You are on page 1of 18

Inr. J. Pm. Ves.

& Pqing 72 (1997) l-18


0 1997 Elsevier Science Limited. All rights reserved
Printed in Northern Ireland
ELSEVfER PII:SO308-0161(97)00014-8 0308~0161/97/$17.00

Nozzles-on external loads and internal


pressure
C. J. Dekker
Continental Engineering B.V., Joan Muyskenweg 22, 1096 CJ Amsterdam, The Netherlands

&

H. J. Bos
Dynaflow Engineering B. V., Tintlaan 73, 2719 AH Zoetermeer, The Netherlands

(Received 30 January 1997; accepted11 February 1997)

Closecomparisonof local load stresscalculation methodsreveals considerable


differences. To investigate we performed many finite element analyses of
nozzles on cylinders concentrating not just on the shell stresses
but alsoon the
stressesin the nozzei wall. Local load stresseswere sometimesfound to be
much higher in the nozzle than in the shell. This led us to formulate a
‘modified improved shrink ring method’ and to devisemultiplication (contour-)
charts for deriving local load nozzle stressesfrom local load shell stresses.
Being important for a proper nozzle assessment, pressure induced stresses
were investigated too. This resulted in non-dimensionalparameter graphsto
determine pressureinduced stressesat nozzles. 0 1997Elsevier ScienceLtd.

NOMENCLATURE pressure is emphasized. This is mainly due to the


prominence that design codes place on this aspect. But
8 Radial thrust load on nozzle reactions from connected piping may give rise to high
MC Circumferential moment load on nozzle stresses too and these stresses are in addition to the
(out-of-plane bending) pressure induced stresses.
Ml Longitudinal moment load on nozzle The stresses due to external loads can be calculated
(in-plane bending) by various analytical methods, e.g. WRC-107,’
P Internal pressure Appendix G of BS 55002 and Wordsworth,3 but their
Y Mean radius of nozzle results may differ up to a factor of 2. For a thorough
Outside radius of nozzle, r, = r + it comparison of these methods see Dekker.4
s Mean radius of vessels To resolve the question which local load calculation
Ro Outside radius of vessel, R,, = R + :T method gives reliable stresses we made numerous
SCF Stress concentration factor finite element analyses of radially placed nozzles on
t Wall thickness of nozzle cylinders. Being important for a proper nozzle
Wall thickness of vessel assessment the pressure-induced stresses were in-
: Relative nozzle size with respect to vessel vestigated too.
size WRC definition: fi = 0.875 X rolR
Y Relative thinness of vessel, WRC defini-
tion: y = R/T
(WRC stands for WRC Bulletin 107, see 2 ASSESSMENT OF NOZZLES
references)
u Stress or stress intensity
2.1 General
1 INTRODUCTION

In the assessment of nozzles the weakening effect of The raison d’&re for vessels are nozzles, afterall who
nozzle openings on vessels with respect to internal needs vessels sealing forever their contents (nuclear
2 C. J. Dekker, H. J. Bos

waste industry?), and these nozzles experience two 2.2 Practical implementation
simultaneously occurring loadings:

Preferably one should superimpose the stress systems


-pressure
due to internal pressure and due to the various
-external loads due to piping reactions.
external load components and only then determine
the maximum stress intensity. However, such an
A proper stress assessment of a nozzle requires the approach is feasible only when using FEM-programs
superposition of the stress systems from both loadings. (Finite Element Method). Having no access to such
Then the maximum stress intensities for the various programs (or the time!) for such an analysis, then one
stress categories are to be determined and compared has to resort to more conventional methods.
with their specific stress limits. One is referred to e.g. To calculate the external load’s stress intensity one
Article 4.1 ‘Mandatory Design Based on Stress could use for example WRC-107’ or the methods
Analysis’ in Appendix 4 of ASME VIII, div. 2,’ given in G.2.2 and G.2.3 of BS 55002 (in Appendix G).
BS 55002 Appendix A ‘Recommendations for design Both WRC-107 as well as Appendix G of BS 5500
where loadings and components are not covered by distinguish between bending stresses and membrane
section three’ or sheet D 1200 ‘General Strength stresses and these stresses are given in 4 different
Assessment by Analysis’ from the Dutch ‘Rules for points, i.e. the two crown points and the two saddle
Pressure Vessels’.6 points.
The various stresses occurring at nozzle/vessel However, the stress distribution due to internal
junctions are to be categorized as follows: pressure at nozzle/vessel junctions is quite another
matter. The only method? known to the authors is
-membrane stresses (local!) due to internal Enquiry Case 5500/19 from BS 5500 and this method
pressure are primary stresses. gives only the maximum stress intensity due to
-the bending stresses due to internal pressure are internal pressure. No distinction in either stress type
secondary stresses. (membrane vs bending) or in position can be made.
-membrane stresses (local) due to external loads Remarkable is that BS 5500 restricts in clause
belong to the local primary membrane stress A.3.3.2 the sum of the stress intensity due to pressure
category whether the origin of the external (calculated in accordance with Enquiry Case 5500/19)
loading is mechanical or thermal. Though in the and the stress intensity due to external loads
latter case it has all the characteristics of a (calculated in accordance with G.2.2 and G.2.3) to
secondary stress. 2.25 X f with f being the basic design stress (note that
-bending stresses due to external loads belong 2.25 X f corresponds with 1.5 times the yield stress).
always to the secondary stress category. Normally the stress intensity of primary and secondary
-stress increments due to concentration in the stresses is limited by 3 xf (or 2 X yield stress). In
transition of vessel to nozzle proper belong to the contrast to BS 5500 the bulletin WRC-107 does
peak stress category and need to be considered neither mention any stress criterion nor provide any
only when a fatigue evaluation is required. The method to calculate the stress intensity due to
amount of stress increment is very much pressure.$
dependent on the actual weld geometry.
2.3 Improved shrink ring method
The stress intensity limit for primary stresses due to
pressure will be deemed to be satisfied by judiciously
following the applicable design code for the Predating the WRC-107 Bulletin and the BS 5500
considered vessel. Assuming that the bending stresses methods is the shrink ring method first published by
due to pressure and due to external loads to be at the MW Kellogg Company in their publication
least as large as the membrane stresses, then one need ‘Design of Piping Systems’.’ The major advantages are
only check the stress intensity due to the sum of
primary stresses and secondary stresses. When this t Admittedly, the Dutch ‘Rules for Pressure Vessels‘” given
latter stress requirement is fulfilled then the other in sheet D 1141--Appendix 1 is an approximate method for
stress requirement (primary membrane stress) is the maximum stress intensity due to pressure but the
fulfilled too. Note that the sum of primary stresses and method’s applicability is restricted.
secondary stresses criterion is to prevent low cycle $ Some people advocate to enter the longitudinal pressure
force in the cylindrical shell forming the nozzle (i.e. x
fatigue. (r(, - t)’ xp) as the radial nozzle load in a WRC-107
If indeed a large number of load cycles occur calculation. The resulting stress intensity from such a
then additionally a fatigue assessment is to be calculation is thought to represent the pressure induced
performed. stresses.
External loads and internal pressure in nozzles 3

