Pressure vessel design manual d moss - 3 ed
Published on: Mar 4, 2016
Transcripts - Pressure vessel design manual d moss - 3 ed
T H I R D E D I T I O N
AMSTERDAM BOSTON HEIDELBERG
LONDON NEW YORK *OXFORD PARIS
SAN DIEGO SAN FRANCISCO SINGAPORE
G p Gulf
+P @ Publishing
ELSEVIER Gulf Professional Publishingis an imprintof Elsevier
Gulf Professional Publishing is an imprint of EIsevier
200 Wheeler Road, Burlington, MA 01803, USA
Linacre House, Jordan Hill, Oxford OX2 8DP, UK
Copyright 02004, Elsevier, Inc. All rights reserved
No part of this publication may be reproduced, stored in a retrieval system, or
transmitted in any form or by any means, electronic, mechanical, photocopying,
recording, or otherwise, without the prior written permission of the publisher.
Permissions may be sought directly from Elsevier’s Science & Technology Rights
Department in Oxford, UK: phone (+44) 1865 843830, fax: (+44) 1865 853333,
e-mail: permissions@elsevier,corn.uk. You may also complete your request
online via the Elsevier Science homepage (http://elsevier.com), by selecting
“Customer Support” and then “Obtaining Permissions.”
00 Recognizing the importance of preserving what has been written, Elsevier prints its
books on acid-free paper whenever possible.
Library of CongressCataloging-in-PublicationData
Moss, Dennis R.
vessel design problems/Dennis R. Moss.-3rd ed.
ISBN 0-7506-7740-6(hardcover: alk. paper)
Pressure vessel design manual: illustrated procedures for solving major pressure
1. Pressure vessels-Design and construction-Handbooks, manuals, etc. I. Title.
681’.76041 4 ~ 2 2
British Library Cataloguing-in-PublicationData
A catalogue record for this book is available from the British Library.
For information on all Gulf Professional Publishing
publications visit our website at www.gulfpp.com
0405060708 10 11 9 8 7 6 5 4 3 2 1
Printed in the United States of America
STRESSES IN PRESSURE VESSELS, 1
Design Philosophy, 1
Stress Analysis, 1
Stress/Failure Theories, 2
Failures in Pressure Vessels, 5
Special Problems, 10
GENERAL DESIGN, 15
Procedure 2-1: General Vessel Formulas, 15
Procedure 2-2: External Pressure Design, 19
Procedure 2-3: Calculate MAP, MAWP, and Test Pressures, 28
Procediire 2-4: Stresses in Heads Due to Internal Pressure, 30
Procedure 2-5: Design of Intermediate Heads, 31
Procedure 2-6: Design of Toriconical Transitions, 33
Procedure 2-7: Design of Flanges, 37
Procedure 2-8: Design of Spherically Dished Covers, 57
Procediire 2-9: Design of Blind Flanges with Openings, 58
Procedure 2-10: Bolt Torque Required for Sealing Flanges, 59
Procedure 2-11: Design of Flat Heads, 62
Procedure 2-12: Reinforcement for Studding Outlets, 68
Procedure 2-13: Design of Internal Support Beds, 69
Procedure 2-14: Nozzle Reinforcement, 74
Procedure 2-15: Design of Large Openings in Flat Heads, 78
Procedure 2-16: Find or Revise the Center of Gravity of a Vessel, 80
Procedure 2-17: Minimum Design Metal Temperature (MDMT),81
Procedure 2-18: Buckling of Thin-Walled Cylindrical Shells, 8.5
Procedure 2-19: Optimum Vessel Proportions, 89
Procedure 2-20: Estimating Weights of Vessels and Vessel Components, 95
vi Pressure Vessel Design Manual
DESIGN OF VESSEL SUPPORTS, 109
Support Structures, 109
Procedure 3-1: Wind Design per ASCE, 112
Procedure 3-2: Wind Design per UBC-97, 118
Procedure 3-3: Seismic Design for Vessels, 120
Procedure 3-4: Seismic Design-Vessel on Unbraced Legs, 125
Procedure 3-5: Seismic Design-Vessel on Braced Legs, 132
Procedure 3-6: Seismic Design-Vessel on Rings, 140
Procedure 3-7: Seismic Design-Vessel on Lugs #1, 145
Procedure 3-8: Seismic Design-Vessel on Lugs #2, 151
Procedure 3-9: Seismic Design-Vessel on Skirt, 157
Procedure 3-10: Design of Horizontal Vessel on Saddles, 166
Procedure 3-11: Design of Saddle Supports for Large Vessels, 177
Procedure 3-12: Design of Base Plates for Legs, 184
Procedure 3-13: Design of Lug Supports, 188
Procedure 3-14: Design of Base Details for Vertical Vessels #1, 192
Procedure 3-15: Design of Base Details for Vertical Vessels #2, 200
SPECIAL DESIGNS, 203
Procedure 4-1: Design of Large-Diameter Nozzle Openings, 203
Procedure 4-2: Design of Cone-Cylinder Intersections, 208
Procedure 4-3: Stresses at Circumferential Ring Stiffeners, 216
Procedure 4-4: Tower Deflection, 219
Procedure 4-5: Design of Ring Girders, 222
Procedure 4-6: Design of Baffles, 227
Procedure 4-7: Design of Vessels with Refractory Linings, 237
Procedure 4-8: Vibration of Tall Towers and Stacks, 244
LOCAL LOADS, 255
Procedure 5-1: Stresses in Circular Rings, 256
Procedure 5-2: Design of Partial Ring Stiffeners, 265
Procedure 5-3:Attachment Parameters, 267
Procedure 5-4: Stresses in Cylindrical Shells from External Local Loads, 269
Procedure 5-5: Stresses in Spherical Shells from External Local Loads, 283
RELATED EQUIPMENT, 291
Procedure 6-1: Design of Davits, 291
Procedure 6-2: Design of Circular Platforms, 296
Procedure 6-3: Design of Square and Rectangular Platforms, 304
Procedure 6-4: Design of Pipe Supports, 309
Procedure 6-5: Shear Loads in Bolted Connections, 317
Procedure 6-6: Design of Bins and Elevated Tanks, 318
Procedure 6-7: AgitatordMixers for Vessels and Tanks, 328
Procedure 6-8: Design of Pipe Coils for Heat Transfer, 335
Procedure 6-9: Field-Fabricated Spheres, 355
TRANSPORTATION AND ERECTION OF PRESSURE
Procedure 7-1: Transportation of Pressure Vessels, 365
Procedure 7-2: Erection of Pressure Vessels, 387
Procedure 7-3: Lifting Attachments and Terminology, 391
Procedure 7-4: Lifting Loads and Forces, 400
Procedure 7-5: Design of Tail Beams, Lugs, and Base Ring Details, 406
Procedure 7-6: Design of Top Head and Cone Lifting Lugs, 416
Procedure 7-7: Design of Flange Lugs, 420
Procedure 7-8: Design of Trunnions, 431
Procedure 7-9: Local Loads in Shell Due to Erection Forces, 434
Procedure 7-10: Miscellaneous,437
Guide to ASME Section VIII, Division 1,443
Design Data Sheet for Vessels, 444
Joint Efficiencies (ASME Code), 445
Properties of Heads, 447
Volumes and Surface Areas of Vessel Sections, 448
Vessel Nomenclature, 455
Useful Formulas for Vessels, 459
Material Selection Guide, 464
Summary of Requirements for 100% X-Ray and PWHT, 465
Material Properties, 466
Metric Conversions, 474
Allowable Compressive Stress for Columns, FA,475
Design of Flat Plates, 478
External Insulation for Vertical Vessels, 480
Flow over Weirs, 482
Time Required to Drain Vessels, 483
Vessel Surge Capacities and Hold-Up Times, 485
Minor Defect Evaluation Procedure, 486
Designers of pressure vessels and related equipment frequently have design infor-
mation scattered among numerous books, periodicals, journals, and old notes. Then,
when faced with a particular problem, they spend hours researching its solution only to
discover the execution may have been rather simple. This book can eliminate those
hours of research by probiding a step-by-step approach to the problems most fre-
quently encountered in the design of pressure vessels.
This book makes no claim to originality other than that of format. The material is
organized in the most concise and functionally useful manner. Whenever possible,
credit has been given to the original sources.
Although eve^ effort has been made to obtain the most accurate data and solutions,
it is the nature of engineering that certain simplifying assumptions be made. Solutions
achie7ed should be viewed in this light, and where judgments are required, they should
be made with due consideration.
Many experienced designers will have already performed many of the calculations
outlined in this book, but will find the approach slightly different. All procedures have
been developed and proven, using actual design problems. The procedures are easily
repeatable to ensure consistency of execution. They also can be modified to incorpo-
rate changes in codes, standards, contracts, or local requirements. Everything required
for the solution of an individual problem is contained in the procedure.
This book may be used directly to solve problems, as a guideline, as a logical
approach to problems, or as a check to alternative design methods. If more detailed
solutions are required, the approach shown can be amplified where required.
The user of this book should be advised that any code formulas or references should
always be checked against the latest editions of codes, Le., ASME Section VIII,
Division 1, Uniform Building Code, arid ASCE 7-95. These codes are continually
updated and revised to incorporate the latest available data.
1am grateful to all those who have contributed information and advice to make this
book possible, and invite any suggestions readers may make concerning corrections or
Dennis H. Moss
Cover Photo: Photo courtesy of Irving Oil Ltd., Saint John, New Brunswick,
Canada and Stone and Webster, Inc., A Shaw Group Company, Houston, Texas.
The photo shows the Reactor-Regenerator Structure of the Converter Section of the
RFCC (Resid Fluid Catalytic Cracking) Unit. This “world class” unit operates at the
Irving Refinery Complex in Saint John, New Brunswick, Canada, and is a proprietary
process of Stone and Webster.
Stresses in PressureVessels
In general, pressure vessels designed in accordance with
the ASME Code, Section VIII, Division 1,are designed by
rules and do not require a detailed evaluation of all stresses.
It is recognized that high localized and secondary bending
stresses may exist but are allowed for by use of a higher
safety factor and design rules for details. It is required, how-
ever, that all loadings (the forces applied to a vessel or its
structural attachments)must be considered. (SeeReference 1,
While the Code gives formulas for thickness and stress of
basic components, it is up to the designer to select appro-
priate analytical procedures for determining stress due to
other loadings. The designer must also select the most prob-
able combination of simultaneous loads for an economical
and safe design.
The Code establishes allowable stresses by stating in Para.
