PRACTICAL 3

Determining the optimum reflux ratio

It is needed to design a distillation column (internal: trays) to separate a liquid
mixture, fed at its bubble point, composed of 50% (mole fraction) benzene and of 50%
(mole fraction) toluene. The column operates continuously at atmospheric pressure and
is equipped with a total condenser and a partial reboiler. The overall efficiency of the
column is estimated equal to 50%, while the efficiency of the reboiler can be considered
equal to 1.
The feed flowrate is 50 kmol/h and the column should produce at the top benzene
with 99% molar purity with a recovery fraction of benzene of 0.995.
Problem description
The problem presented here is to design a distillation column considering the
optimal economic solution which involves finding the optimum reflux ratio in order to
minimize the total costs of the process.
Total costs can be subdivided into two subcategories which are investment costs
and operating costs. In order to solve the problem, we look at them independently.
Investment costs
The evaluation of investment costs is based on the dimensions of the equipment
and they can be determined using the expressions by Guthrie provided in the text of the
exercise. In order to determine those, we need to look at the design of the column.

𝐼. 𝐶 𝑜𝑓 𝑐𝑜𝑙𝑢𝑚𝑛 [$] = (𝑀&𝑆/280)(3.18)(101.9)(9.2015) 𝐷𝑇1.066 𝐻𝑇0.802 [1]

𝐼. 𝐶 𝑜𝑓 𝑐𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑟 [$] = (𝑀&𝑆/280)(3.29)(101.3)(7.691 ∙ 10−3 ) (∆𝐻𝑒𝑣 ∙ 𝑉)0.65 [2]
𝐼. 𝐶 𝑜𝑓 𝑟𝑒𝑏𝑜𝑖𝑙𝑒𝑟 [$] = (𝑀&𝑆/280)(3.29)(101.3)(5.7 ∙ 10−3 ) (∆𝐻𝑒𝑣 ∙ 𝑉)0.65 [3]

In the expressions reported above, ∆Hev is an average value equal to 7,740

[kcal/kmol] specified in the text of this exercise. It’s also given the M&S index which is
equal to 1,536.5.
Expression 1 is related to the costs of the column which is function of the diameter
and height of it.
First, in order to calculate the height of the column, the following formula is
applied:
𝐻𝑇 = 0.6 (𝑁𝑃 − 1) + 4.6 [4]
 Then, in order to calculate the diameter of the column, the following steps are
considered:
First, the vapor flow rates values are imported from Aspen Hysys®. Then, those
are used to calculate the area of the column which is determined by considering the
maximum allowable velocity of vapour, to avoid flooding conditions. In this case, for the
column is told to use a velocity equal to 0.3 m/s. The area is calculated as:
𝑉
𝐴 [𝑚2 ] = [5]
𝑣
After, the diameter is determined:

4
𝐷 [𝑚] = √ ∙ 𝐴 [6]
𝜋

It’s also assumed to amortize the investment costs in three years, so each category
is divided by three.
As it can be seen for both the condenser and reboiler, it is just necessary the vapor
flow rate. At the end, the total investment costs are calculated as below:
𝑇𝑜𝑡𝑎𝑙 𝐼. 𝐶 [$/𝑦𝑒𝑎𝑟] = 𝐼. 𝐶 𝑜𝑓 𝑐𝑜𝑙𝑢𝑚𝑛 + 𝐼. 𝐶 𝑜𝑓 𝑐𝑜𝑛𝑑𝑒𝑛𝑠𝑒𝑟 + 𝐼. 𝐶 𝑜𝑓 𝑟𝑒𝑏𝑜𝑖𝑙𝑒𝑟 [7]

Operating costs
The evaluation of the operating costs is related to the costs of cooling water and
vapor at the condenser and reboiler respectively. In order to determine the costs of them,
the formulas are provided as follows:
$
𝐶𝑜𝑠𝑡 𝑜𝑓 𝑐𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 [ ] = 3.54 ∙ 10−7 ∙ (4.186) ∙ (8150) ∙ (∆𝐻𝑒𝑣 ∙ 𝑉) [8]
𝑦𝑒𝑎𝑟

