for a flush nozzle for obtaining the s.c.f.
unless otherwise indicated (see3.1.6): 6.2.6.3 Maximum st- at a sphediwzzle junction
due to application of radial load or thrmst
K kafactor, Fgme G.34 gives plots of s.c.f. against the
M is theexternalmomentapplied at nozzle non-dimensional parameterp for flush nozzles. The
(in N-mm); maximum stress is calculated by multiplyingthe s.c.f.
P is the internal pressure (in N/mmz); obtained fromf i g u G.34~ by:
F'igure G.% gives similar plots for protruding nozzles. Thus 6.2.4 can be used to calculate the dieaway of
Before using figure G.% a check shouldbe made to stnss, and the reduction factor,at the required
ensure that the internal nozzle protrusionis equal to or distance from the loaded area for the applicationof
greater than $%e if it is not, figure G.34 should be radial loads or e x t e d moments. This reduction factor
used as for aflush nozzle for obtainingthe s.c.f. can then be appliedto the maximum stress calculated
in 6.2.6 to obtain the&reses away from the loaded
6.2.6.4 Maximum s t w s at a Sphew'nozzle junction area
due to a M k t i o n of rmternd momat NOTE. An altemative method may be used, see [B].
For flush nozzles the maximum stress at a If the loadednozzle area is less than 2.5@ from
spheremozzle junctioncan be determined by using another stnss concentrating feature,&reses as
figure G.36. The fhtstep is to read off the s.c.f. for the c a m
l in accordancewith annex G become
appropriate vessel nozzle geometry. The maximum unreliable and some other method of assessing the
stress is then obtained by multiplying the s.c.f.thus total stress, for example Mte element shes3 analysis
obtained bythe factor orprooftest,isrequired
6.2.6 Spherical shells: shakedown loada for
radiai m z z h
G.2.6.1 General
,,a = s.c.f. x- (G.2.H) 6.2.6.1.1 Inlroduction
Figure G.37 gives similar plots for protruding nozzles. All the shakedown loadsgiven in6.2.6.2 to 6.2.6.6 are
Before using figure G.37 a check should be madeto based on the maximum shear stress criteria
ensure that the internalnozzle protrusionis equal to or For vessels subjectedto cyclic loading, beit pressure,
greater than $%e if it is not, figure G.36 should be radial load, external momentor any combination of
these, itis essential to have a knowledge of the
used as for aflush nozzle for obtainingthe s.c.f.
shakedown limit in order to prevent plastic cycling or
6.2.5.6 Maximum s t w s at a spMmzzlejunction incremental collapse.By keeping the cyclic loadings
due to application of shear load within the shakedown limits it ensuresthat, after initial
Figure G.38 should be used fordeteminhg the s.c.f. plastic deformation, further deformation will be in the
for flush nozzles. Themaximum stress, urna, is then elastic range,ie. the vessel has 'shaken down' to
calculated by multiplying the s.c.f. obtained inthe first purely elastic behaviow. The method given does not
step by the factorS L f l , Le.: necessarily implya limited plastic deformation before
S shakedown is achieved.
a m a = S.c.f. X - (G.2.54) The shakedown conditionscan occur after different
27crT
Figure G.39 gives similar plots for prohuding nozzles. numbers of cycles depending on thecyclic conditions
and stress level; in certain cases, the plastic
Before using figure G.39 a check should be madeto defonnation before shakedown might be signiticant.
ensure that the length of the internal nozzle protrusion The method of predicting shakedown factors for
isequaltoorgreaterthan~;ifitisnot,figureG.38 internal pressure,radial nozzle thrust and external
should be usedas for a flush nozzle for obtaining moment at a vessehozzle junction in6.2.6.2
the s.c.f. to 6.2.6.6 is based on data given in [27]. From the data
6.2.6.6 Maximum slress at a spMmzzlejunction shakedown factors forflush and protruding nozzles
under combined loading can be estimated for eachof the aforementioned
For a conservative estimate of the stresses occuring individual loading conditions.
under the actionof combined loading themaximum Where the various loading conditions occur
stresses obtained from eachof the individual loadings simultaneouslya simple formulais given that considers
should be added together.This w i always be
l the interaction betweenany of these loading conditions
conservative because the maximum stresses for (=e [271).
individual loadings may occur at different locations No clear distinction betweena flush and a protruding
and different directions(ae andor ad. nozzle is given. It is considered necessaryto stipdatÆ a
6.2.5.7 Stresses awayfmm #N loaded a m
The method givenin 6.2.6.1 and 6.2.6.6 for
lower limit onthele
i@ of the nozzleinternal
protrusion equal to 2rt. Nozzles with intend
protrusionless than shouldbetreated as flush
calculating localstresses at a spherehozzle junction nozzles. By doing so, some additional conservatismwill
caters for themaximum stress levels only. No be introduced for those protruding nozzles where the
information is given on stresses away from the loaded internal projection does notsatisfy these
area recommendations.
Stress distributions in the vicinity of the spherehozzle The shakedown factors given in6.2.6.2 to 6.2.6.6 are
junction are required in cases where other loaded based on data obtained fora sphere of constant
areas are inproximity to the one under consideration. thickness T, whereas in practiceT is looked upon as
It is proposed to use the data already available in6.2.4 the local shell thickness acljacent to the nozzle, the
to determine thesestresses. The assumption hereis main vessel b e i of smaller thicknessT For these
that,although the magnitudes of local stsesses may curves to be valid the thickness of the shell should not
differ, the plotof stress level versus distance from the be reduced to T within a distance H as defined
loaded area remains basically similar. in 3.5.4.3.4.
Figure 6.32 Maximum stress in sphere for internal pressure (flush nozzles)
Gm6 0 m1 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI P D 5500-ENGL 2000 m L b 2 4 b b 9 0832439 3T0 D
Annex G PD 6500.2000
Figure 6.38 Maximum stress in sphere for thurst loading (protruding nozzles)
o BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
GE7
Licensed by Information Handling Services
'STDmBSI P D 5500-ENGL 2000 1h2'+hh9 0832440 O12 m
PD 65002000 Annex G
f h M
"
0.01 0.1 1.o 10.0
Figure G36 Maximum stress in sphere for moment loading (flush nozzles)
O
0.01 0.1 1.0 10.0
Figure 6.37 Maximum stress in sphere for moment loading (protruding nozzles)
O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
S T D - B S I P D 5500-ENGL 2000 m L624667 0832443 T57 m
Annex G PD 5500.2000
20
15
+
:
10
O
0.01 0.1 1.0
p = LJZ 10.0
R T'
Figure 6.38 Maximum stress in sphere for shear loading (flush nozzles)
20
15
y.:
c!
v)
10
O
0.01 0.1 1.0 10.0
Figure 6.39 Maximum stress in sphere for shear loading (protruding nozzles)
Q B S I 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
6.2.6.1.2 Notation The moment shakedown factorcan be defined as:
For the purposesof 6.2.6 the symbols are as defined
in 6.2.6.
