Class 5

Heat Exchanger Design

and

Separation Tower Design

1
Class 5a
Heat Exchanger Design
Important – the design methodology presented in these slides is a
selective summary of the best initial design approach for a shell and
tube heat exchanger and is not explicitly described in your text in such
detail – it is highly recommended that students be sure to download all ppt
files of “class slides” to your laptops as this information may be useful
for years to come.

2
Simplified View of Heat Exchangers
used in preliminary stages of design
factors ignored: Example 18.1
source/sink for energy transferred
rate of energy transfer
type and size of equipment
only overall heat duty considered
Q  m  H out  H in 
Example from HYSYS Vinyl Chloride simulation

"one-sided" heat exchanger

3
Heat Transfer Media (utilities)
Utility exchangers use media such as cooling water, steam, flue gases,
refrigerants, etc.
Table 18.1 provides list of different media covering range -150ºF to 2000ºF
Cooling water (Huerisitic 27, pg 168) Local abundant source of cool water
120°F in some areas: coastline, riverbanks

90°F
Cooling
Tower

Process streams cooled

to as low as 100°F
Process Stream Heat Exchangers

Direct air cooling to 120°F

If process streams are > 250°F, use waste heat boilers to create process
steam.

4
Heat Transfer Media (utilities)
Cooling below 100°F: refrigerants used, designated by R-number
R-717: ammonia (to -30oF) R-134a: tetrafluoroethane (to – 15oF)
R-290: propane (to -40oF)
Refrigeration cycle required: compressor, condensor, expansion
valve, + utility exchanger (vaporization mode)
Many HFC's ("freons") phased out since 1980 because of
ozone layer damage.
Chilled water can be used for cooling to 45°F (sensible heat)
Chilled brines to 0°F (sensible heat)
Heat medium: commonly steam
LP: 50 psig MP: 150 psig HP: 450 psig
(298°F) (366°F) (459°F)
Steam condenses in heat exchanger. Steam through-flow trapped.
For heating above 450°F to 750°F: Dowtherm (Dow Chemical product)
Molten salts (to 1100oF), molten metals (to 1400oF) for higher temperatures
Fired furnaces (to 2000oF) for high temperatures 5
Temperature Driving Force
When two heat transfer fluids considered, 2-sided heat exchanger
model used.
With negligible heat losses (assumed adiabatic),

T1in T1out
m 1
m 2
T2out T2in Table 18.5

Q  m 1CP1 T1out  T1in   m 2CP 2 T2in  T2out   UATm

energy balances transport
equation
Tm : mean temperature driving force
determines required heat exchange area

smallest ∆T at ends: minimum temperature approach

6
Temperature Driving Force
Common design strategy: specify
inlet conditions of each stream
pressure drops
a minimum approach temperature
Compute end for minimum, exit temperatures, heat duty
Optimal minimum approach temperature depends on conditions:
cryogenic: 1-to-2°F; 10oF or less for temperatures < ambient
Heuristic 26 ambient: 10°F; 20oF for ambient to (and including)300oF
high temperature: 100°F
When one fluid is boiled:
natural convection: ∆T's < 10°F
nucleate boiling: ∆T's 20°F to 45°F high transfer rates
film boiling: ∆T > 100°F
Heuristic 28
transition region: 50°F < ∆T < 100°F
rule of thumb:
maintain ∆T = 45°F (for example, a steam boiler) 7
Temperature Driving Force
Example 18.4
Fluid 1: 100 lbmol/hr ethyl chloride
10 lbmol/hr ethanol ∆P = 5 psi
200°F, 35 psia
Fluid 2: 90 lbmol/hr ethanol
90°F, 100 psia ∆P = 10 psi

Minimum approach temperature: 10°F

HYSYS Solution

8
HYSYS Solution

9
(specify min approach ∆T)
HYSYS Solution

Approach ∆T
spec met

but . . . Fluid 1
temperature
crossover
Fluid 2
Note: condensation
of hot stream

10
HYSYS Solution
Adjust approach ∆T to eliminate crossover (min ∆T inside Xchanger)

Fluid 2

Fluid 1

11
Typical Heat Exchanger Pressure Drops
Low viscosity liquid – no phase change: 5 psi
Vapor streams with no phase change: 3 psi
Condensing streams 1.5 psi Heuristic 31
Boiling streams 1.5 psi
Process streams through furnaces 20 psi