UG-23(c) that the maximum general primary membrane
stress must be less than allowablestresses outlined in material
sections. Further, it states that the maximum primary mem-
brane stress plus primary bending stress may not exceed 1.5
times the allowable stress of the material sections. In other
sections, specifically Paras. 1-5(e) and 2-8, higher allowable
stresses are permitted if appropriate analysis is made. These
higher allowable stresses clearly indicate that different stress
levels for different stress categories are acceptable.
It is general practice when doing more detailed stress
analysis to apply higher allowable stresses. In effect, the
detailed evaluation of stresses permits substituting knowl-
edge of localized stresses and the use of higher allowables
in place of the larger factor of safety used by the Code. This
higher safety factor really reflected lack of knowledge about
A calculated value of stress means little until it is associ-
ated with its location and distribution in the vessel and with
the type of loading that produced it. Different types of stress
have different degrees of significance.
The designer must familiarize himself with the various
types of stress and loadings in order to accurately apply
the results of analysis. The designer must also consider
some adequate stress or failure theory in order to combine
stresses and set allowable stress limits. It is against this fail-
ure mode that he must compare and interpret stress values,
and define how the stresses in a component react and con-
tribute to the strength of that part.
The following sections will provide the fundamental
knowledge for applying the results of analysis. The topics
covered in Chapter 1 form the basis by which the rest of
the book is to be used. A section on special problems and
considerations is included to alert the designer to more com-
plex problems that exist.
Stress analysis is the determination of the relationship
between external forces applied to a vessel and the corre-
sponding stress. The emphasis of this book is not how to do
stress analysis in particular, but rather how to analyze vessels
and their component parts in an effort to arrive at an
economical and safe design-the rllfference being that we
analyze stresses where necessary to determine thickness of
material and sizes of members. We are not so concerned
with building mathematical models as with providing a
step-by-step approach to the design of ASME Code vessels.
It is not necessary to find every stress but rather to know the
governing stresses and how they relate to the vessel or its
respective parts, attachments, and supports.
The starting place for stress analysis is to determine all
the design conditions for a gven problem and then deter-
mine all the related external forces. We must then relate
these external forces to the vessel parts which must resist
them to find the corresponding stresses. By isolating the
causes (loadings),the effects (stress)can be more accurately
The designer must also be keenly aware of the types of
loads and how they relate to the vessel as a whole. Are the
2 Pressure Vessel Design Manual
effects long or short term? Do they apply to a localized
portion of the vessel or are they uniform throughout?
How these stresses are interpreted and combined, what
significance they have to the overall safety of the vessel, and
what allowable stresses are applied will be determined by
1. The strengtwfailure theory utilized.
2. The types and categories of loadings.
3. The hazard the stress represents to the vessel.
Membrane Stress Analysis
Pressure vessels commonly have the form of spheres,
cylinders, cones, ellipsoids, tori, or composites of these.
When the thickness is small in comparison with other &men-
sions (RJt > lo),vessels are referred to as membranes and
the associated stresses resulting from the contained pressure
are called membrane stresses. These membrane stresses are
average tension or compression stresses. They are assumed
to be uniform across the vessel wall and act tangentially to its
surface. The membrane or wall is assumed to offer no resis-
tance to bending. When the wall offers resistance to bend-
ing, bending stresses occur in addtion to membrane stresses.
In a vessel of complicated shape subjected to internal
pressure, the simple membrane-stress concepts do not suf-
fice to give an adequate idea of the true stress situation. The
types of heads closing the vessel, effects of supports, varia-
tions in thickness and cross section, nozzles, external at-
tachments, and overall bending due to weight, wind, and
seismic activity all cause varying stress distributions in the
vessel. Deviations from a true membrane shape set up bend-
ing in the vessel wall and cause the direct loading to vary
from point to point. The direct loading is diverted from the
more flexible to the more rigid portions of the vessel. This
effect is called “stress redistribution.”
In any pressure vessel subjected to internal or external
pressure, stresses are set up in the shell wall. The state of
stress is triaxial and the three principal stresses are:
04 = circumferentialAatitudina1 stress
or= radial stress
In addition, there may be bending and shear stresses. The
radial stress is a direct stress, which is a result of the pressure
acting directly on the wall, and causes a compressive stress
equal to the pressure. In thin-walled vessels this stress is so
small compared to the other “principal” stresses that it is
generally ignored. Thus we assume for purposes of analysis
that the state of stress is biaxial. This greatly simplifies the
method of combining stresses in comparison to triaxial stress
states. For thickwalled vessels (RJt < lo), the radial stress
cannot be ignored and formulas are quite different from
those used in finding “membrane stresses” in thin shells.
Since ASME Code, Section VIII, Division 1,is basically for
design by rules, a higher factor of safety is used to allow for
the “unknown” stresses in the vessel. This higher safety
factor, which allows for these unknown stresses, can impose
a penalty on design but requires much less analysis. The
design techniques outlined in this text are a compro-
mise between finding all stresses and utilizing minimum
code formulas. This additional knowledge of stresses warrants
the use of higher allowablestresses in some cases,while meet-
ing the requirements that all loadings be considered.
In conclusion, “membrane stress analysis’’is not completely
accurate but allows certain simplifymg assumptions to be
made while maintaining a fair degree of accuracy. The main
simplifying assumptions are that the stress is biaxial and that
the stresses are uniform across the shell wall. For thin-walled
vessels these assumptions have proven themselves to be
reliable. No vessel meets the criteria of being a true
membrane, but we can use this tool with a reasonable
degree of accuracy.
As stated previously, stresses are meaningless until com-
pared to some stresdfailure theory. The significance of a
given stress must be related to its location in the vessel
and its bearing on the ultimate failure of that vessel.
Historically, various ‘‘theories” have been derived to com-
bine and measure stresses against the potential failure
mode. A number of stress theories, also called “yield cri-
teria,” are available for describing the effects of combined
stresses. For purposes of this book, as these failure theories
apply to pressure vessels, only two theories will be discussed.
They are the “maximum stress theory” and the “maximum
shear stress theory.”
Maximum Stress Theory
This theory is the oldest, most widely used and simplest to
apply. Both ASME Code, Section VIII, Division 1, and
Section I use the maximum stress theory as a basis for
design. This theory simply asserts that the breakdown of
Stresses in Pressure Vessels 3
material depends only on the numerical magnitude of the
maximum principal or normal stress. Stresses in the other
directions are disregarded. Only the maximum principal
stress must be determined to apply this criterion. This
theory is used for biaxial states of stress assumed in a thin-
walled pressure vessel. As will be shown later it is unconser-
vative in some instances and requires a higher safety factor
for its use. While the maximum stress theory does accurately
predict failure in brittle materials, it is not always accurate
for ductile materials. Ductile materials often fail along lines
4 5 to the applied force by shearing, long before the tensile
or compressive stresses are maximum.
This theory can be illustrated graphically for the four
states of biaxial stress shown in Figure 1-1.
It can be seen that uniaxial tension or compression lies on
tlir two axes. Inside the box (outer boundaries) is the elastic
range of the material. Yielding is predicted for stress
combinations by the outer line.
Maximum Shear Stress Theory
This theory asserts that the breakdown of material de-
pends only on the mdximum shear stress attained in an ele-
ment. It assumes that yielding starts in planes of maximum
shear stress. According to this theory, yielding will start at a
point when the maximum shear stress at that point reaches
one-half of the the uniaxial yield strength, F,. Thus for a
biaxial state of stress where 01 > ( ~ 2 ,the maximum shear
stress will be (al- (s2)/2.
Yielding will occur when
Both ASME Code, Section 1'111, Division 2 and ASME
Code, Section 111, utilize the maximum shear stress criterion.
This theory closely approximates experimental results and is
also easy to use. This theory also applies to triaxial states
of stress. In a triaxial stress state, this theory predicts that
yielding will occur whenever one-half the algebraic differ-
ence between the maximum and minimum 5tress is equal to
one-half the yield stress. Where c1> a2> 03,the maximum
shear stress is (ul-
Yielding will begin when
01 - 0 3 - F,
This theory is illustrated graphically for the four states of
biaxial stress in Figure 1-2.
A comparison of Figure 1-1 and Figure 1-2 will quickly
illustrate the major differences between the two theories.
Figure 1-2 predicts yielding at earlier points in Quadrants
I1 and IV. For example, consider point B of Figure 1-2. It
shows ~ 2 = ( - ) ( ~ 1 ; therefore the shear stress is equal to
c2- (-a1)/2, which equals o2+a1/2 or one-half the stress
r Safety factor boundary
imposed by ASME Code
_ _ _ _ I IV111
+ l . O
I Failure surface (yield surface) boundary
Figure 1-1. Graph of maximum stress theory. Quadrant I: biaxial tension; Quadrant II: tension: Quadrant Ill: biaxial compression; Quadrant IV:
4 Pressure Vessel Design Manual
,-Failuresurface (yield surface boundary)
Figure 1-2. Graph of maximum shear stress theory.
which would cause yielding as predcted by the maximum
Comparison of the TwoTheories
Both theories are in agreement for uniaxial stress or when
one of the principal stresses is large in comparison to the
others. The discrepancy between the theories is greatest
when both principal stresses are numerically equal.
For simple analysis upon which the thickness formulas for
ASME Code, Section I or Section VIII, Division 1,are based,
it makes little difference whether the maximum stress
theory or maximum shear stress theory is used. For example,
according to the maximum stress theory, the controlling
stress governing the thickness of a cylinder is 04,circumfer-
ential stress, since it is the largest of the three principal
stresses. Accordmg to the maximum shear stress theory,
the controlling stress would be one-half the algebraic differ-
ence between the maximum and minimum stress:
The maximum stress is the circumferential stress, a4
04 = PR/t
0 The minimum stress is the radial stress, a,
a, = -P
Therefore, the maximum shear stress is:
ASME Code, Section VIII, Division 2, and Section I11 use
the term “stress intensity,” which is defined as twice the
maximum shear stress. Since the shear stress is compared
to one-half the yield stress only, “stress intensity” is used for
comparison to allowable stresses or ultimate stresses. To
define it another way, yieldmg begins when the “stress in-
tensity” exceeds the yield strength of the material.
In the preceding example, the “stress intensity” would be
equal to 04 - a,.And
For a cylinder where P =300 psi, R =30 in., and t =.5 in.,
the two theories would compare as follows:
Maximum stress theory
o = a4 = PR/t = 300(30)/.5 = 18,000psi
Maximum shear stress the0y
a = PR/t +P = 300(30)/.5 +300 = 18,300 psi
Two points are obvious from the foregoing:
1. For thin-walled pressure vessels, both theories yield
approximately the same results.