6.2.6.2 shakedown.factor for intemal pressure Before using the relevant figures for the protruding
loading nozzles, a check should be carried out onthe nozzle
m e s G.40 and G.41 should be used for determining inner projection. this is less than$22 then the
the shakedown factors under internal pressure corresponding plots for flush nozzles should be used in
conditions for flush and protruding nozzles determining the necessary shakedown factor.
respectively. 6.2.6.6 Intemction between shalcedownfactors under
The pressure shakedown f&tm can be definedas the combined Ming conditions
ratio of the nominal pressure stressin the spherical For the caseof the combined loading condition,
shell to the value of yield stsess in simple tension,ie.: [27] gives the following equation so that the overall
shakedown conditionis obtained
F+ =
G+ =
mS 1
-
= (Gm
6.2.6.3 shakedownfactor for did thrust at a Po Qo mo
mzzk In this equation the valuesof &, go and Eoare read
The relevant shakedown factors for flush and off from figureG.40 to figure G.47 inclusive for the
protrudmg nozzles subjectedto radial loads (radial appropriate vessehozzle geometry, whileF, 5 and Eï
with respect to the vessel) shouldbe determjned from are as derived fromthe relevant equations (G.3), (G.4)
figures G.42, G.44, and G.46 and from figures G.43, G.45 and (G.5).
and G.47 respectively. Where the conditionsare such that the relationship
The radial thrust shakedown factor can be definedas: given by equation(G.6) is not satisfied thena revised
nozzldshell geometry (increasedvessel shell or branch
wall thickness) should be used and the procedure
repeated until equaljon (G.6) is fullilled.
6.2.6.4 shakedownfactorfor e x
- moment
Rgures G.42, G.44 and G.46 should be used for
calculating the moment shakedown factorm for flush
nozzles. For protruding nozzlesthe correspondingplots
for the shakedown factorare given in figures G.43,
G.45 and G.47.
G/60
COPYRIGHT British Standards Institute on ERC Specs and Standards
0 BSI 11-1999
Licensed by Information Handling Services
~~ ~
= 1.0
. = 0.50
-
4
2 .o
1.5
Thrust'
1.0 '
o .S
O
0.01 10.0
Figure 6.42 Shakedown values for thrust and moment loadings (flush
nozzle)
-
m
5.0
'
4.0
3.0
2.o
1.o
O
0.01 0.10 1.0 10.0
R T'
F'igure 6.43 Shakedown values for thrust and moment loadings
(protruding nozzle)
ñì
5.0
-
4
2.0 4.0
1.5 3.0
1.0 2.0
1.0
O
0.01 0.10 1.o
"JT
r
P r R
R T
Figure 6.44 Shakedown values for thrust and moment loadings (flush
nozzle)
O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD=BSI P D 5500-ENGL 2000 W L b 2 4 b 6 9 0832445 hT4 W
Annex G PD 56Mk2000
-
m
5.0
f
4'
2.0
9 G .O
I
Yomen t
1.5 - 3.0
Thrust
1.0 2 .o
0.5 1.0
O O
0.01 1.0 1'0.0
p =L/%
R T'
Figure 6.46 Shakedown valuesfor thrust and moment loadings
(protruding nozzle)
-
m
5.0
!
-
4
2.0 4.O
\
1.!i 3.0
Thrust '
1.0 2.0
0.5 1.o
O O
0.01 0.10 1.0 10.0
oB I 11-1999 G/@
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI PD 5500-ENGL 2000 W 3b24bb9 083244b 530
PD 550052000 Annex G
-
m
5 .O
-
4
2.0 4.0
1.5 3.0
1.o 2 .o
0.5 1.o
O O
0.01 0.10 1.0 10.0
6.2.7 The N e c t sf externalforces and moments this amount should be addedto the value ofM+
at branches calculated in6.25.
Large external forces and moments can be applied to These corrections apply only to circumferential
the branchesof vessels by the thermal movements of moments andare due to the effect of the rigidity of the
pipework. attachment of the branch whichhas little influence on
the effect of longitudinal moments.
The stresses due to these are likely to be greatly The tensionat the inside of the shell dueto the local
overestimated if the forcesin the pipe system are circmferential bending momentM is added to the
determined by assuming that the connectionto the circumferential membranestress &e to internal
vessel is equivalent to an anchor inthe pipe system pressure, butthis stress will not be present whenthe
More accurate valuesof the terminal forces and vessel is under hydraulic test.
moments can be found if the deflection dueto a unit Where nozzle branches with reinforcing platesare
radial load and the slopes dueto unit longitudinal and attached with full peneiralion welds,Le. in accordance
circumferential moments distributed over the area of with figures E.2.27, E.2.28a, and E.2.314 they may be
the branch and its reinforcement are known. assumed (for the purposeof local stresses evaluation)
These can be found fora given vessel and branch by to be integral with the shell andthe stresses evaluated
the methodsgiven in 6.2.2.3 and 6.2.3 for cylindrical in accordance with 6.2. Where nozzle branches with
vessels and by methods given in 6.2.4.2 and 6.2.4.3 reinforcing platesare attached, with partial penetration
for spherical vessels. Experiments in the USA, welds, i.e. in accordance withfigures E.2.28b, E.2.29,
discmsed in [17], have shown that slopes and E.2.30 and E.2.31b, they may be analysed in
deflections calculatedin this way are sufficiently accordance with6.3.1.6.
accurate for practical purposes except that the slope of 6.3 Supports and mountings for pressure
a branch due to a circumferential momentis vessels
about 75 % of the calculated value because of the effect 6.3.1 General considerationsfor supports
of local stiffening by the metal of the branch. 6.3.1.1 Intmdwtkm
When the loads from the pipework are known, the This clause and6.3.2 and 6.3.3 are concerned with
local stresses in the vessel shellcan be found by the the supports for pressure vessels and the supports for
methods given in 6.2, except that, in a branch with an fittings carried from the shellor ends of the vessel,
extemal compemahg ring of thickness Q subject to a with regard to their effect on the vessel. Thetdructural
ciw@erential moment there is an additional design of supports is not included because it can be
circumferential moment in the shellat the edge of the dealt with bythe usual methods of structural design.
reinforcing ring to Nq,Q/4 and [ln recommends that Convenient references for theseare [M] and [41].
