By

Dr Norin Zamiah binti Kassim Shaari

Reference:
Coulson J.M., Richardson J.F. &Sinnot R.K., Chemical Engineering, Vol. 6, Vol.
6, Pergammon, (5th ed).
¡ The design in term of dimensions etc. for each equipment
must be checked with the heuristics.
¡ There are ranges of the dimensions such as height to
diameter ratio for vessel that should be compared to
optimum range as outlines in the heuristics
¡ PRESSURE VESSELS – DISTILLATION
COLUMN, REACTOR, SEPARATOR
¡ HEAT EXCHANGER
¡ Binary and multicomponent mixtures
¡ Binary mixtures – McCabe-Thiele method or Lewis Sorel
method
¡ Stage by stage calculation to get to get the height and
diameter of column
¡ Finally, check with the heuristics on the height to diameter
ratio whether in the optimum range or not.
¡ Tray spacing must be in the acceptable range
¡ If the presence of the other components does not significantly affect the
volatility of the key components, the keys can be treated as a pseudo-
binary pair.
¡ The number of stages can then be calculated using a McCabe-Thiele
diagram, or the other methods developed for binary systems. This
simplification can often be made when the amount of the non-key
components is small, or where the components form near-ideal mixtures.
¡ Where the concentration of the non-keys is small, say less than 10 per
cent, they can be lumped in with the key components
¡ Hints: What are the heavy heavy component and light light key
components entering the distillation column
¡ For higher concentrations the method proposed by Hengstebeck
(1946) can be used to reduce the system to an equivalent binary
system.
¡ To reduce a multicomponent system to an equivalent binary it is
necessary to estimate the flow-rate of the key components
throughout the column. Hengstebeck makes use of the fact that
in a typical distillation the flow-rates of each of the light non-key
components approaches a constant, limiting, rate in the rectifying
section; and the flows of each of the heavy non-key components
approach limiting flow-rates in the stripping section.
¡ Smith and Brinkley developed a method for determining the
distribution of components in multicomponent separation
processes.
¡ Their method is based on the solution of the finite difference
equations that can be written for multistage separation
processes, and can be used for extraction and absorption
processes, as well as distillation.
¡ The two most frequently used empirical methods for estimating
the stage requirements for multicomponent distillations are the
correlations published by Gilliland (1940) and by Erbar and Maddox
(1961). These relate the number of ideal stages required for a given
separation, at a given reflux ratio, to the number at total reflux
(minimum possible) and the minimum reflux ratio (infinite number
of stages).
¡ The Erbar-Maddox correlation is generally considered to give more
reliable predictions.
1. Decide on the configuration of reactor – plug flow, packed bed,
fluidized bed, stirred tank
2. Collect together all the kinetic and thermodynamic data on the
desired reaction and the side reactions either from Hysys or
literature. Values will be needed for the rate of reaction over a
range of operating conditions: pressure, temperature, flow-rate
and catalyst concentration.
3. Collect the physical property data required for the design, which
is normally from the literature or book.
3. Identify the predominant rate-controlling mechanism:
kinetic, mass or heat transfer. Choose a suitable reactor type,
based on literatures on the similar reactions.
4. Make an initial selection of the reactor conditions to give the
desired conversion and yield.
5. Size the reactor (volume, height, diameter) and estimate its
performance. Exact analytical solutions of the design relationships
are rarely possible; semi empirical methods based on the analysis
of idealised reactors will normally have to be used.
6. Select suitable materials of construction (based on chemical
compatibility of the process streams with that material – this is
more important than ability of material to withstand the
operating pressure of the vessel).
7. Make a mechanical design for the reactor: the vessel design
(thickness of the reactor), heat-transfer surfaces, internals and
general arrangement.
¡ Orientation of separators: vertical for gas-liquid separation
and horizontal for liquid-liquid separation
¡ Find diameter and height of the vessel. Process stream (liquid)
normally occupy half the volume of the vessel). Flow rate and
average density for each stream need to be identified for the
calculation.
¡ The logical value of height/diameter ratio must be in the
required range as suggested by heuristic.
¡ For shell and tube heat exchanger
¡ Find the tube and shell side heat transfer coefficient
¡ Find the pressure drop
¡ For the mechanical design, shell side should designed
according to pressure vessel, and for the tube side – tube
length, inside and outside diameter should be identified.
TEMA –American Tubular Exchanger
Manufacturer Association

