INTRODUCTION

TO BASIC DESIGN
(2013/2014)
ASSIGNMENT #2A REPORT
SALVATORE LUISO 0000707745

CHEMICAL AND PROCESS ENGINEERING

A mixture of Benzene – Toluene with zBenz = 0.45 is fed to a distillation column with flow rate 
equal  to  25000  kg/h  (291.72  kmol/h)  and  inlet  temperature  T  =  40°  C.  The  distillation 
column,  operating  at  P  =  1  atm,  is  used  to  obtain  two  liquid  streams  with  top  benzene 
composition xD = 0.995 and bottom benzene composition xW = 0.002  (for every calculus see 
calculus sheet annex). 

1. We start doing an overall mass balance to find the flow rate and the composition of 
the outlet products: 

291.72  0.002  (291.72  0.45  D  0.995)  D

F  W  D   D  131.61kmol / h
   ( 291.72  0.45  D  0.995) 
 F  z F  D  x D  W  xW  W W  160.11kmol / h
0.002
Where D is the top product and W is the bottom one. We have the equilibrium data so we 
can  calculate  the  bubble  point  and  the  dew  point  T  for  the  top,  for  the  feed  and  for  the 
bottom composition with a linear interpolation of data: 

EQUILIBRIUM DATA    

T (°C)  y  x 

110,68  0,0000 0,0000 

109,59  0,0539 0,0250 

…  … ... 

100,74  0,4357 0,2500 

99,86  0,4688 0,2750 

…  … … 

94,09  0,6627 0,4500 

…  … … 

80,65  0,9903 0,9750 

80,13  1,0000 1,0000 

1 
 
The results are:

   FEED =  25000 kg/h       

   z  F (kmol/h)  T(°C / K)  TB(°C / K)  TDP(°C / K) 
BENZENE  0,450 131,27 40,00 94,09  100,36
TOLUENE  0,550 160,44 313,15 367,24  373,51
sum     291,72   
TOP PRODUCT    
   xD  D (kmol/h) TB(°C / K)  TDP(°C / K)
BENZENE  0,995 130,95 80,23  80,40
TOLUENE  0,005 0,66    353,38  353,55
sum     131,61   
BOTTOM PRODUCT   
   xW  W (kmol/h) TB(°C / K)  TDP(°C / K)
BENZENE  0,002 0,32 110,59  110,64
TOLUENE  0,998 159,79    383,74  383,79
sum     160,11         
 

We see that the feed is under cooled liquid. 

We insert the x‐y plot: 

y vs x
1,0
0,9
0,8
0,7
0,6
diagonal 
y 0,5 y=x
0,4
0,3 BENZ‐
0,2 TOL EQ.
0,1
0,0
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0

x (benzene)
 
2. We will follow the McCabe‐Thiele procedure, so by doing the following assumptions: 
 Sensible heats negligible with respect to latent heats (isothermy); 
 Latent heats equal for all compounds in the mixture and constant in T; 
 Mixing heat negligible (it could be verified); 

2 
 
  No heat losses. 

These  hypothesis  are  reasonable  because  the  temperature  differences  in  column  are  few 
tens degree and we are dealing with an hydrocarbons mixture. 

We do now the heat balance on the feed plate, setting the enthalpy of the saturated liquid  
as the reference enthalpy and equal to 0: 

~ ~
H L  H 0  0
 ~ ~ ~
 H G  H 0   (T DP )  c P , L (T DP  TB )
~ ~
 H F  H 0  c~P , L (TF  TB )
 
Where for λ and cp,LIQ we integrate the polynomial equations for each compound from Perry,
8th between TB and TD (at feed composition) for the cP in the second equation and between TB
and TF for the third one. Then we calculate a mean value for both, weighted on composition
of liquid.

Now we can find the q-factor needed to find the q-line:

~ ~ ~ ~
H G  H F H 0   (TDP )  c~P , L (TDP  TB )  H 0  c~P , L (TF  TB ) 32003.92  1047.97  8550.75
q ~ ~  ~ ~   1.259
HG  H L H 0   (TDP )  c~P , L (TDP  TB )  H 0 32003.92  1047.97

Where HG is the enthalpy of the saturated gas at the dew point at the composition of the feed.
So the q-line has the equation:
q z
y x  F  4.865 x  1.739
q 1 q 1

Now we can find the rectifying line and, looking at the intersection with the y-axis, we find
the minimum reflux:

L D R 1  y  0.4235 x
y x  xD  x xD    RMIN  D  1  1.3495
G G R 1 R 1 x  0 y

Actual typical reflux is between 1.2*RMIN and 2*RMIN. We choose 1.5 times, however we
programmed a calculus sheet the provides automatically the following results, only by
changing the ratio between minimum and actual reflux.

Thus, the rectifying line modifies and the equation becomes:

y  0.6693 x  0.329

The stripping line crosses the point (xW,xW) and the intersection between the q-line and the
rectifying line:

3
 y  1.3381x  0.0007

We are now able to calculate the flow rates of liquid and gas circulating respectively in the
top and in the bottom section:

α= 1,5
R = α RMIN =  2,024    
 L= R∙D =  266,41  kmol/h 
TOP 
 G= L+D =  398,02 kmol/h

 Ḻ = L+F∙q = 633,59  kmol/h 

BOTTOM 
 Ḡ = G+F∙(q-1) = 473,49  kmol/h 

y vs x
1,0
xD
0,9

0,8

0,7 diagonal y=x
0,6
y rectifying 
0,5
0,4235 zF line
0,4
0,329 q‐line
0,3

0,2 stripping line
0,1

0,0 xW Rmin
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0

x (benzene)

Now we can draw the stages at maximum reflux and the theoretical stages.

The results are:

 12 stages at maximum reflux;

 23 stages at the actual reflux;

4 
 
 y vs x
1,0 3 21
5 4 xD
6
0,9 7
8
0,8 9
10
0,7 11 diagonal y=x
12
0,6 rectifying line
y 13
0,5 q‐line
14
0,4235 zF
0,4 stripping line
15
0,329
0,3 Theoretical stages
16
0,2 17 Stages at total reflux = 
12
18
0,1 Rmin
19
20
21
22
23
0,0 x W BENZ‐TOL EQ.
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0

x (benzene)
 Feed plate is number 11.

3. We want to calculate the heat duties of both the overhead condenser and the reboiler.
Thus, we start to calculate the heat of vaporization of the mixtures at xD and xW: since the 2
mixtures are almost pure it is not so important that the temperature for the calculus is at the
dew point, because bubble and dew point differ from less than 1°C one another; it is also
worthless to weight the heats of vaporization on the composition again because of the purity.
The results are:

λD (TDP=80,13°) = 30795,88 kJ/kmol

λB (TDP=110.68°) = 33360,03 kJ/kmol

As for the condenser we need the gas flow rate circulating in the top section, as for the
reboiler we need the gas flow rate in the bottom section.

Q D  G   D (TDP  80.13C )  3404.83kW

Then:
Q B  G   B (TDP  110.68C )  4387.67kW

5 
 
We want to know also the flow rates of the cooling water in the condenser and of the steam in
the reboiler. They will be needed when we will estimate the operative costs. So we start
calculating the average cP of the water between 30°C and 40°C [see Assignment #1 report]
and heat of vaporization of steam at 4/5 bar. So:

Cooling water flow rate: mH2O = 81.32 kg/s

Steam flow rate: msteam = 1.94 kg/s

4. The problem suggests us to design a tray column for the operation. We follow the
procedure from Coulson & Richardson’s Chemical Engineering, Vol. 6. We start dividing the
column in 4 critical sections: top, above feed, below feed and bottom. In each section there
are different values of flow rate, composition, temperature, density and other properties, so
we have to find the correct design for each and then we will choose the best that is well-
suited for them all.

Looking at the composition we calculate the molecular weight for each section, so to
determine the mass liquid and vapor flow rates. Then, starting again from composition, we
take the temperature at the bubble point as the temperature of each liquid and the temperature
at the dew point as the temperature of each vapor. Of course the mixtures at the top and at the
bottom are almost pure and the 2 temperatures are quite the same. After we have calculated
the densities at each T, we can find the volumetric flow rate and we can use the following
plot to determine which kind of liquid-flow arrangement we have:

6
With this plot we can have an idea of the structure of the plates. The results are:

INPUT DATA                         
TOP STAGE                         
VW  VW  ρL ρV MMIX( MMIX( FLOW
LW (kg/s)  LW (m3/s)  (kg/s)  (m3/s)  (kg/m3) (kg/m3) L) V) PATTERN
5,78  7,16E‐03  8,63  1,33 807,58 6,50 78,15 78,07  Cross flow 
ABOVE FEED STAGE                      
VW  VW ρL ρV MMIX( MMIX( FLOW
LW (kg/s)  LW (m3/s)  (kg/s)  (m3/s)  (kg/m3) (kg/m3) L) V) PATTERN
6,34  1,07E‐02  9,15  2,01 591,92 4,56 85,7 82,72  Cross flow 
BELOW FEED STAGE                      
VW  VW  ρL ρV MMIX( MMIX( FLOW
LW (kg/s)  LW (m3/s)  (kg/s)  (m3/s)  (kg/m3) (kg/m3) L) V) PATTERN
15,08  2,55E‐02  11,27  2,47 591,92 4,56 85,7 82,72  Cross flow 
BOTTOM 
STAGE                         
VW  VW  ρL ρV MMIX( MMIX( FLOW
LW (kg/s)  LW (m3/s)  (kg/s)  (m3/s)  (kg/m3) (kg/m3) L) V) PATTERN
16,19  3,29E‐02  12,09  4,00 491,46 3,02 91,97 91,90  Double pass 

We decide to select a plate distance lt = 0.61m (24 in). Then, we calculate the liquid-vapor
LW V
flow factor for each section: FLV  . We income with them in the following plot to
VW L
find the K1 factor (line colors are the same of the first plot):

7
The following restrictions apply to the use of the previous plot:

a) Hole size less than 6.5 mm. Entrainment may be greater with larger hole sizes;
b) Weir height less than 15 per cent of the plate spacing;
c) Non-foaming systems;
d) Hole: active area ratio greater than 0.10;
e) Liquid surface tension 0.02 N/m.

 L V
We can estimate the flooding velocity: u f  K1 
V

We want to know the actual velocity but we need the fractional entrainment ψ. As a rough
guide the upper limit of ψ can be taken as 0.1; below this figure the effect on efficiency will
be small. Thus, we insert in the following plot the FLV and at ψMAX = 0.1 we get the
percentage of flooding:

8
 FLV Type K1 from plot uf (m/s) Flood % at max ψ=0,1
Top stage 0,060 1,10E-01 1,2 92

Above feed stage 0,061 1,10E-01 1,2 92

Below feed stage 0,117 9,30E-02 1,1 96

Bottom stage 0,105 9,50E-02 1,2 97

Since %FME of all the sections is above 85: ug = 0.85 uf, where ug is the actual velocity. Now
VW
we can calculate the net area An  and by setting the coefficient fd based on the
W  u g
liquid flow pattern (0.12 for cross flow, 0.24 for double pass) we can know the area of the
An
column Ac  and the diameter of the column dc. The results follow:
1 fd

ug (m/s) An (m2) fd Ac (m2) dc (m)

Top stage 1,0 1,279 0,12 1,454 1,361
 
Above feed stage 1,1 1,890 0,12 2,148 1,654
 
Below feed stage 0,9 2,755 0,12 3,131 1,997
 
Bottom stage 1,0 3,898 0,24 5,129 2,555

Now, to be sure that there will not be flooding anywhere we take the biggest diameter, i. e.
2.555 m. We calculate now the area of the column Ac, the area of the downcomer Ad
(Ad=fd∙Ac, where we take the cross flow pattern so fd=0.12), the active area Aa (Aa=Ac-2Ad),
the perforated area Ap (Ap≈Aa); then we set the holes diameter dh= 3 mm, the pitch equal to
Ph=3.3∙dh, the weir height hw= 50 mm that is good for Patm and the apron spacing hap=40mm;
2
d 
so we calculate also the overall area of holes: Ah  0.907   h   A p . Here there is the final
 Ph 
table:

Ac Ad An Aa Ap dh Ph Ah hw hap
diameter m (m2) (m2) (m2) (m2) (m2) (mm) (mm) (m2) (mm) (mm)
2,555 5,129 0,615 4,513 3,898 3,898 3,0 10 0,325 50 40
To finalize the tray design we need the percentage of the ratio between the downcomer and
A
the column area: d  100  12 . We put it in the following plot and we get the weir length:
Ac

9 
 
→ lw=0.76∙dc=1.942 m. So the clearance area under the downcomer is Ap= lw ∙ hap =0.078 m2.

Now there are some restrictions that we have to control in each section. First we check the
2/3
 L 
weeping. Thus, we calculate the weir crest: how  750   W  , so we can income in the
  L  lw 
following plot with the sum how+hw to get the coefficient K2 needed to calculate the minimum

vapor velocity through the holes (based on the hole area) u h  2
K  0.9  25.4  d h  :
 v  1/ 2

10
 Vw 
To check the weeping it must be u h   u h , where uh is the actual vapor velocity.
W  Ah
The results:

how (mm) uh (m/s) hw +how(mm) Type K2 ŭh (m/s)

Top 17,901 4,090 68 30,5 4,06

Above feed 23,416 6,179 73 30,6 4,89

Below feed 41,722 7,615 92 30,9 5,03

Bottom 49,506 12,332 100 31,0 6,24

At this point we want to estimate the pressure drops in each plate. Let’s call hb = how+hw. For
each section we calculate the total pressure drops between the plates hT = hd + hb + hR, where
 H 2O
hR  12.5  and hd is taken from the following plot incoming with: percentage ratio
L
between the hole area and the perforated area and the ratio between the plate thickness
(typically 3 mm for steel alloy) and the hole diameter.

2
u  v
With C0 we calculate hd  51   h  
 C0  L

11
 hR (mm) hd (mm) hT (mm)
Top 15,5 10,0 93,3

Above feed 21,1 21,8 116,3

Below feed 21,1 33,1 145,9

Bottom 25,4 69,2 194,1

It’s time for the second check: the downcomer flooding. The liquid flow rate in the
downcomer Lwd is calculated taking the real fractional entrainment ψ from FLV-ψ plot,
L
incoming with 85% flooding: Lwd  w . Then we need Am, that is either the downcomer
1 
area Ad or the clearance area under the downcomer Aap; whichever is the smaller. Finally the
2
 Lwd 
head loss in the downcomer is: hdc  166    .
  L  Am 
For safe design the liquid head hb in the downcomer hb = hT + hdc + how+hw should not exceed
half the plate spacing lt, to avoid flooding. Allowing for the weir height:
hb  0.5  (lt  hw )
 L  hb  Ad
Meanwhile the residence time in the downcomer   should be greater than 3 to
Lw
5 seconds. If both the liquid head and the residence time are nearby their limit, try to change
the configuration to get a safer design. The results:

Downcomer ψ LWD (kg/s) hdc (mm) 0,5(hw+lt) hB (mm) τ

Top flooding 0,055 6,1 1,6 330,0 162,8 13,2
 
Above feed 0,055 6,7 3,5 330,0 193,3 10,5
 
Below feed 0,030 15,5 19,0 330,0 256,6 6,0
 
Bottom 0,035 16,8 32,0 330,0 325,7 5,9

Further step is to calculate the column height. To do this, we need the plate efficiency: even if
we chose sieve plates, as a rough estimate the efficiency could be got either from the usual
plot for bubble-cap efficiencies or from the interpolating equation:
 oa  0.492   F   hk  0.245
, where μF is the viscosity of the feed and αhk is the relative
volatility in each section. The results:

12 
 
 Column height αHK  μL (feed) cP αHK μL ηoa
Top 2,213 2,81E-01 6,21E-01 0,55

Above feed 2,405 2,81E-01 6,75E-01 0,54

Below feed 2,405 2,81E-01 6,75E-01 0,54

Bottom 3,518 2,81E-01 9,88E-01 0,49

We choose to take an average efficiency of 0.53, so the actual number of plates in the column
is 43. We are almost at the end.
For the column height we need the top and the bottom height. As for the former we set HT=1
m to allow disengagement between liquid and gas, as for the latter HB it’s the sum of:
 the height that multiplied by the column area gives a volume that can keep inside the
amount of liquid product flow rate for 3 minutes;
 the height that multiplied by the column area gives a volume that can keep inside the
amount of liquid actually inside the column on all plates;
 the diameter of the gas nozzle and his distance from the first plate.

Thus the column height is: H  lt  ( N P  1)  H B  H T

Finally there are the pressure drops: P   v  ht  N P
We made an average for the density and calculating the number of plates before and after the
feed we can estimate with more accuracy the average pressure drops in the top section and in
the bottom one: then, we sum them.
The final results are:
NP ηoa HB (m) H (m) ΔP (Pa) Feed plate
43 0,53 4,53 31,15 35066 21

Our design is going to be completed. Actually nozzle diameters are still missing. We assume
that gases should have a velocity of 10-20 m/s (we choose 10 to avoid flooding), and liquids
of 1 m/s, so we calculate a diameter and then we check on the standard the nearest one:

d F  127 mm u F  0.99m / s
 
d GAS ,TOP  406.4mm u GAS ,TOP  10.54m / s
4G  
d 'b   d LIQ ,TOP  101.6mm so actual velocities are: u LIQ ,TOP  0.93m / s
   u  
d GAS , BOTTOM  711mm u GAS , BOTTOM  10.32m / s
d LIQ , BOTTOM  219.1mm u LIQ , BOTTOM  0.91m / s
 

The design is complete.

5. We want to estimate now the investment and the operating costs of the equipment.
Among the different methods in the literature we choose Guthrie method, because more
reliable than other but still simple. As we are dealing with a tray column Guthrie suggests us
to estimate separately the tray stack cost and the vessel cost. We start with the first one and
we implement this graphical method on a calculus sheet, thus to make the procedure

13 
 
automatic. The following plot is a collection of real values that relate the tray stack height in
foot (with a tray spacing of 24 in) to a tray cost, shop installed.

DATA 
Tray stack height  Diameter  Tray spacing  Tray height 
31,15  m 2,56  m  0,61 m  25,62  m 
102,16  ft 8,38  ft  24,02 in  84,03  ft 

100000 Guthrie Method
Tray cost, shop installed, $

10000
6 y = 66,667x
7 y = 90x
8 y = 125x

9 y = 150x
1000
10 y = 166,4x + 67,8

100
1 10 100
Tray stack height, ft. (24 in spacing)

Different line is to be chosen following the diameter: here we take that one for 8 ft diameter.
Tray base cost     $12.769            
  
 Now we need some adjustment factors to get the equipment cost. As for the material, we 
use stainless steel: 

 
Tray Cost, $ ‐ [Base cost (Fs + Ft + Fm)] Index             
Adjustment factor                      
                          
Tray 
Tray spacing, In.  Fs  type    Ft Tray material  Fm
24     1,0  Grid(no downcomer)  0,0   Carbon steel  0,0
18     1,4  Plate     0,0   Stainless  1,7
12     2,2  Sieve     0,0   Monel  8,9
         Trough or valve  0,4        
Bubble 
         cap     1,8        

14 
 
Here tray spacing is 24  Sieve tray are used   We select stainless steel  
Fd=  1,0  Fp= 0,0 Fm=  1,7
     
Tray cost     $34.478            

Now we calculate the cost of the vessel. To do this we use another plot that relates the
column diameter, the column height, the vessel configuration to a base equipment cost. So:

Base equipment cost (CB‐EQ)  $45.000    Pressure  101325 Pa 

 We need adjustment factors as before:  15 PSI
Vessel Cost, $ ‐ [Base cost *Fm * Fp] Index                
Adjustment factor           Pressure factor       
               PSI Fp       
      Fm Fm Up to 50 1,00       
Shell material  Clad  Solid    100 1,05       
Carbon steel  1,00 1,00    200 1,15       
Stainless 316  2,25 3,67    300 1,20       
Monel     3,89 6,34    400 1,35       
Titanium     4,23 7,89    500 1,45       

15
 All the column is in stainless steel Fm= 3,67  Fp= 1,00    
     

Vessel cost (CEQ)     $165.150            

Guthrie method takes into account also the costs for installation and labor, thus to have the
erected cost.

            Field installation modules      
Module           5A  5B  5C  5D  5E 
Base dollar magnitude, $ 100,000     Up to 2  2 to 4  4 to 6  6 to 8  8 to 10 
Bare module cost        423,0 412,1 406,8 405,7  402,0

Since our equipment cost is below $200000, the factor we are interested in is the first
(divided by 100 in this procedure: 100*f = 423.0).

Bare module cost (erected eq.) 
CEQ,erected=(CEQ‐CB‐EQ)+f ∙ CB‐EQ=  $310.500

Total cost (erected + tray)  $344.978

This is the final equipment cost but it is US $1968 so we need some coefficients to actualize it.
We take the Chemical Engineering Plant Cost index of 1968 for the equipments (=109.9) and
we update it with the CEPCI of September 2012, choosing the tag heat exchangers & tanks:

16
Then we apply the comparative price levels factor from United States to Italy again in
September 2012:

17
Finally, we actualize the Italian cost with index from ISTAT in April 2014:

 CEPCI 2012  I   643.9   103.7 

C IT , 2014  C1968     L$US  € IT , 2012   2014   344978     1.01     2061300€
 CEPCI1968   I 2012   109.9   102.7 

Let’s discuss now about operative costs. It’s difficult to know something about them because
we don’t know the other characteristics of the plant, however we can estimate at least the cost
of pumping cooling water into the condenser, steam into the reboiler and reflux to the top. As
for the first and the second we find the corresponding amount of ECO (Equivalent
combustible oil, a standard liquid fuel with 10000 kcal/kg) per m3 of water and per kg of
steam at 5 bar. Then we need the average industrial fuel price in Italy:

Water flow rate in the condenser:  81,32 kg/s 

Steam flow rate in the reboiler:  1,94 kg/s 
Industrial water:  0,07 kgECO/m3 water
Steam (5 bar) :  0,058 kgECO/kg steam
Average industrial fuel price italy  0,97329 €/kg 

So the annual costs for condenser and reboiler are:

kgECO
m H 2O  C fuel 
m H3 2O s h d
C cond , annual   (3600 )  (24 )  (365 )  176039€
 H 2O h d y

kgECO s h d
C reboil , annual  m steam  C fuel   (3600 )  ( 24 )  (365 )  3455975€
kg steam h d y

As for the reflux pumping we set a total dynamic head H = 36.15, that is the height of the
column allowing 5 m for pressure drops in the line. By considering a system efficiency
QgH
(motor+pump) = 60%, the power needed is P   3.42kW . Now we take the cost of


18
electric energy in Italy from Authority of Energy: CE = 0.2 €/kWh; and we estimate the
annual operative cost:

h d
C reflux ,annual  P * 24 * 365 * C E  5992€
d y

The total annual operative cost is:

C op ,annual  C reflux , annual  C cond ,annual  C reboil ,annual  3638107 €

Last step is to calculate the Net Present Cost assuming an interest rate i = 0,11 and a number
of years n=10:

 (1  i ) n  1 
NPC  C capital  C op ,annual   n 
  23486957€
 i  (1  i ) 

Let’s spend some words on the P&ID. We designed the tray column connected to a BEM
condenser and to a natural circulation reboiler. As for the condenser see Assignment #1
report. From the condenser, the reflux is pumped to the top; a temperature indicator and
control is set in the top section of the column in the most appropriate tray, so to actuate the
control valve on the reflux line. Other temperature indicators are set on the feed line and on
the bottom section to have a look if the process is doing well. As well as in the buffer tank for
the condenser, in the bottom section a level indicator and control acts on a control valve
down the process so to maintain the desirable liquid level in the column. Low alarm level and
level switch are also present on the column: they eventually switch off the pump to prevent
gas bubbles going inside it.

A flow indicator and control is responsible of the flow rate of the feed.

A pressure safety valve and a hand switch on the relief line are set both on the top outlet gas
line and on the buffer tank.

19