G/65
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI PD 5500-ENGL 2000 m L b 2 4 b b 9 0879343 984 m
PD 6500.2000 Annex G
doubtful because of settlement ofthe supports and Nt+vessel shell = N&& + t) (G.3.14)
initial errors of r o u n d n ~of or straightness in the G.3.1.5.2 Radùdly outwuni load and/m moment an a
VesseL Cylinder
These loadsrequire gusset platesas shown on the
6.3.1.4 &.ackets typical arrangementin Sgure G.49b to achieve the load
Brackets are fitted to the shellsof pressure vesselsto transference from the attachment to the vesseL
support either the vesselor some structure which has The stresses in the vessel shellat the edge ofthe
to be carried from it.mical brackets are shown in reinforcing plateare approximat& equal to thaw
figure G.48. calculated byassuming the loador moment to be
distributed over the whole area of the reinforcing
The brackets themselvesare designed by the ordinary plate 2d, X 2d,+and proceeding as described in6.2.2.1
methods used for brackets supportingbeams in for a radial load or in 6.2.3 for a moment Note that
structural engineering. thesignoftheradialloadisreversedtothat
A bracket always appliesan external momentto the in 6.3.1.5.1.
shell equal to W@
wall
-Support
I
' Backing- plate
i
a) Bracket for vessel support
A safe approximation forthe maximum stresses in the b) Flnd the resultant stresses due to these by
reinforcing plate, whichoccur at the edges of the assuming that the vesel wall and the reinforcing
actual loaded area 2Cx X 2C+ is given by the following plate share the momentsM+ and M, in proportion
prOCedUre. to the cubesof their thicknesses and the membrane
a) the moments M+ a d Mx and the forces N+ and Nx in direct ProPortion to their
maximum membrane forces NA and N, for the same thicknwes, as given in G8-1.6.1b.
loading applied to a cylinder oi'thicknk (t + tl) 6.3.1.6.3 Loa&s on spherical vessels
from the charts in 6.2.2.1 for a radial load or The principles of6.3.1.6.1 and6.3.1.6.2 can be
from 6.2.3 fora moment, both applied over the applied using the appropriate chartsof 6.2.4.2
loaded area 2CxX 2C+. and 6.2.4.3.
A 1 L
"
i ' t
Vessel
"_ E
Ø ~
Reinforcing
plate
G/69
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
~~
~
~~
.Support
m g
,Steelwork
Reinforcing pads
if nece:sary
This distance
t o be as small
L
Legs braced
if required
Points of
maximum
torsion
6.M. = O
For a derivation of the basic equations and constants in this clause see (371.
a) Saddlc rupports
Centroid of
L
ring support ,
b) Ring supports
c) Leg supports
8 BSI 11-1999 Gm
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
In the case of largediameter thin-walled vessels, the A positive bending moment found from these equations
most arduous conditions can occur duringfilling. is one causing tension at the lowest pointof the shell
However, the methods presented, based on the full cross section. The momentM4 may be positive in
condition, produce designs whichare satisfactory for vessels of large diameter with supports near the ends
the partially full condition. because of the effect of hydrostatic pressure (see
The included angle ofa saddle support(S in Sgure G.55).
figure G.&) should normally be withinthe When Ur and b h are known, these reduceto:
range 120’5 S 5 150”. This limitalion, which is
imposed by most codes of practice,is an empitical one
M3 = Wl(C1L - A) (G.3.33) I
based on experienceof large vessels. Saddleangles where
outside this range would require careful consideration. C1 is a factor obtainedh m figure G.56,and
When the supportsare near the endsof the vessel
(A Ir> the sWnesses of the ends tendto maintain
circular support crosssections and the shellis said to where C, and C3 are factms obtained from figure G.57.
be stiffened by the ends.
Similar expressions for the longitudinal bending
Where the stressesin the region of the supportare moments can be obtained bythe ordinary methods of
found to exceed the allowabled u e s a thickened statics for vesselsin which the supportsare not
stsake may be used The width of this should not be symmetricallyplaced
less than 5 rh about the centre saddle profile. That is,
a total length equal to, or greater than,the radius of 6.3.3.2.2 Longitudinal stresses at mid-span
the vessel see [B] and [39]. The resultant IongituM stresses at midispan due to
NOTE 1. In providing a thickened strake in the region of the pressure and bendingare given by the following
saddle it is assumed that the high stresses associated with the equatiOIlS
saddle have died away. The longitudinal and shear stresses at the
stepped down thicknessof the vessel may, therefore, be calculated at the highest point of the cross section
using KI a d Kz = 1.0 a d K3 = 0.319.
Although the valuesof the bending moment and shear forceat the
(G.3.3-5)I
stepped down thicknesswill be slightly less than M4
-
and W,[@ 2A)/(L + 4b/3)] respectively, it is recommended that at the lowest point ofthe cross section
I the full values of these are used in equations (G.3.3-7), (G.3.38)
and (G.3.3-9) with the values of the constants quoted above.
NOTE 2. A range of standardized saddle supports weldedto
(G.3.3-6) I
pressure vessels is included in (421. These equations are based on simple beam theory
6.3.3.2.1 Lollgitudinal bendingmoments which assumes that cross sections remain circular.
F’igure G.55shows the loads,redons and longitudid The cdcuIated tende and compressivesfmsses should
bending momentsin a vessel resting ontwo not exceed thevalues permitted inA3.4.2.1
symmetricxdlyplaced saddle supports. The bending and A.3.6.
moments are given by the following equations (see 1191
and [371):
at midispan
(G.3.3-1)
at supports
(G.3.3-2)
+)
26 2 4
'
Load unit Length = 2 w , =W
I L+- 4b
a
I J
l
I A I h
I \
c I \
I \
Hydrostatic"
loads on ends
-
L r
L+-4b
:
ur 12W= r w r
bb
L-
L+ -
3 3
1 I
I I
I I
Positive values of M4 are obtained for the following forms and proportions:
flat ends Alr < 0.707
ends with10 % knuckle radiusAlr 0.44
semiellipsoidal ends 2:l ratio Alr < 0.363
M4 is always negative for hemispherical ends,
The dimension 3bB is an approximation for the distance from the tangent plane to the centre of gravity of the
dished end and its contents for all ranges of dished end covered by this specification.