its simplicity of use and its quickness for assessment 3 NUMERICAL VERIFICATION OF
calculations. There is no need to extract data from EXTERNAL LOADS
numerous graphs which would lead inevitably to
interpolation. 3.1 Introductory notes
However, comparing the original shrink ring
method with the ‘WRC-107’ calculated stress inten-
As mentioned in the Introduction the widely accepted
sities show that the stress intensity could be severely
external load calculation methods do result in quite
underestimated. In ‘External Loads on Nozzles” an
large differences, see Dekker.4 Trying to establish the
improved shrink ring method is proposed together
more reliable method, we embarked upon ‘finite
with a stress limit which takes into account the
internal pressure stresses. element method’ analyses. For these analyses we used
the program FE-Pipe,? a dedicated finite element
Improved shrink ring method: program for nozzles.
The external load on the nozzles always consisted of
VEE a single load component, i.e. either a radial thrust
u due to thrust = 4.5 ~ X thurst load, a circumferential moment or a longitudinal
2mOT
moment. The vessel, or better said the cylinder on
which the nozzles are placed, are either moderately
l/RITXM
u due to M, = 1.5~ 1 thick or relatively thin as in the experience of the
rcrgT authors thick-walled vessels never pose problems with
regard to local load stresses. The y-parameter being
l%E the ratio of the average cylinder radius and the wall
aduetoM,= l+?.05cT grit XM,
i thickness, is taken as 25 for moderately thick-walled
vessels and as 50 for thin-walled vessels.
with (T (thrust) + a(M,) + a(M,) c 1 Xf
3.2 Results of cases analyzed
Application range is for r,JR s 0.6 and 10 d (R/T) 8
100.
In ‘Proper Interface Design for Pressure Vessels” The nozzle geometries analyzed are listed with their
the improved shrink ring method is combined with the resulting stresses in tables, see the Appendix. For both
so called load fraction rule. The resulting formulation y-values a table is given for each specific t/T ratio.
gives the piping/vessel designer an even better insight With four different t/T values (i.e. 0.5, 0.75, 1.0 and
into the relative severity of each loading component: 1.5, respectively) this amounts to eight tables.
In these tables the maximum stress intensity in the
Fry act I MI, act MC,act cylinder proper due to the three different single load
Gl components, are given together with the load causing
F,, max M, , max + MC, max
these stress intensities.
with: In addition, the so called ‘back-calculated’ SCFs are
listed which were derived from the following
F,, act the actual thrust load on the nozzle. equations:
F,, max the maximum allowable thrust load as
calculated by means of the improved WTXF
shrink ring method and in the absence of S.I. due to thrust = SCF x -
2nr,T r
any other external load component.
M,, act the actual longitudinal moment loading on
S.I. due to M, = SCF x -
vRITx M
the nozzle. 1
nr:T
M, , max the maximum allowable longitudinal mo-
ment as calculated by means of the WTXM
S.I. due to MC= SCF X -
improved shrink ring method and in the m$T ’
absence of any other external load
component.
MC,act the actual circumferential moment loading 7 This is a proprietary program of the Paulin Research
on the nozzle. Group, Texas. It is a dedicated FE-program for pipe
MC, max the maximum allowable circumferential configurations, i.e. for many often occuring problems,
moment as calculated by means of the so-calledmesh-generatingtemplatesare provided. For radial
nozzles on cylinders the user can suffice with entering the
improved shrink ring method and in the major geometric data like radii and wall thicknessesof
absence of any other external load cylinder and nozzle, respectively and the program
component. automatically generatesthe finite element mesh.
4 C. J. Dekker, H. J. Bos

To compare with the ‘back-calculated’ SCFs are the 8,O


SCFs according to the improved shrink ring method.
These latter SCFs are either 4.5 (for thrust), 1.5 (for
longitudinal moments) or
7,O

( 1+l.o5&T
>
for circumferential moments, see Section 2.3 60 ‘MIST’

Last entries in the tables are the stress intensities in


the branch (or nozzle) itself and these S.1.s are
expressed as a percentage of the corresponding S.I. in
the vessel.
Obviously, the SCFs are functions of the specific
nozzle/vessel geometry characterized by the non-
dimensional geometry parameters /3 and y and the t/T
ratio. This enables the graphic display of the various
SCFs as functions of the p-parameter:
-The back-calculated SCFs from the finite element
analyses are plotted as distinct points. To show
more markedly the t/T dependency, curves are
drawn through ‘equal t/T’ points.
-The (continuous) SCF-curves as derived from the
WRC-107 bulletin, see Dekker.*
-The SCF functions from the improved shrink rink
method and the modified improved shrink ring

0 0,l 0,2 0.3 0,4 0,5 0,6 0,7


60
-P

Fig. 2. Stress intensity due to thrust for y = 50.


7,o

method are identified by ‘improved’ and ‘mist’


‘MIST respectively. For the modified improved shrink ring
690
method see Section 5.1.
Six separate graphs are given for thrust, longitudinal
5,O
moment and circumferential moment respectively at
both considered y-values (Figs l-6).
‘IMPROVED’

/ :
3.3 Stress intensities in nozzle necks
4,O /..a .+

Studying the tables with numerical results one learns


that the stress intensity in the nozzle neck proper may
3,O
differ considerably from the stress intensity in the shell
at externally loaded nozzles. For nozzle configurations
NRC-1 07 with t/T = 1.0 the stress intensities are about the
same, but for t/T < 1.0 the stress intensity in the
m
nozzle neck becomes larger than the corresponding
5 FE-Pipe results
stress intensity in the shell. For t/T > 1.0 the reverse
A
*for UT = 0.50 is true: the nozzle’s S.I. becomes less than the
14 ‘._. ..--
+for UT = 0.75 corresponding shell’s S.I.
*for VT = 1 .OO The reason for this is of course, that the
+for UT = 1.50 discontinuity bending moments in the nozzle neck and
co in the shell must be necessarily equal at the junction.
0 0,l 0,2 0,3 0,4 0,5 0,6 0,7 Note that this moment is not constant along the length
-P of the junction! The resulting bending stress from this
Fig. 1. Stress intensity due to thrust for y = 25 moment is inversely proportional to the square of the
External loads and internal pressure in nozzles 5

890

60 67’3

50 5,O

4,O 4.0

3,O 3,O

ZO 24

+for VT = 0.50 *for UT = 0.50


14 I,0 4for VT = 0.75
*form= 1.00 +forl/r= 1.00
--fort/l= 1.50 --fort/T= 1.50

w I w
0 0,1 02 0,3 0,4 03 0,6 0,7 0 0,l 0,2 0,3 0,4 0,5 0,6 0,7

-P -P

Fig. 3. Stress intensity due to circular moment for y = 25. Fig. 4. Stress intensity due to circular moment for y = 50.

local thickness (i.e. u = 6M/t2) and hence unequal extrapolate in a linear way between y = 2.5 and y = 50
thicknesses of the nozzle and of the shell cause the for other y-values.
overall stress intensities of the nozzle neck and of the
shell to differ quite substantially. Although not in
proportion to the square of the t/T ratio, as not all 4 NUMERICAL VERIFICATION OF
constituent stress components of the stress intensity INTERNAL PRESSURE STRESS
exhibit this ‘square t/T’ behaviour. 4.1 General
In order to quantify this potential large stress
intensity raising effect due to unequal nozzle/shell For a proper assessment of the stress intensity at a
thicknesses, we prepared contour charts of this factor nozzle/vessel junction the external load on the nozzle
as functions of the t/T ratio and the p-parameter is important, but so is the internal pressure. Afterall,
(Figs. 7-8). As these factors differ only slightly for the the stress intensity due to primary and secondary
three loading types, i.e. radial thrust, longitudinal stresses from both loadings together is limited to 3 X f
moment and circumferential moment, the contour (or twice the yield stress). Being able to assess the
charts give the average value of these three factors. stress intensity due to external loads is not enough. It
Hence the contour charts, given for y = 25 and y = 50, is vital to know too the stress intensity due to internal
respectively, are applicable to all three loading types. pressure.
Having established in one way or another the stress Analytical methods to calculate the stress intensity
intensity in the shell at the nozzle/shell junction of an due to internal pressure are ‘Enquiry Case No.
externally loaded nozzle, the stress intensity in the 5500/19’ from BS 55002 and the ‘approximate method’
nozzle wall proper can be obtained by multiplying the from the Dutch ‘Rules for Pressure Vessels’,6 sheet
former with the factor read from these contour charts. D 1141-Appendix 1. However, the method ‘Enquiry
Where necessary one could interpolate or even Case No. 5500/19’ is apparently regarded with some
6 C. J. Dekker, H. J. Bos

‘390

*for VT = 0.50

.-.m;:
:.-I+
+for VT = 0.75
7,O
*for f/T = 1 .OO *for UT = 1 .OO
+for VT = 1.50 +forVT= 1.50
I

690

.- .- .-..,.