Catalog of Heat Exchange Equipment Types

pipe-in-pipe (double-pipe) small, < 1ft2 transfer area
shell-and-tube 50 to 12,000 ft2
countercurrent vs cocurrent flow vs crossflow
1-2, 1-4, 1-6, 1-8, 2-4, 2-8, 3-6, 4-8 (# shell pass - # tube pass)
air-cooled (fin-fan) area as much as 20,000 ft2
compact
plate-and-frame
spiral-plate
spiral-tube
plate-fin 12
Double-pipe Heat Exchangers
Simplest form: single, straight run
Example: inner: 12-ft long, 1-1/4-in, Sch 40 pipe Area: 5.22 ft2
More area req'd? use return bends, hairpin units
High temperature, high pressure, corrosive fluid in inside pipe
Generally not used when phase change occurs

Table 18.3
Nominal pipe

13
Shell-and-tube Heat Exchangers
Designs standardized by Tubular Exchanger Manufacturers Association
(TEMA)
Many configurations available

1
8
18.9
.
14
9
Air-cooled Heat Exchangers
used where cooling water is scarce

18.
18.11
11

15
Compact Heat Exchangers

18.12 16
Temperature Driving Forces
Q  UATm Tm : mean temperature driving force
Describing Tm is complex combination of factors:
flow configuration, fluid properties, phase change
Assumptions for use of TLM
steady-state flows
countercurrent or cocurrent
constant overall U
sensible enthalpy changes with constant Cp
negligible heat losses (adiabatic)
Stream temperature profiles are linear

T1  T2
TLM 
ln  T1 T2 
For isothermal condensation/boiling on one or both sides,
TLM applies to all heat exchanger configurations
17
Temperature Driving Forces
Correction factor for multiple-pass heat exchangers -- FT

Tm  FT TLM
R and S Factors

THin  THout TCout  TCin

R S
TCout  TCin THin  TCin

1-2 heat exchanger  1 S 

R 2  1  ln 
 1  RS 
FT 

 2  S R  1  R2  1
 R  1  ln 
 
 2  
S R  1  R 2
1  


0.85  FT  1 desirable FT  0.75 unnacceptable

Q = UAFT∆TLM for countercurrent flow 18

Temperature Driving Forces
1-2 heat exchanger

18.14

FT not changed drastically from 1shell pass-2 tube pass (1-2) when
additional tube passes are added (i.e. 1-4, 1-6, or 1-8)
19
FT charts available for multiple-shell-pass and crossflow exchangers
FT summary

When shell and tube heat exchangers with multiple-tube passes,

or multiple shell and tube passes are used, the flow direction of the
two fluids are combinations of countercurrent and cocurrent flow.
The resulting Tm for given values of T1 and T2, based on
countercurrent flow, is less than standard TLM. Hence, FT < 1.

When FT is unsatisfactory, a multiple-shell-pass heat exchanger

is used. The more shell passes, the higher is the value of FT. For
a given number of shell passes, the number of tube passes has
very little effect on FT

20
Heat Transfer Coefficients
Overall heat transfer coefficient, U, based on area of inner wall or
outer wall.
Sum-of-thermal-resistances formulas conduction

1
Uo 
1 tw Ao Ao Ao
R fo     R fi
ho kw Am hi Ai Ai
 L  Do  Di 
Am  film
ln  Do Di 
fouling

1
Ui 
Ai Ai tw Ai 1
R fo     R fi
Ao Ao ho k w Am hi

21
Heat Transfer Coefficients
For preliminary design, use reasonable estimates for U, see Table 18.5

For detailed equipment design, use published correlations

Categories for Correlations

Turbulent Flow, Straight, Smooth Ducts, Pipes, Tubes, Circular Xsection

Turbulent Flow, Annular Region Between Concentric Pipes
Turbulent Flow on Shell Side
Laminar Flow
Condensation
Boiling
Compact Exchangers

22
Heat Transfer Coefficients
Example of correlation
Turbulent Flow on Shell Side
0.14
ho D 1 3  b 
Nu   C  Re  Pr  
n

kb  w  viscosity
Donohue (1949) correction

D: tube o.d. kb : thermal conductivity of

bulk fluid
C: 0.2
G: mass velocity, average of two:
D G 1) parallel in baffle window
Re  2) normal to centerline tubes
b
CPb b n: 0.6
Pr  b : viscosity of bulk fluid
kb
CPb : heat capacity  w : viscosity of fluid at wall
surface temperature 23
of bulk fluid
Heat Exchanger Design Using HYSYS
Example 18.7
Existing 2-8 heat exchanger, carbon steel
shell ID: 39 in (Birmingham Wire Gauge, wall thickness of pipes)
1024 3/4-in 14 BWG, 16-ft-long tubes, 1-inch square pitch
38 segmental baffles, baffle cut: 25%
Tube inlet/outlet nozzles: 4-in, Sch 40 pipe; Shell 2.5 in, Sch 40 pipe
fouling factors: 0.002 (hr-ft2-°F)/BTU on each side
tube side: Toluene feed stream, 125000 lb/hr, 100°F, 90 psia
shell side: Styrene product stream, 150000 lb/hr, 300°F, 50 psia