2. For thin-walled pressure vessels the radial stress is so
small in comparison to the other principal stresses that
it can be ignored and a state of biaxial stress is assumed
Stresses in Pressure Vessels 5
For thick-walled vessels (R,,,/t < lo), the radial stress
becomes significant in defining the ultimate failure of the
vessel. The maximum stress theory is unconservative for
designing these vessels. For this reason, this text has limited
its application to thin-walled vessels where a biaxial state of
stress is assumed to exist.
FAILURES IN PRESSURE VESSELS
Vessel failures can be grouped into four major categories,
which describe why a vessel failure occurs. Failures can also
be grouped into types of failures, which describe how
the failure occurs. Each failure has a why and how to its
history. It may have failed through corrosion fatigue because
the wrong material was selected! The designer must be as
familiar with categories and types of failure as with cate-
gories and types of stress and loadings. Ultimately they are
Categories of Failures
1. Material-Improper selection of material; defects in
2. Design-Incorrect design data; inaccurate or incor-
rect design methods; inadequate shop testing.
3. Fabrication-Poor quality control; improper or insuf-
ficient fabrication procedures including welding; heat
treatment or forming methods.
4. Seruice-Change of service condition by the user;
inexperienced operations or maintenance personnel;
upset conditions. Some types of service which require
special attention both for selection of material, design
details, and fabrication methods are as follows:
b. Fatigue (cyclic)
c. Brittle (low temperature)
d. High temperature
e. High shock or vibration
f. Vessel contents
0 Compressed air
Types of Failures
1. Elastic defi,rmation-Elastic instability or elastic buck-
ling, vessel geometry, and stiffness as well as properties
of materials are protection against buckling.
2. Brittlefracture-Can occur at low or intermediate tem-
peratures. Brittle fractures have occurred in vessels
made of low carbon steel in the 40’50°F range
during hydrotest where minor flaws exist.
3. Excessive plastic deformation-The primary and sec-
ondary stress limits as outlined in ASME Section
VIII, Division 2, are intended to prevent excessive plas-
tic deformation and incremental collapse.
4. Stress rupture-Creep deformation as a result of fa-
tigue or cyclic loading, i.e., progressive fracture.
Creep is a time-dependent phenomenon, whereas fa-
tigue is a cycle-dependent phenomenon.
5. Plastic instability-Incremental collapse; incremental
collapse is cyclic strain accumulation or cumulative
cyclic deformation. Cumulative damage leads to insta-
bility of vessel by plastic deformation.
6. High strain-Low cycle fatigue is strain-governed and
occurs mainly in lower-strengthhigh-ductile materials.
7. Stress corrosion-It is well known that chlorides cause
stress corrosion cracking in stainless steels; likewise
caustic service can cause stress corrosion cracking in
carbon steels. Material selection is critical in these
8. Corrosion fatigue-Occurs when corrosive and fatigue
effects occur simultaneously. Corrosion can reduce fa-
tigue life by pitting the surface and propagating cracks.
Material selection and fatigue properties are the major
In dealing with these various modes of failure, the de-
signer must have at his disposal a picture of the state of
stress in the various parts. It is against these failure modes
that the designer must compare and interpret stress values.
But setting allowable stresses is not enough! For elastic
instability one must consider geometry, stiffness, and the
properties of the material. Material selection is a major con-
sideration when related to the type of service. Design details
and fabrication methods are as important as “allowable
stress” in design of vessels for cyclic service. The designer
and all those persons who ultimately affect the design must
have a clear picture of the conditions under which the vessel
6 Pressure Vessel Design Manual
Loadings or forces are the “causes” of stresses in pres-
sure vessels. These forces and moments must be isolated
both to determine where they apply to the vessel and
when they apply to a vessel. Categories of loadings
define where these forces are applied. Loadings may be
applied over a large portion (general area) of the vessel or
over a local area of the vessel. Remember both general
and local loads can produce membrane and bending
stresses. These stresses are additive and define the overall
state of stress in the vessel or component. Stresses from
local loads must be added to stresses from general load-
ings. These combined stresses are then compared to an
Consider a pressurized, vertical vessel bending due to
wind, which has an inward radial force applied locally.
The effects of the pressure loading are longitudinal and
circumferential tension. The effects of the wind loading
are longitudinal tension on the windward side and lon-
gitudinal compression on the leeward side. The effects of
the local inward radial load are some local membrane stres-
ses and local bending stresses. The local stresses would be
both circumferential and longitudinal, tension on the inside
surface of the vessel, and compressive on the outside. Of
course the steel at any given point only sees a certain level
of stress or the combined effect. It is the designer’s job to
combine the stresses from the various loadings to arrive at
the worst probable combination of stresses, combine them
using some failure theory, and compare the results to an
acceptable stress level to obtain an economical and safe
This hypothetical problem serves to illustrate how cate-
gories and types of loadings are related to the stresses they
produce. The stresses applied more or less continuously and
unqomly across an entire section of the vessel are primary
The stresses due to pressure and wind are primary mem-
brane stresses. These stresses should be limited to the code
allowable. These stresses would cause the bursting or
collapse of the vessel if allowed to reach an unacceptably
On the other hand, the stresses from the inward radial
load could be either a primary local stress or secondary
stress. It is a primary local stress if it is produced from an
unrelenting load or a secondary stress if produced by a
relenting load. Either stress may cause local deformation
but will not in and of itself cause the vessel to fail. If it is
a primary stress, the stress will be redistributed; if it is a
secondary stress, the load will relax once slight deforma-
Also be aware that this is only true for ductile materials. In
brittle materials, there would be no difference between
primary and secondary stresses. If the material cannot
yield to reduce the load, then the definition of secondary
stress does not apply! Fortunately current pressure vessel
codes require the use of ductile materials.
This should make it obvious that the type and category of
loading will determine the type and category of stress. This
will be expanded upon later, but basically each combina-
tion of stresses (stress categories) will have different allow-
0 Primary stress: P, < SE
0 Primary membrane local (PL):
PL = P, +PL < 1.5SE
PL = P,, +Q, < 1.5 SE
0 Primary membrane + secondary (Q):
Pm +Q < 3 SE
But what if the loading was of relatively short duration? This
describes the “type”of loading. Whether a loading is steady,
more or less continuous, or nonsteady, variable, or tempo-
rary will also have an effect on what level of stress will be
acceptable. If in our hypothetical problem the loading had
been pressure + seismic + local load, we would have a
different case. Due to the relatively short duration of seismic
loading, a higher “temporary” allowable stress would be ac-
ceptable. The vessel doesn’t have to operate in an earth-
quake all the time. On the other hand, it also shouldn’t fall
down in the event of an earthquake! Structural designs allow
a one-third increase in allowable stress for seismic loadings
for this reason.
For steady loads, the vessel must support these loads more
or less continuously during its useful life. As a result, the
stresses produced from these loads must be maintained to
an acceptable level.
For nonsteady loads, the vessel may experience some
or all of these loadings at various times but not all at once
and not more or less continuously. Therefore a temporarily
higher stress is acceptable.
For general loads that apply more or less uniformly across
an entire section, the corresponding stresses must be lower,
since the entire vessel must support that loading.
For local loads, the corresponding stresses are confined to
a small portion of the vessel and normally fall off rapidly in
distance from the applied load. As discussed previously,
pressurizing a vessel causes bending in certain components.
But it doesn’t cause the entire vessel to bend. The results are
not as significant (except in cyclic service) as those caused by
general loadings. Therefore a slightly higher allowable stress
would be in order.
Stresses in Pressure Vessels 7
Loadings can be outlined as follows:
IA. Categories of loadings
1. General loads-Applied more or less continuously
across a vessel section.
a. Pressure loads-Internal or external pressure
(design, operating, hydrotest. and hydrostatic
head of liquid).
b. Moment loads-Due to wind, seismic, erection,
c. Compressive/tensile loads-Due to dead weight,
installed equipment, ladders, platforms, piping,
and vessel contents.
d. Thermal loads-Hot box design of skirthead
2. Local loads-Due to reactions from supports,
internals, attached piping, attached equipment,
Le., platforms, mixers, etc.
a. Radial load-Inward or outward.
b. Shear load-Longitudinal or circumferential.
c. Torsional load.
d. Tangential load.
e. Moment load-Longitudinal or circumferential.
f. Thermal load.
B. Typey of loadings
1. Steady load-Long-term duration, continuous.
a. InternaVexternal pressure.
b Dead weight.
c. Vessel contents.
d. Loadings due to attached piping and equipment.
e. Loadings to and from vessel supports.
f. Thermal loads.
g. Wind loads.
a. Shop and field hydrotests.
e. Upset, emergency.
f. Thermal loads.
g. Start up, shut down.
2. Nonsteady loads-Short-term duration; variable.
ASME Code, SectionVIII, Division 1 vs.
ASME Code, Section VIII, Division 1 does not explicitly
consider the effects of combined stress. Neither does it give
detailed methods on how stresses are combined. ASME
Code, Section VIII, Division 2, on the other hand, provides
specific guidelines for stresses, how they are combined, and
allowable stresses for categories of combined stresses.
Division 2 is design by analysis whereas Division 1 is
design by rules. Although stress analysis as utilized by
Division 2 is beyond the scope of this text, the use of
stress categories, definitions of stress, and allowable stresses
Division 2 stress analysis considers all stresses in a triaxial
state combined in accordance with the maximum shear stress
theory. Division 1 and the procedures outlined in this book
consider a biaxial state of stress combined in accordance with
the maximum stress theory. Just as you would not design
a nuclear reactor to the niles of Division 1, you would
not design an air receiver by the techniques of Division 2.
Each has its place and applications. The following discussion
on categories of stress and allowables will utilize informa-
tion from Division 2, which can be applied in general to all
Types, Classes, and Categories of Stress
The shell thickness as computed by Code formulas for
internal or external pressure alone is often not sufficient to
withstand the combined effects of all other loadings.
Detailed calculations consider the effects of each loading
separately and then must be combined to give the total
state of stress in that part. The stresses that are present in
pressure vessels are separated into various cla.~.sr~sin accor-
dance with the types of loads that produced them, and the
hazard they represent to the vessel. Each class of stress must
be maintained at an acceptable leL7eland the combined
total stress must be kept at another acceptable level. The
combined stresses due to a combination of loads acting
simultaneously are called stress categories. Please note
that this terminology differs from that given in Dikision 2,
but is clearer for the purposes intended herc,.
