PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

STRESS CLASSIFICATION IN PRESSURE VESSELS AND PIPING

Arturs Kalnins
Department of Mechanical Engineering and Mechanics, Lehigh University, Bethlehem
PA 18015-3085 USA

Keywords: Design analysis, pressure vessels, piping, stress categories, primary stresses,
secondary stresses, limit analysis, limit load, stress limits, lower bound

Contents

1. Introduction
2. Primary Stress
3. Secondary Stress
Glossary
Bibliography
Biographical Sketch

Summary

The definitions and limits of primary and secondary stress categories are taken from the
ASME Boiler & Pressure Vessel Code. Primary stresses are defined as those that
develop the internal forces and moments needed to equilibrate the applied external
boundary forces and moments and body forces. Secondary stresses are those that
develop the internal forces and moments needed to satisfy internal or boundary
constraints. These definitions are taken as the basic premise for reach of the stress
categories. The general and local primary membrane, primary bending, and secondary
stresses, and their limits, are presented in a way that is consistent with that premise.
Limits on the applicability of the primary stress approach with respect to the wall
thicknesses of pressure vessels and piping are given. Examples are provided that explain
the basic concepts for the calculation of primary and secondary stresses.

1. Introduction

The main goal in pressure vessel design is to assure safe and satisfactory performance of
a vessel or piping. Stress classification provides a vehicle for reaching that goal. It is
recognized that different kinds of stress have different degrees of significance and must
be held to different limits. The purpose of stress classification is to identify these
different kinds of stress. They are placed in stress categories. The classification of
stresses into primary and secondary categories separates the issues regarding overall
strength, which is of primary importance and therefore referred to the realm of primary
stresses, from the issues of local behavior, which is of secondary importance and
therefore referred to the realm of secondary stresses.

The stress categories of interest to this chapter are those of primary stress, and its
subcategories of general and local primary membrane and bending stress, and the
secondary stresses. The category of peak stress relates to the assessment of fatigue
failure of metals and is not included in this chapter.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

In this chapter, primary stresses will be discussed first in Section 2, followed by a

discussion of secondary stresses in Section 3.

2. Primary Stress

2.1. Definition

The basis of the primary stress concept is rooted in the following definition cited in the
ASME Boiler & Pressure Vessel Code:

A primary stress is any normal stress or a shear stress developed by an imposed loading
that is necessary to satisfy the laws of equilibrium in terms of the external and internal
forces and moments.

When applied to the design of a pressure vessel, the important part of the definition is
the idea that equilibrium is required between forces and moments on plane sections
through the wall of a pressure-retaining boundary and not between stresses at all points
through the wall. This is a plate and shell theory concept as illustrated by a plate in
Figure 1. It should be noted at the outset that the use of the primary stress approach as
defined above implies limitations on the thickness of the wall to which it can be applied.

SCP-2 SCP-1

t
x1
t x2
SCL

Figure 1. Stack of cubes equilibrated by internal forces and moments

Adopting the terminology of shell theory, a reference surface is defined to represent the
shell. The analysis is performed with respect to coordinates on this surface, such as x1
and x2 in Figure 1. A normal at a point on this surface will be called in this chapter a
stress classification line, or SCL. It defines a shell element consisting of a stack of
infinitesimal cubes (only six shown in Figure 1) between bottom and top surfaces of the
shell. The SCL also defines two stress classification planes on the sides of the shell
element, denoted by SCP-1 and SCP-2 in Figure 1. The internal forces and moments are
the membrane, bending, and shear stress resultants applied to these two SCPs. The
primary stress definition requires that these stress resultants be in equilibrium with the
external forces and moments that are applied by an external agency. In simple words,
the stress resultants must satisfy the equations of equilibrium of shell theory.
The important point to note is that, according to the definition of primary stress, an SCL
defines only two SCPs. There is no SCP-3 normal to the SCL. Primary stresses do not
equilibrate each of the individual cubes of the stack along the SCL.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

For example, if pressure is applied to the bottom surface of the shell element of Figure
1, primary stress does not recognize any through-thickness normal stress distribution
within the stack because there is no external force that it equilibrates. The pressure is
equilibrated by the forces on the two SCPs by the shear stress resultants and, for a
curved shell, also by the membrane stress resultants. The fact that the bottom cube feels
the pressure is not part of the primary stress framework. According to the above
definition, the primary-stress mission is to keep only the whole stack in equilibrium, not
every individual cube.

