Chemical Engineering Apparatus Design (Cheg 4191)
Chemical Engineering Apparatus Design (Cheg 4191)
Flat-end closures
(a)Flanged plate
(b)Welded plate
(c)Welded plate
(d)Bolted cover
(e)Bolted cover
Cont.….
Domed heads
2. Hemispherical heads 3. Ellipsoidal heads
where
f = the maximum allowable stress (the design stress),
D = the effective plate diameter,
C = a constant, which depends on the edge support. Taking Poisson’s ratio as 0.3, a typical value for steels,
then if the edge can be taken as completely rigid C = 0.43, and if it is essentially free to rotate C = 0.56.
• The design equations used to determine the thickness of flat ends are based on the analysis of
stresses in flat plates;
• The thickness required will depend on the degree of constraint at the plate periphery.
The minimum thickness required is given by:
where
Cp = a design constant, dependent on the edge constraint,
De = nominal plate diameter,
f = design stress.
Cont.…
• Values for the design constant Cp and the nominal plate diameter De are given in the
design codes and standards for various arrangements of flat end closures.
• The values of the design constant and nominal diameter for the typical designs are
given below:
(a) Flanged-only end, for diameters less than 0.6 m and corner radii at least equal to
0.25e, Cp can be taken as 0.45; De is equal to Di.
(b, c) Plates welded to the end of the shell with a fillet weld, angle of fillet 45° and depth
equal to the plate thickness, take Cp as 0.55 and De = Di.
(d) Bolted cover with a full face gasket, take Cp= 0.4 and De equal to the bolt circle
diameter.
(e) Bolted end cover with a narrow-face gasket, take Cp = 0.55 and De equal to the
mean diameter of the gasket.
Design of domed ends
• Design equations and charts for the various types of domed heads are given in the codes and
standards and should be used for detailed design. The codes and standards cover both
unpierced and pierced heads.
• Pierced heads are those with openings or connections.
• The head thickness must be increased to compensate for the weakening effect of the holes
where the opening or branch is not locally reinforced.
• For convenience, simplified design equations which are suitable for the preliminary sizing of
unpierced heads and for heads with fully compensated openings or branches are discussed
here.
1. Hemispherical heads
• For no difference in dilation between the two parts (equal diametrical strain) it can be shown
that for steels (Poisson’s ratio = 0.3) the ratio of the hemispherical head thickness to cylinder
thickness should be 7/17.
• However, the stress in the head would then be greater than that in the cylindrical section; and
the optimum thickness ratio is normally taken as 0.6
Cont.….
2. Ellipsoidal heads
Most standard ellipsoidal heads are manufactured with a major and minor axis ratio of 2 : 1.
For this ratio, the following equation can be used to calculate the minimum thickness required:
3. Torispherical heads
There are two junctions in a torispherical end closure: that between the cylindrical section and
the head, and that at the junction of the crown and the knuckle radii.
The bending and shear stresses caused by the differential dilation that will occur at these points
must be taken into account in the design of the heads.
One approach taken is to use the basic equation for a hemisphere and to introduce a stress
concentration, or shape, factor to allow for the increased stress due to the discontinuity.
The stress concentration factor is a function of the knuckle and crown radii.
Cont.….
where
Cs = stress concentration factor for torispherical heads =
Rc = crown radius
Rk = knuckle radius
• The ratio of the knuckle to crown radii should not be less than 0.06, to avoid buckling; and
the crown radius should not be greater than the diameter of the cylindrical section.
• For formed heads (no joints in the head) the joint factor J is taken as 1.0.
Cont.…
Conical sections and end closures
• Conical sections (reducers) are used to make a gradual reduction in diameter from one cylindrical
section to another of smaller diameter.
• Conical ends are used to facilitate the smooth flow and removal of solids from process equipment;
such as, hoppers, spray-dryers and crystallizers.
• The thickness required at any point on a cone is related to the diameter by the following expression:
where
……..i Dc is the diameter of the cone at the point,
half the cone apex angle.
• This equation will only apply at points away from the cone to cylinder junction. Bending and shear
stresses will be caused by the different dilation of the conical and cylindrical sections. This can be
allowed for by introducing a stress concentration factor, in a similar manner to the method used for
torispherical heads,
……..ii
Cont.…
• A formed section would normally be used for the transition between a cylindrical section and conical
section; except for vessels operating at low pressures, or under hydrostatic pressure only.
• The transition section would be made thicker than the conical or cylindrical section and formed with a
knuckle radius to reduce the stress concentration at the transition.
