13 replies to this topic

Dear All,

I just want to know that for horizontal vessel/ KOD sizing. I have to follow the Velocity criterion ( drop out velocity should be higher than vapor horizontal velocity ) or Residence time criterion( Drop out residence time should be lower than vapor residence time) or both the criterion to be followed.

[Zauberberg]

Not sure if I understood your question well. Normally we employ horizontal separators in those cases when handling of large liquid flows and fairly small gas/vapor flows is required.

Liquid residence time is required for accurate level control, allowing sufficient time for operator to respond. In case of 3-phase separator, you need to maintain certain velocity that is required for proper liquid-liquid phase separation, and that depends on their properties.

Here is one standard available at: http://www.mop.ir/st...pr/e-pr-880.pdf

[S.AHMAD]

Residence time is required for the liquid phase to ensure sufficient time for good separation between gas and liquid. The vapor velocity is required to ensure that no liquid entrainment in the vapor phase. This means that both vapor velocity and residence time are required for sizing disengagement vessel.

Residence time is required for the liquid phase to ensure sufficient time for good separation between gas and liquid. The vapor velocity is required to ensure that no liquid entrainment in the vapor phase. This means that both vapor velocity and residence time are required for sizing disengagement vessel.

Thanks for your reply.......but if i provide the sufficient time for liquid droplet to settle down then why should i keep the velocity of vapor below the liquid dropout velocity. liquid droplet settling velocity depends on density difference of liquid & gas phase , Gravitational constant & physical parameters. if velocity of vapor is high than velocity of dropout then the liquid will settle down or not. this is related to flare KOD sizing.

[daryon]

Hi, they are both important, let me try to explain;

The drop out velocity is the gas or vapour velocity that must be achieved to allow liquid droplets of a certain size to settle out from the gas stream. Above this velocity and liquid droplets will be carried with the gas stream and will not disengage.

When you have slowed the gas stream down sufficiently to achieve your drop velocity all liquid droplets above a certain size will start to fall out of the gas stream, however they won't instantly fall out. You have to provide sufficent residence time to allow the falling drops to reach the liquid level in the vessel before the gas exits the vessel.

Imagine a horizontal FKOD with a diameter of 2.5 m and a liquid level of 1.0 m. The liquid droplets entering near the top of the vessel must fall 1.5 m to the liquid level by the time the gas has past through the FKOD or they won't have had time to settle out. The required residence time is determined by the vertical height the liquid droplet must fall divided by the drop out velocity (i.e. the velocity at which the gas is moving). So if you're drop velocity is 0.7 m/s the required residence time is 2.1 seconds.

Hi, they are both important, let me try to explain;

The drop out velocity is the gas or vapour velocity that must be achieved to allow liquid droplets of a certain size to settle out from the gas stream. Above this velocity and liquid droplets will be carried with the gas stream and will not disengage.

When you have slowed the gas stream down sufficiently to achieve your drop velocity all liquid droplets above a certain size will start to fall out of the gas stream, however they won't instantly fall out. You have to provide sufficent residence time to allow the falling drops to reach the liquid level in the vessel before the gas exits the vessel.

Imagine a horizontal FKOD with a diameter of 2.5 m and a liquid level of 1.0 m. The liquid droplets entering near the top of the vessel must fall 1.5 m to the liquid level by the time the gas has past through the FKOD or they won't have had time to settle out. The required residence time is determined by the vertical height the liquid droplet must fall divided by the drop out velocity (i.e. the velocity at which the gas is moving). So if you're drop velocity is 0.7 m/s the required residence time is 2.1 seconds.

Thanks for your reply........ Please find the attachment of FKOD sizing as per API.

In This example dropout velocity is 0.71 m/s & and in results ( last page) it is showing 4 possible dimensions of the vessel. in that for all cases the vapor velocity is much higher than liquid dropout velocity. it is providing only residence time to liquid dropout to settle down. and in conclusion it stated that " all of the drum sizes above fulfilling the design requirement".

and dropout velocity is the liquid velocity not the vapor velocity....liquid residence time is calculated based on liquid dropout velocity not the vapor velocity.

Attached Files

•  FkOD sizing as per API.pdf   103.88KB   627 downloads

[daryon]

Hi,
Yes your are correct, my apoligies for the poor/wrong explanation, your vapour velocity can be above the liquid drop out velocity provided there is sufficient time for the droplet to fall. If the vapour velocity = the liquid drop out velocity the drop will fall stright down and no residence is required.