$
𝐶𝑜𝑠𝑡 𝑜𝑓 𝑠𝑡𝑒𝑎𝑚 [ ] = 7.78 ∙ 10−6 ∙ (4.186) ∙ (8150) ∙ (∆𝐻𝑒𝑣 ∙ 𝑉) [9]
𝑦𝑒𝑎𝑟

Steam cost is much higher than the cooling water, so in order to reduce costs we
need to pay attention to it mostly. At the end the total operating costs are determined
adding the utilities costs.
𝑇𝑜𝑡𝑎𝑙 𝑂. 𝐶 [$/𝑦𝑒𝑎𝑟] = 𝑂. 𝐶 𝑐𝑜𝑜𝑙𝑖𝑛𝑔 𝑤𝑎𝑡𝑒𝑟 + 𝑂. 𝐶 𝑠𝑡𝑒𝑎𝑚 [10]

Results
The distillation model presented in this exercise is solved by using four different
approaches. The problem is asked to be solved in Aspen Hysys® using three different
thermodynamic packages. It’s also asked to solve the problem in Excel using the Antoine
equation. Below the results of these four methodologies are shown:

Antoine models comparison

The Antoine model is usually applicable for low pressure systems that behave
ideally. This model provides good results regards to the ones provided by rigorous models
such as SRK y PR. The total costs obtained for each case will result slightly different due
to different facts explained below.
In order to solve the problem in Excel we assume that the vapor and liquid that
leave the distillation column are in completely thermodynamic equilibrium, so it’s
possible to use thermodynamic models to relate the concentrations of the flows in
equilibrium. So, in this way, a theoretic column is designed composed by equilibrium
stages instead of trays as used in the simulation environment, in order to simplify the real
calculations.
In this exercise, it’s considered a liquid mixture of benzene and toluene. Benzene
has a higher volatility than toluene. It’s assumed a relative volatility constant and it can
be calculated using the feed, top and bottom of the column temperatures.
Once the temperatures are calculated the partial pressure is determined by using
the Antoine equation which is a semiempirical correlation where A, B and C are
coefficients whose values are specific for each component. The ones corresponding for
the resolution of this exercise are specified in Table 1.

𝐵
ln(𝑃𝑒𝑣 ) = 𝐴 − [11]
𝑇+𝐶
Table 1. Coefficients of the components for the resolution of equation 11.

Component A B C
Benzene 15.9008 2788.51 -52.36
Toluene 16.0137 3096.52 -53.67

Results obtained with equation 11 will be compared to the results obtained by

simulating the process in Aspen Hysys® choosing as fluid package the one called Antoine
which uses an extension version of equation 11:

𝐵
ln(𝑃𝑒𝑣 ) = 𝐴 + + 𝐷 ∙ ln(𝑇) + 𝐸𝑇 𝐹 [12]
𝑇+𝐶

The additional parameters allow the description of the entire vapor pressure curve.
The coefficients used for this exercise are listed in Table 2.
Table 2. Coefficients of the components for the resolution of equation 12.

Component A B C D E F
Benzene 169.6500 -10314.80 0 -23,5895 2.09E-05 2.00
Toluene 76.4511 -6995.00 0 -9,1635 6.23E-06 2.00
 Then the diameter used in each model is analysed in order to see if it cooperates
to create differences in the total costs’ comparison of the two models. The diameter can
be considered as a function of the vapor flowrate just as the investment costs of the
reboiler and condenser are. Figure 1 shows the results obtained for the vapor flow rate in
both cases, and it can be seen that is basically the same for the two procedures, so the
vapor flowrate is not the one that would influence the differences in the total costs as so
on the diameter. The same criteria can be applied for the calculation of the investment
costs of reboiler and condenser which are just function of the vapor flowrate.

Figure 1. Vapor flow rate considering the different Antoine procedures.