Figure G.66 Cylindrical shell acting as beam over supports
o BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Gff6
Licensed by Information Handling Services
STD=BSI P D 5500-ENGL 2000 m L h 2 4 b h 9 0832458 252 m
PD 55002000 Annex G
o. 5 0.5
0.45 0.45
consistent units
0.4 0.4
0.35 O.35
O. 3 0.3
L- LÏ-
u- %
O O
VI
5=I 0.25 0.25 5
d
d
d
>
O >
o. 2 0.2
0.15 0.15
0.1 0.1
0.05 o. o5
O O
1.0 2 .o 3.0 4.0 5.0 6.0 8.0 10 15 20
Q B I 11-1999 GI77
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI P D 5500-ENGL 2000 W Lb24bb70877353 T50 m
PD 66Wm Annex G
6.3.3.2.3 LmgitwLind stresses at the saddles I 'Igble 6.2 Design factors KI and K2 I
Longitudinal &esses at the saddles depend upon the
local stiffness of the shellin the planeof the supports
because, if the shell does notr e d round underload, Shell stiffened by end 120 1 1
a portion of the upper partof its Cros section, as or rings, ie. A S rB or 1%
shown diagrammaticauy in figure G.58, is ineffective 1 1
rings provided
against longitudinal bending (see[19]). 150 1 1
The resultant longitudinalstresses dueto pressure and Shell unsmened by 120 0.107 O. 192
weight should beevaluated at two positions as follows. end or rings, i.e. A > rE
135 0.234 0.132
a) Either and norings provided
150 0.279 0.161
1) at the highest point of the cross section when
the shellis stiffened by rings or by prordmityof
the ends, 6.3.3.2.4 lbngentid shring stresses:
ie.AsrB; Rqential shearing stsesses are given bythe following
equatiOIlS.
m
2) near theequator when the shellis unsti€fened The values ofK3, K4 and the allowable tangential
shearing stress values are given in table G.3.
In both cases1) and 2) the stress is given by: The thickness ofthe saddle plate should not be
(G.3.3-7) includedwhen using equations (G.3.3-9) to (G.3.311). I
a) Saddle not near vessel end(A > M),with or
b) At the lowest pointof the am section. without rings added
(G.3.3-8)
(G.3.39) I
Values of KI and K2 are given in tableG.2.
The thickness of the saddle plate should not be This equation does notapply when A > U4,but such
included in the equations. proportions areunusual.
The calculated tensile and compressivestresses should b) Saddle near vessel end (A 5 TB), without rings
not exceed the values permitted in A.3.4.2.1 added
and A.3.6. Inthiscasethereareshearingstmsesinboththe
shell and vesselend They are given by
2)intheendqe=-&W1
rte
(G.3.3-13) I
d) Saddle near vessel end(A 5 TB), with rings
e a c e n t to the saddle.
Table 6.3 Design factors lij and & and allowable tangential shearingstresses
Component
Vessel shellK3
Condition Saddle angle
0 (degrees) I Factor
A>?72 A s a
-
Shell unstiffened byrings 120 1.171 0.880
stiffened
Shell
by
plane of saddles
rings in I
0.319
10.319
0.319
1 0.319
I N m 2. The nominal maximurn tensile stress in head due to internal pressure,f,(d can be found from figure 3.62using appropriate
values of h,@ and e/D to give p(f and hence f,cd) = p / o where e is the vessel end analysis thickness.
Maximum bending
moment
=Me=K6W,r
-5I
Maximum bending
moment on each ring
stiffener
=Ma = K b W,r
n
where r) is the number
of stiffener5
6.3.3.2.6.1 shell not stiflèned by rings The stresses in the shellat the edge of the saddle plate
The circumferentialstresses are calculatedas follows. should
checked
be using equations (G.3.3-17) I
to (G.3.3-19). The saddle angle6 may now include the
a) At the lowest pointof the cross sectiom angle of the saddle plateup to but not exceeding+ 12".
-KSW1 (G.3.3-17) The value oft should be takenequal to the shell
f 5 =-
tb2 thickness; is assumed to be unchangd
b) At the horn of the saddle (see figureG.s.Qa): In,-S when an extended saddle plateof angle
L 2 (8 + 12') and width 2 b2 = bl + lot is used without
for - 2 8,then& =--- W' 3rq;w1 (G.3.3-18)
r 4tbz -2tz the use of a stiffening rhg, the stressesf5 and& are
L
for - c 8, then& =-- -
calculated using equations (G.3.317) to (G.3.319) as
(6.3.319) follows:
1
r 4th Lt
where b2 = bl + lot. a) At the edge of the saddle; using a thicknw equal
Unless the saddleis welded to the vesseI, values of K5 to (t + t l ) and constantsK5 and & based upon
should correspondto those given in table G.5 for rings upon an angle 8 and b2 = bl + lot. When the saddle
@scent to saddle. When the saddleis welded to the is welded to the vessel the allowable valueof f5 isf
vessel, K5 may be takenas onetenth of this value. and f6 is 1.26J where f is the minimum of the design
Values for & are given in table G.4. sbress values for the shell andfor the saddle plate.
These &eses may be reducedif necessary by b) At the edge of the extended saddle plate,using a
extending the saddle plateas shown in figure G.60. It is thickness t and constantsK5 and & based upon an
recommended that the thickness of the saddle platein angle 6 + 12" and bz = bl + lot. When the saddle is
this case should beequal to the thickness of the shell welded to the vessel thedowable value of f5 isf
Plate. andfi is 1.25J wheref is the design stress value for
the shell.
Ifthewidthofthisplateisnotlessthanb~andit
subtends an angle not less than (8 + 129, the reduced If the stresses are unacceptable then thewidth andor
stressesintheshellattheedgeofthesaddlecanbe the included angleof the saddle should be increased
obtained by substituting (t + tl), the combined and thecalculatons repeated, or alternatively provide
thickness of shell and saddle plate, for t in rings and carry out an analysis in accordance
I equations (G.3.317) to (G.3.3-19). b2 is assumed to be with 6.3.3.2.5.2.
unchanged.
!
Parts of saddle below
I
H = K, W, this line offer no
a) Simple saddle support
2
appreciable resistance
to force H
Saddie J
plate
"L ìL
b) Saddle support with extended plate
H = Kg W , Parts of saddle below
this line offer no
appreciable resistance
t o force H
W
a) Ring stiffener in plane of saddle
I I
thickness
L A J b,+IOt<xbr
b,+lOtSxSr
Q B S I 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI PD 5500-ENGL 2000 W Lb24bb9 0879357 479 m
PD 5600240oO Annex G
In summary, when an extended saddleplate of angle - In the stiflming ring centre p$Tl.e
L (B + 123 and width 2 b2 + 10t is used with C d c U h e f 7 and f8 using a thickness of t, and axial
stiffening rings d a c e n t to the saddle, as in figures length of shell of h + 10t for each ring for I, a,c
G.61(b) and (c), the &esses f5 and f6 are calculated and d and constants K7 and Kg based on an angle
using equations (G.3.3-M), (G.3.3-19) and (G.3.3-22) and of B + 12". The allowable values of f7 andf8 are
stressesf7 and f8 from equations (G.3.3-23) and 1.25J wheref is the minimum of the design stress
(G.3.3-24) as fOll0~:- values for the shell andfor the stiffening ring.
