Horizontal Knockout Drums with very high Liquid Flowrate

[Rospen]

Hi, after reading a few threads about Knouckout vessel sizing here on the Forum, I decided to start my own thread about a specific situation I'm facing right now.

I must design a horizontal gas-liquid separator, separating water from a gaseous oxygen stream (4 barg and 70 C).

But for my application specifically, the flow rate of the liquid in relation to the vapor is "incompatible". With too High Liquid Flowrate and Low Vapor Flowrate.
While the diameters suggested to guarantee the separation of droplets in the gas phase are reasonable and achievable, when using minimum residence time for the liquid phase as a reference, the vessel becomes unreachable (lenght becomes too big, even increasing the diameter to resonable values).

I understand that the residence time of the liquid phase is related to the safety of controlling the vessel that is being sized, but at the moment it is unthinkable to apply values of 10, 5 and even 2 minutes as residence time...

Hence the question, I would like to know if you have already calculated a minimum liquid phase residence time for effective separation of the bubbles or gas within the liquid phase.

In some other threads, I saw people discussing the calculation of the minimum residence time of the liquid phase for separating bubbles of a certain diameter from the liquid phase. Has anyone of you ever faced a similar problem?

 

[LittleInch]

Welcome to E-T.

One of the issues is often which forum to post as only one is allowed.

I think you need to copy and paste this into "Chemical plant design" and delete this one as it has very little to do with heat transfer...
And while you do that please add some data in terms of flow rates and what the pressures are upstream and downstream.

Also I'm not sure if your aim is to de water a gas stream or de gas a liquid stream?

There is often a degree of compromise in these situations as you can't have perfect separation within a budget.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[3DDave]

You can try clicking on the Report button for your post and ask to be relocated by the powers that be.

 

[3DDave]

Would a cyclonic separator be a better device?

 

[Rospen]

I am thankful for your answers.

I will try to copy the thread to "Chemical Plant Design"

Little Inch, the system of this GLS is a little more different than normal I may say.
Considering that I yet need to apply a Heat Exchanger and Coalescing Filter to the vapor stream outlet after this GLS (to remove the water linked with the saturation of the oxygen stream), I believe I can answer that I am a little more worried about sending bubbles on the liquid stream than a little bit of water to the vapor stream.
But yet, as I said on my first message, the minimum diameter for the gas constrain (respecting maximum velocity for droplets of a certain diameter) is really easy to accomplish, the problem is when I consider the residence time of certain references.


Thank you 3DDave, I will try to do your approach first.

Regarding the cyclonic separator, because of the high flow of liquid, I always considered Horizontal Gas liquid Separators, at least at first sight. But I do can study a little more about cyclonic separators, if they are good on handling high liquid flowrates.

But considering my first question, someone ever heard about a minimum residence time for the liquid phase, and not for effective control of the vessel, as always suggested?

 

[LittleInch]

Rospen,

Your terminology needs to be better.

A "knock out drum" is commonly envisaged to be a low pressure drum where mainly gas has some large liquid droplets that need to be "knocked out" of the liquid stream.

This is often a vertical drum.

You seem to be talking about a gas liquid separator or even a degassing drum.

Whatever you do, there will still be considerable gas dissolved / entrained in the liquid at 4bar pressure. This will come out of the liquid once you drop the liquid pressure further.

Residence time is one of those variables that you need to play with to come up with a practical vessel size.

You could always build another one next to it and split flow, or connect a second PV with the liquid phase connected or just drop the pressure again for the liquid phase in a secondary separator.

There are many options to consider but you're only giving us one little bit of the puzzle.


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Rospen]

LittleInch,

Regarding the vertical separator, the equipment is not suitable for the solution right now (size limitations).

That's why I am searching for a horizontal gas-liquid separator. But again, as I said before, the cyclonic separation could be an option, even as an inlet device for the GLS.

Regarding the gas dissolved, this point is going to be my consideration for 100% efficiency (I am going to loose gases in this form), and because of the system itself, the pressure will not be dropped after the separator.

So, as I said, I am wondering about an efficient separation of oxygen and water (70 °C and 4 barg) with a high flow of water when related to the vapor, and for a solution based on a horizontal gas-liquid separator, the minimum diameter is easily reachable, but the length of the vessel, considering the most common residence time of different references, is completely out of range. Even if I play with the vessel diameter, for this application specifically, a residence time for the water of 1, 2, 5, or 10 minutes is unthinkable.

My concern right now is gas at the outlet of the liquid in the form of bubbles and so on.

Again, is there any methodology to estimate the minimum residence time for the bubbles to rise? Even if it's going to be really small, I would like to have this value and work knowing this limitation of the process.

Thanks for the answering.

 

[georgeverghese]

Can you tell us what criterion you used for setting LALL, LAL, LAH and LAHH on this KOD?
Is this vessel continuously monitored at control room ?

 

[FMJalink]

Hi Rospen,
In order to understand your separation problem, somewhat more information seems required on the incoming stream, e.g. flowrate, ratio between gas (oxygen) & water and the phase of the mixture water & gas (e.g. fog, slugs, etc.)

 

[Rospen]

georgeverghese and FMJalink, Thanks for answering.

Regarding the criteria for LALL, LAL, LAH, and LAHH for the KOD, or GLS as LittleInch defined, I am following the next guide:

"Design two-phase separators within the right limits.".