• The simplest method of providing compensation is to weld a pad or collar around the
opening
• The outer diameter of the pad is usually between 1 to 2 times the diameter of the
hole or branch. This method, however, does not give the best disposition of the
welded pad
reinforcing material about the opening, and in some circumstances high thermal
stress can arise due to the poor thermal conductivity of the pad to shell junction.
• Forged reinforcing rings provide the most effective method of compensation, but are
expensive.
• They would be used for any large openings and branches in vessels operating under
severe conditions.
Forged ring
Calculation of reinforcement required
• The “equal area method” is the simplest method used for calculating the amount of
reinforcement required, and is allowed in most design codes and standards.
• The principle used is to provide reinforcement local to the opening, equal in cross-sectional
area to the area removed in forming the opening.
• For short closed vessels, and long vessels with stiffening rings, the critical buckling pressure will be higher than
that predicted by this equation.
• The effect of stiffening can be taken into account by introducing a “collapse coefficient”, K c, into equation above.
where Kc is a function of the diameter and thickness of the vessel, and the
effective length L’ between the ends or stiffening rings; and is obtained from
Figure below.
This equation can be used to determine the critical buckling pressure and hence the thickness required to resist a
given external pressure. A factor of safety of at least 3 should be applied to the values predicted using equation.
Cont.…
Plain vessel
I-section
stiffening
rings
With stiffeners (use smaller of L’
and Ls)
Jacketed vessel
Cont.….
• The critical distance between stiffeners, Lc, beyond which stiffening will not be effective is given by
For vessels under external pressure this should not normally exceed 1.5 per
cent.
Design of stiffness rings
• The usual procedure is to design stiffening rings to carry the pressure load for a distance of L s on each side of
the ring, where Ls is the spacing between the rings. So, the load per unit length on a ring F r will be given by:
where Pe is the external pressure.
• The critical load to cause buckling in a ring under a uniform radial load F c is given by the following expression
where Ir = second moment of area of the ring cross-section,
Dr = diameter of the ring (approximately equal to the shell outside diameter)
Combining these two equations will give an equation from which the required dimensions of the ring can be
determined:
• In calculating the second moment of area of the ring some allowance is normally made for the vessel wall;
the use of Ir calculated for the ring alone will give an added factor of safety.
• In vacuum distillation columns, the plate-support rings will act as stiffening rings and strengthen the vessel;
Vessel heads
• The critical buckling pressure for a sphere subject to external pressure is given by:
where Rs is the outside radius of the sphere.
• The pressure to cause a “dimple” to form is about one-quarter of that given by equation above, and is given by:
A generous factor of safety is needed when applying this equation to the design of heads under
external pressure.
• A value of 6 is typically used, which gives the following equation for the minimum thickness:
• Torispherical and ellipsoidal heads can be designed as equivalent hemispheres. For a torispherical head the radius R s is
taken as equivalent to the crown radius Rc.
• For an ellipsoidal head the radius can be taken as the maximum radius of curvature; that at the top, given by:
3. Bending stresses resulting from the bending moments to which the vessel is
subjected. Bending moments will be caused by the following loading conditions:
a) The wind loads on tall self-supported vessels
b) Seismic (earthquake) loads on tall vessels
c) The dead weight and wind loads on piping and equipment which is attached to
the vessel, but offset from the vessel centre line
d) For horizontal vessels with saddle supports, from the disposition of dead-
weight load
Cont.…
• The bending stresses will be compressive or tensile, depending on location, and are given by:
where
Mv is the total bending moment at the plane being considered and
Iv is the second moment of area of the vessel about the plane of bending.
4. Torsional shear stresses resulting from torque caused by loads offset from the vessel axis. These loads will
normally be small, and need not be considered in preliminary vessel designs.
The torsional shear stress is given by:
where T = the applied torque,
Ip = polar second moment of area =
Principal stresses
The principal stresses will be given
by:
w should be counted as positive if tension and negative if compressive. is not usually significant.
Cont.…
• The third principal stress, that in the radial direction 3, will usually be negligible for thin-
walled vessels. As an approximation it can be taken as equal to one-half the pressure loading
3 will be compressive (negative).
Allowable stress intensity
• The maximum intensity of stress allowed will depend on the particular theory of failure
adopted in the design method. The maximum shear-stress theory is normally used for
pressure vessel design.
Using this criterion the maximum stress intensity at any point is taken for design purposes as
the numerically greatest value of the following:
• The vessel wall thickness must be sufficient to ensure the maximum stress intensity does not
exceed the design stress (nominal design strength) for the material of construction, at any
point.
Compressive stresses and elastic stability
• Under conditions where the resultant axial stress z due to the combined loading is compressive, the vessel may
fail by elastic instability (buckling).