2.2. Objective of Primary Stress Approach

The objective of primary stress limits is to prevent the loss of load-carrying capacity of
a vessel or piping, which will be referred to as collapse. The loading that produces
collapse is defined as the Limit Load that is defined according to the rules of the theory
of limit analysis. The calculation of the Limit Load for complicated geometries is
difficult. For practical design purposes, a lower bound to the Limit Load (henceforth
lower-bound load) is calculated. This is based on the lower-bound theorem of limit
analysis, which can be stated briefly for a 3-dimensional continuum (i.e., 3-dimensional
solid elements) as follows:

Theorem (1): Any stress field that is in equilibrium with the applied loading, satisfies
the stress boundary conditions and prescribed body forces (i.e., gravity), and lies on or
within the yield surface of the material renders the applied loading a lower bound to the
Limit Load.

In this form, the theorem applies to cases in which equilibrium of each individual,
infinitesimal cube (see Figure 1) is satisfied. It can also be stated for a 2-dimensional
surface with assigned thickness (i.e., beam, plate, and shell elements) in terms of the
resultant forces and moments.

Theorem (2): Any field of internal forces and moments that is in equilibrium with the
external forces and moments and lies on or within the yield surface written in terms of
forces and moments renders the applied loading a lower bound to the Limit Load of the
external forces and moments.

In this form, the theorem applies to cases in which equilibrium of only the stack of
cubes shown in Figure 1 is satisfied. Equilibrium of each individual, infinitesimal cube
is not required for a lower bound load. Since primary stresses are defined in terms of
internal forces and moments, Theorem (2) provides the basis for the primary stress
approach. Its objective is to produce a lower bound load according to Theorem (2). The
following points may be noted:

1. A lower bound according to Theorem (1) applies to a structure of any geometry,

including vessel walls of any thickness.
2. A lower bound according to Theorem (2) applies only to thin vessel walls and will
approach the lower bound according to Theorem (1) as the wall thickness tends to
zero.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

This distinction will be encountered in the discussion in Section 2.7.2.2 of the calculated
collapse load of a cylindrical shell by elastic-plastic, 3-dimensional continuum analysis
and primary stress analysis.

2.3. Primary Stress Subcategories

After the internal membrane, bending, and shear stress resultants are determined from
equilibrium, corresponding stresses across the wall, which develop these forces and
moments, are calculated. This leads to the following subcategories of primary stress:

1. Primary membrane stress

2. Primary bending stress
3. Primary shear stress

For design purposes, the primary membrane stress is further divided into general
primary membrane stress and local primary membrane stress subcategories. A general
primary membrane stress is one that is so distributed in the structure that no
redistribution of stress occurs as the result of yielding. It occurs typically remote from
discontinuities and is responsible for the kind of collapse that occurs in a sufficiently
long thin cylindrical shell. When the general primary membrane stress reaches the yield
surface, higher pressure cannot be equilibrated and collapse occurs.

A local primary membrane stress is a primary membrane stress that occurs over a local
region. It is of interest in pressure vessel design when it exceeds the primary membrane
stress outside the local region remote from discontinuities. A typical example is a
pressurized cylindrical shell with a reduced wall thickness over a part of its length, as
shown in Figure 2.

t t/2
T

P
Rm

Figure 2. Illustration of occurrence of local primary stress

The length of the local region of reduced thickness is important. If it is sufficiently long,
then the primary membrane stress inside this region will affect the collapse of the shell
in the same way as a general primary membrane stress outside the local region and must
be subjected to the same limit, except that the stress must be calculated used the reduced
thickness of T . If it is shorter than the length limit given in design codes, then the rise

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

in the primary membrane stress is a local detail that does not influence the collapse of
the vessel using the thickness of t . However, if not limited in the latter case, it may
produce unacceptable distortion of the shell caused by the reduced section. For this
reason, a limit on the local primary membrane stress is applied, but it is higher than that
placed on the general primary membrane stress.

2.4. Applicability

It has already been stated in Sections 2.1 to 2.3 above, that primary stress is a concept of
shell theory. Furthermore, the names assigned to the primary stress subcategories in
Section 2.3, membrane and bending, imply application to a special structure, namely
one that has a recognizable wall, like that of a shell. In pressure vessel design, that
would typically be a wall of a pressure-retaining boundary. All this means that the
primary stress approach is applicable to reasonably thin walls, as is shell theory, perhaps
no thicker than a shell with a radius/thickness ratio of 25.