Classes of stress, categories of stress, and allowable
stresses are based on the type of loading that produced
them and on the hazard they represent to the structure.
Unrelenting loads produce primary stresses. Relenting loads
(self-limiting) produce secondary stresses. General loadings
produce primary membrane and bending stresses. Local
loads produce local membrane and bending stresses.
Primary stresses must be kept l o ~ e rthan secondary stresses.
8 Pressure Vessel Design Manual
Primary plus secondary stresses are allowed to be higher
and so on. Before considering the combination of stresses
(categories), we must first define the various types and
classes of stress.
Types of Stress
There are many names to describe types of stress. Enough
in fact to provide a confusing picture even to the experienced
designer. As these stresses apply to pressure vessels, we
group all types of stress into three major classes of stress,
and subdivision of each of the groups is arranged according
to their effect on the vessel. The following list of stresses
describes types of stress without regard to their effect on
the vessel or component. They define a direction of stress
or relate to the application of the load.
12. Load induced
13. Strain induced
Classes of Stress
The foregoing list provides examples of types of stress.
It is, however, too general to provide a basis with which
to combine stresses or apply allowable stresses. For this
purpose, new groupings called classes of stress must be
used. Classes of stress are defined by the type of loading
which produces them and the hazard they represent to the
1. Primay stress
0 Primary general membrane stress, P,
0 Primary general bending stress, Pb
b. Primary local stress, PL
a. Secondary membrane stress, Q,
b. Secondary bending stress, Q b
2. Seconday stress
3. Peak stress, F
Definitions and examples of these stresses follow.
Primary general stress. These stresses act over a full
cross section of the vessel. They are produced by mechanical
loads (load induced) and are the most hazardous of all types
of stress. The basic characteristic of a primary stress is that it
is not self-limiting. Primary stresses are generally due to in-
ternal or external pressure or produced by sustained external
forces and moments. Thermal stresses are never classified as
Primary general stresses are divided into membrane and
bending stresses. The need for divilng primary general
stress into membrane and bending is that the calculated
value of a primary bending stress may be allowed to go
higher than that of a primary membrane stress. Primary
stresses that exceed the yield strength of the material can
cause failure or gross distortion. Typical calculations of
primary stress are:
PR F MC
Jt ’ A ’ I ’
Primary general membranestress, P,. This stress occurs across
the entire cross section of the vessel. It is remote from dis-
continuities such as head-shell intersections, cone-cylinder
intersections, nozzles, and supports. Examples are:
a. Circumferential and longitudmal stress due to pressure.
b. Compressive and tensile axial stresses due to wind.
c. Longitudinal stress due to the bending of the horizontal
vessel over the saddles.
d. Membrane stress in the center of the flat head.
e. Membrane stress in the nozzle wall within the area of
reinforcement due to pressure or external loads.
f. Axial compression due to weight.
Primary general bending stress, Pb. Primary bending stresses
are due to sustained loads and are capable of causing
collapse of the vessel. There are relatively few areas where
primary bending occurs:
a. Bending stress in the center of a flat head or crown of a
b. Bending stress in a shallow conical head.
c. Bending stress in the ligaments of closely spaced
Local primary membrane stress, PL. Local primary
membrane stress is not technically a classificationof stress but
a stress category, since it is a combination of two stresses. The
combination it represents is primary membrane stress, P,,
plus secondary membrane stress, Q,, produced from sus-
tained loads. These have been grouped together in order to
limit the allowable stress for this particular combination to a
level lower than allowed for other primary and secondary
stress applications. It was felt that local stress from sustained
(unrelenting) loads presented a great enough hazard for the
combination to be “classified” as a primary stress.
A local primary stress is produced either by design
pressure alone or by other mechanical loads. Local primary
Stresses in Pressure Vessels 9
stresses have some self-limiting characteristics like secondary
stresses. Since they are localized, once the yield strength of
the material is reached, the load is redistributed to stiffer
portions of the vessel. However, since any deformation
associated with yielding would be unacceptable, an allowable
stress lower than secondary stresses is assigned. The basic
difference between a primary local stress and a secondary
stress is that a primary local stress is produced by a load that
is unrelenting; the stress is just redistributed. In a secondary
stress, yielding relaxes the load and is truly self-limiting. The
ability of primary local stresses to redistribute themselves
after the yield strength is attained locally provides a safety-
valve effect. Thus, the higher allowable stress applies only to
a local area.
Primary local membrane stresses are a combination of
membrane stresses only. Thus only the “membrane” stresses
from a local load are combined with primary general
membrane stresses, not the bending stresses. The bending
stresses associated with a local loading are secondary
stresses. Therefore, the membrane stresses from a WRC-
107-type analysis must be broken out separately and com-
bined with primary general stresses. The same is true for
discontinuity membrane stresses at head-shell junctures,
cone-cylinder junctures, and nozzle-shell junctures. The
bending stresses would be secondary stresses.
Therefore, PL=P, +Qlllrwhere Q,, is a local stress from
a sustained or unrelenting load. Examples of primary local
membrane stresses are:
a. PI,,+membrane stresses at local discontinuities:
1. Head-shell juncture
2. Cone-cylinder juncture
3. Nozzle-shell juncture
5. Head-slurt juncture
6. Shell-stiffening ring juncture
b. P,, +membrane stresses from local sustained loads:
1. support lugs
2. Nozzle loads
3. Beam supports
4. Major attachments
Secondarystress. The basic characteristic of a second-
ary stress is that it is self-limiting. As defined earlier, this
means that local yielding and minor distortions can satisfy
the conditions which caused the stress to occur. Application
of a secondary stress cannot cause structural failure due
to the restraints offered by the body to which the part is
attached. Secondary mean stresses are developed at the junc-
tions of major components of a pressure vessel. Secondary
mean stresses are also produced by sustained loads other
than internal or external pressure. Radial loads on nozzles
produce secondary mean stresses in the shell at the junction
of the nozzle. Secondary stresses are strain-induced stresses.
Discontinuity stresses are only considered as secondary
stresses if their extent along the length of the shell is limited.
Division 2 imposes the restriction that the length over which
the stress is secondary is m.Beyond this distance, the
stresses are considered as primary mean stresses. In a cylin-
drical vessel, the length arepresents the length over
which the shell behaves as a ring.
A further restriction on secondary stresses is that they may
not be closer to another gross structural Qscontinuity than
a distance of 2 . 5 m . This restriction is to eliminate the
additive effects of edge moments and forces.
Secondary stresses are divided into two additional groups,
membrane and bending. Examples of each are as follows:
Seconday membrane stress, Q,,,.
a. Axial stress at the juncture of a flange and the hub of
b. Thermal stresses.
c. Membrane stress in the knuckle area of the head.
d. Membrane stress due to local relenting loads.
Secondary bending stress, QL.
a. Bending stress at a gross structural discontinuity:
b. The nonuniform portion of the stress distribution in a
c. The stress variation of the radial stress due to internal
d. Discontinuity stresses at stiffening or support rings.
nozzles, lugs, etc. (relenting loadings only).
thick-walled vessel due to internal pressure.
pressure in thick-walled vessels.
Note: For b and c it is necessary to subtract out the average
stress which is the primary stress. Only the varymg part of
the stress distribution is a secondary stress.
Peak stress, E Peak stresses are the additional stresses due
to stress intensification in highly localized areas. They apply
to both sustained loads and self-limiting loads. There are no
significant distortions associated with peak stresses. Peak
stresses are additive to primary and secondary stresses pre-
sent at the point of the stress concentration. Peak stresses are
only significant in fatigue conditions or brittle materials.
Peak stresses are sources of fatigue cracks and apply to
membrane, bending, and shear stresses. Examples are:
a. Stress at the corner of a discontinuity.
b. Thermal stresses in a wall caused by a sudden change
c. Thermal stresses in cladding or weld overlay.
d. Stress due to notch effect (stress concentration).
in the surface temperature.
Categories of Stress
Once the various stresses of a component are calculated,
they must be combined and this final result compared to an
10 Pressure Vessel Design Manual
allowable stress (see Table 1-1). The combined classes of
stress due to a combination of loads acting at the same
time are stress categories. Each category has assigned
limits of stress based on the hazard it represents to the
vessel. The following is derived basically from ASME
Code, Section VIII, Division 2, simplified for application to
Division 1vessels and allowable stresses. It should be used as
a guideline only because Division 1 recognizes only two
categories of stress-primary membrane stress and primary
bending stress. Since the calculations of most secondary
(thermal and discontinuities) and peak stresses are not
included in this book, these categories can be considered
for reference only. In addition, Division 2 utilizes a factor
K multiplied by the allowable stress for increase due to
short-term loads due to seismic or upset conditions. It also
sets allowable limits of combined stress for fatigue loading
where secondary and peak stresses are major considerations.
Table 1-1sets allowable stresses for both stress classifications
and stress categories.
Allowable Stresses for Stress Classifications and Categories
Stress Classification or Cateaorv
General primary membrane, P,
General primary bending, Pb
Local primary membrane, PL
Secondary membrane, Q,
Secondary bending, Qb
p m f Pb +em+Qb
p m +Pb +Q& +Qb
Pm +Pb +Q& +Qb +F
1.5SE < .9Fy
1.5SE 4 .9Fy
1.5SE < .9Fy
3SE < 2Fy UTS
3SE < 2Fy < UTS
1.5SE < .9Fy
3SE < 2Fy < UTS
Q,, =membrane stresses from sustained loads
W, =membrane stresses from relenting, self-limiting loads
S=allowable stress per ASME Code, Section VIII, Division 1, at design
F,= minimum specified yield strength at design temperature
UTS=minimum specified tensile strength
S,=allowable stress for any given number of cycles from design fatigue curves.
This book provides detailed methods to cover those areas
most frequently encountered in pressure vessel design. The
topics chosen for this section, while of the utmost interest to
the designer, represent problems of a specialized nature. As
such, they are presented here for information purposes, and
detailed solutions are not provided. The solutions to these
special problems are complicated and normally beyond the
expertise or available time of the average designer.
The designer should be familiar with these topics in order
to recognize when special consideration is warranted. If
more detailed information is desired, there is a great deal
of reference material available, and special references have
been included for this purpose. Whenever solutions to prob-
lems in any of these areas are required, the design or analysis
should be referred to experts in the field who have proven
experience in their solution.