Class R à severe duties of petroleum

Class C à moderate duties in commercial
and general purposes
Class B à chemical process industries

3 letter code à 1st denotes tube side head

type (front end)
à 2nd shell type
à 3rd rear end
DESIGN OF SHELL AND TUBE HEAT
EXCHANGER (KERN method)
 • Heat transfer rate
• Fluid flow rates
Duty • Temperatures

• Density
Fluid • viscosity
propertie • Thermal conductivity
s

• No of shell passes
Type of
HE • No of tubes

Calculate • Mean Temperature Difference

∆Tm

Overall • Trial value of overall coefficient

coefficien (From Fig. 12.1)
t (U)
 • Q = UAF∆T (F from Fig.
Area
12.19)

• Decide exchanger layout - tube outside and inside diameter, tube

length (page 827-828), and baffle spacing, tube pitch, bundle
Layout diameter, hydraulic diameter (page 834, 855 and Table 12.4)

• Individual coefficient (hi and ho/hs) – find Re and Pr for

Coefficient tube and shell section respectively

Overall • Calculate overall coefficient and compare with trial value.

coefficient • If differs significantly, calculate pressure drop for tube and
(U) shell sections respectively.

• Pressure drop must be within the range as shown in section

Pressure
12.7.4.
drop • If pressure drop is too high for shell side, increase the baffle
spacing so that the shell side velocity is reduced – pressure
drop decrease
Factor Tube-side Shell-side
Corrosion More corrosive fluid Less corrosive fluids

Fouling Fluids with high Low fouling and scaling

fouling and scaling

Fluid temperature High temperature Low temperature

Operating Fluids with low Fluids with high pressure

pressure pressure drop drop

Viscosity Less viscous fluid More viscous fluid

Stream flow rate High flow rate Low flow rate

21
¡ High velocities will give high heat-transfer coefficients but
also a high-pressure drop.
¡ The velocity must be high enough to prevent any suspended
solids settling, but not so high as to cause erosion. High
velocities will reduce fouling.
¡ Liquids
Tube-side, process fluids: 1 to 2 m/s, maximum 4 m/s if required to
reduce fouling; water: 1.5 to 2.5 m/s.
Shell-side: 0.3 to 1 m/s.
¡ Vapours
¡ For vapours, the velocity used will depend on the operating
pressure and fluid density; the lower values in the ranges given
below will apply to high molecular weight materials.
Vacuum 50 to 70 m/s
Atmospheric pressure 10 to 30 m/s
High pressure 5 to 10 m/s
¡ The closer the temperature approach used (the difference between the
outlet temperature of one stream and the inlet temperature of the other
stream) the larger will be the heat-transfer area required for a given duty.
¡ As a general guide the greater temperature difference should be at least
20oC, and the least temperature difference 5 to 7oC for coolers using
cooling water, and 3 to 5oC using refrigerated brines.
¡ The maximum temperature rise in recirculated cooling water is limited to
around 30oC.
¡ When the heat exchange is between process fluids for heat recovery the
optimum approach temperatures will normally not be lower than 20oC.
 T1