For thin walls, membrane and bending stresses define adequately the deformation
process, while for thick walls the through-wall normal stress (radial stress in a
cylindrical shell) comes into play, which, as stated in Section 2.1, is not a primary
stress. A limitation of the primary stress approach to less than a radius/thickness ratio of
25 may not be that prohibitive. Even for low strength carbon pressure vessel steels it
would amount to a limit on pressure of about 7 MPa (1000 psi) for a cylindrical shell,
which is likely to be higher than the pressures in most applications. The limit would be
even higher for high strength steels. A back up to this argument is also given in Section
2.7.2.2 below.

It is important to note that this limit on wall thickness cannot be avoided by using 2-D
or 3-D solid elements in FEA instead of shell elements. The wall-thickness limit applies
to the wall geometry and not to the method or elements used in the analysis.

If the wall thickness exceeds that of the above limit, there is a practical way to deal with
the assessment of collapse, which applies to any wall thickness and arbitrary geometries
that have no recognizable wall. That way is to calculate a lower bound load by elastic-
plastic FEA. For this FEA, elastic-perfectly plastic (EPP) material law and linear
deformation theory must be used. The loading magnitude at which equilibrium can no
longer be maintained, typically signaled by an error message that convergence cannot
be achieved, gives the highest lower-bound load that can be obtained for the vessel by
the FE model and program used in the analysis. If Sm is used for the yield stress of the
EPP material law, then, according to Theorem (1) of Section 2.2, the calculated lower-
bound load is a conservative allowable load for prevention of collapse.

2.5. Primary Stress Limits

When applied to a pressure vessel, the limits placed on primary stresses are meant to
prevent gross plastic distortion and subsequent burst of a vessel under a single
application of pressure. To achieve that goal, a theory of strength must be identified.
Within the primary stress framework, that theory is the theory of limit analysis. This
theory identifies the threshold that separates gross plastic deformations that threaten the
serviceability of a structure from those that do not. That threshold is called collapse and

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

the loading at collapse is called the Limit Load of the vessel.

The purpose of limiting primary stresses is to achieve an assessment of collapse that is

conservative with respect to the allowable load that is determined by the theory of limit
analysis. The calculated Limit Load of a structure is proportional to the yield stress that
is specified in the limit analysis. This means that if the minimum specified yield
strength of each material under consideration is reduced by an agreed-upon design
margin, which is then used as the yield stress in the limit analysis, the calculated Limit
Load is the maximum allowable load of the structure with respect to the failure mode of
collapse. In Section III of the ASME Boiler & Pressure Vessel Code, this reduced yield
stress is denoted by Sm and is called design stress intensity. In the 2007 Section VIII-
Division 2, it is denoted by S and called allowable stress. It provides the basis of the
limits set for the primary stresses. Design codes typically provide the S m for each
material directly or give a formula from which it can be calculated, usually in terms of
the minimum specified yield strength and ultimate strength of the material.

The yield conditions commonly used in limit analysis for metals of sufficient ductility
are based either on the maximum shear stress theory, which uses the Tresca yield
condition, or on the distortion energy theory, which uses the Mises yield condition.

According to the Tresca yield condition, yielding begins in an element when twice the
maximum shear stress in that element reaches the yield strength of a tension-test
specimen of the same material. This yield strength is denoted by S y . If the calculated
principal stresses are S1, S 2 , S3 , and the Tresca stress intensity is defined by

Sint = max(| S1 − S2 |,| S2 − S3 |,| S3 − S1 |) (1)

then yielding will occur when Sint = S y .

According to the Mises yield condition, yielding occurs in an element when the Mises
equivalent stress, defined by Eq. (2), reaches S y .

1
Seq = [( S1 − S2 ) 2 + ( S2 − S3 ) 2 + ( S3 − S1 ) 2 ] (2)
2

For a case in which S3 = 0 , the Tresca and Mises yield surfaces can be represented in
the form of the hexagon and ellipse, respectively, as shown in Figure 3. The material is
yielding when the S1, S2 coordinates lie on the yield surface and not yielding when
they lie inside the yield surface.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

S2

Sy

S1
0 Sy

Figure 3. Tresca and Mises yield surfaces

After the primary stress components that belong to the primary stress subcategories
have been calculated, design limits can be applied to either the stress intensity (Eq. (1))
or the equivalent stress (Eq. (2)), depending on the yield condition adopted by the
design code. The ASME Boiler & Pressure Vessel Code includes the following stress
parameters to which the indicated design limits are applied:

1. General primary membrane stress intensity or equivalent stress is limited to S m

2. Local primary membrane stress intensity or equivalent stress is limited to 1.5Sm
3. Primary membrane (general or local) plus primary bending stress intensity or
equivalent stress is limited to 1.5S m

These limits are intended to prevent collapse, which is exhibited by gross plastic
deformation from a single application of pressure, and to provide a nominal factor of
safety on the ductile burst pressure.