~ ~ ~ ~
As discussed previously, the equations used for design of
thin-walled vessels are inadequate for design or prediction of
failure of thick-walled vessels where R,,/t < 10. There are
many types of vessels in the thick-walled vessel category as
outlined in the following, but for purposes of discussion here
only the monobloc type will be discussed. Design of thick-
wall vessels or cylinders is beyond the scope of this book, but
it is hoped that through the following discussion some insight
will be gained.
In a thick-walled vessel subjected to internal pressure, both
circumferential and radlal stresses are maximum on the
inside surface. However, failure of the shell does not begin
at the bore but in fibers along the outside surface of the shell.
Although the fibers on the inside surface do reach yield first
they are incapable of failingbecause they are restricted by the
outer portions of the shell. Above the elastic-breakdown pres-
sure the region of plastic flow or “overstrain” moves radially
outward and causes the circumferential stress to reduce at the
inner layers and to increase at the outer layers. Thus the
maximum hoop stress is reached first at the outside of the
cylinder and eventual failure begins there.
The major methods for manufacture of thick-walled
pressure vessels are as follows:
1. Monobloc-Solid vessel wall.
2. Multilayer-Begins with a core about ‘/z in. thick and
successivelayersareapplied. Each layerisvented (except
the core) and welded individually with no overlapping
3. Multiwall-Begins with a core about 1%in. to 2 in.
thick. Outer layers about the same thickness are suc-
cessively “shrunk fit” over the core. This creates com-
pressive stress in the core, which is relaxed during
pressurization. The process of compressing layers is
called autofrettage from the French word meaning
4.Multilayer autofirettage-Begins with a core about
‘/z in. thick. Bands or forged rings are slipped outside
Stresses in Pressure Vessels 11
and then the core is expanded hydraulically. The
core is stressed into plastic range but below ultimate
strength. The outer rings are maintained at a margin
below yield strength. The elastic deformation resi-
dual in the outer bands induces compressive stress
in the core, which is relaxed during pressurization.
5. Wire wrapped z)essels--Begin with inner core of thick-
ness less than required for pressure. Core is wrapped
with steel cables in tension until the desired auto-
frettage is achieved.
6. Coil wrapped cessels-Begin with a core that is subse-
quently wrapped or coiled with a thin steel sheet until
the desired thickness is obtained. Only two longitudinal
welds are used, one attaching the sheet to the core and
the final closure weld. Vessels 5 to 6ft in diameter for
pressures up to 5,OOOpsi have been made in this
Other techniques and variations of the foregoing have been
used but these represent the major methods. Obviously
these vessels are made for very high pressures and are very
For materials such as mild steel, which fail in shear rather
than direct tension, the maximum shear theory of failure
should be used. For internal pressure only, the maximum
shear stress occurs on the inner surface of the cylinder. At
this surface both tensile and compressive stresses are max-
imum. In a cylinder, the maximum tensile stress is the cir-
cumferential stress, 06. The maximum compressive stress is
the radial stress, or.These stresses would be computed as
Therefore the maximum shear stress, 5 , is :
ASME Code, Section VIII, Division 1, has developed
alternate equations for thick-walled monobloc vessels. The
equations for thickness of cylindrical shells and spherical
shells are as follows:
0 Cylindrical shells (Para. 1-2 (a) (1))where t > .5 Ri or
P > ,385 SE:
S E + P
SE - P
Figure 1-3. Comparision of stress distribution between thin-walled (A)
and thick-walled (B) vessels.
0 Spherical shells (Para. 1-3)where t > ,356 Rior P >.665SE:
2SE - P
The stress distribution in the vessel wall of a thick-walled
vessel varies across the section. This is also true for thin-
walled vessels, but for purposes of analysis the stress is
considered uniform since the difference between the inner
and outer surface is slight. A visual comparison is offered
in Figure 1-3.
Whenever the expansion or contraction that would occur
normally as a result of heating or cooling an object is
prevented, thermal stresses are developed. The stress is
always caused by some form of mechanical restraint.
12 Pressure Vessel Design Manual
Thermal stresses are “secondary stresses” because they
are self-limiting. That is, yielding or deformation of the
part relaxes the stress (except thermal stress ratcheting).
Thermal stresses will not cause failure by rupture in
ductile materials except by fatigue over repeated applica-
tions. They can, however, cause failure due to excessive
Mechanical restraints are either internal or external.
External restraint occurs when an object or component is
supported or contained in a manner that restricts thermal
movement. An example of external restraint occurs when
piping expands into a vessel nozzle creating a radial load
on the vessel shell. Internal restraint occurs when the tem-
perature through an object is not uniform. Stresses from
a “thermal gradient” are due to internal restraint. Stress is
caused by a thermal gradient whenever the temperature dis-
tribution or variation within a member creates a differential
expansion such that the natural growth of one fiber is
influenced by the different growth requirements of adjacent
fibers. The result is distortion or warpage.
A transient thermal gradient occurs during heat-up and
cool-down cycles where the thermal gradient is changing
Thermal gradients can be logarithmic or linear across a
vessel wall. Given a steady heat input inside or outside a tube
the heat distribution will be logarithmic if there is a tem-
perature difference between the inside and outside of the
tube. This effect is significant for thick-walled vessels. A
linear temperature distribution occurs if the wall is thin.
Stress calculations are much simpler for linear distribution.
Thermal stress ratcheting is progressive incremental
inelastic deformation or strain that occurs in a component
that is subjected to variations of mechanical and thermal
stress. Cyclic strain accumulation ultimately can lead to
incremental collapse. Thermal stress ratcheting is the result
of a sustained load and a cyclically applied temperature
The fundamental difference between mechanical stresses
and thermal stresses liesin the nature of the loading. Thermal
stresses as previously stated are a result of restraint or tem-
perature distribution. The fibers at high temperature are
compressed and those at lower temperatures are stretched.
The stress pattern must only satisfy the requirements for
equilibrium of the internal forces. The result being that
yielding will relax the thermal stress. If a part is loaded
mechanically beyond its yield strength, the part will continue
to yield until it breaks, unless the deflection is limited by
strain hardening or stress redistribution. The external load
remains constant, thus the internal stresses cannot relax.
The basic equations for thermal stress are simple but
become increasingly complex when subjected to variables
such as thermal gradents, transient thermal gradients,
logarithmic gradients, and partial restraint. The basic equa-
tions follow. If the temperature of a unit cube is changed
Figure 1-4. Thermal linear gradient across shell wall.
from TI to Tz and the growth of the cube is fully
where T1= initial temperature, O F
Tz=new temperature, O F
(11=mean coefficient of thermal expansion in./in./”F
E =modulus of elasticity, psi
v =Poisson’s ratio =.3 for steel
AT =mean temperature difference, O F
Case 1: If the bar is restricted only in one direction but free
to expand in the other drection, the resulting uniaxial
stress, 0,would be
0 = -Ea(Tz - TI)
0 If Tt > TI, 0 is compressive (expansion).
0 If TI > Tz, 0 is tensile (contraction).
Case 2: If restraint is in both directions, x and y, then:
0,= cy= -(~IEAT/1- o
Case 3: If restraint is in all three directions, x, y, and z, then
0,= oy= 0,= -aE AT11 - 2~
Case 4: If a thermal linear gradient is across the wall of a
thin shell (see Figure 14),then:
0, = O+ = f(11EAT/2(1- V)
This is a bending stress and not a membrane stress. The hot
side is in tension, the cold side in compression. Note that this
is independent of vessel diameter or thickness. The stress is
due to internal restraint.
Vessel sections of different thickness, material, dameter,
and change in directions would all have different displace-
ments if allowed to expand freely. However, since they
are connected in a continuous structure, they must deflect
and rotate together. The stresses in the respective parts at or
near the juncture are called discontinuity stresses. Disconti-
nuity stresses are necessary to satisfy compatibility of defor-
mation in the region. They are local in extent but can be of
Stresses in Pressure Vessels 13
very high magnitude. Discontinuity stresses are “secondary
stresses” and are self-limiting. That is, once the structure
has yielded, the stresses are reduced. In average application
they will not lead to failure. Discontinuity stresses do
become an important factor in fatigue design where cyclic
loadlng is a consideration. Design of the juncture of the
two parts is a major consideration in reducing discontinuity
In order to find the state of stress in a pressure vessel, it is
necessary to find both the membrane stresses and the dis-
continuity stresses. From superposition of these two states
of stress, the total stresses are obtained. Generally when
combined, a higher allowable stress is permitted. Due to
the complexity of determining dlscontinuity stress, solutions
will not be covered in detail here. The designer should be
aware that for designs of high pressure (>1,500psi), brittle
material or cyclic loading, discontinuity stresses may be a
Since discontinuity stresses are self-limiting, allowable
stresses can be very high. One example specifically
addressed by the ASME Code, Section VIII, Division 1,
is discontinuity stresses at cone-cylinder intersections
where the included angle is greater than 60”. Para. 1-5(e)
recommends limiting combined stresses (membrane + dis-
continuity) in the longitudinal direction to 4SE and in the
circumferential direction to 1.5SE.
ASME Code, Section VIII, Division 2, limits the com-
bined stress, primary membrane and discontinuity stresses
to 3S,,, where S, is the lesser of %FFyor ‘/,U.T.S.,whichever
There are two major methods for determining dis-
1. Displacement Method-Conditions of equilibrium are
2. Force Method-Conditions of compatibility of dis-
See References 2, Article 4-7; 6, Chapter 8; and 7,
Chapter 4 for detailed information regarding calculation of
expressed in terms of displacement.
placements are expressed in terms of forces.
ASME Code, Section VIII, Division 1, does not speci-
fically provide for design of vessels in cyclic service.
Although considered beyond the scope of this text as well,
the designer must be aware of conditions that would require
a fatigue analysis to be made.
When a vessel is subject to repeated loading that could
cause failure by the development of a progressive fracture,
the vessel is in cyclic service. ASME Code, Section VIII,
Division 2, has established specific criteria for determining
when a vessel must be designed for fatigue.
It is recognized that Code formulas for design of details,
such as heads, can result in yielding in localized regions.
Thus localized stresses exceeding the yield point may be
encountered even though low allowable stresses have been
used in the design. These vessels, while safe for relatively
static conditions of loading, would develop “progressive frac-
ture” after a large number of repeated loadings due to these
high localized and secondary bending stresses. It should be
noted that vessels in cyclic service require special considera-
tion in both design and fabrication.
