12 replies to this topic

[hkarimi]

Dear all
I want to simulate a liquid-liquid separator by HYSYS. In this separator a stream includes 60wt% Water and 40wt% Furfural (C4H3O-CHO) enters a vessel and get residence time. The Furfural density is about 1.13 kg/lit and goes down and water that is lighter goes up and two liquid phases are separated.
I simulate this separator in HYSYS by a 3-phase separator. I used “Kabadi Danner” as the fluid pakage but I have a problem! The light and heavy liquid phases leave separator reversed! The organic phase (The Furfural rich phase) leaves the separator as light phase and the watery phase leaves as heavy liquid phase!
What’s my mistake?! Please help me…

[demank]

I simulate this separator in HYSYS by a 3-phase separator.

You will separate 2 phase of liquid then try to simulate with 3-phase separator?

I think you should try another vessel for simulation.
Or HYSYS does not support 2-phase of liquid/liquid separation?

[Technocrat]

Dear hkarimi,

Please design the vessel as a two phase separator.

Which is the dispersed phase, furfural or water? This you can check easily with a formula given by SM Walas in his book 'Chemical Process Equipment' page no. 613. Based on this result we can discuss further.

Regards.

[Padmakar Katre]

I simulate this separator in HYSYS by a 3-phase separator.

I think you should try another vessel for simulation.
Or HYSYS does not support 2-phase of liquid/liquid separation?

Dear,
Please check with the proper refernce before you quote the above statement.

[Padmakar Katre]

Dear all
I want to simulate a liquid-liquid separator by HYSYS. In this separator a stream includes 60wt% Water and 40wt% Furfural (C4H3O-CHO) enters a vessel and get residence time. The Furfural density is about 1.13 kg/lit and goes down and water that is lighter goes up and two liquid phases are separated.
I simulate this separator in HYSYS by a 3-phase separator. I used “Kabadi Danner” as the fluid pakage but I have a problem! The light and heavy liquid phases leave separator reversed! The organic phase (The Furfural rich phase) leaves the separator as light phase and the watery phase leaves as heavy liquid phase!
What’s my mistake?! Please help me…

Dear,
You are modeling the right way.Just want to ask why yor using KD Fluid Package why not any activity coeffiecient model like Uniquac, NRTL. The concern is to check the BIPs for these 2 components in KD fluid package and the densities given by the KD fluid package. Waiting for your comments.

[joerd]

The Hysys Simulation Basis Manual says:

The Peng-Robinson and SRK property packages will always force the water rich phase into the heavy liquid phase of a three phase stream. As such, the aqueous phase is always forced out of the bottom of a three phase separator, even if a light liquid phase (hydrocarbon rich) does not exist. Solids are always carried in the second liquid phase.

It looks like this applies to more than just PR and SRK. Interesting; good observation.

[hkarimi]

Dear Friends,
Thanks for your attention and excuse my long delay in reply to your messages. In the previous days I tried to get more information about the Furfural and its usages.
Furfural is a common solvent that used to extract the aromatic components from the feed of the lubricating oil production. In the solvent recovery steps of this process the mixture of Furfural and Water is the overhead product of stripper towers.
Furfural is partial immiscible in Water. Also, its density is about 1.13 kg/lit and is heavier than water. Thus, based on these properties a liquid-liquid separator vessel is used to separate water from furfural. The attachment here is a schematic picture that indicates the vessel and its input and outputs.
This separator consists of a horizontal tank divided into two compartments by a solid baffle. The feed stream enters the first compartment, pass through a calming baffle (perforated baffle) and settle into two phases.
The heavy phase containing about 93wt% Furfural and 7wt% water is trapped in front of the baffle. The lighter phase containing about 10wt% Furfural and 90wt% water flows over the baffle into the second compartment.
I simulated this separator in HYSYS with a 3-phase separator and the concentration of light and heavy liquid phases is approximately correct.
But, so that I told you before, the light and heavy phases leave the separator reversely! The aqueous phase (Water Rich Phase) leaves the separator as Heavy phase and the heavy phase (Furfural Rich Phase) leaves the separator from top position!
I tested other fluid packages such as UNIQUAC or NRTL but this problem don’t solve! Two liquid produced phases leave the separator reversely even when I use other fluid packages.
Sincerely Yours,
Hossein Karimi

Edited by hkarimi, 15 December 2009 - 04:40 PM.

[Technocrat]

Dear hkarimi,

Please give us all data like: Flow rate, density, viscosity for both the light and heavy phases at operating conditions, interfacial tension between heavy & light liquid to judge the ease or difficulty of separation and operating temperature and pressure. If you know the droplet size distribution in feed it is better.

Based on the above data I will design a separator for you.

Regards.

[hkarimi]

Dear Technocrat,
I'm sorry but I could only get this information.
Furfural rich phase (90wt %) flow rate is 14 cubic meters per hour.
Water rich phase (7wt %) flow rate is 4 Cubic meters per hour.
Operating temperature is 90 degrees centigrade.
Operating Pressure is 0.02 barg.
I think this information isn’t enough for you.
However, many thanks for your attention...
All the best.

[duaneloh]

Hi,

Did we have an answer for the issue described above? I am facing the same problem with separating out Furfural and Water out using the Separator in Hysys. Thanks for your help

Regards,
Duane

[Technical Bard]

HYSYS doesn't separate the streams based on density - it separates them based on aqueous and non-aqueous. If your non-aqueous stream is heavier than water, then the separator block in HYSYS gives you the results backwards. But it's ok because as long as you know which is which you can connect the streams appropriately.

[chemdoc]

may I suggest a different approach ?
A true multiphase flash and a excel page for simulating the liquid liquid separator.
Your simulator should be able to solve this phase equilibria, if not you can consider a suitable process library as Prode Properties , once you have solved the phase equilibria and know composition and properties of different phases it should be easy in Excel to size the separator,
I use the formulations given in "Gas liquid and liquid-liquid separators" by Steward and Arnold and results are reasonably accurate.
I wonder why many engineers do not use Excel for these tasks, well, it's ok to use a simulator for converging streams but then tools as Excel or Matlab are much more flexible for equipment sizing.

[duaneloh]

Thanks for the input.