DTlm =
(T1 - t 2 ) - (T2 - t1 )
æ T1 - t 2 ö
lnçç ÷÷
t2
T2

t1
è T2 - t1 ø
¡ Steam (available at various temperatures and pressures)
¡ Cooling water (15oC)
¡ Chilled water (5oC)
¡ Brines (calcium chloride/water fp. -18 deg cent. at 20% by
mass; sodium chloride/water fp. -16.5 deg cent. at 20 % by
mass)
¡ Methanol/water mixtures
¡ Ethylene glycol/water mixtures
¡ Propylene glycol/water mixtures (fp. -22 deg cent at a
concentration of 40% by mass)
¡ Silicone oils (‘syltherm’)
¡ The correction factor is a function of the shell and tube
fluid temperatures, and the number of tube and shell
passes. It is normally correlated as a function of two
dimensionless temperature ratios: R and S
¡ An economic exchanger design cannot normally be
achieved if the correction factor Ft falls below about
0.75.
¡ In these circumstances an alternative type of exchanger
should be considered which gives a closer approach to
true counter-current flow. The use of two or more shells
in series, or multiple shell-side passes, will give a closer
approach to true counter-current flow, and should be
considered where a temperature cross is likely to occur.
¡ Ft from Fig. 12.19 -12.22 depending on how many shell and
tube passes selected.
DTm = Ft DTlm
¡ Prior to that, R and S to be determined through equations:

T1 - T2 t 2 - t1
R= S=
t 2 - t1 T1 - t1
¡ In many applications the pressure drop available to drive the fluids through the
exchanger will be set by the process conditions, and the available pressure drop
will vary from a few milibars in vacuum service to several bars in pressure systems.
¡ Liquids
Viscosity <1 mN s/m2 35 kN/m2
1 to 10 mN s/m2 50 - 70 kN/m2
¡ Gas and vapours
High vacuum 0.4-0.8 kN/m2
Medium vacuum 0.1 x absolute pressure
1 to 2 bar 0.5 x system gauge pressure
Above 10 bar 0.1 x system gauge pressure
¡ The fluid physical properties required for heat-exchanger design are: density,
viscosity, thermal conductivity and temperature-enthalpy correlations (specific
and latent heats).
¡ In the correlations used to predict heat-transfer coefficients, the physical
properties are usually evaluated at the mean stream temperature. This is
satisfactory when the temperature change is small, but can cause a significant
error when the change in temperature is large.
TUBE SIDE HEAT TRANSFER
COEFFICIENT AND PRESSURE DROP
¡ TEMA design standard allow tubes diameter between 6.4 mm
and 50 mm
¡ Tube range 16 mm to 50 mm often used
¡ Smaller diameter are preferred for most duties at it gives
more compact and cheaper HE
¡ Larger tube for heavily fouling material
¡ Tube thickness must with stand internal and external pressure
and allow corrosion allowance
 Preferred length of tube :- 1.83m, 2.44 m, 3.66m,
4.88m, 6.10m, 7.32m
Longer tubes will reduce shell diameter à lead to
increase in pressure drop and pump work
Tube length to shell diameter ratio fall within 5-10

Arrangements Triangular & Rotated square give higher heat

transfer but higher pressure drop
Square & Rotated suitable for heavily fouling à
easily clean tube
Tube pitch to be recommended 1.25 tube outside
diameter
Square à used for ease of cleaning min clearance
6.4mm
Passes
Passes direct the flow, to increase
the path length
No of passes selected to give tube
side velocity
¡