Fatigue failure can also be a result of thermal variations as
well as other loadings. Fatigue failure has occurred in boiler
drums due to temperature variations in the shell at the feed
water inlet. In cases such as this, design details are of
Behavior of metal under fatigue conrlltions vanes signifi-
cantly from normal stress-strain relationships. Damage
accumulates during each cycle of loading and develops at
localized regions of high stress until subsequent repetitions
finally cause visible cracks to grow, join, and spread. Design
details play a major role in eliminating regions of stress
raisers and discontinuities. It is not uncommon to have
the design strength cut in half by poor design details.
Progressive fractures develop from these discontinuities
even though the stress is well below the static elastic strength
of the material.
In fatigue service the localized stresses at abrupt changes
in section, such as at a head junction or nozzle opening,
misalignment, defects in construction, and thermal gradients
are the significant stresses.
The determination of the need for a fatigue evaluation is
in itself a complex job best left to those experienced in this
type of analysis. For specific requirements for determining if
a fatigue analysis is required see ASME Code, Section VIII,
Division 2, Para. AD-160.
For additional information regarding designing pressure
vessels for fatigue see Reference 7, Chapter 5.
14 Pressure Vessel Design Manual
ASME Boiler and Pressure Vessel Code, Section VIII,
Division 1, 1995 Edition, American Society of
ASME Boiler and Pressure Vessel Code, Section VIII,
Division 2, 1995 Edition, American Society of
Popov, E. P., Mechanics of Materials, Prentice Hall,
Bednar, H. H., Pressure Vessel Design Handbook,
Van Nostrand Reinhold Co., 1981.
Harvey, J. F., Theory and Design of Modern Pressure
Vessels,Van Nostrand Reinhold Co., 1974.
Hicks, E. J. (Ed.), Pressure Vessels-A Workbook for
Engineers, Pressure Vessel Workshop, Enera-
Sources Technology Conference and Exhibition,
Houston, American Society of Petroleum Engineers,
January 19-21, 1981.
Pressure Vessel and Piping Design, Collected Papers
1927-1959, American Society of Mechanical
Brownell, L. E., and Young, E. H., Process Equipment
Design, John Wiley and Sons, 1959.
Roark, R. J., and Young, W. C., Formulasfor Stress and
Strain, 5th Edition, McGraw Hill Book Co., 1975.
Burgreen, D., Design Methods for Power Plant
Structures, C. P. Press, 1975.
Criteria of the ASME Boiler and Pressure Vessel Code
for Design by Analysis in Sections I11and VIII, Division
2, American Society of Mechanical Engineers.
GENERALVESSEL FORMULAS 11, 21
P =internal pressure, psi
D,, D, =insidehtside diameter, in.
S =allowable or calculated stress, psi
E =joint efficiency
L =crown radius, in.
K,, R, =insidehutside radius, in.
K, M =coefficients (See Note 3)
crx =longitudinal stress, psi
crTd=circumferential stress, psi
R,,, =mean ra&us of shell, in.
t =thickness or thickness required of shell, head,
r =knuckle radius, in.
or cone, in.
1. Formulas are valid for:
a. Pressures < 3,000psi.
b. Cylindrical shells where t 5 0.5R, or P 5 0.385 SE.
For thicker shells see Referencr 1,Para. 1-2.
c. Spherical shells and hemispherical heads where
t 5 0.356 R, or P 5 0.665 SE. For thicker shells see
Reference 1,Para. 1-3.
2, All ellipsoidal and torispherical heads having a mini-
mum specified tensile strength greater than 80,000psi
shall be designed using S =20,000psi at ambient tem-
perature and reduced by the ratio of the allowable
stresses at design temperature and ambient tempera-
ture where required.
Figure 2-1. General configuration and dimensional data for vessel
shells and heads.
3. Formulas for factors:
$Table 2-1 v)
GeneralVessel Formulas 2
Stress, SThickness, t Pressure, P
I.D. OB. I.D. 0.0. 1.0. 0.0.Part Stress Formula
[I, Section 1-1(a)(2);
[ l, Section 1-4(c)]
[l, Section UG-32(d)]
Ro - 1.4t
P(RiEt+0.6t) P(R0Et- 0.4t)SEt
Ro - 0.4t
P(R0 - 0.8t)
Ro - 0.8t
ax= a+ =-
2t 2SE - 0.2P
2SE - 0.2P
KDo -2t(K - 0.1)
Do - 1.8t
2SE +2P(K - 0.1)
2SE - 0.2P
Torispherical Urc 16.66
1,M - t(M - 0.2)
P b M
2SE +P(M- 0.2)
2SE - 0.2P
Longitudinal P(Di - 0.8tCOS a) P(Do-2.8tc0~CX)
4Etcos cx 4Etcos 0:
Do - 2.8tCOS O:
Di - 0.8tCOS cx
PRm PDi PDo
2tcos 0: 4COS cx (SE +0.4P) 4c0S cx (SE+1.4P)
P(Di +1CO COS K) P(Do- 0.8tc0~CX)
2Etcos cx2Etcos cx
Do- 0.8tCOS cx
PRm PDi PDo
2COS O: (SE -0.6P) 2 ~ 0 sN (SE+0.4P)
General Design 17
24 36 48 60 72 84 96 108 120 132 144 156 168
Figure 2-la. Required shell thickness of cylindrical shell.
18 PressureVessel Design Manual
WWm 0 N d W
2 9 2 ? 7
Vessel Diameter, Inches
Figure 2-la. (Continued)
General Design 19
EXTERNAL PRESSURE DESIGN
A =factor “A,” strain, from ASME Section TI, Part
A, =cross-sectional area of stiffener, in.2
D, Subpart 3, dimensionless
R =factor “B,” allowable compressive stress, from
D =inside diameter of cylinder, in.
Do=outside diameter of cylinder, in.
]I1,=outside diameter of the large end of cone, in.
D, =outside diameter of small end of cone, in.
E =modulus of elasticity, psi
I =actual moment of inertia of stiffener, in.
I, =required moment of inertia of stiffener, in.4
I: =required moment of inertia of combined shell-
ring cross section, in.
L, =for cylinders-the design length for external
pressure, including k the depth of heads, in.
For cones-the design length for external pres-
snre (see Figures 2-lb and 2-lc), in.
ASME Section 11, Part D, Subpart 3, psi
L,, =equivalent length of conical section, in.
L, =length between stiffeners, in.
I,, - T =length of straight portion of shell, tangent to
P =design internal pressure, psi
P;,=allowable external pressure, psi
P, =design external pressure, psi
R,, =outside radius of spheres and hemispheres,
t =thickness of cylinder, head or conical section, in.
crown radius of torispherical heads, in.
t,, =equivalent thickness of cone, in.
c( =half apex angle of cone, degrees
Unlike vessels which are designed for internal pressure
alone, there is no single formula, or unique design, which
fits the external pressure condition. Instead, there is a range
of options a~ail~bleto the designer which can satisfy the
solution of the design. The thickness of the cylinder is only
one part of the design. Other factors which affect the design
are the length of cylinder and the use, size, and spacing of
stiffening rings. Designing vessels for external pressure is an
iterative procedure. First, a design is selected with all of the
variables included, then the design is checked to determine
if it is adequate. If inadequate, the procedure is repeated
until an acceptable design is reached.
Vessels subject to external pressure may fail at well below
the yield strength of the material. The geometry of the part is
the critical factor rather than material strength. Failures can
occur suddenly, by collapse of the component.
External pressure can be caused in pressure vessels by a
variety of conditions and circumstances. The design pressure
may be less than atmospheric due to condensing gas or
steam. Often refineries and chemical plants design all of
their vessels for some amount of external pressure, regarcl-
less of the intended service, to allow fbr steam cleaning and
the effects of the condensing steam. Other vessels are in
vacuum service by nature of venturi devices or connection
to a vacuum pump. Vacuums can be pulled inadvertently by
failure to vent a vessel during draining, or from improperly
External pressure can also be created when vessels are
jacketed or when components are within iririltichairibererl
vessels. Often these conditions can be many times greater
than atmospheric pressure.
When vessels are designed for bot11internal arid external
pressure, it is common practice to first determine the shell
thickness required for the internal pressure condition, then
check that thickness for the maximum allowable external
pressure. If the design is not adequate then a decision is
made to either bump up the shell thickness to the next
thickness of plate available, or add stiffening rings to
reduce the “L’dimension. If the option of adding stiffening
rings is selected, then the spacing can be determined to suit
the vessel configuration.
Neither increasing the shell thickness to remove stiffening
rings nor using the thinnest shell with the Inaximum number
of stiffeners is economical. The optimum solution lies some-
where between these two extremes. Typically, the utilization
of rings with a spacing of 2D for vessel diairictcrs up to abont
eight feet in diameter and a ring spacing of approximately
“D” for diameters greater than eight feet, provides an eco-
The design of the stiffeners themselves is also a trial and
error procedure. The first trial will be quite close if the old
APT-ASME formula is used. The forinula is its follows:
Stiffeners should never be located over circurnferentlal
weld seams. If properly spaced they may also double as insu-
lation support rings. Vacuum stiffeners, if coinbined with
other stiffening rings, such as cone reinforcement rings or
saddle stiffeners on horizontal vessels, must be designed for
the combined condition, not each independently. If at all
20 Pressure Vessel Design Manual
possible, stiffeners should always clear shell nozzles. If una-
voidable, special attention should be given to the design of a
boxed stiffener or connection to the nozzle neck.
Design Procedure For Cylindrical Shells
Step 1:Assume a thichess if one is not already determined.
Step 2: Calculate dimensions “L’and “D.” Dimension “L’
should include one-third the depth of the heads. The over-
all length of cylinder would be as follows for the various
L = LT-T +0.333D
L = LT-T +0.16661)
W/(2) 100% - 6% heads L = h - ~+0.112D
Step 3: Calculate UD, and D,Jt ratios
Step 4:Determine Factor “A’from ASME Code, Section 11,
Part D, Subpart 3, Fig G: Geometric Chart for
Components Under External or Compressive Loadings
(see Figure 2-le).
Step 5: Using Factor “A’ determined in step 4, enter the
applicable material chart from ASME Code, Section 11,
Part D, Subpart 3 at the appropriate temperature and
determine Factor “B.”
Step 6: If Factor “A’falls to the left of the material line, then
utilize the following equation to determine the allowable
Step 7: For values of “A’ falling on the material line of the
applicable material chart, the allowable external pressure
should be computed as follows:
Step 8: If the computed allowable external pressure is less
than the design external pressure, then a decision must be
made on how to proceed. Either (a)select a new thickness
and start the procedure from the beginning or (b) elect
to use stiffening rings to reduce the “L’ hmension. If
stiffening rings are to be utilized, then proceed with the
Step 9: Select a stiffener spacing based on the maximum
length of unstiffened shell (see Table 2-la). The stiffener
spacing can vary up to the maximum value allowable for
the assumed thickness. Determine the number of stiffen-
ers necessary and the correspondmg “L’dimension.
