14 replies to this topic

[batchreactor82]

Hi everyone. I'm new of this forum.

I hope someone can help me as soon as possible.

I made an Aspen simulation of a system composed by ethanol, water and hydrocarbons ,such as pentane, hexane, p-xylene and so on. They refers to typical component of gasoline.

My target is to reduce the quantity of this hydrocarbon component because I have to feed the entire strem to a molecular sieve units for dehydrating ethanol. The zeolite bed used as sieve could be affected by the presence of this strange components.

The system is particular because these components take part as ternary or binary azeotropies with ethanol and also with water.

The resulting boiling points are various because sometimes the mixture boiling point is lower than ethanol's one, and sometimes upper.

The simulation is quite easy and I can reach also distillate and bottom composition that I desire.

The problem is linked to the profile of the column that is very far from that one we always observe in this system:

1 97
2 97
3 97.1
4 97.2
5 97.2
6 97.3
7 97.3
8 97.4
9 97.5
10 97.5
11 97.6
12 97.6
13 97.7
14 97.8
15 97.8
16 97.9
17 97.9
18 98
19 98.1
20 98.1
21 98.2
22 98.3
23 98.4
24 98.6
25 98.9
26 99.3
27 99.8
28 100.5
29 101.2
30 101.8
31 102.4
32 102.8
33 103.1
34 102.9
35 102.4
36 102
37 100.4
38 101
39 96.1
40 100.4
41 97.4
42 101.7
43 84.7
44 90.7
45 80.2
46 82.9
47 78.4
48 78.5
49 77.5
50 75.8
51 77
52 74.1
53 76.6
54 73
55 76.4
56 72.3
57 76.2
58 72.8
59 73.5
60 50.6

I'm scared about the discontinuty points in the profile. Someone can explaine how verify if this profile is physically correct?

We alredy build this kind of column but in system without gasoline impuritis, only ethanol and water and other impurities, and we observed that the typical profile is of temperature at the top of 78°c (nearly the B.P. of ethanol) and the profile became with greater temperature from the top to the last 4 stages before the reboiler, in which the temperature i quite close to water B. P.

Sorry for my bad English, I'm an Italian Girl!

Thanks a lot ihf anyone can help me find the right way to understand.

Disposable to give any other details if necessary.

[abhi_agrawa]

Hello,

First of all, have you checked if the thermodynamic method you have used is compatible with the system? You should be extra cautious about this. I can assure you by changing the thermodynamic method you can reach a different solution.

You say that the profile (I assume you mean Temperature profile) is far from what you observe in the system. Assuming that you have accurate and reliable temperature data I'd say that it is very important that you should be able to match the temperature profile.

Assuming that the temperature profile is correct, the profile would indicate that you have more than one feeds to the column and the feed location may not be the optimum.

The temperature difference between stage #59 and #60 is quite odd. The temperature of Stage #60 should be higher than that of #59. Please check.

Unless you give us more information about you simulation like feed composition, feed conditions, tower pressure, recovery specification etc. we can not say anything more.

Hope this help,
-abhishek

[Esteban Lopez]

As it was said maybe the problem is the thermodynamics package.

It can be possible that the feed(s) stage(s) are not the corrects ones.

Also you have to consider if the system Ethanol-Water-impurities has azeotropes and which of them are unstables and stable nodes to analyse the temperatures at the condenser and the reboiler.

And as it was said previously there is an odd change on the temp between stage 59 and 60

If it is possible can you give us info abaout the stream(s) and column to see whta is the problem?

[batchreactor82]

Thanks a lot for your analysis!!!

If I send you the simulation date can You take a look?

I'm not really sure using Aspen because at the beginning I used only Hysys.

During my last evaluation I found that the worst is the temperature in the last 4 stages.

Thanks a lot!

My name is Daniela!

Attached Files

•  Simulation Dates.doc   539.5KB   27 downloads

[batchreactor82]

Thanks a lot!

I send you the same file as above with my specification!

Brest regards.

Daniela

Attached Files

•  Simulation Dates.doc   539.5KB   11 downloads

[batchreactor82]

Hello,

First of all, have you checked if the thermodynamic method you have used is compatible with the system? You should be extra cautious about this. I can assure you by changing the thermodynamic method you can reach a different solution.

You say that the profile (I assume you mean Temperature profile) is far from what you observe in the system. Assuming that you have accurate and reliable temperature data I'd say that it is very important that you should be able to match the temperature profile.

Assuming that the temperature profile is correct, the profile would indicate that you have more than one feeds to the column and the feed location may not be the optimum.

The temperature difference between stage #59 and #60 is quite odd. The temperature of Stage #60 should be higher than that of #59. Please check.

Unless you give us more information about you simulation like feed composition, feed conditions, tower pressure, recovery specification etc. we can not say anything more.

Hope this help,
-abhishek

[batchreactor82]

Hi everyone!

I'm changing the thermodynamic package to try to understand if there's any problem about it!

I usually choose the UNIQUAC method because my system are always related to ethanol and water and I found in literature that it is always used for this system.

I re-load column temperature profile and, according our experience, the profile is good enough to well describe the system but the problem is still between 58° and 59° stage and then between the 59° and the 60°.

At the bottom of the column I should obtaine a stream, that I called Stillage, that must be water with traces of ethanol ( the desiderable value of mass purity should be 0.007, but I obtained o.15) and of the high boiling substances that enter the system as gasoline impurities.