Nt – no of tubes
K1 – from table 12.4
Db – bundle diameter
Do – outside diameter
ni – from table 12.4
pt – tube pitch
¡ The fluid in the tube is usually directed to flow back and forth in a
number of “passes” through groups of tubes arranged in parallel,
to increase the length of the flow path.
¡ The number of passes is selected to give the required tube-side
design velocity.
¡ Exchangers are built with from one to up to about sixteen tube
passes. The tubes are arranged into the number of passes required
by dividing up the exchanger headers (channels) with partition
plates (pass partitions). The arrangement of the pass partitions for
2, 4 and 6 tube passes are shown in Figure 12.11.
¡ BS3247 cover HE from
150 mm to 1067 mm
diameter
¡ Minimal shell diameter
must be selected to give
closed fit to tube
bundle; to reduce
bypassing round outside
the bundle
¡ Clearance between
tubes in bundle and
shell inside diameter
will depends on HE type
¡ Used to :-
àdirect the fluid stream across tubes,
à increase the fluid velocity and
à improved the heat transfer rate
¡ Baffle spacing 0.2 to 1.0 shell diameter
¡ Closed baffle spacing give higher heat
transfer coefficient but lead higher pressure
drop
¡ Optimum 0.3-0.5 times shell diameter
¡ It is often convenient to correlate heat-transfer data in terms of a heat transfer “j” factor, which is
similar to the friction factor used for pressure drop.
¡ The use of the jh factor enables data for laminar and turbulent flow to be represented on the same
graph; Figure 12.23.
¡ Equation 12.14 can be rearranged to a more convenient form:
¡ In some texts the equivalent (hydraulic mean) diameter is defined differently for
use in calculating the heat transfer coefficient in a conduit or channel, than for
calculating the pressure drop.
¡ The perimeter through which the heat is being transferred is used in place of the
total wetted perimeter. In practice, the use of de calculated either way will make
little difference to the value of the estimated overall coefficient; as the film
coefficient is only, roughly, proportional to de-0.2
¡ Besides eqn. 12.15, a more accurate estimate can be made by using equations
developed specifically for water. The physical properties are conveniently
incorporated into the correlation.
¡ Two major sources of pressure loss on the tube-side of a shell
and tube exchanger:
- the friction loss in the tubes
- losses due to the sudden contraction and expansion
¡ Heat transfer in heat exchangers involve the flow of heat from
one fluid through a retaining wall to another fluid.
¡ Heat flows from warmer to colder fluid through several
thermal resistances in series.
¡ The total resistance comprises of due to convective heat
transfer in the 2 fluids, the resistance due to fouling on the
wall surface, the resistance through the retaining wall.
 æ do ö
d o lnçç ÷÷
1 1 1 è di ø æ do 1 ö æ do 1 ö
= + + + çç x ÷÷ + çç x ÷÷
U O hO hOd 2k w è d i hid ø è d i hi ø

ho = hs , hod = outside dirt coefficient (shell side fouling factor )

hid = inside dirt coefficient (tube side fouling factor )