-Attheedgeofthesaddle a)Aringinthepluneofthesadde
Calculatef5 and f6 using a thickness (t + tl), At the horn of the saddle, in the shek
&, and K6 based upon an angle 8,with the
values & for A/r 5 0.50, and b2 = bl + lot. When
:.
(G.3.3-20) I
the saddle is welded to the vessel the allowable I
value of f5 isf and f6 is 1.25J where f is the At the horn of the saddle in the flange or tip of the
minimum of the design stress values for the shell ring remote from the shek P'
and for the saddle plate.
-Attheedgeoftheextmdedsa&dEeplate
c d C U h e f 5 andf6 Using a t h i C k I l e s S Of t, C O I l S h l t S b) Rings a&bxmt to the saddle
K5 and KGbased upon an angle 8 + 120, with the At the lowest point of the cross section:
value of & for A/r 5 0.50, and b2 = bl + lot. When
the saddle is welded to the vessel the allowable (G.3.3-22) I
value of f5 isf and f6 is 1.25f, where f is the
minimum of the design stress for the shell. Near the equator, in the shell:
h =c d I l I7 w l m a
K8Wl (G.3.3-23)
Near the equator, in the flange or tip of the ring
remote from the shell:
(G.3.3-24)
6.3.3.2.6 D&gn of saddles The longitudinal bending moments inthe shell andthe
The widthbl of steel saddles(seefigure G.54a) should corresponding skesses can be found inthe sameway
bechosentosatisfythecircumferential~limitsas
I defined by equations (G.3.3-17) to (G.3.3-24),
where
as for saddle supportsh m equations (G.3.31)
to (G.3.3-8).
The tangentialshear sbresses in the shella x e n t to
I
applicable. For mostcases a width equal to
(where D is the mean diameterof the vessel in mm) the ring support are given by:
will be saIisfa&xy.
The minimum section at the low point ofa saddle (see (G.3.3-26) I
figureG.60)hastoresistaforceHequaltothesumof
the horizontal componentsof the reactions on onehalf The allowable tangential shearingstress values are
of the saddle. The effectivecross sectionresisting this given in tableG.3.
load should belimited to the metalcross section
within a distance equal to v3 below the shell andthe The maximum circumferentialstress in the ring, due to
average direct stma on this cross section should be dead loads is given by:
limited to twdhirds of the allowable design stress. fio=TKlOWlrZ h W 1
+- U (G.3.326) I
H = &Wl
where Itcanbeassumedthatalengthofshelll(=fi
+ contacting widthof support) acts with thering
support to form a combined section andthat,72 is the
radius through the centroidof this section, 2 is the
I 0.231 I 0.259 I 0.288 I least section modulus andu is an effective area of the
section. The constantsKlo and K11 are found h m the
The upper and lower flanges of a steel saddle should table G.6.
be thick enough to resist the longitudinal bending over
the web or webs due to the bearing loads as in any hble G.6 Values of Kloand Kll
machine support. The web should be stiffened against
buckling dueto vertical shear forcesas for structural Klo r11
The stress in the ringfi0 should not exceedf.In the S i c e the thicknessof branch and shellwill usually be
case of category 1 and 2 vessels therings are in general dissimilar, there will be differential expansionof the
of the same material as the vessel and constructed to branch and shell during the transient, whichwill
the same categoryas the vessel withthefvalue produce additional discontinuityS.-
obtained from tables2.3-2 to 2.3-12. In the case of the A rigorous stress analysis would needthe use of finite
rings associated with category 3 vessels, it is element computer methods which, the in case of a
considered acceptableto use the corresponding branch on a cylindrical shell, would involvea complex
category 1and 2 vesselfvalues as given in tables2.3-2 threedimensional approach. It would be djfficult to be
to 2.3-12, providedthe radial weld seamsjoining the equally precise in specifying the heat tramfer rates
segments of the rings are located in the regionof low operating, which have beenshown experimentally to
bending stress in therings. The distributionof the vary considerably aroundthe circumference of
bending moment in a typical ring support is shown branches in cylinders. The cost of one.such rigorous
in (371. analysis would be prohibitive in mostcases and
Where the ring is made of a Merent material íi-om usually the designer will need to consider. . . several
that of the vessel, thef value for the weaker &r
i
a l transient operating conditions.
should beused For mild steelring girders used on Of more valuein general p m u r e vessel work are
category 3 vessels and not subjectto above ambient more simple methods which give realistically
temperatures, it is acceptable to use the allowable conservative maximum stress levels foruse in a fatigue
stsesses from (401. In this case thering should be assessment.
designed as a separatestructure w i k t the benefit of
the length of the shell. 6.4.2 Outline of the suggested design method
Unless a vessel withring supports worksat The method described in6.4.3 to 6.4.6 first uses well
atmospheric temperature and pressure,at least one known analytical methods for determining
ring support has to be provided with some formof through-thickness temperature distribution and stresses
sliding bearingat its connectionto the foundationor in the branch and shell material during a fluid
supporting structure. transient The average temperatureof each component
NOTE. The values ofKlo and KI, are derived from the absolute is then used ina thin-shell discontinuityanalysis at the
maximum circumferential moment and the a b s o l u t e maximum junction of branch and shell. Thetotal stress is taken
direct force in a ring support as shown in figure G.54b. The to be the sum of the temperature and discontinuity
influence of shear forces in the ring due to reactiom W1/2is not stress.
taken into account andthe designer should satisfy himself that the
ring section is sufficient in cmsectional area and lateral The solution yieldsa conservative estimate of the gross
stiffness to resist these forces. Itis not necessaryto take into section stresses from which the maxjmum equivalent
consideration secondary shellben- stresses induced by the sbress intensitycan be calculated In applying the
rigidity of, for example, a support ring, when evaluating, except
where fatigue is a governing criterion when the permissible stress results in a fatigue analysis, stress concentration
is a matter for individual consideration. factors would be applied to allow for the effectof
welds or local geomehy.