: ‘IMPRQVED’ + ‘I.+1

0,3 0.4 05 0.6 0.7 0,3 OS4 0,5 096 0,7

Fig. 5. Stress intensity due to longitudinal moment for Fig. 6. Stress intensity due to longitudinal moment for
y = 25. y = 50.

suspicion because, as already mentioned in Section pressure as a function of the p-parameter (= relative
2.2, BS 5500 limits then the total stress intensity to nozzle size) and the t/T parameter (thickness ratio)
2.25 Xf instead of 3 Xf Comparing the resulting (Figs 9-12). As in general the pressure stress intensity
stress intensities from both methods revealed fairly absorbs the lion’s share of the available stress limit of
large differences of about 25%. 3 XL an error in this pressure stress intensity has a
All this combined with the results from some earlier more serious consequence than a similar (relative)
finite element analyses made us pursue internal error in the external load’s stress intensity. So it is
pressure in a similar systematic way with FE Pipe as imperative to estimate the pressure’s stress intensity as
was the case with external loadings. accurately as possible, hence the separate graphs for
y = 20, 30, 40 and 50, respectively.
4.2. Results and graphical representation The graphs are based on the highest occurring stress
intensity due to internal pressure in either the
cylindrical shell or in the nozzle wall. Reading off the
The results from the internal pressure calculations are appropriate SCF and multiplying this with p X R/T
presented in tables, see the Appendix. In these tables
gives directly the highest occurring stress intensity due
are listed the geometric data, resulting non- to pressure.
dimensional parameters, the actual stress intensity in
both the cylindrical shell as well as the nozzle (branch)
and an abstracted SCF factor. This abstracted SCF Remark
factor for internal pressure is derived from The SCFs given in these graphs can easily be
compared with the corresponding SCFs from both
SI. pressure = SCF,,,,,,,, X P.R
- mentioned analytical methods. Enquiry Case No.
T 5500119 from BS 5500 can be transformed into:
For each of the four different y (= R/T) values graphs
were made of these FEM-determined SCFs for internal
External loads and internal pressure in nozzles 7

096

w 0,2 0,3 0,4 0,5 0,6 0,7 038 099 1 lo1 182
0.2 0.3 0,4 0,5 0,6 0,7 0,8 0,9 1 I,1 1,2
-t/T -t/T

Fig. 7. Factor for nozzle stress, y = 25. Fig. 8. Factor for nozzle stress, y = 50.

where
equal in strength to the vessel material, then one finds:

-Using the graphs as presented here:


is a function depending on non-dimensional para- SCF,,,,,,re = 4.74
meters. The approximate method from the ‘Rules’ is: -Based upon enquiry case no. 5500119:
SCF,,,,,,,e = 3.40
2.5 p. R -Approximate method from the ‘Rules’:
~pressure =
[ 1
--
z T SC$m,,,re = 4.08
-Nozzle’s longitudinal pressure force as thrust load in
a WRC-107 calculation and transforming the found
where z is the strength reduction coefficient of the
nozzle opening to be determined in accordance with stress intensity into an SCF: SCF,,.,,,,,, = 4.91
chapter D 0501 of the ‘Rules’.
The expression between square brackets can be
interpreted as a stress concentration factor for the 5 DISCUSSION AND CONCLUSIONS
stress intensity due to internal pressure at nozzles.
5.1 MIST
P.R
S.1. pressure = SCF,,,,,,,, x -
T
Though we think that the performance of the
To illustrate the differences between these various improved shrink ring method is fair, certainly when
determined pressure SCFs, let us consider a vessel, one includes from the contour charts as given in 3.3,
O.D. = 2020 mm and T = 20 mm (i.e. y = 50), and a the factor for the stress intensity raising effect that
nozzle with O.D. nozzle = 685.7 mm and t = 20 mm thin nozzle walls have, it may underestimate
(i.e. p = 0.3). A ssuming the nozzle material to be nevertheless the stress intensity in the shell wall by
C. J. Dekker, H. J. Bos

IO,0

84

7,O

60 630

5,O 5,O

480 4,O

p 3,0

2,0 1

I,0 1,o
0,05 0,l 0,15 0,2 0,25 0,3 0,35 0,4 0,45 0,5 0,05 0,l 0,15 0,2 0,25 0,3 0,35 0,4 0,45 005
-P -P

Fig. 9. Maximum stress intensity due to pressure for Fig. 10. Maximum stress intensity due to pressure for
R/T = 20. R/T =30.

about 25%. In case one prefers to use a simple local 5.2 Proposed design and assessmentmethod for
load stress method which is conservative over its nozzles
entire range, then we recommend using the Modified Nozzles which will experience external loads from
Improved Shrink ring method (or MIST for short), connecting (process) piping should of course conform
which was devised just for that purpose: to the applicable design code.
Next, with the internal pressure SCF-graphs one
VEE should establish that the S.I. due to internal pressure
u due to thrust = 6.0 X ~ x thrust is not more than 2 Xf (twice the design stress) and so
2nroT
leaving a stress margin of 1 Xf for external load
stresses. Where necessary, nozzle and/or vessel
v’Ii7Tx, thicknesses are to be increased to achieve this.
u due to A4r = l-5 X- 1 Then once the external load on the nozzle is known
7i7$T
(by means of a pipe stress analysis or otherwise) the
stress intensity due to this load is to be assessed by
means of a suitable method, e.g. improved shrink ring
u due to M, = (1415m) X2 X M, method or the MIST-approach. Hereby one should
0
take into account the sometimes higher stress levels in
the nozzle necks, i.e. to include the multiplication
with u (thrust) + a(M1) + a(M,) s 1 Xflfactor factor from the contour charts. This external load’s
S.I. is not to exceed 1 Xf as this is the margin left by
setting the pressure’s S.I. at 2 Xc
and factor obtained from the contour charts as given
in 3.3, but not less than 1.0. Remark
Range of application of this MIST approach is It is possible to reserve a larger stress range than just
10 d y(= R/T) < 100 with r,/R 6 0.8 (or p G 0.7). 1 Xf for the external load’s S.I., but then one has to
External loads and internal pressure in nozzles 9

10,o

7,O

60 630

50 50

4,O 4,O

3,O 8rn 3,O I

zo

l,O
0,05
-P

Fig. 11. Maximum


0,l 0,15 0,2 0,25 0,3

stress intensity
RIT =4Q.
0,35 0,4

due to pressure
0,45 0,5

for
I1 2,0

18
0,05

Fig. 12. Maximum


-P
0,l 0,15 0,2 0,25

stress intensity
R/T = 50.
0,3 0,35 0,4 0,45

due to pressure
0,5

for

lower simultaneously the stress range for the internal improved shrink ring method made us formulate the
pressure’s S.I., as the limit for the combined S.I. is modified improved shrink ring method.
fixed at 3 X & In our experience the suggested 2 X ,f for -The stress intensity from local load stresses due to
the internal pressure is not a bad choice as the S.1.s of externally loaded nozzles is sometimes larger in the
realistic nozzle loads seldom exceed 1 Xc In this way wall of the nozzle than in the shell’s wall. This is the
the vessel design can be finalized before the pipe case when the thickness of the nozzle wall is less
stress analysis reveals the actual external load on the than the thickness of the shell. This stress raising
nozzle. In the rare cases that the nozzle load proves effect is more or less the same for thrusts,
too large then pipe lay-out changes have to solve the longitudinal moments and circumferential moments.
problem. To quantify this effect we devised contour charts for
this magnification factor which apply equally to all
5.3 Conclusions three load types. Note that these contour charts are
independent of the used local load stress calculation
-The local load stresses in cylindrical shells at nozzles method and could be used, if so desired, in
with external loads are dependent too on the ratio conjunction with e.g. the ‘Appendix G’ method of
of nozzle wall thickness and shell wall thickness. BS 5500. Needless to say that we recommend the far
The thicker the nozzle wall is in comparison with easier (modified) improved shrink ring method.
the shell wall, the more the nozzle behaves like a -The bending stresses are always larger than the
rigid insert and, consequently, locally steeper (local) membrane stresses for both internal pressure
deformation gradients in the shell wall occur and as well as external loads. Though the membrane
hence the larger the stresses in the shell wall are. stresses are not included separately in the tables
This effect is reflected neither in WRC-107 nor in with FE-Pipe results we do confirm the correctness
the improved shrink ring method. The sometimes of this assumption made in Section 2.1 for the whole
up to 25% underestimation of stresses by the range of nozzles considered here.
10 C. J. Dekker, H. J. Bos