2.6. Calculation of Primary Stresses

According to Section 2.1, the calculation of primary stresses includes the determination
of the internal forces and moments that are necessary to satisfy equilibrium, and a
lower-bound load is obtained by applying the primary stress limits of Section 2.5. If this
can be achieved without a solution involving specific material parameters, for example
elasticity, the lower-bound load will then be the optimum for the primary stress
approach.

The method for calculating the optimum lower bound has been developed but is too
complicated for general design purposes. However, such a calculation can be performed
in cases in which the SCP cuts off a free body, which is bounded by free and loaded
surfaces and the selected SCP. If the internal forces and moments are constant on the
SCP along its direction within the x1, x2 plane (see Figure 1), then they can be
calculated from simple laws of equilibrium of a statically determinate free body.
Illustrative examples of such cases are given in Section 2.7. If this can be done, then the
optimum lower-bound load will be obtained.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

Another option, more practical for general use, is to take the primary stresses from the
solution of an elastic FEA and a shell element model and apply the limits of Section 2.5.
This option is consistent with the primary stress definition and has been widely used.
Typical finite element programs write the membrane forces and the bending moments,
or the corresponding stresses, directly on standard output files, making the stresses over
the whole model readily available for the evaluation of their limits. The allowable
loading will, in general, be lower than the optimum lower-bound load. An example
showing that is given for the flat head in Section 2.7.3.

The thickness limitation of Section 2.4 must be observed when calculating primary
stresses by elastic FEA and applying the primary stress limits. For wall thicknesses
exceeding the stated limit on wall thickness, the calculation of a lower-bound to the
Limit Load by elastic-plastic FEA with solid elements, which is described in Section
2.4, provides a practical answer.

2.7. Primary Stresses from Free Bodies

Two examples are selected for which primary stresses can be calculated from statically
determinate free bodies described in Section 2.6. The purpose of the plate example is to
outline the procedure of the calculation. The point is that internal forces and moments
are calculated first from equilibrium and then stress distributions are chosen that
develop those forces and moments. The purpose of the second example is to show that
only the stresses that satisfy the primary stress definition should be included in primary
stress limits.

2.7.1. Plate

The example is a bar or plate of uniform rectangular cross section, shown in Figure 4.
The bar has a thickness t and width b . Its left end is fixed to a rigid wall and its right
end is subjected to some external surface action, which is represented in Figure 4 by the
shear force V , membrane force F , and bending moment M . An SCL is selected,
which defines only one relevant SCP because the stresses on planes parallel to the plane
of the paper are zero. Primary stresses are to be determined on the SCP.

SCP a b
V F M
t

Figure 4. External resultant forces and moment, and stress classification plane

As per definition of primary stress in Section 2.1, the internal forces and moment can be
obtained from equilibrium of the free body between the SCP and the right end shown in
Figure 5. They are given by Eqs. (3), (4), and (5).

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

M internal Finternal
V F M

Vinternal

Figure 5. Internal forces and moment applied to SCP

Finternal = F (3)
Vinternal = V (4)
M internal = M + aV (5)

According to the definition of Section 2.1, the internal forces and moments on the
selected SCP are dictated by equilibrium alone and are independent of material and
analysis method. While they are unique for a selected SCP, the distributions of the
primary stresses over the SCP are not. It is only necessary that the stress distributions
develop the internal F, V, and M. According to the primary stress subcategories defined
above, commonly assumed distributions are given by Eqs. (6), (7), and (8), which are
also shown in the sketches of Figure 6.

Finternal
Pmembrane = (6)
bt
Vinternal
Pshear = (7)
bt
6
Pbending = 2 M internal (8)
bt

Primary Primary Primary

Membrane Shear Bending
Stress Stress Stress

Figure 6. Commonly assumed primary stress distributions

The primary stress distributions shown in Figure 6 are those that are consistent with the
primary stress limits of Section 2.5. Other distributions could have been chosen. The
only requirement is that they develop the calculated internal forces and moments. If
other distributions were chosen, their limits would have to be adjusted accordingly.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

2.7.2. Axisymmetric Vessel

The example is a pressure vessel consisting of a cylindrical shell, capped by flat heads,
and subjected to inside pressure. Two locations are considered in the vessel. One
location is in the cylindrical shell remote from the joint and the other is in the flat head.
The use of free bodies to calculate the internal forces on the corresponding SCPs is
illustrated next.