6.4 Simpliffed method for assessing transient
thermal stress at a pressure vessel nozzle Graphs and tablesare included which reducethe
overall solutionto the simple use of t h e d and stress
6.4.1 Introduction factors which are applied in a final set of stress
It is often necessaryto consider thestresses that will €"tiOns.
arise at the junction of a nozzle with a cylindrical or
spherical shell when the fluid contained in the vessel is
subject to a rise or fall in temperature. The value of
these stresses may decide the number of temperature
transients which can be accommodated withoutthe
risk of fatigue failure or, altematively,the stress levels
may dictate the rates of temperature variation which
can safely be permitted
During such variationin operating conditions, shell and
branch material will be subjectto stresses developed
by transient through-thickness temperature distribution
The intensity ofthese stsesses will be dependent upon
the rate of fluid temperature rise or fall, the surface
h&transfer coefficient and also upon the metal
thicknesses and properties.
e
is the radial discontinuity (see
equations (G.4.3-6) and (G.4.3-7));
is the time h m start of transient (in S);
I
are branch influence coefficients,
P is thedensity of the material (in kg4n3).
are shell influence coefficients; 6.4.3.2 Derivatùm of method
are stress fadors ( h m tables G.7, G.8, Consider a cylinder-tmphere assemblyas shown in
G.9, and G.lO); íigureG.62withafluidsubjecttoariseintemperature
on the inside. Assumethat heat transfer
is the specific heatof material coefficients(hand h.$ apply at the branch and shell
(
inJ&!.K)); inner surfaces. The fluid velocityin the branchwill
is the diffusivity of material (m m%) (see usuallybegreaterthanthatintheshellandhbmaybe
equation (G.4.36)); severaltimeslargerthan&.Duringaramprisein
is the modulusof elasticity ( i N/m2); temperature thetimetemperature behaviour of branch
shear force(in N); andshellmaterialwillbesimilartothatshownin
is the surface heattransfer coefficient Sgure G.63.
(in W/(m2.K)); Branch and shellmaterial away h m the discontinuity
is the surface heat transfer coefficient at w libe subject to thermal stress proportional to the
the branch inner surface (iin W/(m2.K)); difterence between the surfacetemperahre (G or To)
is the surface heat transfer coefficient at and the mean temperature(Td. These
the shd inner surface(in W/(m2.K)); through-thickness temperaturestresses will generally
be Merent in branch and shell.
is the conductivity of vesselmaterial
( nW/(m2-K));
i Solutions are given in [29] for stresses in a flat plate
subject to a ramp rise in fluid temperature at one
are branch thermalfactors ( h m surface.
Sgures G.64 and G.65);
Taking Poisson’s ratio as equal to 0.3, these solutions
are shell thermalfadom ( h m figures G.64 may be plottedin the form of figures G.64 and G.65,
and G.65); where
are branch and shell mean temperature Si = - KlEaTf (G.4.3-1) I
factors ( h m figure G.66);
is the mean temperatureMerence f a d q S, = K2EaTf (G.4.3-2) I
is a thermal parameter (see 6 and So are the thermal&eses at inner and outer
equation (G.4.3-4)); surfaces.
shear moment (inN-m); The values of K1 and K2 are plotted in figures G.64
is a thermal parameter (see
and G.65 against the parameters
de
equation (G.4.3-3));
N=F (G.4.33) I
is the mean radius of branch (ii m);
k
is the outerradius of branch ( i i m); m=-
ht
(G.4.3-4) I
is the inner radius of branch (in m);
is the mean radius of shell (in m); where
is stress (in N/m2) (seetext for specific d = Wcp (G.4.3-6) I
Symbow; Also, h m the solutions given in 111, curves may be
is the nominal branch thickness(in m) drawn as shown in figure G.66 which give the ratio of
(see 1.6); rise in mean metal temperatureto the rise in fluid
is the nominal shell thickness(in m) temperature (T.,,&) using the same parametersN
(see 1.6); and m.
is the fluid temperaturerise from start of Assuming that the thermal expansionof the branch
transient (in K); and the shell openingis proportional to the respective
is the inner surface temperature (in K); average metal temperatures, the radial discontinuity
introduced atthe junction would be
is the outer surface temperature (inK);
is the mean temperature (in K);
d = (Kb - Ks)arTf (G.4.3-6) I
is the discontinuity of edge rotation(see
where Kb and& are obtained from figure G.66.
equation (G.4.38)); In addition to the relative horizontal displacementof
the two parts, a rotational discontinuity (V) wi also be
l
produced by edge rotation of the shellopening.
(G.4.3-10)
o4
- Time
L- Time "--L
" "
0B SI11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
GEI
Licensed by Information Handling Services
STDmBSI PD 5500-ENGL 2000 9 l b 2 4 b b 9 0832472 b22
PD 6 5 0 0 . ~ Annex G
%III
N
L
GM O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI P D 5500-ENGL 2000 m lb24bb9 0832473 5b9 m
Annex G PD 56005HMO
O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI P D 5500-ENGL 2000 W Lb24bb708793b283b
Annex G PD 6600.~O00
6.4.4 lbtal stress equattons 6.4.6 Use and limitations @the method
6.4.4.1 Junction sEresses The M stress equations should providea designer
a) withasimplemeansofestimatingstressrangesina
branch dueto t h e d cycling. Although based upon
Inner suface the analysis of the rotationally symmetric
circumferential cylinder-here connection, theresults should be
(G.4.41) sufiïciently accurate for use in the fatigue assessment
of branches in cylindrical shells.
meridional The analysis for through-thickness temperaturestress
is based upon a "pate solution andis reasonably
(G.4.42) ac- for cylinders wherer& is less than about 1.1.
Outer surface For branches thickerthan this, the stress factor (kl)
may be multiplied by theratio r/q for a conservative
circumferential result
(G.4.4-3) The analysis for discontinuity stress w i give
l
consenative results since it neglects the effects of edge
meridional rotation produced byany axial temperature gradient
on the branch and by temperature gradientthrough the
(G.4.44) shell waU. Both effects would in practice tend to
b) Bmmh reduce the junction forces and moments.
Inner surface Such a simple eype of analysii cannot, of course,
circumferential predict the peakstresses which would occur due to
local changes in geometry at the junction and the
S'hi = [&(C1 + 0.3c2 - 1.0) - k&!i'aTf (G.4.4-5) designer would needto apply appmpriatestress
longitudinal concentration factors beforeapplying stress results in a
S'h = [&C2 - kl]EaTf (G.4.4-6) w e analysis.
Outer surface In practical use the tabulated stress factors Cl,C2,C3
circumferential and C4 w libe found to plot as fairly straight lines
S'ho = [h+ &(C1 - 0.3Cz - 1.O)IEaTf (G.4.47) against the various patarnetem, and interpolation for
longitudinal intermediate geometricratios can be made with
reasonable a~curacy.