Hi,
Yes your are correct, my apoligies for the poor/wrong explanation, your vapour velocity can be above the liquid drop out velocity provided there is sufficient time for the droplet to fall. If the vapour velocity = the liquid drop out velocity the drop will fall stright down and no residence is required.

Thanks

but i have 1 more query related to that............if only residence time is required to settle down then why do we do split entry or exit.......due to split entry or exit only vapor velocity is reduced. residence time is same for both the cases. then wat is the significance of split entry or exit?

[Art Montemayor]

The subject of liquid-vapor separators – whether vertical or horizontal – has been discussed and kicked around our Forums for many years. The subjects of maximum allowable vapor velocity and “Residence Time” (as it relates to the separated liquid) have also been discussed in the same threads.

I have designed, built, and operated many separators during my career – both vertical and horizontal – and I have never come across the terms being loosely used in this thread, such as “drop out velocity”, “Drop out residence time”, “split entry or exit”. None of these terms are defined here, so I don’t know what is meant by them. What I do know regarding knock-out drums (KODs) is the following:

• The equation universally used is the Brown-Souders equation; this equation is clearly dependent on only three (3) factors: the liquid density, the vapor density, and an EMPIRICAL constant.
• The best articles and explanations on liquid-vapor separators are written by Milton Beychok and found at:
• http://en.wikipedia.org/wiki/Souders-Brown_equation
• http://en.citizendium.org/wiki/Souders-Brown_equation
• Residence time is NOT required for the liquid phase to ensure sufficient time for good separation between gas and liquid.
• You DO NOT have to provide sufficient residence time to allow the falling drops to reach the liquid level in the vessel before the gas exits the vessel.
• You do require a reasonable and common sense distance from the feed mixture inlet and the top of the internal liquid level; this is to avoid further entrainment and turbulence and is an experience factor.
• You can vary your liquid level within the separator within reasonable, common sense limits. You do not require a liquid “residence time” in order to achieve separation. As Zauberberg clearly points out: “Liquid residence time is required for accurate level control, allowing sufficient time for operator to respond.”

The reason for the accurate level control is to ensure that you maintain a positive seal between the two, internal phases --- i.e., you don’t allow the gas phase to exit through the bottom liquid outlet.

I highly recommend everyone to read both of Milton Beychok’s great articles on this subject.

[daryon]

Apologies for the lack of clarity and definition in this thread.

My understanding is that we are talking specifically about the design of Horizontal Flare Knock Out Drum (FKOD) using the methodology presented in API Standard 521.

The dropout velocity (U_C) is referring to the velocity at which a liquid droplet will drop out of the vapor stream and is expressed in meters per second. It is a function of the differential density between the gas and liquid, the droplet diameter, and the drag coefficient.

The residence time is referring to the residence time of the gas phase only, for the discussion in this thread the liquid residence time is of no concern, and agreed it has no impact on separation between gas and liquid.

What we are trying to ensure is that when a large liquid relief occurs into the flare system, the FKOD has sufficient capacity to accommodate the liquid and still provide sufficient gas volume to knock out liquid droplets above a certain size. There is no demister and this has to be achieved by gravity separation. The API recommendation is that the FKOD should be of sufficient capacity to accommodate a liquid relief for 20 to 30 mins, and that the vapor space shall still be sufficient at the maximum liquid level in the FKOD to remove droplets larger than 300 to 600 microns.

If the area available for vapor flow when the FKOD is at the maximum liquid level is less than the dropout velocity then residence time of the vapor is of no concern as all droplets above the specified diameter will fall. However if the vapor space at the maximum liquid level is insufficient to slow the vapour down to less than the dropout velocity then residence time is applicable to avoid potential entrainment of liquid particles in the vapor stream. API Std. 521 says "residence time of the vapor must be equal to or greater than the time required to travel the available vertical height at the dropout velocity of the liquid particles".

Going back to Lokesh question a split entry or spilt exit can reduce the required diameter of the FKOD for high vapor flowrates because it reduces the vapor velocity. Vapor velocities above a certain value may be unacceptable as the risk of entrainment of liquid droplets from the liquid / gas interface increases with vapor velocity. If you have a requirement to keep your vapor velocity below say 3 m/s to avoid liquid re-entrainment from the liquid surface using a spilt entry or spilt exit design can reduce the vapor velocity without requiring an excessively large diameter.

Apologies for the lack of clarity and definition in this thread.

My understanding is that we are talking specifically about the design of Horizontal Flare Knock Out Drum (FKOD) using the methodology presented in API Standard 521.

The dropout velocity (U_C) is referring to the velocity at which a liquid droplet will drop out of the vapor stream and is expressed in meters per second. It is a function of the differential density between the gas and liquid, the droplet diameter, and the drag coefficient.

