Coefficients from

Shell-nozzle-pressure drop
The pressure loss in the shell nozzles will normally only be significant with
gases.
The nozzle pressure drop can be taken as equivalent to 112 velocity heads for
the inlet and 12 for the outlet, based on the nozzle area or the free area
between the tubes in the row immediately adjacent to the nozzle, whichever
is the least.

Optimization using a CAD program

The use of a proprietary computer program (HTFS, M-TASC) to find the lowest
cost design that meets the specification resulted in the design set out below.
The program selected longer tubes, to minimize the cost.

CAD design
Split ring, floating head, 1 shell pass, 2 tube passes.
168 carbon steel tubes, 6096 mm, 19.05 mm o.d., 14.83 mm i.d., triangular
pitch,
pitch 23.18 mm.
Heat transfer area 61 m2.
Shell i.d. 387, baffle spacing 77.9 mm, 15% cut.
Tube-side coefficient 851 W/m2 0C-1, clean.
Shell-side coefficient 1191 W/m2 0C-1, clean.
Overall coefficient estimated 484Wm2 0C-1 clean.
Overall coefficient estimated 368Wm2 0C-1 dirty.

Pressure drops, including drop over nozzles:

Tube-side, estimated 0.5 bar.
Shell-side, estimated 0.5 bar.
12.9.4. Bells method
The heat-transfer coefficient and pressure drop are estimated from
correlations for flow over ideal tube-banks, and the effects of leakage,
bypassing and flow in the window zone are allowed for by applying correction
factors.
It will give more satisfactory predictions of the heat-transfer coefficient and
pressure drop than Kerns method
It can be used to investigate the effects of constructional tolerances and the
use of sealing strips.

Heat-transfer coefficient
The shell-side heat transfer coefficient is given by:

hoc, ideal cross-flow coefficient

The heat-transfer coefficient for an ideal cross-flow tube bank can be
calculated using the heat transfer factors given in Figure 12.31.

The Reynolds number for cross-flow through a tube bank is given by:

Fn, tube row correction factor

The mean heat-transfer coefficient will depend on the number of tubes
crossed. Figure 12.31 is based on data for ten rows of tubes.
For turbulent flow the correction factor Fn is close to 1.0.
 In laminar flow the heat-transfer coefficient may decrease with increasing
rows of tubes crossed, due to the buildup of the temperature boundary layer.

1. Re >2000, turbulent; take Fn from Figure 12.32.

2. Re >100 to 2000, transition region, take F n=1.0;
3. Re <100, laminar region, Fn (Nc)-0.18

Where Nc is the number of rows crossed in series from end to end of the
shell, and depends on the number of baffles. The correction factor in the laminar
region is not well established, and Bells paper, or the summary given by Mueller
(1973), should be consulted if the design falls in this region.

Fw, window correction factor

This factor corrects for the effect of flow through the baffle window, and is a
function of the heat-transfer area in the window zones and the total heat-
transfer area.
The correction factor is shown in Figure 12.33 plotted versus R w, the ratio of
the number of tubes in the window zones to the total number in the bundle,
determined from the tube layout diagram.

Fb, bypass correction factor

This factor corrects for the main bypass stream, the flow between the tube
bundle and the shell wall, and is a function of the shell to bundle clearance, and
whether sealing strips are used:

FL, Leakage correction factor

This factor corrects for the leakage through the tube-to-baffle clearance and
the baffle-toshell clearance.
Pressure drop
The pressure drops in the cross-flow and window zones are determined
separately, and summed to give the total shell-side pressure drop.

Cross-flow zones
The pressure drop in the cross-flow zones between the baffle tips is
calculated from correlations for ideal tube banks, and corrected for leakage and
bypassing

Pi ideal tube bank pressure drop

`The number of tube rows has little effect on the friction factor and is
ignored. Any suitable correlation for the cross-flow friction factor can be used; for
that given in Figure 12.36, the pressure drop across the ideal tube bank is given by:

Fb, bypass correction factor for pressure drop

Bypassing will affect the pressure drop only in the cross-flow zones. The
correction factor is calculated from the equation used to calculate the bypass
correction factor for heat transfer, equation 12.30, but with the following values for
the constant.

Laminar region, Re<100,=5.0

Transition and turbulent region, Re>100, =4.0
The correction factor for exchangers without sealing strips is shown in Figure 12.37.

FL, leakage factor for pressure drop

Leakages will affect the pressure drop in both the cross-flow and window
zones. The factor is calculated using the equation for the heat-transfer leakage-
correction factor, equation 12.31, with the values for the coefficient L taken from
Figure 12.38.

Window-zone pressure drop

Any suitable method can be used to determine the pressure drop in the
window area; see Butterworth (1977). Bell used a method proposed by Colburn.
Corrected for leakage, the window drop for turbulent flow is given by

End zone pressure drop

There will be no leakage paths in an end zone (the zone between tube sheet
and baffle). Also, there will only be one baffle window in these zones; so the total
number of restrictions in the cross-flow zone will be N cv+ Nwv. The end zone pressure
drop Pe will therefore be given by:

Total shell-side pressure drop

Summing the pressure drops over all the zones in series from inlet to outlet gives:

End zone lengths

The spacing in the end zones will often be increased to provide more flow
area at the inlet and outlet nozzles. The velocity in these zones will then be lower
and the heat transfer and pressure drop will be reduced slightly.

12.9.5. Shell and bundle geometry

 The bypass and leakage areas, window area, and the number of tubes and
tube rows in the window and cross-flow zones can be determined precisely from the
tube layout diagram
Ra is obtained from Figure 12.41, for the appropriate baffle cut Bc

Where ct is the diametrical tube-to-baffle clearance; the difference between the hole
and tube diameter, typically 0.8 mm.