I am also using as main references the following documents (I will not attach the documents to this message):

Stewart, Maurice, and Ken Arnold. Gas-liquid and Liquid-liquid Separators. Gulf Professional Publishing, 2008.
Shoeibi Omrani, P., and A. Eijk. "EFRC guidelines on how to avoid liquid problems in reciprocating compressor systems." (2014).
Specification, A. P. I. "12J." Specification for Oil and Gas Separators (2009).
Couper, James R. Chemical process equipment: selection and design. Gulf professional publishing, 2005.
Towler, Gavin, and Ray Sinnott. Chemical engineering design: principles, practice and economics of plant and process design. Butterworth-Heinemann, 2021.
Lucia, A., 2008. Chemical engineering design principles, practice, and economics of plant and process design by G. Towler and R. Sinnott.Kirveskari, Lauri.
"Design of horizontal phase-separators using computational fluid dynamics." Master's thesis, 2016.

All the methods above are based on settling or Souders-Brown velocity.

FMJalink, I am sorry for not sharing all the information; I actually asked open-ended questions to get different ideas than mine.

Here is the flowrate of water and oxygen in the inlet line:

Water (4 barg, 70 °C): 162 m³/h
Oxygen (4 barg, 70 C): 57,60 m³/h

Regarding the formation of slugs, fog, and so on, this is a specific topic that I am searching for. If you have some documents or calculations concerning this phenomenon, that would help me a lot.

Any documents or experience regarding different configuration, like multiple inlet (more than one inlet along the vessel), split entry (usually with two inlets, one in each side of the vessel, and the outlets in the center) or split exit (inlet in the center of the vessel and outlets at the sides of the vessel) is completely welcome and will help me a lot too.

I know my answers and communications are not so objective; I just want to hear different ideas or opinions from someone more experienced on this topic.

Again, thanks for the answers, I hope I made the problem a little more clear.

 

[LittleInch]

Rospen,

I'm no expert in this, but the key for getting gas bubble out of liquids is residence time and surface area AFAIK.

So keeping the vessel half full would be a good idea.

162m3/hr is 2.7 m3/min. That's not a lot and I would expect any sort of decent sized vessel to hold at least 4-5 mins residence time, but you are saying even 1 minute is unfeasible. That's a very small separator if it can't handle 2.7m3.
Also nothing is 100% in these cases unless you have infinite time.
Bubble size is also critical so not getting the liquid very agitated is important.

What is the pressure upstream the separator? You say there is no lower pressure system so what happens to the water? Does it go back into the start point again?

Some vendors probably have more experience, but you could also look at tilting plate - works for dissimilar liquids but might do the same for de gassing a liquid.

A different type of separator might be much more economic for you, e.g. cyclonic.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[georgeverghese]

Typical residence time criteria for liquid side, where near complete degassing is not required:
a)LALL : 6-8inches from vessel bottom, and at least 2-3inches above top of any vortex breaker
b)LALL to LAL : 1minute
c)LAL to LAH : 2minutes standard
d)LAH to LAHH : 1minute

Height at LAHH should be no more than say 0.7x vessel dia, and keep a clearance of at least 4inches to the bottom of any gas side demisting device.
Split entry / other variants is only done when gas side is controlling vessel dimensions.

 

[FMJalink]

Ìt is a nice puzzle you have. With horizontal separation vessels you want a nice stratified laminar flow in the gas phase to allow droplets to fall in the water below and a nice stratified laminar flow in the liquid phase to allow gas bubbles to raise to the surface. I did not yet do any calculations, but your composition seems to ask for splitting by means of a vertical separator into a gas phase with droplets which can enter the horizontal separator on the upper side in the gas phase and a liquid phase with gas bubbles which can enter the horizontal separator on the lower side in the liquid phase. Assuming use of economical inlet pipe diameter, having this initial separation in the horizontal separator will ask a substantial part of the vessel and will make the liquid quite turbulent, maybe even requiring a stratifying section.

 

[Rospen]

LittleInch, georgeverghese and FMJalink,

I am thankful for all your answers.

For this problem, for example, it may be a little bit strange and go completely against what everybody is saying here on this thread.

We work here with a vessel with a residence time for the liquid much smaller than this type of consideration; something around 10 s of liquid residence time works pretty well, and I am talking about the same conditions and flowrate that I shared previously:

Water (4 barg, 70 °C): 162 m³/h.
Oxygen (4 barg, 70 C): 57,60 m³/h.

Liquid residence time is not a primary parameter for a separation to occur, at least in my understanding of the reality that I face right here.

The holdup and surge time that georgeverghese shared with us on this thread (1 and 2 minutes) are completely based on control aspects of the level interface and empirical assumptions, as some references expose. And that's why I am searching for a way to calculate this parameter.

Nowadays, we are about to face a problem regarding the size of the whole gas-liquid separator; a vertical one with the dimensions to separate, for example, 5 or more times the flowrate that I am sharing, is not going to be possible. As FMJalink shared, there is the possibility of multiple separator applications; any works or references are welcome. It's going to help me a lot.

As I already shared, an option is to play with the streams and the inlet and outlet, like considering multiple inlet and outlet. With this approach, there is the possibility to accomplish the separation with a higher value of slenderness (L/D).

Answering LittleInch, the water is going to be back into the starting point again; it's reflux water. Also, tilting plates could be a possibility too; with a quick research, I understand I could use it even as an internal device. Is that correct? A cyclonic separator works well with high flow rates of water over gas?