2.7.2.1. Prevention of Collapse of Cylindrical Shell by Primary Stress Limits

For axisymmetric deformation of a cylindrical shell, the shell element is shown in

Figure 7.

SCP-1
t

SCL
SCP-2
dx1
P
dx2
x1
x3

x2
Figure 7. Cylindrical shell element showing the SCL and SCPs

The SCL is a radial line and SCP-1 and SCP-2 are planes normal to meridional and
circumferential (hoop) directions, respectively.

An SCL is selected remote from the shell/head junction, so that the shear force and
moments of Figure 7 are zero. The only non-zero forces are shown in Figure 8.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

NAXIAL NAXIAL
SCP-1

t R P SCP-2
Pressure

NHOOP
NHOOP

(a) (b)

Figure 8. Free bodies for determination of primary stresses in cylindrical shell

SCP-1 in sketch (a) is normal to the axis. The internal axial force on the SCP-1, per unit
length of the circumference, is determined from equilibrium with the resultant force of
the applied pressure in the axial direction and given by Eq. (9). R is the inside radius of
the shell, t its thickness, and Rm = R + t / 2 its mean radius.

PR 2
N AXIAL = (9)
2 Rm

SCP-2 in sketch (b) is a slice of the shell between two planes that are a unit length apart
along the axis and normal to the axis. That slice is cut to a semi-circular shape by
passing an SCP through the axis. The internal hoop force per unit axial length applied to
SCP-2 is determined from equilibrium with the resultant force of the applied pressure in
the vertical direction and given by equations

N HOOP = PR (10)

The stresses on the two SCPs are given by Eqs. (11) and (12), in which t is the
thickness of the shell wall.
N AXIAL PR
SAXIAL = = (11)
t ⎛ t ⎞
2t ⎜1 + ⎟
⎝ 2R ⎠
N HOOP PR
SHOOP = = (12)
t t
According to Section 2.1, the general primary membrane stress intensity is given by Eq.
(13). and its equivalent stress by Eq. (14).

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

PR
Pm-intensity = (13)
t
3 PR
Pm-equivalent = (14)
2 t

t
where the approximation 1 + ≈ 1 has been used.
2R
To evaluate protection against collapse, the limits of Section 2.5 are applied. The only
limit that matters is on the general primary membrane (Tresca) stress intensity or
(Mises) equivalent stress, depending on which yield surface has been adopted. When
limited to S m , the corresponding maximum allowable working pressures are given by
Eqs. (15) and (16).

t
PMAWP-Tresca = Sm (15)
R
2 t
PMAWP-Mises = Sm (16)
3R

These are the design pressures according to primary stress limits. It is seen that the
pressure is higher by about 15% if Mises yield condition is used.

User must exercise caution when applying Eqs. (15) and (16) to wall thicknesses of
pressure vessels and piping that exceed the limit of Section 2.4. Guidance for the
assessment of collapse of thick shells is given next.

2.7.2.2. Collapse of Cylindrical Shell of Unlimited Thickness

The limit pressure for a sufficiently long cylindrical shell is given by Eq. (17).

⎛ t ⎞
PMAWP-3D = Sm ln ⎜1 + ⎟ (17)
⎝ R⎠

It has been derived from the theory of limit analysis using a 3-dimensional continuum
formulation and not a shell theory formulation. In the derivation, Tresca yield condition
was used and the yield stress was set to S m . As explained in Section 2.5, the setting of
the yield stress to S m means that the calculated limit pressure equals the design
pressure. It is valid for any wall thickness, thin or thick.

Expanding the ln term in infinite power series, and assuming ( t / 2 R )  1 , Eq. (17)
2

can be rewritten as Eq. (18). This pressure errs on the conservative side with respect to
the exact value for all t / R ratios. The error reaches 2% at the t / R ratio of 0.65.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

t 1
PMAWP-2 = Sm (18)
R ⎛ t ⎞
⎜1 + ⎟
⎝ 2R ⎠

Assuming (t / 2 R ) << 1 in the series, Eq. (17) can be rewritten as Eq. (19). This
pressure errs on the unconservative side with respect to the exact value of Eq. (17). The
error is less than 2% up to the t / R ratio of 0.04.

t
PMAWP-1 = Sm (19)
R
All three are limit analysis formulas that have been included in design codes. For
practical purposes, only that of Eq. (17) is needed. Just as in the relationship between
Eqs. (15) and (16), the corresponding formulas for the Mises yield condition are
obtained by multiplying the formula of Eqs. (17)-(19) by 2 / 3 .