The residence time is referring to the residence time of the gas phase only, for the discussion in this thread the liquid residence time is of no concern, and agreed it has no impact on separation between gas and liquid.

What we are trying to ensure is that when a large liquid relief occurs into the flare system, the FKOD has sufficient capacity to accommodate the liquid and still provide sufficient gas volume to knock out liquid droplets above a certain size. There is no demister and this has to be achieved by gravity separation. The API recommendation is that the FKOD should be of sufficient capacity to accommodate a liquid relief for 20 to 30 mins, and that the vapor space shall still be sufficient at the maximum liquid level in the FKOD to remove droplets larger than 300 to 600 microns.

If the area available for vapor flow when the FKOD is at the maximum liquid level is less than the dropout velocity then residence time of the vapor is of no concern as all droplets above the specified diameter will fall. However if the vapor space at the maximum liquid level is insufficient to slow the vapour down to less than the dropout velocity then residence time is applicable to avoid potential entrainment of liquid particles in the vapor stream. API Std. 521 says "residence time of the vapor must be equal to or greater than the time required to travel the available vertical height at the dropout velocity of the liquid particles".

Going back to Lokesh question a split entry or spilt exit can reduce the required diameter of the FKOD for high vapor flowrates because it reduces the vapor velocity. Vapor velocities above a certain value may be unacceptable as the risk of entrainment of liquid droplets from the liquid / gas interface increases with vapor velocity. If you have a requirement to keep your vapor velocity below say 3 m/s to avoid liquid re-entrainment from the liquid surface using a spilt entry or spilt exit design can reduce the vapor velocity without requiring an excessively large diameter.

Thanks for ur valuable reply........if my dropout velocity is .7 m/s. then upto what vapor velocity i can design the vessel.
Is there any reference available for that? because i have no requirement of vapor velocity limit. thats why upto what size i can optimize the vessel diameter.

[fallah]

If you have a requirement to keep your vapor velocity below say 3 m/s to avoid liquid re-entrainment from the liquid surface using a spilt entry or spilt exit design can reduce the vapor velocity without requiring an excessively large diameter.

As per API 521 "A split entry or exit configuration can be used to reduce diameter (but increase the length) for large flowrates and should be considered if the vessel diameter exceeds 12 ft (3.66 m)".

Obviously,the length has to be increased due to not increasing the diameter and also for keeping the residence time of the vapor in proper level.

[daryon]

Thanks for ur valuable reply........if my dropout velocity is .7 m/s. then upto what vapor velocity i can design the vessel.
Is there any reference available for that? because i have no requirement of vapor velocity limit. thats why upto what size i can optimize the vessel diameter.

I don't know any standard or guideline that specifies an upper limit on the vapor velocity but I would not go too high. Petronas technical standard (which is the same as the Shell DEP) advises that the Maximum Liquid Level shall remain below the horizontal center line of the vessel. In the past I have found this a very stringent requirment. API and NORSOK do not have this requirement.

I've looked through some previous projects FKODs, these are for FPSOs:
(1) HP FKOD - Dropout velocity - 1.15 m/s; droplet diameter above which will be knocked out 450 microns, vapor velocity at maximum liquid level was 2 m/s
(2) HP FKOD - Dropout velocity - 0.67 m/s; droplet diameter above which will be knocked out 400 microns, vapor velocity at maximum liquid level was 1.32 m/s

(1) LP FKOD - Dropout velocity - 1.04 m/s; droplet diameter above which will be knocked out 300 microns, vapor velocity at maximum liquid level was 7.8 m/s
(2) LP FKOD - Dropout velocity - 1.7 m/s; droplet diameter above which will be knocked out 400 microns, vapor velocity at maximum liquid level was 9.3 m/s

Please note... i'm not saying these are the right answers but these are figures I have used in the past. I have specified the maximum liquid level upto 60% of vessel diameter. This is just advise and please don't take any of this as gospel. I suggest you talk to your client and see if they are happy with the basis of your sizing calculation.

[daryon]

On further digging round in design standards PTS say vapour velocity should be no higher that 1.7 x dropout velocity for horizontal drums. Another design guidline says it should be a little lower at 1.5 x dropout velocity.

Please note; The vapour velocity I presented in the earlier post for the LP FKOD were very high, but these were calculated using the maximum liquid level and the maximum vapour relief rate which in fact could not occur co-incidentally. The actual vapour velocity based on the worst case vapour relief whilst the LP FKOD could be at maximum liquid level was much lower.