-The stress intensity due to internal pressure is larger and vessel. The plant manager will call in a weld
than predicted by analytical methods like Enquiry specialist and that is quite understandable! However,
Case No. 5500/19 of BS 5500. Here too the as minor defects and blemishes are always present in a
maximum stress intensity in the nozzle wall is weld, the weld specialist is indeed able to pinpoint
sometimes higher than in the shell wall, this is the such a weld irregularity and blame it for having caused
case when the thickness of the nozzle wall is less the failure. After welding a new nozzle in the vessel,
than the thickness of the shell. As in general the the vessel will operate well for many years to come as
internal pressure induced S.I. takes up the larger the plant’s operators are experienced by now. The
part of the allowable stress range of 3 x f, it is smooth way of operating the plant results in few
imperative to use accurate pressure S.1.s. For this (start-stop) load cycles and low cycle fatigue does not
purpose we made graphs of the (maximum) SCF get a second chance.
due to internal pressure for y = 20, 30, 40 and 50, It is our opinion that many nozzle failures
respectively. masquerade as weld defects while in reality the nozzle
-Following our proposed design/assessment method design (or better said: the nozzle design method) is to
for nozzles, i.e. using the here-given pressure graphs blame. One would be wise to also consult a stress
and contour charts in conjunction with either the specialist at nozzle failures to establish the true cause
improved shrink ring method or the MIST- of the failure!
approach, will result in soundly designed nozzles.
Admittedly, our proposed design method is
ACKNOWLEDGEMENT
sometimes a bit conservative in that it adds up
algebraically the various load component’s maxi- Much of the here-presented material is from studies
mum stress intensities which do not necessarily occur undertaken in cooperation with the authors’ com-
at the same position, but the resulting nozzle design panies under commission by NAM (Assen, The
will be safe from low cycle fatigue. In the exceptional Netherlands). We would like to thank Mr W. J.
case that you are designing a pressure vessel from gold Stikvoort of NAM, business unit Groningen, for
then use FE-techniques in order to save as much allowing us to use this material and for his
material as possible, but otherwise use the relatively encouragement in preparing this paper.
simple and quick design method for nozzles as
proposed by us.
-As shown, the stresses at nozzle/vessel junctions REFERENCES
may be much higher than predicted by ‘conven- 1. Wichman, K. R., Hopper, A. G. and Mershon, J. L.,
tional’ methods. Especially due to the stress raising Local stressesin spherical and cylindrical shellsdue to
effect of a nozzle with a wall much thinner than the external loadings. WRC Bulletin 107/August 1965,
vessel wall. This applies both to pressure induced Revision March 1979.
2. BS 5500: 1991, Specification for Unfired Fusion Welded
stresses as well as to extenal load induced stresses. PressureVessels.British StandardsInstitution, London,
One might wonder why in the daily practice of plant 1991.
operating this has never been noticed through 3. Wordsworth, A.C., Stressesin cylindrical pressurevessels
nozzle failures due to overstressing. Well, in our due to local loads. In Structural Zntegrity Assessment, ed.
opinion the following explanation deals with that P. Standley. Elsevier Applied Science,London, 1992.
4. Dekker, C. J., Comparisonof local load stresscalculation
objection (plus operating conditions being less methods for nozzles on cylinders. Znt. J. Pres. Ves. &
severe than the design conditions and the yield Piping, 1994, 58, 203-213.
stresses of materials often being higher than their 5. ASME Boiler and Pressure Vessel Code, Section
specified minimum yield stresses). VIII-Division 2, 1995edition. The American Society of
Mechanical Engineers,New York, 1995.
In case such a vessel with a thin walled nozzle does 6. Rules for PressureVessels, Issue 96-02. Published on
not experience many pressure-cum-load cycles, then behalf of Stoomwezen B.V. bv SDU Publishers, The
failure need not occur as the failure mode is low cycle Hague.
fatigue for too large ‘primary-and-secondary’ stress I. The MW Kellogg Co., Design of Piping Systems, 2nd
edition. John W%.y, New York, 1956.
intensities. But if enough cycles occur to result in 8. Dekker, C. J., External loadson nozzles.Znt. J. Pres. Ves.
failure of the nozzle then the true nature of the failure & Piping, 1993, 53,335-350.
is often not recognized. At the point of failure (or 9. Stikvoort, W. J., Proper interface design for pressure
very near to it) will also be the weld between nozzle vessels. Chemical Engineering, 1994,133-134.

APPENDIX: TABLES WITH NUMERICAL RESULTS FROM FE-PIPE CALCULATIONS


Tables with the main results from our FE-Pipe analyses, for both external loads as well as internal pressure, are
included here for reference purposes.
There are eight external load tables and four internal pressure tables.
External loads and internal pressure in nozzles

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gamma ratio: 25.00 Mean shell radius: 400.00mm
t/T ratio: 0.50 Thicknessof shell: 16.00mm
Thicknessof nozzle: 8.00mm

Identification Resultsfrom ‘FE-pipe’ calculations SCF act. Stressin branch


improved
Load type Load SCF (back shrink S.I. Stressin
(kN, kNm) (;II’a) calculated) ring method (MPa) header (%)

beta 0.100 Thrust 50 171 3.143 4,500 211 123.4


r-0 [mm] 45.714 M-circ. 50 3997 1.697 1.600 5338 133.6
M-long. 50 3358 1.411 1.500 4254 126.7
beta 0.150 Thrust 50 135 3.723 4.500 187 138.5
r-o [mm] 68.571 M-circ. 50 2471 2.336 1.900 3606 145.9
M-long. 50 1681 1.589 1.500 2354 140.0
beta 0.200 Thrust 50 103 3.787 4.500 176 170.9
r-o [mm] 91.429 M-circ. 50 1647 2.768 2.200 2852 173.2
M-long. 50 949 1.595 1.500 1428 150.5
beta 0.250 Thrust 50 80 3.677 4.500 14.5 181.3
r-o [mm] 114.286 M-circ. 50 1134 2.987 2.500 2078 183.2
M-long. 50 617 1.620 1.500 979 158.7
beta 0.350 Thrust 50 54 3.474 4.500 116 214.8
r-0 [mm] 160.000 M-circ. 50 622 3.202 3.100 1375 221.1
M-long. 50 303 l-560 1.500 516 170.3

