3 replies to this topic

[JakeVincent]

Dear all,

 

I am trying to operate a distillation column in Hysys to separate a feed stream with composition (in kmol/hr) as follows:

 

EthylBenzene              158.63

Styrene                        1.77

Hydrogen                     3.413e-004

Benzene                      7.611

Ethylene                      1.415e-003

Toluene                       9.392

Methane                      3.546e-004

H₂O                             8.026e-002

CO                               5.153e-006

CO₂                              1.577e-002

 

Feed T/P: 41^oC/127 KPa

Condenser Pressure: 120 KPa

Reboiler Pressure: 140 KPa

 

The LK for this column is Toluene, and the HK is Ethylbenzene. Column is made to converge to a top leak of 0.01 Ethylbenzene and a bottom leak of 0.01 Toluene. Upon running the case, a Hysys error pops out saying "Two liquid phases were found in column T-1 on stage Condenser".

 

I was told that this is due to "sizeable" amounts of H₂O being present in the feed stream, but at a mole fraction of 0.0005, I don't see how this is going to hinder Hysys operations or real life operations for that matter.

 

To go about this problem, I found 2 possible way to make this column converge, each with respective things that I am unsure about.

 

1. Turn on the "Auto water draw" function of the distillation tower. When this function is turned on, I am able to add a "water draw" stream in the column environment. However, I am unable to control the amount of water to be drawn (it is done by Hysys automatically). I am also unable to understand how does a "water draw" work industrially, or is there even such a function of reflux drums in the first place? My Googling thus far has not turned up any reasonable answers.

 

2. Add a vent to the distillation tower. This vent will, presumably, vent out any excess water that will cause the formation of 2 phases in the reflux stream (assuming that water is very lowly soluble in oil since I am using Peng Robinson package). In operating this partial condenser, I cannot figure out on a temperature that I should set for the condenser, and make the column converge to that temperature. How do I determine the optimum temperature for this partial condenser such that I only condense any excess water that will cause phase separation?

[I am assuming that the less I vent, the better because I do not want to be venting away my valuable raw materials. Also, the short cut column gave me a condenser temperature of ~78^oC.]

 

 

Any help and advice regarding my simulation will be greatly appreciated. Thank you!

 

Best regards,

Jake

[PingPong]

However, I am unable to control the amount of water to be drawn (it is done by Hysys automatically). I am also unable to understand how does a "water draw" work industrially, or is there even such a function of reflux drums in the first place?

Don't worry, it is normal that a reflux drum has a boot to drain water if there is a chance that column feed can contain water. Control is by interface level control.

You cannot control the water drawoff rate in the simulation simply because exactly all the water from the feed will need to be drawn. It is simply dictated by the water balance over the column.

 

Continue with Option 1. Forget Option 2.

[JakeVincent]

Hi PingPong,

 

Following your suggestion, I went to read up on Hysys operation manual to understand the workings of the Auto Water Draw (AWD) function. When Hysys detects the presence of a 3rd phase at the top stage, it will check the next stage (and so on) until it finds a stage with only 2 phases. From that stage upwards, it places AWDs to remove the 3rd phase from the column. Comparing with real life mechanism in a reflux drum, this seems to be somewhat different and I am wondering how closely does this emulate what actually happens, and how I can justify my using of AWDs.

 

My professor recommended using a vent to achieve column conversion. When I tried this the column converged and I am venting a small amount of ~3.4 kg/s (half of which is water, the other half is benzene, no raw material/product is dumped) at the top. Is this a more justifiable approach to the problem, or is it not worth to construct a vent just to throw this small amount of water out?

 

What I am concerned with is that, both method gives me column convergence but that isn't the main point of my studies. I am hoping to simulate a plant that can emulate as closely as possible, what happens in real life. Hence, I seek to find a justification for whichever approach that I take. How do people in the industry who uses Hysys approach this problem?

 

Thank you for your help!

 

Best regards,

Jake

[PingPong]

I don't use Hysys so I cannot comment on what that software does.

 

I will repeat what I said before: it is normal that a reflux drum has a boot to drain water if there is a chance that the column feed can contain water.

Hundreds of all kinds of columns, all over the world, have that.

 

Nobody creates a hydrocarbon vent gas with only purpose to get rid of water, as your professor is suggesting. He clearly does not know what he is talking about. Forget option 2.

 

The amount of water in the feed (1.5 kg/h) is so small that you could simply leave it out of the simulation, as it does not influence anything. That is how we did it in the good old days when simulators did not yet include dissolved water in hydrocarbons coming from a 2 or 3 phase separator. We knew anyway that any water in feed would end up in the reflux drum, so we designed it with a water boot. Nowadays simulators do include dissolved water and all you get is: convergence problems. And ignorent professors. That's progress....