SI10 = [h- &CzlEaTf (G.4.4-8)
where 6.4.6 Worked example
& = Kb - K, (read from figure G.66) Problem. A branch 300 mm mean diameter and50 mm
Ki,K2, kl, are temperaturefactorsfrom thick is welded to a steel vessel3 m diameter
figures G.64 and G.65 for shell(K) and branch (IC) and 100 mm thick The contained fluidis subject to a
ramp rise in temperature of 200 "Cin 10 min. The
Cl, C2, C3 are stress factors from tablesG.7, G.8 average heat transfer coefficients to shell and branch
and G.9. are estimated as 570 W/(m2.Q and2850 W/(m2.K)
The maximum equivalent stress intensity w l
l
iusually respectively. Calculate thermal stress in the assembly at
occur at the junction between branch and shell to the end of the transient.
which point the above stress equations refer. We:
hbximum bending stress in the branchmay occur at
k = 41.6 W/(m-Kj
a distance 0.62@ from the junction.At this point the
total thermal stsesses w ibe given byequations
l C = 420 J/(kg.K)
in 6.4.4.2. p=7700kgh3
6.4.4.2 h& sEresses
Inner surface
circumferential
I S",,, = [Kd(0.322(C1 - 1) t 0.192C2 - 0.3c1) - kll EaTf (G.4.49)
longitudinal
I s ' h = [&(0.644Cz - cd - kl] EaTf (G.4.410)
Outer surface
circumferential m =- k = 41.5 = 0.73
I .Tho = [Kd(0.322(C1- 1) + 0.3C1 - 0.192C2) + kx] EaTf (G.4.4-11) hT 570 X 0.1
longitudinal KI = 0.32 ( h m figure G.64)
I S'1, = [&(C4 - 0.64462) + h] EaTf (G.4.412) K.= 0.14 (from figure G.65)
where C4 is a stress factor from tableG.lO. K, = 0.30 (from figure G.66)
sheu:
Total Stress
stress (m/rn2)
factor
-0.157 -83.2
-0.296 -157.2
O. 248 131.4
-0.019 -9.9
Branch:
Total Stress
stress (Wrn2)
factor
S, = 0.52[0.26 + (0.3 X 0.7) - 1.01 - 0.17 = -0.446 -236.4
S', = (0.52 X 0.7) - 0.17 = O. 194 102.8
S'ho= 0.08 + 0.52 [0.26 - (0.3X 0.7) - 1.01 = -0.414 -219.4
S'la = 0.08 - (0.62 X 0.7) = -0.284 -150.5
0 B S I 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
G/95
Licensed by Information Handling Services
Table G.9 Meridional stress factor C3
RIT = 15
r/R 2=3 2=
24 = 2 2 = 5Z = 1.5 Z =1 2 = 0.66 2 = 0.5
0.05 0.07 0.09 o.12 O. 18 0.22 0.29 0.37 0.44
o.1 0.05 0.06 0.09 O. 13 0.16 0.21 0.29 0.35
0.03 0.2 0.04 0.06 0.08 O. 14 0.10 0.25 0.20
0.3 0.03 0.03 0.05 0.06 0.08 O. 15 0.11 0.19
0.4 0.02 0.03 0.04 0.05 0.09 0.06 o. 12 O. 15
0.5 0.02 0.03 0.02 0.04 0.07 0.05 o. 10 O. 13
RIT= 50
rlR 2=5 2=4 2=3 2=2 Z = 1.5 Z =1 2 = 0.66 2 = 0.5
0.05 0.04 o. 10 0.07 O.0.05
13 o. 18 0.25 0.31
o.1 0.03 0.03 0.05 0.07 0.09 o. 12 O. 18 0.23
0.03 0.02 0.02
0.04 0.2 0.06 0.08 o.12 O. 15
0.02 0.02 0.01 0.3 0.04 0.09 0.11
RIT = 100
0.01 0.5
0.01 0.01 0.4
0.02 0.01 1 0.03
0.03 I i:f 0.07
0.06
0.09
0.07
-+
0.4 0.95 0.88 O. 76 0.58 0.46 0.35 0.23 0.27
0.5 0.94
_ _ _ _ _ _ ~ ~ ~
~ ~~~~~
0.88
~ ~~~ ~
O. 76
~ ~
0.57 0.45 0.33 0.26 0.22
R/T= 100
rlR 2=5 2=4 2=3 2=2 Z = 1.5 2=1 2 = 0.66 2 = 0.5
0.05 1.01 0.95 0.84 0.67 0.56 0.44 0.35 0.30
o.1 0.99 0.92 0.81 0.64 0.53 0.42 0.35 0.30
0.2 0.97 0.90 0.79 0.61 0.50 0.39 0.32 0.27
0.97 0.3 0.89 0.77 0.59 0.48 0.37 0.28 0.25
0.4 0.94 0.89 O. 75 0.58 0.46 0.35 0.28 0.24
0.5 0.94 0.85 O. 75 0.57 0.45 0.34 0.26 0.22
GM O BSI 11-1999
COPYRIGHT British Standards Institute on ERC Specs and Standards
Licensed by Information Handling Services
STD-BSI P D 5500-ENGL 2000 m L b 2 4 b b 9 0832479 T87
Annex G PD 6 6 0 0 "
G.5 Bibliography
1. ESHBACH. Hanclbook of engàneeringfi~mtu ls.
Wiley.
2. TIMOSHENKO, S. í"gof elastic instability. 2nd ed. McGraw HiU, 1961.
3. DONNELL, L H. and WAN,C. C. Effect of imperfections on bucklug of thin cylinder underaxial compression.
J. appl. Mech. 1950, March
4. SEIDE, F! A x i m e t r i c a l buckling of circular cones under axial compression.J. a&. Mech. 1956,
December, 625.
5. HARW and LEXLAND. Conical vessels subjectto external pressure. %m. I. C'hm. E. 1952: 30,65 - 74.
6. SEIMON, K. Pressure vessel ?nunual. Edwards Bra., 1942
7. FREESE, C. E. Vibrations of vertical pressure vessels. J. E w . Ind. 1959, February.
8. BLJLAARD, F! F! Local stresse in spherical shells fromradial or moment l m . Weld.J. (Research
Supplement). 1957,May.
9. BIJLAARD, F! F! On the stresses from local loadson spherical pressure vessels and pressure vessel
heads. 1957. Weldmg Research Council BulletinNo. 34.
10. BLTLAARD, P. P Stresses from radial loads in cylindrical pressure vessels.Wé" J 19W, 33, 615s - 623s.
11. BLJLAARD, P.F! Stresses from radial loads and external momentsin cylindrical pressure vessels.Weld. J.
1965, December,608s - 617s.