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gammaratio: 25.00 Mean shellradius: 400.00mm
t/T ratio: 0.75 Thicknessof shell: 16.00mm
Thicknessof nozzle: 12.0mm
Identification Resultsfrom ‘FE-pipe’ calculations SCF act. Stressin branch
improved
Load type Load SCF (back shrink Stressin
(kN, kNm) (;;a) calculated) ring method (;:a) header(%)

beta 0.250 Thrust 50 81 3.723 4500 113 139.5


r-0 [mm] 114.286 M-circ. 50 989 2.597 2.500 1452 146.8
M-long. 50 511 1,342 1.500 657 128.6
beta 0.350 Thrust 50 59 3.796 4.500 85 144.1
r-0 [mm] 160.000 M-circ. 50 614 3.160 3.100 927 151.0
M-long. 50 254 1.307 1.500 349 137.4
beta 0.450 Thrust 50 44 3.640 4.500 65 147.7
r-0 [mm] 205.714 M-circ. 50 405 3.446 3.700 623 153.8
M-long. 50 147 1.251 1.500 213 144.9
beta 0.550 Thrust 50 34 3.438 4.500 50 147.1
r-0 [mm] 251.429 M-circ. 50 276 3.508 4.300 428 155.1
M-long. 50 93 1.182 1.500 138 148.4
2 C. J. Dekker, H. J. Bos

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gammaratio: 25.00 Mean shellradius: 400.00mm
t/T ratio: 1.00 Thicknessof shell: 16.00mm
Thicknessof nozzle: 16.00mm
Identification Resultsfrom ‘FE-pipe’ calculations SCF act. Stressin branch
improved
Load type Load SCF (back shrink Stressin
(kN, kNm) (I’&I’a) calculated) ring method (;:a) header(%)
beta 0.200 Thrust 50 110 4.044 4.500 97 88.2
r-o [mm] 91.429 M-circ. 50 1415 2.378 2.200 1309 92.5
M-long. 50 801 1.346 1.500 651 81.3
beta 0.250 Thrust 50 93 4.274 4.500 79 84.9
r-o [mm] 114.286 M-circ. 50 1076 2.826 2.500 968 90.0
M-long. 50 535 1.405 1.500 448 83.7
beta 0.350 Thrust 50 67 4.311 4.500 60 89.6
r-o [mm] 160.000 M-circ. 50 671 3.454 3.100 628 93.6
M-long. 50 267 1.374 1.500 233 87.3
beta 0.450 Thrust 50 51 4.219 4.500 46 90.2
r-o [mm] 205.714 M-circ. 50 454 3.863 3.700 426 93.8
M-long. 50 154 1.310 1.500 144 93,5
beta 0.550 Thrust 50 38 3.842 4.500 34 89.5
r-o [mm] 251.429 M-circ. 50 306 3.889 4.300 281 91.8
M-long. 50 98 1.246 1.500 93 94.9
beta 0.700 Thrust 50 28 3603 4.500 27 96.4
r-o [mm] 320.000 M-circ. 50 190 3.912 5.200 183 96.3
M-long. 50 63 1.297 1.500 63 100.0

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gammaratio: 25.00 Mean shell radius: 400.00mm
t/T ratio: 1.50 Thicknessof shell: 16.00mm
Thicknessof nozzle: 24.00mm

Identification Resultsfrom ‘FE-pipe’ calculations SCFact. Stressin branch


improved
Load type Load SCF (back shrink Stressin
(kN, kNm) (;:a) calculated) ring method (&;a) header (%)
beta 0.200 Thrust 50 117 4.302 4.500 54 46.2
r-o [mm] 91.429 M-circ. 50 1412 2.373 2.200 691 48.9
M-long. 50 789 1.326 1.500 356 45.1
beta 0.250 Thrust 50 98 4.504 4.500 45 45.9
r-o [mm] 114.286 M-circ. 50 1085 2.849 2.500 520 47.9
M-long. 50 533 1.400 1.500 250 46.9
beta O-350 Thrust 50 72 4.632 4.500 34 47-2
r-o [mm] 160.000 M-circ. 50 707 3.639 3,100 339 47.9
M-long. 50 274 1.410 1.500 133 48.5
beta 0.450 Thrust 50 54 4.467 4.500 26 48.1
r-o [mm] 205.714 M-circ. 50 475 4,042 3.700 232 48.8
M-long. 50 156 1.327 1.500 83 53.2
beta 0.550 Thrust 50 41 4.145 4.500 19 46.3
r-o [mm] 251.429 M-circ. 50 326 4.144 4.300 157 48.2
M-long. 50 103 1.309 1.500 55 53.4
beta 0.700 Thrust 50 33 4.246 4.500 16 48.5
r-o [mm] 320.000 M-circ. 50 221 4.550 5.200 106 48.0
M-long. 50 71 1.462 1.500 40 56.3
External loads and internal pressure in nozzles 13

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gammaratio: 50.00 .Mean shellradius: 800.00mm
t/T ratio: 0.50 Thicknessof shell: 16.00 mm
Thicknessof nozzle: 8.00mm

Identification Resultsfrom ‘FE-pipe’ calculations SCF act. Stressin branch


improved
Load type Load SCF (back shrink S.I. Stressin
(kN, kNm) (;;a) calculated) ring method (MPa) header (%)

beta 0.100 Thrust 50 142 3.692 4.500 231 162.7


r-o [mm] 91,429 M-circ. 50 1654 1.966 1.849 2917 176.4
M-long. 50 1269 1.508 1.500 1968 155.1
beta 0.150 Thrust 50 101 3.939 4.500 202 200.0
r-o [mm] 137.143 M-circ. 50 980 2.621 2.273 2014 205.5
M-long. 50 584 1.562 1.500 988 169.2
beta 0.195 Thrust 50 76 3.853 4500 163 214.5
r-o [mm] 178.286 M-circ. 50 620 2.802 2.655 1398 225-5
M-long. 50 337 1.523 1.500 620 184.0
beta 0.250 Thrust 50 56 3.640 4.500 139 248.2
r-0 [mm] 228.571 M-circ. 50 410 3.045 3.121 1049 255.9
M-long. 50 215 1.597 1.500 380 176.7
beta 0.350 Thrust 50 40 3.640 4.500 101 252.5
r-o [mm] 320.00 M-circ. 50 230 3.348 3.970 647 281.3
M-long. 50 105 1.529 1.500 207 197.1

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gammaratio: 50.00 Mean shell radius: 800.00 mm
t/T ratio: 0.75 Thicknessof shell: 16.00mm
Thicknessof nozzle: 12.00mm

Identification Resultsfrom ‘FE-pipe’ calculations SCFact. Stressin branch


improved
Load type Load SCF (back shrink S.I. Stressin
(kN, kNm) (;;a) calculated) ring method NW header(%)
beta 0.250 Thrust 50 70 4.550 4.500 101 144.3
r-o [mm] 228.571 M-circ. 50 468 3.476 3.121 697 148-9
M-long. 50 194 1.441 1.500 257 132.5
beta 0.350 Thrust 50 50 4.550 4.500 72 144.0
r-0 [mm] 320.000 M-circ. 50 286 4.164 3.970 427 149.3
M-long. 50 95 1.383 1.500 136 143.2
beta 0.450 Thrust 50 36 4.212 '4.500 52 144.4
r-o [mm] 411,429 M-circ. 50 184 4.428 4.818 275 149.5
M-long. 50 57 1.372 l-500 80 140.4
beta 0.550 Thrust 50 27 3.861 4.500 39 144.4
r-o [mm] 502.857 M-circ. 50 122 4.386 5.667 182 149.2
M-long. 50 37 1,330 1.500 53 143.2
14 C. J. Dekker, H. J. Bos