k w = thermal conductivity of the tube material
d i = tube inside diameter
d o = tube outside diameter
¡ Deposited material will lowering the thermal conductivity and
reduce overall coefficient
¡ Duplicate HE should be considered for severe fouling HE
¡ Velocity is important before area been added to avoid fouling
acceleration.
¡ Lower the tube side velocity, reduce the shear inside the tubes
and increase the rate of tube side fouling
¡ Alternative than increase area tube is à increase the length,
which increase the pressure drop
Shell and tubes can handle fouling but it can be
reduced by
¡ keeping velocities sufficiently high to avoid
deposits
¡ avoiding stagnant regions where dirt will collect
¡ avoiding hot spots where coking or scaling might
occur
¡ avoiding cold spots where liquids might freeze or
where corrosive products may condense for gases
¡ A vessel must be designed to withstand the maximum pressure to which it is
likely to be subjected in operation.
¡ For vessels under internal pressure, the design pressure is normally taken as the
pressure at which the relief device is set, which is normally be 5 to 10 per cent
above the normal working pressure, to avoid spurious operation during minor
process upsets.
¡ When deciding the design pressure, the hydrostatic pressure in the base of the
column should be added to the operating pressure, if it is significant.
¡ Vessels subject to external pressure should be designed to resist the maximum
differential pressure that is likely to occur in service. Vessels likely to be
subjected to vacuum should be designed for a full negative pressure of 1 bar,
unless it is fitted with an effective, and reliable, vacuum breaker.
¡ The strength of metals decreases with increasing temperature
so the maximum allowable design stress (S) will depend on
the material temperature.
¡ The design temperature at which the design stress is
evaluated should be taken as the maximum working
temperature of the material, with due allowance for any
uncertainty involved in predicting vessel wall temperatures.
¡ Pressure vessels are constructed from plain carbon steels, low
and high alloy steels, other alloys, polymers, and reinforced
plastics.
¡ Selection of a suitable material must take into account the
suitability of the material for fabrication (particularly welding)
as well as the compatibility of the material with the process
environment (process streams – reactant and products).
¡ For design purposes it is necessary to decide a value for the maximum allowable
stress (nominal design strength) that can be accepted in the material of
construction.
¡ This is determined by applying a suitable “design stress factor” (factor of safety)
to the maximum stress that the material could be expected to withstand without
failure under standard test conditions. The design stress factor allows for any
uncertainty in the design methods, the loading, the quality of the materials, and
the workmanship.
¡ For materials not subject to high temperatures the design stress is based on the
yield stress (or proof stress), or the tensile strength (ultimate tensile stress) of the
material at the design temperature.
¡ For materials subject to conditions at which the creep is likely to be a
consideration, the design stress is based on the creep characteristics of the
material: the average stress to produce rupture after 105 hours, or the average
stress to produce a 1 per cent strain after 105 hours, at the design temperature.
¡ The strength of a welded joint will depend on the type of joint and the quality of the welding.
¡ The soundness of welds is checked by visual inspection and by non-destructive testing
(radiography).
¡ The possible lower strength of a welded joint compared with the virgin plate is usually allowed for
in design by multiplying the allowable design stress for the material by a “welded joint factor” E.
¡ The value of E used in design will depend on the type of joint and amount of radiography required
by the design code. Typical values are shown in Table 13.3. Taking the factor as 1.0 implies that the
joint is equally as strong as the virgin plate; this is achieved by radiographing the complete weld
length, and cutting out and remaking any defects.
¡ The use of lower joint factors in design, though saving costs on radiography, will result in a thicker,
heavier, vessel, and the designer must balance any cost savings on inspection and fabrication
against the increased cost of materials.
¡ The “corrosion allowance” is the additional thickness of metal added to
allow for material lost by corrosion and erosion, or scaling
¡ The allowance to be used should be agreed between the customer and
manufacturer. The allowance should be based on experience with the
material of construction under similar service conditions to those for the
proposed design.
¡ For carbon and low-alloy steels, where severe corrosion is not expected, a
minimum allowance of 2.0 mm should be used; where more severe
conditions are anticipated this should be increased to 4.0 mm.
¡ Most design codes and standards specify a minimum allowance of 1.0
mm.
¡ A structure must be designed to resist gross plastic
deformation and collapse under all the conditions of loading.
¡ They can be classified as major loads, that must always be
considered in vessel design, and subsidiary loads.
¡ Formal stress analysis to determine the effect of the
subsidiary loads is only required in the codes and standards
where it is not possible to demonstrate the adequacy of the
proposed design by other means; such as by comparison with
the known behaviour of existing vessels.
1. Design pressure: including any significant static head of
liquid.
2. Maximum weight of the vessel and contents, under operating
conditions.
3. Maximum weight of the vessel and contents under the
hydraulic test conditions.
4. Wind loads.
5. Earthquake (seismic) loads.
6. Loads supported by, or reacting on, the vessel.
¡ Wind loading will only be important on tall columns installed
in the open. Columns and chimney-stacks are usually free
standing, mounted on skirt supports, and not attached to
structural steel work. Under these conditions the vessel
under wind loading acts as a cantilever beam, Figure 13.19.
¡ For a uniformly loaded cantilever the bending moment at
any plane is given by: 2
Wx
Mx =
2

¡ 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.
Hint: Make sure to design the tall vessel with thickness
¡ For tall columns the bending stress due to wind loading will
increasing from top to bottom of the vessel
often be greater than direct stress due to pressure, and will
determine the plate thickness required.
1. Local stresses caused by supports, internal structures and connecting
pipes.
2. Shock loads caused by water hammer, or by surging of the vessel
contents.
3. Bending moments caused by eccentricity of the centre of the working
pressure relative to the neutral axis of the vessel.
4. Stresses due to temperature differences and differences in the
coefficient expansion of materials.
5. Loads caused by fluctuations in temperature and pressure.

However, a vessel will not be subject to all these loads simultaneously.