12. HOFF, N. J., KEMF'NER, J., NARDO, S.V. and POHIE, E V. Defonnation and stresses in circular cylindrical
shells caused by pipe attachments. Part 1, Summary of investigation. Knolls Atomic Power Laboratory.
Schenectady. 1953.KAPL -921.
13. HOFF, N. J., KEMPNER, J., and POHLE, E V. Line load applied along generators of thin-walled circular
cylindrical shellsof finite length. Q.a&. Math, 1954: XI(4), 411- 425.
14. KEMF'NER, J.,SHENG,J., P O H L E , F .V. Tables and curvesfor deformation andstresses in circular cylindrical
shells under localized loadmgs.J. aermutut. Scà. 1957, Febmuy, 119 - 129.
15. SHOESSOW, G. J., and KOOISTRA, L. E Stresses in a cylindrical shell due to nozzle or pipe connection. %m.
A.S.M.E., 67,1945. A-107.
16. GARTNER, A I. Nomograms for the solution of anchor bolt problems.Petroleum &$iner 1951, July.
17. BIJLAARD, F! F! and CRANCH, E. T. Stsesses and deflections dueto local loadings on cylindrical shells.
W&. J. (Research Supplement). 1960, July.
18. WEIL, N. A and M U R P H Y , J. J. Design and analysis of welded pressure vesselskirt supports. J. Engng.Ind.
1960, Febnmy.
19. ZICK, L I? Stresses in large horizontal cylindricalpressure vessels on two saddle supports.W&. J. (Research
Supplement). 1951, September.
20. KETCHUM, M. S. 117te design of W&, bins and grain ekmtors. McGraw Hill, 1929.
21. FORBES, F! D. and TOO", A S. An analysis for twin saddle supported unstiffened cylindrical vessels. Joint
British C w m e m Stress Analysis, 1968.
22. BF¿ITISH STANDARDS INSllTUTION. A wuiew of th^ methods of 0alCulatin.g scresseS due to local loads and
local attudments of pmmm vessels. 1969. PD 6439.
23. BLJLAARD, F! P. Stsesses in spherical vessels fromradial loads and external momentsacting on a pipe. 1959.
Weld. Res. Colm. Bull. No. 49.
2 4 . BLJLAARD, F! F! Stsesses in spherical vessels from localloads transferred by a pipe. 1960.
Weld. Res. C o m Bull. No. 50.
25. LECKE, E A and PENNY, R. K. Solutions for the stresses in nozzles in pressure vessels. 1963.
Weld. Res. C o m BuU No. W.
26. RODABURGH, E. C., Wl'IT, E J. and CLOUD, R. L., Stresses at nozzles in spherical shells loaded with pressure,
moment andthrust. 1966. US.Atomic Energy CommissionPhase Report No. 2.
27. LECKE, E A and PENNY, R K. Shakedown loads forradial nozzles in spherical pressure vessels.
Inst. J. Solids and Structures. 1967 3,743.
28. WILSON, J. D. and T O O T H ,A S. The support of unstiffened cylindrical vessels.2nd Int. Co@ Pressure Vessel
Tmhml.ASME. 1973.
29. HEISLER, M. P. "kamient thermal&-esses in slabs and circular pressure vessels.J. appl. M A . 1953.
- I .
30. WICHMAN, K. R., HOPPER, k G. and MERSHON, J. L. LOA stresses in spherical and cylindrical shells due to
external loadings.1965. Weld. Res. Coun. Bull. No. 107.
31. ROSE, R.T. New design methods for pressure vessel nodes. The Enginew, 214, July 20,1962, p.90.
32. DUTI-IIE, G. and TOOTH, A S. The analysis of horizontal cylindrical vessels supported bysaddles welded to
the vessel - a comparison of theory and experiment.
3rd Int. Con$ IT V Tokyo 1977.
33. KANNAS, A, KITCHING, R. and GILL, S. S. A design procedure for pad reinforced flush nozzles in spherical
pressure vessels.Int. J.Pres. Ves.Piping, 1978,6,2.
34. ASHTON, J. N., McINTYRE, H. and GILL, S. S. A design procedure based on limit analysis for a pad reinforced
nozzle in a spherical pressure vessel.Int. J.Mech. sei., 1978, 20,747 - 757.
35. S O L I M A N , S. E and GILL, S. S. Stress concentration factors for integral and pad reinforced nodes in
spherical pressure vessels subjectedto radial load and moment.Int. J. h. Ves.P i p i n g , 1979, 7,275 - 307.
36. HSU,I. M., KETI’LEWELL, J. and GILL, S. S. Shear -1 of pad reinforced nozzlesin spherical pressure
vessels - a theoretical investigation.Int. J. Pres. V ? .Piping, 1980,8,461- 486.
37. BRITISH STANDARDS INSTITUTION. Stresses in horizontal cylind&d pressure WS& supported on twin
:-S a derivation of the basic equations and constants wed in G.3.3 of BS5500.1982. PD 6497.
38. DUTHIE, G., WHITE, G. C., TOOTH, A. S. An analysis for cyhdrical vessels under local loading-
an application to saddle supported vessel problems. J. Stmin A d p s i s 1982, 17,157 - 167.
39. T O O T H , A S., DUTHIE, G., WHITE, G. C., CARMICHAEL, J. Stresses in horizontal storage vessels-
a comparison of theory and experiment.J. S h i n A d u s i s . 1982, 17,169- 176.
40. BRITISH STANDARDS INSTITUTION. S’fkatim for Ute we of structural steel in building. BS 449.
41. INSTITUTE OF WELDING. Handbook for welded stmctuml st~dwmk.
42. BRITISH STANDARDS INSTITUTION. Fressure vessel details [dimensions). Part 2. S’fhtion for saddle
supportsfor horizontal cylindrical pressure vessels. 1983. BS 5276.
43. TEMERA, MCLEISH, G I L A simplified approachto c a l c w stresses due to radial loads and moments
applied to branches in cylindrical pressure vessels.J. Stmin A d u s i s , 1981, 16, No. 4.
44. KENDRICK, S., T O O T H , A. S. The behaviour of a horizontal vessel on loose saddles- a buckling assessment
of the support region. J. Stmin Analysis. 1986,21,45- 50.
45. NADARAJAH, C., TOOTH, k S. and SPENCE, J. The radial loading of cylindrical vessels- influence of
attachment rigidity. Int. J. Pres. Ves. & Piping, 1996,67,6579.
46. NADARAJAH, C., TOOTH, k S. and SPENCE, J. The radial loading of cylindrical vessels- influence of large
displacements.Int. J. Pres. Ves. & Piping, 1996, 67,81-94.