Step 10:Assume an approximate ring size based on the fol-
Step 11:Compute Factor “B” from the following equation
utilizing the area of the ring selected:
g = -
Step 12: Utilizing Factor “B” computed in step 11,find the
corresponding “A’ Factor from the applicable material
Step 13:Determine the required moment of inertia from the
following equation. Note that Factor “A” is the one found
in step 12.
Ls(t+As /Ls 1-41
Step 14: Compare the required moment of inertia, I, with
the actual moment of inertia of the selected member. If
the actual exceeds that which is required, the design is
acceptable but may not be optimum. The optimization
process is an iterative process in which a new member
is selected, and steps 11 through 13 are repeated
until the required size and actual size are approximately
1. For conical sections where c( < 22.5 degrees, design the
cone as a cylinder where Do=DL and length is equal
2. If a vessel is designed for less than 15psi, and the
external pressure condition is not going to be stamped
on the nameplate, the vessel does not have to be
designed for the external pressure condition.
Portionpf a cone
General Design 21
Case B Case c I
Case D Case E
Figure 2-1b. External pressurecones 22 1/2"< a<60".
For Case B, L,.=L
For Cases A, C, D, E:
t,. =tcos c(
22 Pressure Vessel Design Manual
Large End Small End
Figure 2-lc. Combined shelkone
Design stiffener for large end of cone as cylinder
Do = DL
t = tI>
Design stiffener for small end of cone as cylinder where:
Do = Ds
t = t,
L --+- 1+-
" 2 L3 "[2 3
Ro= 0.9 Do
R o w = Do
Sphere/Hemisphere 2:l S.E. Head
Figure 2-ld. External pressure -spheres and heads.
General Design 23
Figure 2-le. Geometric chart for componentsunder external or compressiveloadings (for all materials). (Reprintedby permission from the ASME
Code. Section VIII, Div. 1.)
Design Procedure For Spheres and Heads
Step 1:Assume a thickness and calculate Factor “A.”
A = -
Step 2: Find Factor ‘‘€3” from applicable material chart.
Step 3: Compute Pa.
Figure2-If. Chart for determining shell thickness of components under external pressure when constructed of carbon or low-alloy steels (specified
minimum yield strength 24,OOOpsi to, but not including, 30,OOOpsi). (Reprinted by permissionfrom the ASME Code, Section VIII, Div. 1.)
o.oooo1 a m i 0.001
0.01 0 1
Figure2-1g. Chart for determining shell thickness of components under external pressurewhen constructed of carbon or low-alloysteels (specified
minimum yield strength 30,OOOpsi and over except materialswithin this range where other specific charts are referenced) and type 405 and type 410
stainless steels. (Reprinted by permission from the ASME Code, Section VIII, Div. 1.)
General Design 25
P, = ~A to left of material line
P. - -
A to right of material line
1. As an alternative, the thickness required for 2:l S.E.
heads for external pressure may be computed from the
formula for internal pressure where P =1.67P, and
E = 1.0.
Maximum Length of Unstiffened Shells
891 1,017 1,152
846 966 1,095
806 919 1,042
1. All values are in in.
2. Values are for temperatures up to 500°F.
3. Top value is for full vacuum, lower value is half vacuum.
4. Values are for carbon or low-alloy steel (Fv>3O,O0Opsi) based on Figure 2-19,
26 Pressure Vessel Design Manual
Moment of Inertia of Bar Stiffeners
Thk Max, Height, h, in.
t, in. ht, in.
1 1% 2 2% 3 3% 4 4% 5 5% 6 6% 7 7% 8
'I4 2 0.020 0.070 0.167
0.250 0.375 0.5
~ r [ T h S ,
5116 2.5 0.026 0.088 0.208 0.407
0.313 0.469 0.625 0.781
"I8 3 0.031 0.105 0.25 0.488 0.844
0.375 0.563 0.75 0.938 1.125 th3I = -12
7/16 3.5 0.123 0.292 0.570 0.984 1.563
0.656 0.875 1.094 1.313 1.531
4 0.141 0.333 0.651 1.125 1.786 2.667
0.75 1.00 1.25 1.50 1.75 2.00
'lI6 4.5 0.375 0.732 1.266 2.00 3.00 4.271
1.125 1.406 1.688 1.969 2.25 2.53
"8 5 0.814 1.41 2.23 3.33 4.75 6.510
1.563 1.875 2.188 2.50 2.813 3.125
1.55 2.46 3.67 5.22 7.16 9.53
2.063 2.406 2.75 3.094 3.438 3.78
'I4 6 1.69 2.68 4.00 5.70 7.81 10.40 13.5
2.25 2.625 3.00 3.375 3.75 4.125 4.50
l3lI6 6.5 2.90 4.33 6.17 8.46 11.26 14.63 18.59
2.844 3.25 3.656 4.063 4.469 4.875 5.281
7 4.67 6.64 9.11 12.13 15.75 20.02 25.01
3.50 3.94 4.375 4.813 5.25 5.688 6.125
5.33 7.59 10.42 13.86 18.00 22.89 28.58 35.16 42.67
4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00
Note: Upper value in table is the moment of inertia. Lower value is the area
General Design 27
Moment of Inertia of Composite Stiffeners
Moment of Inertia of Stiffening Rings
Figure 2-1h. Case 1: Bar-type stiffening ring. Figure 2-li. Case 2: T-type stiffening ring.
28 Pressure Vessel Design Manual
STIFFENING RING CHECK FOR EXTERNAL PRESSURE
If B> 2,500 psi,
determine A from
I Moment of inerliawlo shell
Moment of inertia w/ shell
D2LO +A S / W1: =
From Ref. 1, Section UG-29.
CALCULATE MAP. MAWP. AND TEST PRESSURES
Sa=allowable stress at ambient temperature, psi
SDT =allowable stress at design temperature, psi
SCA =allowable stress of clad material at ambient tempera-
SCD =allowable stress of clad material at design tempera-
SBA =allowable stress of base material at ambient tempera-
SBD =allowable stress of base material at design tempera-
C.a. =corrosion allowance, in.
t,, =thickness of shell, corroded, in.
t,, =thickness of shell, new, in.
thc=thickness of head, corroded, in.
thn =thickness of head, new, in.
tb =thickness of base portion of clad material, in.
t, =thickness of cladding, in,
R, =inside radius, new, in.
R, =inside radius, corroded, in.
R, =outside radius, in.
D, =inside diameter, new, in.
D, =inside diameter, corroded, in.
D, =outside diameter, in.
P =design pressure, psi
Ps =shop hydro pressure (new and cold), psi
PF =field hydro pressure
(hot and corroded), psi
E =joint efficiency, see Procedure 2-1
and Appendix C
Maximum Allowable Working Pressure (MAWP):The
MAWP for a vessel is the maximum permissible pressure at
the top of the vessel in its normal operating position at a
specific temperature, usually the design temperature.
When calculated, the MAWP should be stamped on the
nameplate. The MAWP is the maximum pressure allowable
in the “hot and corroded’ condtion. It is the least of the
values calculated for the MAWP of any of the essential parts
of the vessel, and adjusted for any difference in static head
that may exist between the part considered and the top of
the vessel. This pressure is based on calculations for every
element of the vessel using nominal thicknesses exclusive of
corrosion allowance. It is the basis for establishing the set
pressures of any pressure-relieving devices protecting the
vessel. The design pressure may be substituted if the
MAWP is not calculated.
The MAWP for any vessel part is the maximum internal or
external pressure, including any static head, together with
the effect of any combination of loadings listed in UG-22
which are likely to occur, exclusive of corrosion allowance
at the designated coincident operating temperature. The
MAWP for the vessel will be governed by the MAWP of
the weakest part.
The MAWP may be determined for more than one de-
signated operating temperature. The applicable allowable
General Design 29
stress value at each temperature would be used. When more
than one set of conditions is specified for a given vessel, the
vessel designer and user should decide which set of condi-
tions will govern for the setting of the relief valve.
Maximum Allowable Pressure (MAP): The term MAP
is often used. It refers to the maximum permissible pressure
based on the weakest part in the new (uncorroded) and cold
condition, and all other loadings are not taken into consid-
Design Pressure: The pressure used in the design of a
vessel component for the most severe condition of coinci-
dent pressure and temperature expected in normal opera-
tion. For this condition, and test condition, the maximum
difference in pressure between the inside and outside of a
vessel, or between any two chambers of a combination unit,
shall be considered. Any thichess required for static head or
other loadings shall be additional to that required for the
Design Temperature: For most vessels, it is the tem-
perature that corresponds to the design pressure.
However, there is a maximum design temperature and a
minimum design temperature for any given vessel. The mini-
mum design temperature would be the MDMT (see
Procedure 2-17). The MDMT shall be the lowest tempera-
ture expected in service or the lowest allowable temperature
as calculated for the individual parts. Design temperature for
vessels under external pressure shall not exceed the maxi-
mum temperatures given on the external pressure charts.
Operating Pressure: The pressure at the top of the
vessel at which it normally operates. It shall be lower than
the MAWP, design pressure, or the set pressure of any pres-
sure relieving device.
Operating Temperature: The temperature that will be
maintained in the metal of the part of the vessel being con-
sidered for the specified operation of the vessel.
e MAWP, corroded at Design Temperature P,.
2:l S.E. Head:
e MAP, new and cold, P M
R, - 0.4tsn
2:l S.E. Head:
e Shop test pressure, Ps.
P, = 1 . 3 P ~or 1 . 3 P ~-
e Field test pressure, PF.
PF = 1.3P
e For clad vessels where credit is takenfor the clad material,
thefollowing thicknesses may be substituted into the equa-
tionsfor MAP and MAWP:
1. Also check the pressure-temperature rating of the
2. All nozzles should be reinforced for MAWP.
3. The MAP and MAWP for other components, i.e.,
cones, flat heads, hemi-heads, torispherical heads,
etc., may be checked in the same manner by using
the formula for pressure found in Procedure 2-1 and
substituting the appropriate terms into the equations.
4. It is not necessary to take credit for the cladding thick-
ness. If it is elected not to take credit for the cladding
thickness, then base all calculations on the full base
metal thickness. This assumes the cladding is equiva-
lent to a corrosion allowance, and no credit is taken for
the strength of the cladding.
flanges for MAWP and MAP.