Then, it’s appreciated that in Hysys is used the fractional number of trays while
in Excel the number is rounded to the next entire number. As a consequence, in Excel we
part from a higher number of trays as it is shown in figure 1. This will affect the results
of the investment costs of the column obtaining higher values.

Figure 2. Actual number of trays applying Antoine methodologies.

 Recapitulating, in both procedures what makes the total investment costs to be
slightly different is the height of the column. This is because the height is function of the
number of trays which higher values are achieved by applying Excel procedure. As a
consequence, higher investment costs of the column are achieved selecting the Excel
procedure. Figure 3 shows the investment costs of the column for both procedures.

Figure 3. Investment costs of the column for both procedures.

In addition, the operating costs would not vary from one procedure to another
because as explained before they are function of the vapor flowrate which is basically the
same in each case.
Once the differences are identified, we need to remember that our aim was to
determine the optimum reflux ratio in order to minimize the total costs for a given
separation. We know that for a given separation an increase in the reflux yields to a
decrease in the number of stages. We know that for a given separation high values of the
number of stages correspond to low values of reflux ratio. However, as the number of
stages approaches infinity, a pinch zone develops somewhere in the column, and the
reflux cannot be reduced further. For a binary separation the pinch usually occurs at the
feed stage.
We know that operating costs increase as increasing R because we need more
steam to vaporize. Steam is much more expensive than the cooling water used in the
condenser so as a result the relevant operating costs is the one where steam is used.
By decreasing the reflux ratio separation can be achieved using a higher number
of stages (N) so investment costs are higher.
At the end we obtained the optimum reflux ratio that minimizes the total costs.
The results obtained for Excel resolution are shown in Figure 4:
 Figure 4. Optimum reflux ratio using Excel procedure.

The results obtained for Aspen Hysys® procedure are shown in Figure 5:

Figure 5. Optimum reflux ratio using Aspen Hysys® procedure.

Due to the differences established previously, we know that the minimum costs
are higher when using Excel procedure. This can be verified in Figure 6.
 Figure 6. Total costs of Antoine models.

Rigorous methods comparison: SRK and Peng Robinson.

The exercise was also simulated using two more rigorous VLE computations
which are PR and SRK which are based on equations of state. In this case, SRK model
provides comparable results to Peng-Robinson, but its range of application is more
limited.

First, simulating using the fluid package SRK in Aspen Hysys® was carried out.
The results for finding the optimum are displayed in Figure 7.

Figure 7. SRK total costs and optimum reflux ratio.

After, simulation using the fluid package Peng-Robinson in Aspen Hysys was
carried out. The results for finding the optimum are displayed in Figure 8.
 Figure 8. PR total costs and optimum reflux ratio.

Peng Robinson gives for each value of the reflux ratio a higher value of the actual
number of trays in the distillation column compared to SRK results as it is shown in
Figure 9. It can also be appreciated that as the reflux ratio increases both results become
more similar.

Figure 9. Actual number of trays when using rigorous methods.

So as discussed in Antoine, it would derive in slightly higher investment costs in

the column which are displayed in Figure 10.
 Figure 10. Investment costs of the column when using rigorous methods.

As explained before the vapor flowrate is the same for both procedures so same
operating costs for both cases and same investment costs for the reboiler and condenser
are obtained.

Figure 11. Vapor flow rates when using rigorous methods.

As the reflux ratio increases, it’s needed to vaporize more water, so the operating
costs increases linearly.

The optimum reflux ratio is smaller using as fluid package SRK because
investment costs are slightly slower. As a consequence, lower total costs would be
achieved using this fluid package. This can be appreciated in Figure 12.
 Figure 122. Total costs when using rigorous methods.

Comparison among the procedures.

The results obtained for the different procedures collected in Table 3 confirm that
a column design based on simple methods like Antoine results in very close values from
much more rigorous computations.
Table 3. Results obtained using the different procedures.

Procedure Optimum reflux ratio Minimum Total Costs [$/year]

Antoine Hysys 1.43 309,647
Antoine Excel 1.50 318,407
SRK 1.64 338,956
PR 1.67 342,172