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gammaratio: SO-00 Mean shellradius: 800.00mm
t/T ratio: 1.00 Thicknessof shell: 16.00mm
Thicknessof nozzle: 16.00mm
Identification Resultsfrom ‘FE-pipe’ calculations SCF act. Stressin branch
improved
Load type Load SCF (back shrink Stressin
(kN, kNm) (;;a) calculated) ring method (;;a) header(%)
beta 0.200 Thrust 50 99 5.147 4.500 82 82.8
r-0 [mm] 182.857 M-circ. 50 706 3.356 2.697 616 87.3
M-long. 50 322 1.531 1.500 257 79.8
beta 0.250 Thrust 50 80 5.199 4.500 69 86.3
r-o [mm] 228.571 M-circ. 50 512 3.803 3.121 461 90.0
M-long. 50 203 1.508 1.500 173 85.2
beta 0.350 Thrust 50 56 5.095 4.500 49 87.5
r-o [mm] 320.000 M-circ. 50 313 4.557 3.970 285 91.1
M-long. 50 103 1.500 1.500 90 87.4
beta 0.450 Thrust 50 41 4.796 4.500 34 82.9
r-0 [mm] 411.429 M-circ. 50 205 4.934 4.818 176 85.9
M-long. 50 60 1.444 1.500 53 88.3
beta 0.550 Thrust 50 31 4.433 4.500 24 77.4
r-0 [mm] 502.857 M-circ. 50 138 4.961 5.667 111 80.4
M-long. 50 38 1.366 1.500 34 89.5
beta 0.700 Thrust 50 22 4.004 4.500 18 81.8
r-0 [mm] 640.000 M-circ. 50 79 4.600 6-940 68 86.1
M-long. 50 25 1.456 1.500 22 88.0

Local load stresses, i.e. FE-pipe versus the improved shrink ring method
Gammaratio: 50.00 Mean shell radius: 800.00mm
t/T ratio: 1.50 Thicknessof shell: 16.00mm
Thicknessof nozzle: 24.00mm

Identification Resultsfrom ‘FE-pipe’ calculations SCFact. Stressin branch


improved
Load type Load SCF (back shrink Stressin
(kN, kNm) (;:a) calculated) ring method (&!‘a) header (%)

beta 0.200 Thrust 50 107 5.563 4.500 44 41.1


r-o [mm] 182.857 M-circ. 50 735 3.494 2.697 322 43.8
M-long. 50 327 1.554 1.500 140 42.8
beta 0.250 Thrust 50 87 5.654 4.500 37 42.5
r-o [mm] 228.571 M-circ. 50 545 4.048 3.121 241 44.2
M-long. 50 212 1.575 1.500 93 43.9
beta 0.350 Thrust 50 61 5.550 4.500 27 44.3
r-o [mm] 320.000 M-circ. 50 335 4.877 3.970 152 45.4
M-long. 50 107 1.558 1.500 48 44.9
beta 0.450 Thrust 50 45 5.264 4.500 18 40.0
r-o [mm] 411.429 M-circ. 50 220 5.295 4.818 94 42.7
M-long. 50 62 1.492 1.500 31 50.0
beta 0.550 Thrust 50 34 4.861 4.500 13 38.2
r-o [mm] 502-857 M-circ. 50 149 5.357 5.667 61 40.9
M-long. 50 42 1.510 1.500 21 50.0
beta 0.700 Thrust 50 25 4.550 4.500 10 40.0
r-o [mm] 640+000 M-circ. 50 92 5.358 6.940 38 41.3
M-long. 50 27 1.572 1.500 13 48.1
External loads and internal pressure in nozzles 15

Nozzles in cylindrical shells under internal pressure


R/T-parameter = 20
Internal pressure = 10 MPa
All dimensions in table below are in [mm] and all stresses are given in [MPa]

Nozzle dimensions Vessel dimensions Geometry parameters FE-pipe results Maximum


stress
O.D. t O.D. T dlD beta tlT Stress in SCF back Stress in Percentage intensity
header calculated branch of stress as SCF
in header

13.14 3.20 658.0 16.0 0.1089 0.0997 0.20 546 2.721 395 72.3 2.721
109.72 3.20 658.0 16.0 0.1659 0.1495 0.20 674 3.360 538 79.8 3.360
182.86 3.20 658.0 16.0 0.2798 0.2492 0.20 966 4.815 875 90.6 4.815
256.00 3.20 658.0 16.0 0.3938 0.3489 0.20 1219 6.076 1249 102.5 6.226
329.14 3.20 658.0 16.0 0.5077 0.4486 0.20 1447 7.212 1586 109.6 7.905
13.14 5.60 658.0 16.0 0.1052 0.0997 0.35 490 2.442 313 76.1 2.442
109.72 5.60 658.0 16.0 0.1622 0.1495 0.35 562 2.801 471 83.8 2.801
182.86 5.60 658.0 16.0 0.2761 0.2492 0.35 793 3.953 805 101.5 4.012
256.00 5.60 658.0 16.0 0.3900 0.3489 0.35 1009 5.029 1150 114.0 5.732
329.14 5.60 658.0 16.0 0.5040 0.4486 0.35 1202 5.991 1453 120.9 7.242
73.14 8.00 658.0 16.0 0.1015 0.0997 0.50 447 2.228 356 19.6 2.228
109.72 8.00 658.0 16.0 0.1584 0.1495 0.50 507 2.527 430 84.8 2.521
182.86 8.00 658.0 16.0 0.2724 0.2492 0.50 681 3.394 115 105.0 3.564
256.00 8.00 658.0 16.0 0.3863 0.3489 0.50 856 4.267 1014 118.5 5.054
329.14 8.00 658.0 16.0 0.5002 0.4486 0.50 1029 5.129 1279 124.3 6.375
13.14 12.00 658.0 16.0 0.0952 0.0997 0.75 396 1.974 337 85.1 1.974
109.72 12.00 658.0 16.0 0.1522 0.1495 0.75 435 2.168 398 91.5 2.168
182.86 12.00 658.0 16.0 0.2661 0.2492 0.75 582 2.901 544 93.5 2.901
256.00 12.00 658.0 16.0 0.3801 0.3489 0.75 739 3.683 755 102.2 3.763
329.14 12.00 658.0 16.0 0.4940 0.4486 0.75 891 4.441 943 105.8 4.700
73.14 16.00 658.0 16.0 0.0890 0.0997 1.00 356 1.774 317 89.0 I.774
109.72 16.00 658.0 16.0 0.1460 0.1495 1.00 393 1.959 380 96.7 1.959
182.86 16.00 658.0 16.0 0.2599 0.2492 1.00 483 2.407 458 94.8 2.407
256.00 16.00 658.0 16.0 0.3738 0.3489 1.00 625 3.115 585 93.6 3.115
329.14 16.00 658.0 16.0 0.4878 0.4486 1.00 761 3.794 714 93.8 3.793
73.14 20.00 658.0 16.0 0.0828 0.0997 l-25 320 1.595 293 91.6 1.595
109.72 20.00 658.0 16.0 0.1398 0.1495 1.25 356 1.774 362 101.7 1.804
182.86 20.00 658.0 16.0 0.2537 0.2492 1.25 405 2.019 435 107.4 2.168
256.00 20.00 658.0 16.0 0.3676 0.3489 1.25 516 2.572 492 95.3 2.572
329.14 20.00 658.0 16.0 0.4815 0.4486 1.25 634 3.160 580 91.5 3.160
73.14 24.00 658.0 16.0 0.0765 0.0997 1.50 294 1.465 267 90.8 1.465
109.72 24.00 658.0 16.0 0.1335 0.1495 1.50 330 1.645 340 103.0 1.695
182.86 24.00 658.0 16.0 0.2474 0.2492 1.50 368 1.834 414 112.5 2.064
256.00 24.00 658.0 16.0 0.3614 0.3489 1.50 421 2.128 466 109.1 2.323
329.14 24.00 65X.0 16.0 0.4753 0.4486 1.50 524 2.612 514 98.1 2.612
73.14 32.00 658.0 16.0 0.0641 0.0997 2.00 258 1.286 214 82.9 1.286
109.72 32.00 658.0 16.0 0.1211 0.1495 2.00 291 1.450 295 101.4 1.470
182.86 32.00 658.0 16.0 0.2350 0.2492 2.00 336 1.675 369 109.8 1.839
256.00 32.00 658.0 16.0 0.3489 0.3489 2.00 367 1.829 418 113.9 2.083
329.14 32.00 658.0 16.0 0.4628 0.4486 2.00 385 1.919 460 119.5 2.293
16 C. J. Dekker, H. J. Bos