Determine:
(1) Exit Ts
(2) Q
(3) P

24
Heat Exchanger Design Using HYSYS
Example 18.7

25
Heat Exchanger Design Using HYSYS
Example 18.7

26
Heat Exchanger Design Using HYSYS
Example 18.7

27
Heat Exchanger Design Procedure
(assuming inlet T, P, composition, phase, and flow rate are known
for both streams and an exit T or equivalent spec is given for one of
two streams)

1.) If utility stream is needed, select it from Table 18.1, together with its
entering and leaving T
2.) Decide which stream is tube side/shell side
3.) Estimate shell & tube pressure drops (Heuristic 31)
4.) Calculate heat duty and remaining exiting conditions of streams from
overall energy balance using eqn 18.1
5.) If a heating/cooling utility is used, calculate its flow rate from an overall
energy balance (eqn 18.1)
6.) Assume a 1-1 heat exchanger (1 shell pass/1 tube pass)
7.) Check for valid solution and that a reasonable T exists on both sides;
if a phase change occurs, a heating/cooling curve is calculated
to make sure that no temperature crossover is computed
8.) Make a preliminary estimation of the heat exchanger area, A, using an
assumed overall heat transfer coefficient, U, from Table 18.5
9.) Compute the mean driving force for heat transfer using heating/cooling
curves or eqn (18.3) 28
10.) Use eqn 18.7 to estimate A with FT = 1; if A > 8,000 ft2, multiple
exchangers of the same area are used in parallel
11.) From the estimated A, preliminary estimates are made of the exchanger
geometry
12.) A tube side velocity in the range of 1 to 10 ft/s is selected (4 ft/s typical)
13.) Calculate total tube inside cross sectional area
14.) Select a tube size (e.g. ¾ in. O.D. 14 BWG)
15.) Calculate number of tubes per pass per exchanger
16.) Select a tube length (e.g. 16 ft typical) and then calculate the number
of tube passes per exchanger
17.) Adjust the tube side velocity and tube length to obtain an integer number
of tube passes
18.) If more than one tube pass is necessary, use Figures 18.14 – 18.16 to
correct the log-mean temperature driving force; this may require
more than one shell pass (see Example 18.5)
19.) Select a tube sheet layout from Table 18.6 and a baffle design & spacing
for the shell side
20.) This completes the preliminary design of the heat exchanger
21.) A revised design is made by using the geometry of the preliminary design
to estimate an overall heat transfer coefficient from the calculated
individual coefficients and estimated fouling factors, as well as P; 29
the entire procedure is iterated to achieve satisfactory tolerance
Heat Exchanger Design Workshop
A hot stream is to be cooled from 300 to 100oC by exchange with
a cold stream being heated from 60 to 200oC in a single shell &
tube heat exchanger.  The duty of the exchanger is 35 MW and 
the overall heat transfer coefficient is estimated to be 
100 W/m2‐K.

(1) Specify the number of shell/tube passes? 
(2) What is the area required for the exchanger in m2?

(Assume minimum approach ∆T met)
Class 5b

Separation Tower Design

Typical Design Parameters
Operating Pressures: Typically, 1 – 415 psia
For Temperature Sensitive Materials: Vacuum Distillation (P > 5 mm Hg)
Condensers: Typically Total Condenser (unless low-boiling components)

Important – the design methodology presented in these slides is a

selective summary of the best initial design approach for a distillation
column and is not explicitly described in your text in such detail – it is
highly recommended that students be sure to download all ppt
files of “class slides” to your laptops as this information may be useful
for years to come.
Algorithm for Establishing Column P and Condenser Type

