Stresses in Pressure Vessel Design: Maximum Principal Stress Theory
Stresses in Pressure Vessel Design: Maximum Principal Stress Theory
Stress analysis is the determination of the relationship between external forces applied to a
vessel and the corresponding stress. In the design of pressure vessels, the components are
analyzed, so as to arrive at an economical and safe design. This is done by analyzing stresses
where necessary to determine the thickness of the material and size of the members. The
starting place for stress analysis is to determine all the design conditions for a given problem
and then determine all the external forces. The external forces are then related to the vessel
parts which must resist them by generating the corresponding stresses. There were many
theories developed to predict elastic failures in vessel design, however, the three most
commonly used are as follows:
This theory considers failure to have occurred when any one of the three principal stresses has
reached a value equal to the elastic limit as determined from a uni-axial tension or compression
test. This theory is the oldest, most widely used and simplest to apply. This theory is used for
bi-axial states of stresses assumed in thin-walled pressure vessels, and ASME Section VIII
Division-1 uses this theory as the basis for designing the equipment. While the maximum
principal stress theory does accurately predict failure in brittle materials, it is not always
accurate for ductile materials.
This theory is also called as Tresca Criterion. This theory considers failure to have occurred
when the maximum shear stress equals the shear stress at the elastic limit as determined by a
pure shear test. The maximum shear stress is defined as one-half the algebraic difference
between the largest and smallest of the three principal stresses. ASME Section VIII Division-
II prior to 2007 Edition utilized this maximum shear stress theory for designing equipment.
This theory is very closely approximates to the experimental results and is also easy to use.
This theory is called as Von Mises Stress Criterion theory. This theory considers failure to have
occurred when the distortion energy accumulated in the part under stress reaches the elastic
limit as determined by the distortion energy in a uni-axial tension or compression test. ASME
Section VIII Division-II from 2007 Edition started utilizes the distortion energy criterion for
designing equipment.
The ever increasing use of equipment industrial processing and power generation under
extreme conditions of pressure, temperature and environment has given special emphasis to
analytical and experimental methods for determining their operating stresses. Therefore, it is
necessary to understand the meaning and significance of these stresses. This includes the means
of determining the value and extent of the stresses and strains, establishing the behavior of the
material involved and evaluating the compatibility of these two factors in the media or the
environment to which they are subjected. Knowledge of material behavior is required not only
to avoid failures, but equally to permit maximum economy of material choice and amount
used.
The stresses produce changes in their dimensions known as strains. The determination of the
relationship between the external forces [P] applied to the vessel, and the stresses [σ] and strains
[e] within the vessel form the basis of this field of stress analysis. Engineering materials
commonly used in the design of structures and pressure vessels have an initial stress-strain
relationship which, for practical purposes, may be assumed linear, indicating that stress is
directly proportional to strain and is represented by the equation:
The value of the measure of material stiffness. Referring to the equation [1] above, the
accompanying strain will be less for a material with a high for one with a low E-Value. This
property of stiffness is very important in designs where deformation must be kept small, as, for
example, gasket joints, and the control rod portions of pressure vessels.
Approximate room temperature values of this modulus for some materials are:
An examination of the stress-strain diagram of Figure-1 above shows that they are made of two
general parts an initial elastic range for which hooks law generally applies, and the following
plastic range where the strains become large and this law no longer applies for design validation
of equipment.
Elasticity is the property of the material to return to its original shape after removal of the
external and internal loads.
Proportional limit is the greatest stress that a material can withstand without deviating from
the direct proportionality of stress to strain.
Elastic limit is the maximum stress which a material upon complete release of the stress.
Determination of the elastic limit is very difficult; however, it closely approximates the more
readily determined value of the proportional limit.
The yield point is the stress at which there occurs a marked increase in a strain without an
increase in stress.
Ultimate strength is the maximum stress that the material can withstand. This stress equals
the maximum load divided by the original cross-sectional area of the specimen.
Ductility is the property of a material to undergo deformation. Frequently it is necessary to
heat the material to high temperatures to increase its ductility during the forming operation.
Toughness is the ability of a material to absorb energy during plastic deformation. It is often
measured by the energy absorbed per unit of volume in stressing to rupture and is called the
modulus of toughness.
Significance of Stresses
Analytical formulas for the evaluation of stresses are usually based on elastic theory and elastic
behavior of the material, i.e., material which conforms to Hooke’s law, and at first glance, it
may appear that materials which follow this behavior right up to the breaking point would be
the most desirable for use. This is not the case, however, for instance, plaster of pairs has a
perfectly straight stress-strain curve up to the breaking point but, of course, is not a suitable
material for pressure vessels, because it is totally elastic and not partially plastic in its behavior.