The designer must determine what combination of possible loads gives
the worst situation, and design for that loading condition.
1. The longitudinal and circumferential stresses due to pressure (internal or external),
given by:
PDi
sh = longitudinal
2t
PDi
sL = circumferential
4t

2. The direct stress s W due to the weight of the vessel, its contents, and any
attachments.
The stress will be tensile (positive) for points below the plane of the vessel
supports, and compressive (negative) for points above the supports. The dead-
weight stress will normally only be significant, compared to the magnitude of the
other stresses, in tall vessels.
 WZ
sW =
p (Di + t )t

WZ - Total weight supported by the

vessel wall at the plane considered
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
¡ The bending stresses will be compressive or tensile M æ Di ö
sb = ± ç + t ÷
p IV è 2 ø
IV =
64
(D 4
0 - Di4 )
where M is the total bending moment
IV is the second moment of area of the vessel
D0 = Di + 2t vessel
Principal Stresses

σz

σh σh

σz
Allowable stress intensity
s h -s Z

¡ Must be < S of the material for vessel

¡ If not, failure will occur and the thickness of vessel wall need
to be revised
Compressive stresses and elastic stability
- The material may fail by elastic instability (buckling)
- For thin-walled vessel, failure can occur under an axial compressive
load by buckling of the complete vessel or by local buckling or
wrinkling of the shell plates.
- Therefore, in the column design, the maximum value of the resultant
axial stress (s + s ) should not exceed the critical buckling stress (s C )
W b

æ t ö N
sC = 2 x 10 ç
çD ÷
4
÷ mm 2
è 0ø
¡ There will be a minimum wall thickness required to ensure
that any vessel is sufficiently rigid to withstand its own
weight, and any incidental loads.
¡ As a general guide the wall thickness of any vessel should not
be less than the values given in the next table (the values
include a corrosion allowance of 2 mm)
¡ Flat plates are used as covers for manways, and as the channel
covers of heat exchangers.
¡ Formed flat ends, known as “flange-only” ends, are
manufactured by turning over a flange with a small radius on a
flat plat.
¡ “Flange-only” heads are the cheapest type of formed head to
manufacture, but their use is limited to low-pressure and
small-diameter vessels.
DOMED HEADS
1. Flat Plates and Formed Flat
Heads
HEAD AND CLOSURES
¡ Standard torispherical heads (dished ends) are the most commonly used end
closure for vessels up to operating pressures of 15 bar.
¡ They can be used for higher pressures, but above 10 bar their cost should be
compared with that of an equivalent ellipsoidal head.
¡ Above 15 bar an ellipsoidal head will usually prove to be the most economical
closure to use.
¡ A hemispherical head is the strongest shape; capable of resisting about twice the
pressure of a torispherical head of the same thickness.
¡ The cost of forming a hemispherical head will, however, be higher than that for a
shallow torispherical head. Hemispherical heads are used for high pressures.
Hemispherical > Ellipsiodal > Torispherical

Hint: To choose the head, the thickness of head must be closer to vessel thickness
¡ Saddle supports – for horizontal vessels
¡ Skirt supports – for tall and vertical vessel
¡ Brackets or lugs – for all types of vessel
¡ To carry the weight of the vessel and its content as well as any
superimposed loads such as wind loads.
¡ Supports will impose localized loads on the vessel wall, and
the design must be checked to ensure the resulting stress
concentrations are below the maximum allowable design
stress (S).
¡ The design of supports should also consider the allowance for
easy access to the vessel and fittings for inspection and
maintenance.
¡ A skirt support consists of a cylindrical or conical shell welded to the base of the
vessel.
¡ A flange at the bottom of the skirt transmits the load to the foundations.
¡ Typical designs are shown in Figure 13.23.
¡ Openings must be provided in the skirt for access and for any
¡ connecting pipes; the openings are normally reinforced. The skirt may be welded
to the bottom head of the vessel. Figure 13.27a; or welded flush with the shell,
Figure 13.27b; or welded to the outside of the vessel shell, Figure 13.27c. The
arrangement shown in Figure 13.27b is usually preferred.
¡ Skirt supports are recommended for vertical vessels as they do not impose
concentrated loads on the vessel shell; they are particularly suitable for use with
tall columns subject to wind loading.
¡ Design of flanges – choose the standard ones
¡ AUTOCAD drawing must be provided – dimensions must
according to the calculation.