Nozzles in cylindrical shells under internal pressure


R/T-parameter = 30
Internal pressure = 10 MPa
All dimensions in table below are in [mm] and all stresses are given in [MPa]

Nozzle dimensions Vessel dimensions Geometry parameters FE-pipe results Maximum


stress
O.D. t O.D. T dlD beta t/T Stress in SCF back Stress in Percentage intensity
header calculated branch of stress as SCF
in header

109.72 3.20 976.0 16.0 0.1110 0~1000 0.20 942 3.140 706 74.9 3.140
164.58 3.20 976.0 16.0 0.1681 O~lSOO 0.20 1213 4.043 990 81.6 4.043
274.28 3.20 976.0 16.0 0.2824 0.2500 0.20 1759 5.863 1773 100.8 5.910
384.00 3.20 976.0 16.0 0.3967 0.3500 0.20 2204 7.347 2422 109.9 8.073
493.72 3.20 976.0 16.0 0.5110 0.4500 0.20 2613 8.710 3006 115.0 10.020
109.72 5.60 976.0 16.0 0.1085 0~1000 0.35 772 2.573 635 82.3 2.573
164.58 5.60 976.0 16.0 0.1656 0.1500 0.35 1000 3.333 981 98.1 3.333
274.28 5.60 976.0 16.0 0.2799 0.2500 0.35 1450 4.833 1758 121.2 5.860
384.00 5.60 976.0 16.0 0.3942 0.3500 0.35 1822 6.073 2394 131.4 7.980
493.72 5.60 976.0 16.0 0.5085 0.4500 0.35 2159 7.197 2953 136.8 9.843
109.72 8.00 976.0 16.0 0.1060 0~1000 0.50 712 2.373 591 83.0 2.373
164.58 8.00 976.0 16.0 0.1631 0.1500 0.50 871 2,903 869 99.8 2.903
274.28 8.00 976.0 16.0 0.2774 0.2500 0.50 1212 4.040 1517 125.2 5-057
384.00 8-00 976.0 16.0 0.3917 0.3500 0.50 1531 5.103 2059 134.5 6-863
493.72 8.00 976.0 16.0 0.5060 0.4500 0.50 1839 6.130 2536 137.9 8.453
109.72 12.00 976.0 16.0 0.1018 0~1000 0.75 616 2.053 560 90.9 2.053
164.58 12.00 976.0 16.0 0.1589 0.1500 0.75 758 2.527 669 88.3 2.521
274.28 12.00 976.0 16.0 0.2732 0.2500 0.75 1075 3.583 1110 103.3 3.700
384.00 12.00 976.0 16.0 0.3875 0.3500 0.75 1370 4.567 1486 108.5 4.953
493.72 12.00 976.0 16.0 0.5018 0.4500 0.75 1644 5.480 1818 110.6 6.060

109.72 16.00 976.0 16.0 0.0976 0~1000 1.00 562 1.873 541 96.3 1.873
164.58 16.00 976.0 16.0 0.1548 0.1500 1.00 646 2.153 608 94.1 2.153
274.28 16.00 976.0 16.0 0.2690 0.2500 1.00 931 3.103 853 91.6 3.103
384.00 16.00 976.0 16.0 0.3833 0.3500 1.00 1203 4.010 1113 92.5 4.010
493.72 16.00 976.0 16.0 0.4976 0.4500 1.00 1454 4,847 1344 92.4 4.847

109.72 20.00 976-O 16.0 0.0935 0~1000 1.25 510 1.700 519 101.8 1.730
164.58 20.00 976.0 16.0 0.1506 0.1500 1.25 551 1.837 585 106.2 1.950
274.28 20.00 976.0 16.0 0.2649 0.2500 1.25 785 2.617 704 89.7 2.617
384.00 20.00 976.0 16.0 0.3792 0.3500 1.25 1028 3.427 891 86.7 3.421
493,72 20.00 976.0 16.0 0.4935 0.4500 1.25 1253 4.177 1056 84.3 4.177

109.72 24.00 976.0 16.0 0.0893 0~1000 1.50 473 1.577 492 104.0 1.640
164.58 24.00 976.0 16.0 0.1464 0.1500 1.50 512 1.707 560 109.4 1.867
274.28 24.00 976.0 16.0 0.2607 0.2500 1.50 654 2.180 652 99.7 2.180
384.00 24.00 976.0 16.0 0.3750 0.3500 1.50 864 2.880 753 87.2 2.880
493.72 24.00 976.0 16.0 0.4893 0.4500 1.50 1066 3.553 877 82.3 3.553

109.72 32.00 976.0 16.0 0.0810 0~1000 2.00 419 1.397 430 102.6 1.433
164.58 32-00 976.0 16.0 0.1381 0.1500 2.00 458 1.527 504 110.0 1.680
274.28 32-00 976.0 16.0 0.2524 0.2500 2.00 523 1.743 589 112.6 1.963
384.00 32.00 976.0 16.0 0.3667 0.3500 2.00 643 2.143 657 102.2 2.190
493.72 32.00 976.0 16.0 0.4810 0.4500 2.00 760 2.533 718 94.5 2.533
External loads and internal pressure in nozzles 17

Nozzles in cylindrical shells under internal pressure


R/T-parameter = 40
Internal pressure = 10 MPa
AI1 dimensions in table below are in [mm] and all stresses are given in [MPa]

Nozzle dimensions Vessel dimensions Geometry parameters FE-pipe results Maximum


stress
O.D. t O.D. T d/D beta tlT Stress in SCF back Stress in Percentage intensity
header calculated branch of stress as SCF
in header