• Failure can occur in a thin-walled process column under an axial compressive load by buckling of the complete
vessel, by local buckling, or wrinkling, of the shell plates.
• For a curved plate subjected to an axial compressive load the critical buckling stress is given by
• By applying a suitable factor of safety, this equation can be used to predict the maximum allowable
compressive stress to avoid failure by buckling. A large factor of safety is required, as experimental work has
shown that cylindrical vessels will buckle at values well below that given by this equation.
• For steels at ambient temperature E = 200,000 N/mm 2, and this equation with a factor of safety of 12 gives:
The maximum compressive stress in a vessel wall should not exceed that given by this
equation; or the maximum allowable design stress for the material, whichever is the
least.
Cont.…
Stiffening
• As with vessels under external pressure, the resistance to failure buckling can be increased
significantly by the use of stiffening rings, or longitudinal strips. Methods for estimating the
critical buckling stress for stiffened vessels are given in the standards and codes.
Loading
• The loads to which a vessel may be subjected will not all occur at the same time. For
example, it is the usual practice to assume that the maximum wind load will not occur
simultaneously with a major earthquake.
• The vessel must be designed to withstand the worst combination of the loads likely to
occur in the following situations:
1. During erection (or dismantling) of the vessel.
2. With the vessel erected but not operating.
3. During testing (the hydraulic pressure test).
4. During normal operation.
Weight loads
• The major sources of dead weight loads are:
1. The vessel shell. 5. Auxiliary equipment which is not self-supported;
2. The vessel fittings: manways, nozzles. condensers, agitators.
3. Internal fittings: plates (plus the fluid on the 6. Insulation.
plates); heating and cooling coils. 7. The weight of liquid to fill the vessel. The vessel will be
4. External fittings: ladders, platforms, piping. filled with water for the hydraulic pressure test; and may
fill with process liquid due to misoperation.
• For vessels on a skirt support, the weight of the liquid to fill the vessel will be transferred directly to the skirt.
• For preliminary calculations the approximate weight of a cylindrical vessel with domed ends, and uniform wall thickness, can
be estimated from the following equation:
where Wv = total weight of the shell, excluding internal fittings, such as plates, N,
Cv = a factor to account for the weight of nozzles, manways, internal supports, etc; which can be taken as
= 1.08 for vessels with only a few internal fittings,
= 1.15 for distillation columns, or similar vessels, with several manways, and with plate support rings, or equivalent
fittings,
Hv = height, or length, between tangent lines (the length of the cylindrical section), m,
g = gravitational acceleration, 9.81 m/s 2,
t = wall thickness, mm
= density of vessel material, kg/m3,
D = mean diameter of vessel = (D + t 10-3), m.
Cont.…
For a steel vessel, this equation reduces to:
The following values can be used as a rough guide to the weight of
fittings;(a) caged ladders, steel, 360 N/m length,
(b) plain ladders, steel, 150 N/m length,
(c) platforms, steel, for vertical columns, 1.7 kN/m 2 area,
(d) contacting plates, steel, including typical liquid loading, 1.2 kN/m 2 plate area.
Typical values for the density of insulating materials are (all kg/m 3):
Foam glass 150
Mineral wool 130
Fibreglass 100
Calcium silicate 200
These densities should be doubled to allow for attachment fittings, sealing, and moisture absorption.
Wind loads (tall vessels)
• So the bending moment, and hence the bending stress, will vary parabolically
from zero at the top of the column to a maximum value at the base.
• For tall columns the bending stress due to wind loading will often be greater
than direct stress due to pressure, and will determine the plate thickness
required. Wind load on tall column
• The most economical design will be one in which the plate thickness is progressively increased from the
top to the base of the column.
• The thickness at the top being sufficient for the pressure load, and that at the base sufficient for the
pressure plus the maximum bending moment.
Cont.…
• Any local increase in the column area presented to the wind will give rise
to a local, concentrated, load. The bending moment at the column base
caused by a concentrated load is given by:
• The term (e/g) is called the seismic constant Ce, and is a function of the natural period of vibration of the vessel and
the severity of the earthquake.
• Values of the seismic constant have been determined empirically from studies of the damage caused by earthquakes,
and are available for those geographical locations which are subject to earthquake activity.
Eccentric loads (tall vessels)
• Ancillary equipment attached to a tall vessel will subject the vessel to a bending moment if the
centre of gravity of the equipment does not coincide with the centre line of the vessel.
offset equipment
can be neglected.
• That produced by heavy equipment, such as reflux condensers and side platforms, can be
significant and should be considered. The moment is given by: