Optimizing Distillation Column Design for Chomical Chemical Company, Study Guides, Projects, Research of Chemistry

Download Optimizing Distillation Column Design for Chomical Chemical Company and more Study Guides, Projects, Research Chemistry in PDF only on Docsity!

ChE 305 - Process Separations Project

Ian Lutz

Tom Salerno

Michael Ocbo

Jimmy George

Stephen Tomczewski

Introduction:

A firm by the name of Chomical Chemical Company is planning to distill

dichloromethane, trichloromethane and carbon tetrachloride. The Separations Group has

been charged with the assignment to come up with a number of column designs and

associated operating conditions to see which is the most cost effective. From this data the

vice president will choose the design which is most economical.

Your manager, however, wants to get moving on this project immediately instead

of waiting a few years to get all of the data from the Separations team, and then need to

go through each design to find out which one is best. Thus he has charged us with the

assignment to get a rough idea of how this column will operate on his desk in a few days,

so that he might continue with his work while the research team is finalizing a more exact

analysis of the column. From the data he distributed to us, we came up with the following

schematic of what the column will look like.

Figure 1: Schematic of distillation column for Chomical Chemical Company

However, for the external balances there were still six unknowns and only 5

independent equations. Thus our group made the assumption that all of the HNK (carbon

tetrachloride) would go to the bottoms. With this assumption we were able to write,

4

,

D CCl

x 

We then solved Equations 1, 2, 3, 4 and 5 simultaneously and were able to

completely specify the feed stream, the bottom stream, and the distillate stream. After this

was completed we found the values for L, the liquid rate flowing down the column, and

V, the vapor rate flowing up the column, by using an overall balance around the

condenser and the definition of the reflux ratio.

(6)  Reflux ratio^  and

L

L D D V L D

D

We realized that although our assumption of all of the HNK going to the bottoms

is a good approximation, it is probably incorrect. The real value might be

-4 -

10 , 10 etc.

Consequently, the percent error in D CCl , 4

x (^) is probably very large, but the percent error in

4

B CCl ,

x is probably very small. Thus we decided to step off stages starting from the

bottom, which is more accurately specified, and moving up.

Before we could step off stages, we needed a method to correlate the

compositions of passing streams, such a function would allow us to move from one stage

to the next. We found such a relation for the part of the column above the feed (the

enriching section) by performing mass balances around the top of the column and an

arbitrary stage j. From this we were able to construct the top equilibrium line as,

, 1 , ,

i j i j i dist

L L

y x x

V V



We obtained another relation for the part of the column below the feed (the stripping

section) by performing mass balances around the bottom of the column and an arbitrary

stage j. From this we obtain our bottom equilibrium line as,

, 1 , ,

i j i j i bot

L L

y x x

V V



From our external balances we were able to completely specify the liquid stream.

Thus our first step was to calculate the vapor compositions leaving this stage. To do this

we had to use our ideal gas relationship,

i

i

tot

p

K

p

where our vapor pressures (

i

p (^) ) for each species at each temperature where found using

the Antoine Equation, to make the calculations in Excel easier. The constants for the

Antoine Equation were found from the text, Elementary Principles of Chemical Process.

The functional form of this equation is shown below,

(10) (^)  

10 i

B

Log p A

T C

With Equation 10, our process was to guess a temperature, calculate the vapor pressures

of each component, find the K values of each component, find the vapor composition of

each component using,

i i i

y  K x

and then see if 1

C

i

i

y 

. We performed this process through the use of Goalseek, a

function in Excel.

A summary of our Excel calculations are shown below

Table 1: Summary of results from Excel calculations

The raw data from our Excel file and the associated hand calculations follow

immediately.

Aspen (Simulator) Calculations:

After finishing our Excel calculations, our group wanted to obtain more

confidence in our rough sketch of the distillation column for Chomical Chemical

Company, and thus decided to try and solve their separation problem using Aspen.