Where cs is the baffle-to-shell clearance, see Table 12.5.b can be obtained from
Figure 12.41, for the appropriate baffle cut, Bc

Where lB is the baffle spacing.

12.9.6. Effect of fouling on pressure drop

Bells method gives an estimate of the shell-side pressure drop for the
exchanger in the clean condition. In service, the clearances will tend to plug
up, particularly the small clearance between the tubes and baffle, and this
will increase the pressure drop.
Devore (1961) has estimated the effect of fouling on pressure drop by
calculating the pressure drop in an exchange in the clean condition and with
the clearance reduced by fouling, using Tinkers method.

12.9.7. Pressure-drop limitations

Though Bells method will give a better estimate of the shell-side pressure
drop than Kerns, it is not sufficiently accurate for the design of exchangers
where the allowable pressure drop is the overriding consideration.
 For such designs, a divided-flow model based on Tinkers work should be
used. If a proprietary computer program is not available, the ESDU Design
Guide, ESDU 83038 (1984) is recommended. Devores method can also be
considered, providing the exchanger layout conforms with those covered in
his work.

12.10. CONDENSERS
-the construction of a condenser will be similar to other shell and tube
exchangers, but with a wider baffle spacing, typically BDDs.

Four condenser configurations are possible:

1. Horizontal, with condensation in the shell, and the cooling medium in the tubes.
2. Horizontal, with condensation in the tubes.
3. Vertical, with condensation in the shell.
4. Vertical, with condensation in the tubes.

Most commonly used types of condenser:

a. Horizontal shell-side and
b. Vertical tube-side

Horizontal exchanger with condensation in the tubes

- is rarely used as a process condenser, but is the usual arrangement for
heaters and vaporisers using condensing steam as the heating medium.

12.10.1. Heat-transfer fundamentals

Filmwise condensation
-is the normal mechanism for heat transfer in commercial condensers

Dropwise condensation
-will give higher heat-transfer coefficients, but is unpredictable; and is not yet
considered a practical proposition for the design of condensers for general
purposes.

Physical properties
The physical properties of the condensate for use in the following equations,
are evaluated
at the average condensate film temperature: the mean of the condensing
temperature and
the tube-wall temperature.

12.10.2. Condensation outside horizontal tubes

In a bank of tubes the condensate from the upper rows of tubes will add to
that condensing on the lower tubes. If there are N r tubes in a vertical row and the
condensate is assumed to flow smoothly from row to row, Figure 12.42a, and if the
flow remains laminar, the mean coefficient predicted by the Nusselt model is related
to that for the top tube by:
Using Kerns method, the mean coefficient for a tube bundle is given by:

12.10.3. Condensation inside and outside vertical tubes

For condensation inside and outside vertical tubes the Nusselt model gives:

for a tube bundle

Equation 12.51 will apply up to a Reynolds number of 30; above this value
waves on the condensate film become important. The Reynolds number for the
condensate film is given by:

The presence of waves will increase the heat-transfer coefficient, so the use
of equation 12.51 above a Reynolds number of 30 will give conservative (safe)
estimates. The effect of waves on condensate film on heat transfer is discussed by
Kutateladze (1963).
Above a Reynolds number of around 2000, the condensate film becomes
turbulent. The effect of turbulence in the condensate film was investigated by
Colburn (1934) and Colburns results are generally used for condenser design,
Figure 12.43. Equation 12.51 is also shown on Figure 12.43. The Prandtl number for
the condensate film is given by:
 Figure 12.43 can be used to estimate condensate film coefficients in the
absence of appreciable vapour shear. Horizontal and downward vertical
vapour flow will increase the rate of heat transfer, and the use of Figure 12.43
will give conservative values for most practical condenser designs.
Boyko and Kruzhilin (1967) developed a correlation for shear-controlled
condensation in tubes which is simple to use. Their correlation gives the
mean coefficient between two points at which the vapour quality is known.
The vapour quality is the mass fraction of the vapour present. It is convenient
to represent the Boyko-Kruzhilin correlation as:

and the suffixes 1 and 2 refer to the inlet and outlet conditions
respectively.h0i is the tube side coefficient evaluated for single-phase flow of the
total condensate (the condensate at point 2). That is, the coefficient that would be
obtained if the condensate filled the tube and was flowing alone; this can be
evaluated using any suitable correlation for forced convection in tubes; see Section
12.8.
Boyko and Kruzhilin used the correlation:

In a condenser the inlet stream will normally be saturated vapour and the
vapour will be totally condensed.
For these conditions equation 12.52 becomes:

For the design of condensers with condensation inside the tubes and
downward vapour
flow, the coefficient should be evaluated using Figure 12.43 and equation 12.52,
and the
higher value selected.

Flooding in vertical tubes

When the vapour flows up the tube, which will be the usual arrangement for a
reflux condenser, care must be taken to ensure that the tubes do not flood. Several
correlations have been published for the prediction of flooding in vertical tubes, see
Perry et al. (1997).
One of the simplest to apply, which is suitable for use in the design of
condensers handling low-viscosity condensates, is the criterion given by Hewitt and
Hall-Taylor (1970); see also Butterworth (1977). Flooding should not occur if the
following condition is satisfied:

Where uv and uL are the velocities of the vapour and liquid, based on each phase
flowing in the tube alone; and di is in metres. The critical condition will occur at the
bottom of the tube, so the vapour and liquid velocities should be evaluated at this
point.