146.28 3.20 1296.0 16.0 0.1118 0~1000 0.20 1414 3.535 1054 74.5 3.535
219.42 3.20 1296.0 16.0 0.1689 0.1500 0.20 1832 4.580 1660 90.6 4.580
365.72 3.20 1296.0 16.0 0.2832 0.2500 0.20 2655 6.638 2842 107.0 7.105
512.00 3.20 1296.0 16.0 0.3975 0.3500 0.20 3314 8.285 3791 114.4 9.478
658.28 3.20 1296.0 16.0 0.5118 0.4500 0.20 3940 9.850 4663 118.4 11.658
146-28 5.60 1296.0 16.0 O-1099 0~1000 0.35 1171 2.928 1081 92.3 2,928
219.42 5.60 1296.0 16.0 0.1670 0.1500 0.35 1520 3.800 1712 112.6 4.280
365.72 5.60 1296.0 16.0 0.2813 0.2500 0.35 2191 5.478 2935 134.0 7.338
512.00 5.60 1296.0 16.0 0.3956 0.3500 0.35 2736 6.840 3896 142.4 9.740
658.28 5.60 1296.0 16.0 0.5099 0.4500 0.35 3244 8.110 4756 146.6 11.890
146.28 8.00 1296.0 16.0 0.1080 0~1000 0.50 1045 2.613 958 91.7 2.613
219.42 S-00 1296.0 16.0 0.1652 0.1500 0.50 1290 3.225 1486 115-2 3-715
365.72 8.00 1296.0 16.0 0.2795 0.2500 0.50 1838 4.595 2496 135.8 6.240
512.00 8.00 1296.0 16.0 0.3938 0.3500 0.50 2320 5.800 3307 142.5 8.268
658.28 8.00 1296.0 16.0 0.5080 0.4500 0.50 2787 6.968 4032 144.7 10.080
146.28 12.00 1296.0 16.0 0.1049 0~1000 0.75 914 2.285 781 85.4 2.285
219.42 12.00 1296.0 16.0 0.1620 0.1500 0.75 1161 2.903 1109 95.5 2.903
365.72 12.00 1296.0 16.0 0.2763 0.2500 0.75 1665 4.163 1790 107.5 4.475
512.00 12.00 1296.0 16.0 0.3906 0.3500 0.75 2115 5.288 2348 111.0 5.870
658.28 12.00 1296.0 16.0 0.5049 0.4500 0.75 2529 6.323 2846 112.5 7.115
146.28 16.00 1296.0 16.0 0.1018 0~1000 1.00 778 1.945 758 97.4 1.945
219.42 16.00 1296.0 16.0 0.1589 0.1500 1.00 1000 2.500 880 88.0 2.500
365.72 16.00 1296.0 16.0 0.2732 0.2500 1.00 1477 3.693 1347 91.2 3.693
512.00 16.00 1296.0 16.0 0.3875 0.3500 1.00 1895 4.738 1732 91.4 4.738
658.28 16-00 1296-O 16.0 0.5018 0.4500 I .oo 2276 5.690 2077 91.3 5.690
146.28 20.00 1296.0 16.0 0.0987 0~1000 1.25 700 1.750 733 104.7 1.833
219.42 20.00 1296.0 16.0 0.1558 0.1500 1.25 846 2.115 809 95.6 2.115
365.72 20.00 1296.0 16.0 0.2701 0.2500 1.25 1271 3.178 1085 85.4 3.178
512.00 20.00 1296.0 16.0 0.3844 0.3500 1.25 1651 4.128 1362 82.5 4.128
658.28 20.00 1296.0 16.0 0.4987 0.4500 1.25 2000 5.000 1610 80.5 5.000
146.28 24.00 1296.0 16.0 0.0955 0-1000 1.50 652 1.630 704 108-O I.760
219.42 24.00 1296.0 16.0 0.1527 0.1500 1.50 718 1.795 777 108.2 1.943
365.72 24.00 1296.0 16.0 0.2670 0.2500 1.50 1074 2.685 926 86.2 2.685
512.00 24.00 1296.0 16.0 0.3813 0.3500 1.50 1410 3,525 1133 80.4 3.525
658.28 24.00 1296.0 16.0 0.4955 0.4500 1.50 1734 4.335 1318 76.0 4.335
146.28 32.00 1296.0 16.0 0.0893 0~1000 2.00 582 1.455 633 108.8 1.583
219.42 32.00 1296.0 16-O 0.1464 0.1500 2.00 634 1.585 708 Ill.7 I.770
365.72 32.00 1296.0 16.0 0.2607 0.2500 2.00 815 2.038 817 100.2 2.043
512.00 32.00 1296.0 16.0 0.3750 0.3500 2.00 1025 2.563 916 89.4 2.563
658.28 32.00 1296.0 16.0 0.4893 0.4500 2.00 1259 3.148 1003 79.7 3.148
18 C. J. Dekker, H. J. Bos

Nozzles in cylindrical shells under internal pressure


R/T-parameter = 50
Internal pressure = 10 MPa
All dimensions in table below are in [mm] and all stresses are given in [MPa]

Nozzle dimensions Vessel dimensions Geometry parameters FE-pipe results Maximum


stress
O.D. t O.D. T a/o beta r/T Stress in SCF back Stress in Percentage intensity
header calculated branch of stress as SCF
in header

182.86 3.20 1616.0 16.0 0.1123 0~1000 0.20 1930 3.860 1582 82.0 3.860
274.28 3.20 1616.0 16.0 0.1694 0.1500 0.20 2508 5.016 2456 97.9 5.016
457.14 3.20 1616.0 16.0 0.2837 0.2500 0.20 3621 7.242 4183 115.5 8.366
640.00 3.20 1616.0 16.0 O-3980 0.3500 0.20 4524 9.048 5480 121.1 10.960
822.86 3.20 1616.0 16.0 05123 0.4500 0.20 5404 10.808 6663 123.3 13.326
182.86 5.60 1616.0 16.0 0.1108 0~1000 0.35 1610 3.220 1672 103.9 3.344
274.28 5.60 1616.0 16.0 0.1679 0.1500 0.35 2088 4.176 2581 123.6 5.162
457.14 5.60 1616.0 16.0 0.2822 0.2500 0.35 2992 5.984 4276 142.9 8.552
640.00 5.60 1616.0 16.0 0.3965 0.3500 0.35 3734 I.468 5601 150.0 11.202
822.86 5.60 1616.0 16.0 0.5108 0.4500 0.35 4444 8.888 6805 153.1 13,610
182.86 8.00 1616.0 16.0 0.1093 0~1000 0.50 1403 2.806 1456 103.8 2.912
274.28 8.00 1616.0 16.0 0.1664 0.1500 0.50 1690 3.380 1972 116.7 3.944
457.14 8.00 1616.0 16.0 0.2807 0.2500 0.50 2518 5.036 3605 143.2 7.210
640.00 8.00 1616.0 16.0 0.3950 0.3500 0.50 3203 6.406 4715 147.2 9.430
822.86 8.00 1616.0 16.0 0.5093 0.4500 0.50 3854 7.708 5724 148.5 11.448
182.86 12.00 1616.0 16.0 0.1068 0~1000 0.75 1255 2.510 1106 88.1 2.510
274.28 12.00 1616.0 16.0 0.1639 0~1500 0.75 1615 3.230 1619 100.2 3.238
457.14 12.00 1616.0 16.0 0.2782 0.2500 0.75 2328 4.656 2556 109.8 5.112
640.00 12.00 1616.0 16.0 0.3925 0.3500 0.75 2950 5.900 3313 112.3 6.626
822.86 12.00 1616.0 16.0 0.5068 0.4500 0.75 3524 7.048 4002 113.6 8.004

2i4,28 16.00 1616.0 16.0 0.1614 0~1500 1 .oo 1421 2.842 1258 88.5 2.842
457.14 16.00 1616.0 16.0 0.2754 0.2500 1.00 2063 4.126 1713 83.0 4.126
640.00 16.00 1616.0 16.0 0.3900 0.3500 1.00 2673 5.346 2421 90.6 5.346
822.86 16.00 1616.0 16.0 0.5043 0.4500 1.00 3210 6.420 2940 91.6 6.420
182.86 20.00 1616.0 16.0 0.1018 0~1000 1.25 913 1.826 945 103.5 1.890
274.28 20.00 1616.0 16.0 0.1589 0~1.500 1.25 1210 2.420 1051 86.9 2.420
457.14 20.00 1616.0 16.0 0.2732 0.2500 1.25 1832 3.664 1510 82.4 3.664
640.00 20.00 1616.0 16.0 0.3875 0.3500 1.25 2360 4.720 1887 80.0 4.720
822.86 20.00 1616.0 16.0 0.5018 0.4500 1.25 2858 5.716 2229 78-O 5.716

182.86 24.00 1616.0 16.0 0.0993 0~1000 1.50 826 1.652 910 110.2 1.820
274.28 24.00 1616.0 16.0 0.1564 0~1500 1.50 1021 2.042 999 97.8 2.042
457.14 24.00 1616.0 16.0 0.2707 0.2500 1.50 1569 3.138 1271 81.0 3.138
640.00 24.00 1616.0 16.0 0.3850 0.3500 1.50 2042 4.084 1555 76.2 4.084
822.86 24.00 1616.0 16.0 0.4993 0.4500 1.50 2511 5.022 1812 72.2 5.022

182.86 32.00 1616-O 16.0 0.0943 0~1000 2.00 741 1.482 829 111.9 1.658
274.28 32.00 1616.0 16.0 0.1514 0~1500 2.00 802 1.604 912 113.7 1.824
457.14 32.00 1616.0 16.0 0.2657 0.2500 2.00 1168 2.336 1058 90.6 2.336
640.00 32.00 1616.0 16.0 0.3800 0.3500 2.00 1504 3.008 1193 79.3 3.008
822.86 32.00 1616.0 16.0 0.4943 0.4500 2.00 1864 3.728 1346 72.2 3.728

You might also like