The reason for citing the approximate formulas is to point out that the formulas of Eqs.
(18) and (19) came from limit analysis, not from primary stress limits.

The fact that Eq. (19) equals Eq. (15), which did come from the general primary
membrane stress limit, supports the restriction of the primary stress approach to thin
walls, for which the case was made in Section 2.4 above.

2.7.3. Flat Head

tPlate

Mθ SCP SCP

P yP
yN
RPlate R Rm

NAXIAL

Figure 9. Free body for determination of primary stresses in flat head

For primary stress analysis, the flat head is modeled as a circular plate with thickness

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

tPlate and radius Rplate = R + t ( t is shell thickness), as shown in Figure 9. The

external forces and moments applied to the plate are those of pressure and N AXIAL ,
shown in Figure 8 and Figure 9. The shear force and moment that the shell applies to the
plate at their junction are self-equilibrated and not taken as sources of primary stress.
An SCP is selected through the center at any angle as shown in Figure 9, revealing a
semi-circular free body of diameter 2RPlate . To get primary stresses, it is necessary to
satisfy the laws of equilibrium between the external and internal forces and moments
applied to the free body.

The resultant force of the external pressure applied to the semi-circular free body has
the magnitude of

1
FPRESSURE = π R 2 P (20)
2
It is applied at the centroid of a semi-circular area within a radius of R , which is located
at a vertical distance of

4R
yP = (21)
3π

from the center of the circle (Figure 9). The resultant force of N AXIAL has the
magnitude of

FN = π Rm N AXIAL (22)

and it is applied at the centroid of a semi-circular arc with radius Rm , which is at a

vertical distance of
2Rm
yN = (23)
π
from the center of the circle. It is seen from Figure 8 and the derivation of Eq. (9) that
the forces given by Eqs. (20) and (22) are equal ( FPRESSURE = FN ) and in opposite
directions, so that no internal forces appear on the SCP.

However, an internal moment distributed along the diameter of the semi-circular free
body (SCP) is necessary to equilibrate the external moments. Since only equilibrium
matters for primary stresses, the distribution of this moment is assumed constant along
the diameter. Taking moments about the diametric line through the SCP, and defining
M θ as moment per unit length, equilibrium of the semi-circular free body is ensured by
Eq. (24).

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

1 ⎛ 2R 4R ⎞
2( R + t ) M θ = π R 2 P ⎜ m + ⎟ (24)
2 ⎝ π 3π ⎠
⎛ 3t ⎞
1+
1 2 ⎝ 2 R ⎟⎠
⎜
M θ = PR
6 t
1+
This results in Eq. R (25).
⎛ 3t ⎞
1+
1 2 ⎝ 2 R ⎟⎠
⎜
M θ = PR (25)
6 t
1+
R
Assuming the stress distribution of Eq. (8), both the maximum primary bending stress
intensity or equivalent stress are given by Eq. (26).

⎛ 3t ⎞
1+
⎛ R ⎞ ⎝ 2 R ⎟⎠
⎜ 2
Pb = P ⎜ ⎟ (26)
⎝ tPlate ⎠ 1 + t
R

The limit of 1.5S m on Pb leads to the maximum allowable working pressure given by
Eq. (27).

t
2 1+
⎛t ⎞ R
PMAWP = 1.5Sm ⎜ Plate ⎟ (27)
⎝ R ⎠ ⎛1 + 3t ⎞
⎜ ⎟
⎝ 2R ⎠

t
For a thin cylindrical shell, for which 1 + ≈ 1 , the maximum allowable working
R
pressure for the plate is then

2
⎛t ⎞
PMAWP = 1.5Sm ⎜ Plate ⎟ (28)
⎝ R ⎠

Just as in the examples of Section 2.7.1, these results are dictated by equilibrium alone.
They are independent of any analysis method and, except for the S m , independent of
material properties. They represent the optimum primary stresses.

As stated in Section 2.6, another option, less accurate but perhaps more practical in
some cases, is to take the primary stresses from an elastic solution. According to the

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

t
elastic solution of a thin, simply-supported plate, for which 1 + Plate ≈ 1 , the
RPlate
maximum primary bending stress intensity or equivalent stress Pb at the center of the
plate is given by Eq. (29).