Now my question is : why the profile athe the last stages is in that way?

P-SPEC STAGE 1 PRES, ATM 2.00000

TEMP-EST STAGE 1 TEMP, C 97.0000
2 97.0000
3 97.1000
4 97.2000
5 97.2000
6 97.3000
7 97.3000
8 97.4000
9 97.5000
10 97.5000
11 97.6000
12 97.7000
13 97.7000
14 97.8000
15 97.8000
16 97.9000
17 98.0000
18 98.0000
19 98.1000
20 98.2000
21 98.2000
22 98.3000
23 98.3000
24 98.4000
25 98.5000
26 98.5000
27 98.6000
28 98.7000
29 98.7000
30 98.8000
31 98.9000
32 99.0000
33 99.0000
34 99.1000
35 99.2000
36 99.3000
37 99.4000
38 99.6000
39 99.7000
40 99.9000
41 100.300
42 100.300
43 100.400
44 100.400
45 100.500
46 100.600
47 100.600
48 100.700
49 100.800
50 100.900
51 101.000
52 101.100
53 101.300
54 101.700
55 102.200
56 103.100
57 104.900
58 106.400
59 87.3000
60 90.8000
This one above is my profile


Thanks a lot!!

[batchreactor82]

here is the last version of the file with all column data.

Thanks!  Column Date.doc   886.5KB   10 downloads

[abhi_agrawa]

Daniela,

You have used Murphee Efficiency of 0.001 for stages 1 to 21. This is too low. In actual practice you will definitely get much more efficiency than this. The recommended way to handle tray efficiency is to simulate by using 100% efficiency and then convert theoretical stage to actual stages by using efficiency (Actual stages = theoretical stages / efficiency).

Looking at the components in your system, I'd say that UNIQUAC is a decent choice to begin with. However, a word of caution, try to check the results with any actual data that you have. Remember that UNIQUAC is an activity coefficient based model, and is suitable for modeling liquid phase non-ideality. However UNIQUAC is extremely poor at predicting the vapor phase non-ideality. Do have a look at the vapor phase model.

Still this does not answer the strange temperature profile at stages 59 and 60. If you can post your simulation then perhaps we may be able to investigate a bit more into this. Also please let us know which version of ASPEN you are using.

Best regards,
-abhishek

[batchreactor82]

Thanks a lot for your analysys!!!

Well I can't attach here my simulation!!!

If you give me your e-mail address I'll send you immediately!!!!

I need to understand if there's any conceptual mistake in my simulation. For this reason I joined this forum to ask to people who have much more experience.

my e- mail adress is batchreactor82@hotmail.com!!! if you prefare A private message!

Best regards.

Daniela

[batchreactor82]

For what concernes the efficiency I agree with you, but the value of 0.001 is the only one that gives to the column the possibility to converge... Or almost is the only one I found!

Now I will try another time to find a better efficiency!!!

For what concernes the thermodynamic package, UNIQUAC is the best for Ethanol and Water system, as adviced in Hysys user's guide!

The problem is linked to gasoline impurities !!! I tried all the package but i didn't find any solution! Can you give me an advice for better choosing this one? I have no idea!!

the version of aspen is 2006.20.0.3595.

Thanks a lot for helping me!!!

best regards another time...

[batchreactor82]

Yesterday I tried to change the efficiencies of the column.

I set the efficiency of stages from 1 to 59 to a value of 0.60 while the last one is 1 because of the presence of the reboiler.

Hysys always put 1 to the efficiency of condenser and reboiler of the column.

Waiting for the e-mail adress for sending the simulation!

Best regards!

Daniela

[abhi_agrawa]

Daniela,

I had a look at your problem again and did a small simulation myself. Attached is the .inp file for the same (you will see the extension as .txt, download this file and change the extension to .inp). You should be able to import this file in ASPEN and run the simulation. I have reduced the number of stages to 15, the pressure drop per stage is now 0.1 psi and all the stages have 100% efficiency. This simulation meets your requirement of 0.15 weight fraction in Stillage.

My simulation also uses UNIQUAC method for liquid phase non-ideality with the option to estimate the missing UNIQUAC parameters by UNIFAC and vapor phase non-ideality is modeled by Redlich Kwong equation of state.

Regarding you concern with column efficiency, I usually simulate the column with 100% efficiency and then while specifying the actual number of trays I divide the stages by the efficiency. I believe that this is also the accepted industry practice.

Hope this would sole some of your problems.

-abhishek

Attached Files

•  simu1.txt   9.52KB   10 downloads

Edited by abhi_agrawa, 15 November 2009 - 08:17 PM.

[batchreactor82]

 sim_1.txt   812.87KB   12 downloadsDear Abhishek,

thanks a lot for having checked my simulation. Unfortunately our process is strictly based on that number of stages (58+2)and we can't change this assumption.

If the stages of your simulation have an efficiency of 1, I have to put my efficiency at least of 0.25 to have 60 real stages.

I've just tried to change the binary interactions coefficients but my profile is the same....!!!

My results is to obtain in stream Stillage 0.0001 of ethanol.

Is this possible???

I attach my simulation!!!now I have no hope to resolve my problem!

Thanks a lot !!! Best regards, Daniela.

[zorn81]

Ciao Daniela,

i can't open your sim_1.txt file, even if i changed the extension to .inp, what's the original extension of that file? Alternatively you can try to zip and post the original one.

grazie,
zorn