30 PressureVessel Design Manual
STRESSES IN HEADS DUE TO INTERNAL PRESSURE [e,31
L =crown radius, in.
r =knuckle radius, in.
h =depth of head, in.
RL=latitudinal radius of curvature, in.
R, =meridional radius of curvature, in.
r ~ #=latitudinal stress, psi
ox=meridional stress, psi
P =internal pressure, psi
Lengths of RL and R, for ellipsoidal heads:
RL = R
At center of head:
a At any point X:
1. Latitudinal (hoop) stresses in the knuckle become com-
pressive when the R/h ratio exceeds 1.42. These heads
will fail by either elastic or plastic buckling, depending
on the R/t ratio.
2. Head types fall into one of three general categories:
hemispherical, torispherical, and ellipsoidal. Hemi-
spherical heads are analyzed as spheres and were
Figure 2-2. Direction of stresses in a vessel head.
:Figure 2-3. Dimensional data for a vessel head.
covered in the previous section. Torisphericz (also
known as flanged and dished heads) and ellipsoidal
head formulas for stress are outlined in the following
General Design 31
I TORISPHERICAL HEADS I
In Crown I
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
.5 .57 .65 .73 .81 .9 .99 1.08 1.18 1.27 1.36
In Knuckle I
At Tangent Line I
I ELLIPSOIDAL HEADS I
I At Any Point X I
At Center of HeadI I
I At Tangent Line I
DESIGN OF INTERMEDIATE HEADS [1, 31
A =factor A for external pressure
A, =shear area, in.2
B =allowable compressive stress, psi
F =load on weld(s), lb/in.
5 :shear stress, psi
E =joint efficiency
S =code allowable stress, psi
HL)=hydrostatic end force, lb
P, =maximum differential pressure on concave side of
El =modulus of elasticity at temperature, psi
P, =maximum differential pressure on convex side of
K =spherical radius factor (see Table 2-2)
L =inside radius of hemi-head, in.
=0.9D for 2:l S.E. heads, in.
=KD for ellipsoidal heads, in.
=crown radius of F & D heads, in.
Spherical Radius Factor, K
32 Pressure Vessel Design Manual
Figure24. Dimensional data for an intermediate head.
A, = t, + lesser of t2 or t,
sin e= 2 ~
A, = lesser of t2or t3
E= 0.7 (butt weld)
Required HeadThickness, t,
0 Znternal pressure, Pi.Select appropriate head formula
based on head geometry. For dished only heads as in
Figure 2-5, Case 3:
E = 0.7
E = 0.55
Design the weld attaching
the head as in Case 3 and
the welds attachingthe
reinforcingplate to share full
Case 3 Alternate
Figure 2-5. Methods of attachment of intermediate heads.
General Design 33
0 External pressure, P,. Assume corroded head thickness, th
Factor A = ____
Factor B can be taken from applicable material charts in
Section 11, Part D, Subpart 3 of Reference 1.
Alternatively (or if Factor A lies to the left of the material/
B = -
The required head thickness shall be the greater of that
required for external pressure or that required for an in-
ternal pressure equal to 1.67x P,. See Reference 1, Para.
where P = 1.5x greater of Pi or P,. (See Reference 1,
0 Shear loads on welds, F.
Note: sin8 applies to Figure 2-5, Case 3 head at-
0 Shear stress, 5.
0 Allowable shear stress, SE.
0 Hydrostatic end force, HD.
DESIGN OF TORICONICALTRANSITIONS 11,31
P =internal pressure, psi
S =allowable stress, psi
E =joint efficiency
PI, P2=equivalent internal pressure, psi
ul.uz =circumferential membrane stress, psi
fi, f2=longitudinal unit loads, lb/in.
a =half apex angle, deg
m=code correction factor for thickness of large
P, =external pressure, psi
MI, M2 =longitudinal bending moment at elevation, in.-lb
WI, Wz=dead weight at elevation, lb
Anyplace on cone
Figure 2-6. Dimensionaldata for a conical transition.
34 Pressure Vessel Design Manual
Calculating Angle, cc
I p& - 0'
0 > 0'
tan@ = -
a = + + @
L = COS 4dA2+B2
0' > 0
t a n @ = -
(Y = 90- 0 - 4
L = sin ~ J A ~+BZ
General Design 35
DI D - 2(R - R COS^)
2 cos a
Large End (Figure2-7)
Figure 2-7. Dimensionaldata for the large end of a conical transition.
e Maximum longitudinal loady,f i
( + ) tension; ( - ) compression
e D&rmine equiualent prmsiire!,PI
e Circumferential stress, D1
PLl '1,"' [;;Ic1=--- -
e Circumferential stress at DI withotit londs, c/1
e Thickness required knuckle, t,-k [I,section 1-4(d)].
2SE - 0.2P
e Thickness required cone, t,, [l,section UG-32(g)].
2 cosa(SE - 0.6P1)
2 COS u(SE - 0.W)
Small End (Figure 2-8)
e Maximum longitudinal loah,f 2 .
( + ) tension; ( - ) compression
e Determine equi.L;alentpressure, P2.
36 Pressure Vessel Design Manual
Figure 2-8. Dimensional data for the small end of a conical transition.
Circumferential stress at DZ.
Circumferential stress at 0 2 without loads, 02.
Thickness required cone, at DZ, t,, [l,section UG-32(g)].
2 cosa(SE - O.6P2)
2 COS a(SE - 0.6P)
Thickness required knuckle. There is no requirement for
thickness of the reverse knuckle at the small end of the
cone. For convenience of fabrication it should be made
the same thickness as the cone.
Additional Formulas (Figure2-9)
Thickness required of cone at any diameter D', tDl.
tD' = 2 cos a(SE - 0.6P)
Figure 2-9. Dimensional data for cones due to external pressure.
a Thickness required for external pressure [l,section UG-
te = tcosa
D ~ = D 2 + 2 b
D, = D1+ 2te
L = X - sina(R +t) - sina(r - t)
L - -e-:( 1+-::)
Using these values, use Figure 2-le to determine Factor A.
Allowable external pressure, Pa.
where E =modulus of elasticity at design temperature.
1. Allowable stresses. The maximum stress is the com-
pressive stress at the tangency of the large knuckle
and the cone. Failure would occur in local yielding
rather than buckling; therefore the allowable stress
should be the same as required for cylinders. Thus
the allowable circumferential compressive stress
should be the lesser of 2SE or F,. Using a lower allow-
able stress would require the knuckle radius to be
made very large-well above code requirements. See
2. Toriconid sections are mandatory if angle 01 exceeds
30" unless the design complies with Para. 1-5(e)of the
General Design 37
ASME Code [l]. This paragraph requires a discontinu-
ity analysis of the cone-shell juncture.
3. No reinforcing rings or added reinforcement is required
at the intersections of cones and cylinders, providing a
knuckle radius meeting ASME Code requirements is
used. The minimum knuckle radius for the large end
is not less than the greater of 3t or 0.12(R+t). The
knuckle radius of the small end (flare)has no minimum.
(See [Reference 1,Figure UG-361).
4. Toriconical transitions are advisable to avoid the high
discontinuity stresses at the junctures for the following
a. High pressure-greater than 300pig.
b. High temperature-greater than 450 or 500°F.
c. Low temperature-less than -20°F.
d. Cyclic service (fatigue).
DESIGN OF FLANGES [1,41
A =flange O.D., in.
AI, =cross-sectional area of bolts, in.
A,,, =total required cross-sectional area of
a =nominal bolt diameter, in.
B =flange I.D., in. (see Note 6)
B1=flange I.D., in. (see Note 6)
B, =bolt spacing, in.
b =effective gasket width, in.
b,, =gasket seating width, in.
C =bolt circle diameter, in.
d =hub shape factor
dl =bolt hole diameter, in.
E, hI), hc;, hT, R =radial distances, in.
e =hub shape factor
F =hub shape factor for integral-type
FL,=hub shape factor for loose-type flanges
f =hub stress correction factor for integral
G =diameter at gasket load reaction, in.
%=thickness of hub at small end, in.
gl =thickness of hub at back of flange, in.
H =hydrostatic end force, lb
HD=hydrostatic end force on area inside of
HG=gasket load, operating, lb
H, =total joint-contact surface compression
HT=pressure force on flange face, lb
h =hub length, in.
h, =hub factor
MD=moment due to HD,in.-lb
MG=moment due to HG, in.-lb
M, =total moment on flange, operating, in.&
ML =total moment on flange, seating
MT =moment due to HT, in.-lb
m, =unit load, operating, lb
mg=unit load, gasket seating, lb
m =gasket factor (see Table 2-3)
N =width of gasket, in. (see Table 2-4)
w =width of raised face or gasket contact width,
n =number of bolts
u =Poisson’s ratio, 0.3 for steel
P =design pressure, psi
S, =allowable stress, bolt, at ambient temperature,
S,, =allowable stress, bolt, at design temperature,
Sf,=allowable stress, flange, at ambient tempera-
Sf,,=allowable stress, flange, at design temperature,
SH=longitudinal hub stress, psi
SR=radial stress in flange, psi
ST=tangential stress in flange, psi T, U, Y
Z =K-factors (see Table 2-5)
T,, U,, Y, =K-factors for reverse flanges
t =flange thickness, in.
in. (see Table 2-4)
t, =pipe wall thickness, in.
V =hub shape factor for integral
VL=hub shape factor for loose flanges
W =flange design bolt load, lb
Wml=required bolt load, operating, lb
Wrrr2=required bolt load, gasket seating, Ib
y =gasket design seating stress, psi
38 Pressure Vessel Design Manual
_ _ _ _ _ _ _ _ _ _ _ _ ~
C - dia. HD
C - dia. HT
C - G
HT = H - HD
HG=operating = Wm, - H
gasket seating = W
(1 - u2)(K2 - l)U
(1 - U) +(1+v)K2
K2 - 1
Y = (1- v2)U
K2(I+4.6052 ( 1 + ~ / l-~)log,,K)-l
1.0472(K2- 1)(K - 1)(1+u)
B1 = loose flanges = B +g,
= integral flanges, f < 1 = B +g,
= integral flanges,f 2 1= B +g,
d = loose flanges = ~
= integral flanges = -
= reverse flanges = -
e = loose flanges = -
= integral flanges = -
G = (if b, 5 0.25 in.) mean diameter of gasket face
= (if bo > 0.25 in.) O.D. of gasket contact face - 2b
Stress Formula Factors
a = t e + l
f f f f
y = - or - for reverse flanges