2
3(3 + ν ) ⎛ RPlate ⎞
Pb = P⎜ ⎟ (29)
8 ⎝ tPlate ⎠

where ν is the Poisson’s ratio of the material. The limit of 1.5Sm on Pb leads to the
design pressure of Eq. (30).

2
4 ⎛ tPlate ⎞
PMAWP = ⎜ ⎟ Sm (30)
(3 + ν ) ⎝ RPlate ⎠

Assuming Poisson’s ratio ν = 0.3 , it is seen that the elastic solution gives a maximum
allowable working pressure that is almost 20% lower than that of Eq. (28), which was
obtained from the optimum primary stresses.

3. Secondary Stress

3.1. Definition

Secondary stresses are developed by constraints that are internal to a vessel or imposed
at its boundaries. If the constraints were removed, the vessel walls would tend to assume
a shape governed by the membrane theory of shells or, in the case of thermal action, by
unrestricted, stress-free, expansion. Secondary stresses are those that develop the forces
and moments that are required to change these shapes to satisfy the constraints. All
thermal stresses fall in the secondary stress category.

Secondary stresses typically appear at gross structural discontinuities, such as head-to-

shell joints, nozzles, and junctions between shells of different diameters, thicknesses, or
material properties, including coefficients of thermal expansion. Moments play a major
role in satisfying the constraints that arise at these discontinuities (see the examples in
Section 3.2 below). The resulting bending stresses can produce local yielding near the
surfaces of the vessel wall, while the core remains below yield. The following two
examples illustrate secondary stresses.

3.2. Examples

Consider the two vessels shown in Figure 10.

The solid lines in the sketches show two cylindrical vessels. One is capped with a
spherical head and the other is built in between two boundary supports. The head and
shell have equal thicknesses but the coefficient of thermal expansion of the shell is

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

greater than that of the head. If pressurized or heated slowly to a uniform temperature,
the shell would expand. The head in sketch (a) would expand less, and the boundary
supports would stay fixed. Constraints at the edges of the shells will keep it connected
to the head in (a) and fixed to the wall in (b). In these examples, only the secondary
stresses developed from the radial constraints will be considered while axial constraints
will be left aside.

gap gap

(a) (b)

Figure 10. Constraints by adjacent part and boundary

If these radial constraints were removed, the dashed lines would show the expansions
that would result. Under pressure, the shell and head would assume the shapes dictated
by membrane theory (zero bending stiffness) and, upon heating, the form would be that
which is normally assumed under a rise of temperature. In both cases the gaps shown in
Figure 10 would appear.

Secondary stresses are developed from the vertical shear force and bending moment that
are needed to satisfy the constraints; that is, to close the gap and to make the slopes of
the head, and shell in sketch (a) and the shell and boundary support in (b), match.

Secondary stresses are self-equilibrated and self-limiting by yield stress. If yield stress
is reached, the gap can keep closing without exceeding the yield stress and without
losing stability. Secondary stresses are controlled by the displacements and rotations
needed for closing the gap and matching the slopes. This displacement-controlled
behavior is produced by both pressure and thermal action.

3.3. Stress Limit Involving Secondary Stresses

A stress limit is applied to the elastically calculated intensity or equivalent stress of the
sum of primary membrane (either general or local), plus primary bending, plus
secondary stresses, commonly called the P+Q stress.. The limit placed on the P+Q stress

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

is set at 3S m or 2S y , where S y is the yield strength of the material. The purpose of

this limit is to achieve shakedown to elastic action after a few repetitions of loading,
thereby preventing accumulation of excessive plastic deformation in cyclic service,
which may lead to progressive distortion of the overall dimensions of a vessel.

3.4. Calculation of P+Q Stress

It is not necessary to calculate the secondary stresses separately. The P+Q stress can be
obtained directly from the solution of an elastic FEA and a shell-element model.
Standard outputs of typical finite element programs contain the membrane forces and
the bending moments, and the corresponding stresses, Tresca stress intensities and
Mises equivalent stresses, making the stresses over the whole model readily available
for the evaluation of the limit.