It is this plastic property of the material, with its ability to give or yield under high peak or
local stress and so accommodate the applied loading by a more favorable distribution of
internal stress that is the most important property of a pressure vessel material. The elastically
computed or actual strength of most members, considering the pressure vessel as a whole,
would be considerably reduced if it were not accompanied by plastic deformation at various
relatively small portions of the member where high local stresses occur.
1. The type and nature of the applied loading, and the resulting stress distribution within the
member. For instance, is the applied loading mechanical or thermal, of a steady (static) or
unsteady (variable or cyclic) nature, and is the resulting stress pattern uniform or does it have
high peak values?
2. The ductile and plastic properties of the material. For instance, are the properties of the
material such that internal yielding or readjustment of strain can reduce the effects of local
stress concentrations?
The strength of a member does not depend only on the value of the maximum stress or strain
in the member, but also on the external shape readjustment that the member itself can make to
one more favorable than that assumed in the design, and on the plastic property of the material
to permit internal stress adjustment.
Pressure vessels are mainly subject to three types of loadings: namely, steady or static, unsteady
(variable or cyclic) and impact. Although practically all vessels encounter variable or cyclic
loadings, most pressure vessels may be assumed to be statically loaded without introducing
serious error. These vessels made of ductile materials and subject to static loads fail by gross
yielding. The ductility of the material allows a redistribution of stresses by plastic flow to
attenuate points of high local values toward a pattern more favorable to maximum resistance.
However, when the loading is such that the member is subjected to a considerable number of
stress cycles, even though the material is ductile, appreciable error can be introduced by
considering a static loading condition to exist in appraising integrity on the basis of simple
elastic formulas. Under such conditions, failure occurs due to a condition known as fatigue. In
fatigue, the failure is due to highly localized stresses which cause a minute fracture that
gradually spreads until the member is ruptured.
The stress distribution near the point of load applications, such as the support brackets on a
vessel, can vary greatly from the assumed pattern on which the ordinary equations are based,
and these local stresses may be relatively high. Even though the material is ductile and a
measure of stress redistribution can occur, these local stresses can be significant ones and are
frequently responsible for failures. In relatively brittle materials, or in ductile material
subjected to cyclic loading, stresses at point of load application may control the strength of the
member rather than the stresses given by the ordinary equations. This is particularly important
in vessels which are designed as membranes and cannot resist large bending moment
perpendicular to their surface, and yet for practical purposes must have support brackets, lifting
lugs, nozzles, etc., attached to them.
The basic equations for determining stresses are based on the assumption that the stresses in a
member are caused only by external loads and the residual stresses set up in the fabrication of
construction processes, such as weld shrinkage, casting cooling, metal heat treatment, etc., are
not considered. Although these stresses are secondary, since their value is self-limiting (they
are not produced by unrelenting external loads), they may be of great importance in brittle
materials, and even in ductile materials when the material is subjected to fatigue loading.
Equally important is the danger of creating, in conjunction with the applied loading stresses, a
three dimensional stress pattern in thick sections that is restrictive to the redistribution of high
localized peak stresses through yielding. It is for this reason that stress relieving of thick vessels
is much more important than thin ones in which the state of stress is essentially two-
dimensional.
Figure-2: Stress Discontinuities in Pressure Vessel Nozzle Junction
Shape of Member
The basic assumption for continuity of action in a member on which ordinary formulas for
direct stress and bending stress are based, requiring that a plane section remain plane after
bending, cannot hold near points of abrupt changes in section due to the restraining influence
of this stiffer portion on adjacent sections. Figure 2 illustrates this condition. The stresses in
the region influenced by these geometrical shape discontinuities are higher than that predicted
by the assumed mathematical law of distribution on which the ordinary stress formulas are
based, and are known as localized or concentration stresses. The errors introduced by the use
of ordinary formulas for the design of members with abrupt changes of section are generally
not serious if the load is static and the material ductile so as permit a slight measure of plastic
flow; hence, the member acts more nearly as assumed.
These localized stresses are, however, most important in brittle materials, even under static
loads, since under such conditions a redistribution or transfer of stress from the highly
overstressed material to the adjacent lower stressed material does not take place and the rupture
of member results. They are equally significant when cyclic loading is involved, even when
the material is ductile, since the region of high stress acts as a focal point from which fatigue
failure can stem.
The problem of evaluating localized stresses in vessels has assumed major importance lately
as the engineering advances have placed great pressure, temperature, and environmental
demands on pressure vessels. The petroleum and chemical processes require operating
pressures in the range of 5,000 psi to 10,000 psi range. The rapidly expanding cryogenics
industry has introduced low temperature conditions to minus 425 degF. The nuclear power
industry has given rise to high pressure, high temperature and special cyclic and material
irradiation operating conditions. All these requirements have focused considerable attention on
stress analysis, materials of construction, and economics of design of vessels for these services.
Membrane Stresses
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 dimensions
(Rm/t > 10), 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 resistance to bending.
When the wall offers resistance to bending, bending stresses occur in addition to the membrane
stresses.