rule of thumb:
add 10 psia to PD

Figure 8.9
Combined Manual/Simulator Design Method
1) Estimate the distillate and bottoms compositions and flow rates via a hand
material balance and recognizing light and heavy key components
2) Use the graphical algorithm in figure 8.9 (page 216 of text) to establish column
pressure and condenser type
Using Distillate Composition
a) use HYSYS to help calculate the estimated bubble point pressure of the
distillate, PD, at 120oF (the idea is to use cooling water which typically exits at
120oF).
b) If this calculated bubble point PD < 215 psia, use a total condenser; however
, if the bubble point PD < 30 psia, set the condenser outlet pressure at 20 < PD < 30
psia to avoid vacuum operation.
c) However, if bubble point PD > 215 psia, calculate the distillate dew point
pressure at 120oF; if that dew point pressure is < 365 psia, use a partial condenser;
if the dew point pressure is > 365 psia, select a refrigerant that gives a minimum
approach temperature of 5 to 10oF (select this in place of cooling water) for the partial
condenser such that the distillate dew point pressure < 415 psia. Up to this point, the
tower operating P has been determined by the estimated composition of the distillate.
Checking Bottoms Composition
d) Using the determined condenser outlet pressure from (a) to (c), assume a
condenser P of 0 to 2 psia; assume a tower pressure drop of from 5 to 10 psia;
hence, a P at the bottom of the column, PB, will be 5 to 12 psia greater than the
condenser outlet pressure (almost all reboilers are partial reboilers that produce a
bottoms product at or close to the bubble point).
e) Determine the bottoms temperature, TB, by a bubble point calculation (using
HYSYS) based on the estimated bottoms composition and pressure; if this exceeds
the decomposition, polymerization, or critical temperature of the bottoms, then
compute the bottoms P based on a bottoms T safely below the limiting T
f) Then, using the assumed P, calculate a new condenser outlet P and T – this may
require a change in the coolant used in the condenser and the type of condenser;
the new condenser outlet P may be < 15 psia, in which case the tower will be
operating under a vacuum
g) For vacuum operations, the vapor distillate is sent to a vacuum pump
3.) Use Fenske Equation to determine the minimum number of equilibrium stages

ln [xLK/xHK)D (xHK/xLK)B]
Nmin = --------------------------------------------
ln [αLK/HK]av

where xLK = mole fraction of light key component

xHK = mole fraction of heavy key component
αLK/HK av = avg geometric relative volatility
= [(αLK/HK)D (αLK/HK)B]1/2 where (αLK/HK)D = KLK/KHK
D = distillate, B = bottoms products
and using HYSYS to calculate the relative volatility
4.) Use Underwood Equations to determine the minimum reflux ratio, Rmin

n (α )(x )
i F,i
Σ ------------ = 1 - q
i= 1 (αi – Θ)
(solve for Θ by trial and error)
n = number of individual components in the feed
αi = avg geometric relative volatility of component i in the
mixture relative to the heavy key component
xF,i = mole fraction of component i in the feed
q = moles of saturated liquid on the feed tray per mole of feed
Θ lies between the relative volatilities of the two key components

Then, solve for the minimum reflux ratio, Rmin:

n (αi)(xD,i)
Rmin + 1 = Σ -------------------
i=1 (αi – Θ)
xD,i = mole fraction of component i in the distillate
5) Select a Reflux Ratio that is R = (1.1 to 1.5) Rmin

6) Use the Gilliland Correlation to calculate the actual number of equilibrium stages, N
N – Nmin R - Rmin
------------- = 0.75 [ 1 – ( ---------)0.566]
R+1
N+1
(note that this equation is a best fit of the curve in Figure 19.1)
7) Use the Kirkbride equation to determine the ratio of trays above and below the
feed point:

ln (NB/ND) = 0.206 ln{(B/D) (xHK/xLK)F [(xLK)B/(xHK)D]2}

B = molar flow rate of bottoms

D = molar flow rate of distillate
ND = number of equilibrium stages above feed tray
NB = number of equilibrium stages below feed tray

(note: N = ND + NB which is solved simultaneously with Kirkbride

equation, using N calculated from the Gilliland Correlation)
8) Calculate non-key components if needed to estimate any additional compositions
for simulator

(xD,i/xB,i) = [(αi)avg]Nmin [(xHK)D/(xHK)B]

where:

xB,i = mole fraction of component i in bottoms

xD,i = mole fraction of component i in distillate
Nmin = minimum number of stages from Fenske equation
(αi)avg = average geometric relative volatility of component i relative
to heavy key component
Thus, prior to simulation, given a feed stream, we have estimated:

(1) molar flow rates of distillate, D, and bottom, B, via simple

material balance assuming a given split (desired split)
(2) column pressure, P, and condenser type via algorithm in
Figure 8.9
(3) minimum number of stages, Nmin, via Fenske equation
(4) minimum reflux ratio, Rmin, via Underwood equations
(5) selected a reflux ratio, based on multiple of Rmin
(6) calculated the actual number of equilibrium stages, N from
Gilliland Correlation (or Figure 19.1)
(7) calculated the feed stage from the Kirkbride equation
(8) estimated composition for any non-key components in distillate
and bottoms
9) Select Column Internals (trays, random or structural packing)
- trays favored for high operating P or liquid flow rate & large column diameter
(sieve trays preferred due to low cost; 60 to 85% efficient)