Glossary

SCP : Stress classification plane

SCL : Stress classification line
t : Thickness
P : Pressure
MAWP : Maximum allowable working pressure
Pm : General primary membrane stress
Pb : Primary bending stress
Sm : Design stress intensity
Sy : Yield strength
P+Q : Primary plus Secondary stress
R : Radius
S : Stress
V : Shear force
F or N : Force
M : Moment
Bibliography

ASME Boiler and Pressure Vessel Code (2007), American Society of Mechanical Engineers, New York.
[This is a design code that defines and uses stress categories derived from stress classification.]
ASME Criteria Document (1969). “Criteria of the ASME Boiler and Pressure Vessel Code for Design by
Analysis in Sections III and VIII, Division 2”, Library of Congress Catalog Card Number 56-3934,
ASME Press, American Society of Mechanical Engineers, New York. [Gives the background for the
stress category concept.]
Hollinger, G. L., and Hechmer, J. L., (2000). “Three-Dimensional Stress Criteria—Summary of the
PVRC Project”, ASME Journal of Pressure Vessel Technology, 122, 105–109. [Provides guidance for
stress classification.]
Kalnins, A., and Updike, D. P. (1993). “On Primary Stress Calculations”, ASME Design Analysis,
Robust Methods, and Stress Classification, PVP-265, 167-174. [Provides basic concepts for assessing
collapse by primary stress approach.]
Kalnins, A., and Updike, D. P. (1997). “Primary Stresses from Simple Laws of Equilibrium”, ASME

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

Pressure Vessel and Piping Codes and Standards, PVP-353, 249-259. [Illustrates calculation of primary
stresses from free-body diagrams.]
Kalnins, A., and Updike, D. P. (2001). “Limit Pressures of Spherical and Cylindrical Shells”, ASME
Journal of Pressure Vessel Technology, 123, pp. 288-292. [Provides limit pressures for comparison with
the lower bounds obtained by primary stress approach.]
Pastor, T. P. and Hechmer, J., (1997). “ASME Task Group Report on Primary Stress”, Journal of Pressure
Vessel Technology, 119, 61-67. [Gives basics for primary stress calculation.]

Biographical Sketch

Arturs Kalnins joined the faculty at Lehigh University (Bethlehem, Pennsylvania, USA) as an associate
professor in 1965 after serving as assistant professor at Yale University (New Haven, Conn.) for five
years. He served as professor of mechanics at Lehigh University from 1967 to 2004 and is currently
professor emeritus of mechanics. His main research interest has been in the behavior and failure of metal
plates, shells, and pressure vessels. He has developed the KSHEL computer programs for the analysis of
shells of revolution, and has served as a consultant to many companies in the United State and abroad. Dr.
Kalnins also lectured at the University of Mexico, Mexico City (1972); was a Fulbright-Hayes Fellow
(1977) at the University of Innsbruck (Tirol, Austria); and gave seminars at Pennsylvania State
University, University Park (1977-80), the Petrobras Company in Rio de Janeiro, Brazil (1977-78), and at
AREVA NP GmbH in Erlangen, Germany (2009). He is the author/co-author of 124 research papers and
one book. Within the activities of the American Society of Mechanical Engineers, Dr. Kalnins joined the
Working Group on Shells in 1976 and has served as member of the Subgroup on Design
Analysis/Subcommittee on Design and the Working Group on Vessels/Section III since 1987 and 2003,
respectively. In this capacity, he developed new design tools for torispherical pressure vessel heads, limit
load analysis, and fatigue analysis. Other contributions include three Pressure Vessels and Piping (PVP)
Conference tutorials (2000, 2001, 2003) on plastic analysis, shakedown, ratcheting, fatigue analysis, and
the finite element method. He has served as topic organizer and session developer at a number of PVP
conferences. Dr. Kalnins was honored with two certificates of appreciation (1987, 1999), a certificate of
recognition (1989) and an Outstanding Service Award (2005). He is a Fellow of ASME and the recipient
of the 2008 ASME PVP Medal.

©Encyclopedia of Life Support Systems (EOLSS)

PRESSURE VESSELS AND PIPING SYSTEMS - Stress Classification In Pressure Vessels And Piping- Arturs Kalnins

To cite this chapter

Arturs Kalnins, (2010), STRESS CLASSIFICATION IN PRESSURE VESSELS

AND PIPING, in Pressure Vessels and Piping Systems, [Eds. Yong W.Kwon, Poh-Sang
Lam], in Encyclopedia of Life Support Systems (EOLSS), Developed under the Auspices
of the UNESCO, Eolss Publishers, Paris, France, [http://www.eolss.net]

©Encyclopedia of Life Support Systems (EOLSS)