The first step in this process was to find the number of stages and the optimal feed

stage for the required separation of Figure 1. To do this we designed a DIST column in

Aspen. This design required us to first solve all of the external balances, which we had

previously done in Excel. We also specified Aspen to use IDEAL relationships in its

calculations and allowed it to use its own stored thermodynamic data. Once this data was

plugged into Aspen, it generated results for the total number of equilibrium contacts, the

optimal feed stage, and the Distillate to Feed ratio. Along with this output came many

other parameters, which were not of vital importance in our rough design. However, this

design did not provide us with the Temperature and Liquid Composition Profiles we also

needed for our design.

Thus we needed to design another column, this time we used RADFRAC. The

inputs required for this column were: first, a completely specified feed stream which was

given in the problem statement; next the number of stages; operating pressure; reflux

ratio; and the Distillate to Feed ratio. All of this data was collected from either the

problem statement or from our previous Aspen column using DIST. After this data was

plugged in and we again specified using IDEAL relationships for their calculations,

Aspen was able to spit out the results of this column. The results here included our final

set of data, the temperature and liquid composition profiles of our column.

The results of these calculations in Aspen are as follows.

Liquid Composition Profile - Mole Fractions

Temperature Profile - degrees Celsius

• Stage DICHL-01 CHLOR-01 CARBO-
  • 6 0.786 0.208 0.
  • 5 0.613 0.369 0.
  • 4 0.408 0.543 0.
  • 3 0.231 0.662 0.
  • 2 0.115 0.701 0.
  • 1 0.051 0.657 0.
  • 6 9. Stage Temperature - Aspen
  • 5 12.
  • 4 16.
  • 3 20.
  • 2 23.
  • 1 26.

Comparison of Methods

The two methods used to solve this distillation problem both have their

advantages and disadvantages. Both solutions methods used the same assumptions:

Constant Molal Overflow, the system is adiabatic, and that all of the LNK’s and HNK’s

go to the top or bottom. The results from both methods accurately agree showing that the

actual number of stages is 6 and the optimum feed location is at Stage 4, when numbered

from the top. As it is said, the results are only as good as the equilibrium data given.

Therefore, the simulator’s result is thought to be more accurate because it retrieves its

data from the program’s highly precise properties package whereas the Excel method

uses the Antoine equation to calculate the K-values.

In comparison, the main difference between the Excel and simulator solution is

that of convenience. Engineers who are well acquainted with programs such as HYSYS,

SIMSCI, and Aspen can yield results quickly and accurately by just inputting initial

conditions and letting the program do all the calculations. Of course longer and more

complex processes will require a greater amount of time even with these simulators, but it

would still be faster than manually typing in all the calculations into Excel. Another

advantage to simulator programs is the wide variety of scientific tools that are readily

accessible within their interfaces. With the click of a mouse, property packages provide

the user with physical, thermodynamic, and transport data which would have to be looked

up if Excel was being used. The output of results that simulators compile shows

everything from flow rate values to estimated power requirements for the system’s

condenser and reboiler. However, the main drawback to simulator programs is their price

tags. Companies and universities spend thousands of dollars to have these tools and keep

them updated. Many times companies spend money on training program for employees

which add to the overall cost.

The Excel solution method is somewhat tedious in setting up the problem,

however, after the first stage calculation is completed, a macro can be written to solve the

rest of the problem. Another disadvantage to using Excel is the error-making

use the average of the two as a good estimate for the temperature and liquid composition

profiles of our column.

Our overall solution is as follows: the column has 6 equilibrium contacts, 5

stages plus a partial reboiler; the feed stage should be located on the fourth stage from

the top; the temperature profile lies somewhere in between the following curves (Excel

solution, Aspen solution, average)

Tem perature Profile

0

5

10

15

20

25

30

0 1 2 3 4 5 6 7

Stage (Reboiler = Stage 1)

Temperature (deg C)

Temperature - Excel

Temperature - Aspen

Temperature - Average

Figure 3: Temperature profile

and the liquid composition profiles lie somewhere between the following curves (Excel

solution, Aspen solution, average)

Liquid Composition Profiles

0 1 2 3 4 5 6 7

Stage (Reboiler = Stage 1)

Mole Fraction

CH2Cl2 - Excel

CH2Cl2 - Aspen

CH2Cl2 - Average

CHCl3 - Excel

CHCl3 - Aspen

CHCl3 - Average

CCl4 - Excel

CCl4 - Aspen

CCl4 - Average

Figure 4: Liquid composition profiles