Typical Sieve Tray Geometry

- random packings recommended for small column diameters, corrosion,

foaming, or for batch columns (ceramic, metal, or plastic)
- structural packings used for low pressure or vacuum operations; also
for low P across column, or low liquid hold-up
Plate Efficiency and HETP
(Height Equivalent to a Theoretical Plate)
Trayed Columns
N actual plates = N equilibrium plates / Eo
(note: N equilibrium plates = N stages from prior calculations)
Typical Eo = 0.70 for distillation
= 0.50 for strippers
= 0.30 for absorbers
Packed Columns
Random Packings with low viscosity liquids (typical)
HETP (in feet) = 1.5 (Dp, inches)
(Dp = nominal diameter of a random packing)
For Vacuum Service
HETP (in feet) = 1.5 (Dp, inches) + 0.5

Packed height = N equilibrium (HETP)

Geometry and Efficiency of Random Packings

Geometry and HETP (height equivalent to a theoretical plate)

of Some Structural Packings
9) Calculate Column Diameter

Column Diameter
-Depends on vapor and liquid flow rates and properties up/down column
-Computed to avoid flooding (liquid begins to fill tower & leave with vapor)

Tray Towers
Packed Towers
Tray (Plate) Tower Column Diameter Calculation
- Diameter is calculated to avoid entrainment flooding

1/2
4G
DT =
(fUf) π [1 –(Ad/AT)]ρG

Where: G = mass flow rate of vapor (e.g. kg/s)

Ad = downcomer area (e.g. m2)
ρg = vapor density (e.g. kg/m3)
AT = tower inside cross-sectional area (e.g. m2)
f = fraction of the vapor flooding velocity (~ 0.75 to 0.85)
Uf = vapor flooding velocity (e.g. m/s)
DT = tower diameter (e.g. m)
Uf = C[(ρL – ρG)/ρG]1/2
ρL = liquid density (e.g. kg/m3)
C = Empirical Capacity Parameter
C = CSB FST FF FHA

Flooding correlation

Note: convert CSB to m/s before using

(L, G; liquid, gas mass flow rates)

Flooding Correlation for Sieve Valve, and Bubble-cap trays
FST = surface tension factor = (σ/20)0.20
where σ is surface tension in (dyne/cm)
FF = foam factor (=1 typical for distillation systems;
= 0.5 to 0.75 for foaming systems)
FHA = hole factor (=1 for valve and bubble-cap trays;
=1 for sieve trays for (Ah/Aa >0.10)
= [5(Ah/Aa) + 0.5)] for 0.06 < (Ah/Aa) < 0.1)
where Ah = hole area of tray
Aa = active area of tray = (AT – 2Ad)
where bubbling occurs
Ad, downcomer area
AT, tower inside cross-sectional area
Packed Tower Column Diameter Calculation
The diameter of a packed tower is calculated from an estimated flooding
velocity with a continuity equation similar to that for a tray tower:

1/2
4G
DT =
(fUf) π ρg

Where: G = mass flow rate of vapor (e.g. kg/s)

ρg = vapor density (e.g. kg/m3)
f = fraction of the vapor flooding velocity (~ 0.7 for packed tower)
Uf = vapor flooding velocity (e.g. m/s)
DT = tower diameter (e.g. m)
Uf for packed towers calculated from generalized correlation

Y = exp[-3.7121 – 1.0371 (ln FLG) – 0.1501 (in FLG)2

– 0.007544 (ln FLG)3]

Y is dimensionless
Flow Ratio Parameter: FLG = (L/G)(ρG/ρL)1/2

f{ρL} = -0.8787 + 2.6776 (ρH2O(l) / ρL) – 0.6313 (ρH2O(l) / ρL)2

for density ratios (ρH2O(l) / ρL) from 0.65 to 1.4

f{µL} = 0.96 (µL0.19) for liquid viscosities from 0.3 to 20 cP

and random packings of 1 inch nominal diameter or more
1/2
(Y)(g) (ρH2O(l) / ρG)
Uf =
Fp f{ρL} f{µL}

Important (units)!

g = 32.2 ft/s2
Y = 0.01 to 10 from correlation
Uf = flooding velocity (ft/s)
Fp packing factor in (ft2/ft3) from Table 19.1

(Tower inside diameter should be at least 10 times nominal

packing diameter and preferably closer to 30 times)
Table 19.1
Separations Tower Design

Workshop

(Workshop was put on D2L for you to download)