The following list of stresses describes types of stress without regard to their effect on vessel
or component.
Types of stresses that are present in pressure vessels are separated into various classes in
accordance with the types of loads that produced them, and the hazard they represent to the
vessel. Each class of stress must be maintained below an acceptable level, and the combined
total stress must be kept under another acceptable level. The combined stresses due to a
combination of loads acting simultaneously are called stress categories.
Classes of Stresses
The pressure vessel codes define three important ‘classes’ of stress according to the types of
loading which produced them and the hazard they represent to the vessel.
Primary Stresses
Primary stress is related to mechanical loading directly and is the most hazardous of all types
of loads. These stresses act over a full cross section of the vessel. Any yielding through the
entire shell thickness will not distribute the stress, but will result in gross distortions, often
carried to failure. The main characteristic of primary stresses is that they are not self-limiting,
i.e., they are not reduced in magnitude by the deformation they produce. Primary stresses are
generally due to internal or external pressure or produced by sustained external forces and
moments. Thermal stresses are never classified as primary stresses.
Primary general stresses are divided into membrane and bending stresses. The need for
dividing primary general stress into membrane and bending is that the calculated value of the
primary general bending stress may be allowed to go higher than that of primary general
membrane stress.
Primary general membrane stress, Pm, occurs across the entire cross section of the vessel
due to mechanical loads such as internal pressure, dead weight, or wind loads. It is remote from
discontinuities such as head-shell intersections, cone-cylinder intersections, nozzles and
supports. It is so distributed in the structure that no redistribution of load occurs as a result of
yielding. ASME Section VIII Division-1 establishes allowable stresses by stating that the
maximum general primary membrane stress must be less than allowable stresses provided in
the material sections. Examples of such stresses are;
c) Longitudinal Stress due to the Bending of the Horizontal vessel over saddles,
Primary general bending stresses, Pb, are due to sustained loads and are capable of causing
collapse of the vessel. ASME Section VIII Division-1 states that the maximum primary
membrane stress plus primary bending stress may not exceed 1.5 times the allowable stress
provided in the material sections. There are relatively few areas where primary bending occurs;
a) Bending stress in the center of flat head or in the crown of a dished head,
Local primary stresses are not technically a classification of stress but a stress category, since
it is a combination of two stresses. The combination it represents is the primary membrane
stress, Pm, plus secondary membrane stress produced from sustained loads. These have been
grouped together in order to limit the allowable stress for this combination to a level lower than
allowed for other primary and secondary stress combinations. It is felt that local stress from
sustained (unrelenting) loads presents 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 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 stress 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 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 combined 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.
b) Pm + Membrane stresses from local sustained loads (Support lugs, Nozzle loads, Beam
supports, Major attachments)
Secondary Stresses
Secondary stresses are those arising from geometric discontinuities or stress concentrations. If
a member is subjected to a stress attributable to a thermal expansion load, such as bending
stresses in shell at a nozzle connection under thermal expansion of the piping, a slight,
permanent, local deformation in the shell wall will produce relaxation in the expansion forces
causing the stress. The basic characteristic of a secondary stress is that it is self-limiting; that
is, once the yield point has been passed locally around the stress concentration, the direct
relationship between load and stress is broken, due to the reduced post-yield stiffness of the
material. This means that local yielding and minor distortions can satisfy the conditions which
caused the stress to occur. The most important self-limiting stresses in the design of pressure
vessels are the stresses produced by thermal expansion and by internal pressure at shell
structure discontinuities. Application of a secondary stress can not cause structural failure due
to the restraints offered by the body to which the part is attached.
Secondary stresses are divided into two additional groups, membrane and bending. Examples
of secondary membrane stresses, Qm, are;
a) Axial stress at the juncture of a flange and the hub of the flange,
b) Thermal stresses,
d) Membrane stress due to local relenting loads. Examples of secondary bending stresses, Qb,
are ;
The practical difference between primary and secondary stresses is obvious and the criteria
used to evaluate the safety of primary stresses should not be applied to the calculated values of
stresses produced by self-limiting loads. Also, 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!
In a region away from any discontinuities, only primary stress will arise. The secondary stress
cannot arise alone however - at a discontinuity, the secondary stress will be superimposed on
the underlying primary stress.When we say away from discontinuities we are talking about
areas of high local stresses such as nozzle-to-shell or cone-to-cylinder junctions. It is
recognized in Section VIII Division-1 that high localized discontinuity stresses may exist in
vessels constructed to this standard. These stresses are not directly calculated but are controlled
to a safe level consistent with experience through design rules and mandatory fabrication
details,
Peak Stresses
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
present at the point of 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;
LITERATURE REFERENCE
2) Pressure Vessel Design Manual by Dennis or Moss Second Edition, Gulf International
Publication 1.
3) Pressure Vessel Design Handbook by Bedner Second Edition, Krieger Publishing Company
1998.