23 replies to this topic

[R NESAMANI]

I am in one pumping station of 1300 km long, LPG pipeline system.

We have flare line connected with the Cold flare & hot flare. But we dont have the Knock-out drum in the flare line.
But i heard that knockout drum is used in flare line just before the Flare chimney.
What is the purpose of knock-out drum...?
What is the standard, which mentioned the use of knockout drum in flare line...?

As per your views, we will install the knockout drum in the flare line.

[fallah]

Hi,

Purpose of knock out drum in flare line is to separate liquid droplets larger than 300-600 micro meter in diameter to avoid entrance of such droplets into flare stack/tip.

In LPG flare line the knock out drum is usually being considered.

For more detail please refer to API STD 521.

Fallah

Edited by fallah, 06 October 2011 - 02:02 AM.

[R NESAMANI]

Thank you Mr.Fallah.

But LPG will evaporate in atm pressure. So liq. LPG comes out in flare tip as LPG Gas.
Am i right...? then what is the effect of liquid droplets larger than 300-600 micro meter in flare line. Anyway it will come out of flare tip in gas state.

Do you know, any plant using knockout drum for LPG flaring application....?

[fallah]

Dear NESAMANI,

You are generally right...... the LPG will evaporate in atmospheric pressure (of course in temperature not so low)..... but liquid LPG from relif point up to stack hasn't enough time to absorb needed latent heat to become vapor completely. Even in most cases it is needed to install electrical (or steam) heater in knock our drum to assist LPG evaporation in order to avoid LPG liquid carry over into the stack!

Entrance of droplets larger than 300-600 micro meter into the flare tip can result in incomplete combustion along with excessive smoking that will lead to burning rain and flame out.

Yes, in one plant i was involved in its design knock out drum has been considered for LPG flaring.

Fallah

Edited by fallah, 06 October 2011 - 03:10 AM.

[kkala]

In 1984 we designed an horizontal LPG vaporizer (heating medium: steam). LPG was directed to flare in some cases and a knock out drum was foreseen. But I cannot remember the reason. Probably we were afraid of probable cooling of LPG gas in the flare header, or liquid droplet entrainment along with LPG gas from vaporizer.
LPG can contain quite small amounts of pentane (b.p. of n-C5 = 36 oC), or even traces of heavier hydrocarbons. Probably you have detected condensates in hot or (especially) in cold flare line. All these indicate a good reason to install knock out drum upstream of flare. But have you observed condensate or do you estimate liquid formation under certain (not normal) conditions?

Edited by kkala, 06 October 2011 - 02:24 PM.

[R NESAMANI]

Thank You for your reply.

--> Mr Fallah --> My observation is LPG burns in black, incomplete combustion. Also Liquid falls near the flare stack also occurs. Due to this incomplete combustion, CO2 emission will also be more, i suppose.


Hi Kkala, i have seen water condensate (even ice formation) on the flare line. The reason is the LPG expansion. am i right? does this have any bad effect (or) its normal...?

Knockout drum will be installed just before the flare stack. So how it will prevent the condensate formation...?

Can you give me any link or file, showing the process of knockout drum..?

[kkala]

Goggling "flare knock out drum" will result in a lot of interesting data. Just few of these links ( you may personally like another), often referring to whole flare system.
http://www.geekinter...ckout-drum.html
http://kolmetz.com/p...re%20Rev1.1.pdf
http://www.gasflare..../Flare_Type.pdf.
A knock out drum collects the liquid droplets already formed, it does not prevent droplet formation.
You are right, as LPG expands (from e.g. a PSV) into flare header, fall of temperature can create liquid droplets or even ice. This indicates necessity of knock out drum to collect these droplets.
Since LPG specifications call for almost zero water content (http://www.iocl.com/...cifications.pdf), conclusion is that condensate is not composed of water but of liquid hydrocarbons in most of the cases. Water is carefully monitored in LPG http://www.kinetic.c...%20Analysis.pdf.
All data found out seem to support knock out drum installation.

Edited by kkala, 07 October 2011 - 08:00 AM.

[R NESAMANI]

Thank you kkala,

In our plant, distance from flare stack to process area is 30 mtr. If we want to drain filter, we open the filter's blowdown valve, LPG flow to the flare. I was talking about the water droplets & ice formation in the outer area of the pipe. Are you also talking the same...?

Also the knockout drum will be at 28 mtr from filter. Will it help...?

In knockout drum design, they say that Vap.LPG will go to flare. what abt Liq. LPG?

Slowly, the liq.LPG also vaporises and go to flare.

[fallah]

--> Mr Fallah --> My observation is LPG burns in black, incomplete combustion. Also Liquid falls near the flare stack also occurs. Due to this incomplete combustion, CO2 emission will also be more, i suppose.

Nezamani,

Considering your above mentioned observations:

Why do you hesitate regarding the installation of a knock out drum before LPG flare stack?

Fallah

[R NESAMANI]

Thank you Mr.Fallah.

I shall have submit a proofs like any Standards (OISD,ASME,etc), code(API,etc) for making my seniors accept this proposal.
That's why i am asking any references.

I have gone thru API 521, it shows that liq. particles of more than 600 mic. will be in the bottom of the KO Drum. This will be pumped to the process line. This is like we can recover some LPG being flared. Am i right...?

[kkala]

1. In our plant, distance from flare stack to process area is 30 mtr. If we want to drain filter, we open the filter's blowdown valve, LPG flow to the flare. I was talking about the water droplets & ice formation in the outer area of the pipe. Are you also talking the same...?
-- Unfortunately I worked in Operations (concerning a fertilizer plant) long ago (1976-81), mentioned LPG vaporizer with flare was in an engineering design of another fertilizer plant (1985, no filter). The question is whether you have detected condensates / liquid droplets inside pipe. Resulting low temperatures inside the pipe combined with air humidity can create ice outside during discharge. This is usual in branch lines to flare line, in your case it occurs also in main flare header. Valves in potentially 'frozen' lines should be heat traced, so that frost will not block them.
2. Also the knockout drum will be at 28 mtr from filter. Will it help...?
-- It should be quite close to flare stack, placed at 28 m (of 30 m total length) seems OK.
3. In knockout drum design, they say that Vap.LPG will go to flare. what abt Liq. LPG? Slowly, the liq.LPG also vaporises and go to flare.
-- You can pump condensates out (as you already said). Operation is critical in the sense that drum should never be full enough to allow liquid entrained to flare. This can mean standby pump, fed from critical power supply.
Alternatively drum content can be heated (http://www.eng-tips....d=146638&page=7), apparently under similar safety provisions.

Edited by kkala, 08 October 2011 - 03:54 AM.

[smalawi]

Hi

in addition to what already been contributed, a good reason to have a KO drum is preventing two phase flow. how unlikely it might sound but in case you do have two phase flow and your lines are not design for that it will damage the lines. I've seen an entire flare header jump over it support, luckily it did not harm people or major equipment.

cheers,
sm

[R NESAMANI]

Hi smalawi, Are you talking about the knockout drum in LPG flare line...?
If you have any doc or links, kindly share with me.

Hi KKALA, How can we see the condensate formation in inner surface of the LPG flare pipe ?

Also the reason for outer area ice formation is, any process equipment(Pump, filter, etc..) are in the pr of more than 20 Kg/cm2. When we open the blowdown valve, LPG flows from 20 to the flare tip, which is at atm. pr. This pr. drop make the Liq.LPG to observe the latent heat and becomes Vap.LPG. So water droplets or ice formation occurs in the outer surface of the pipe. Do you agree with me...?

[kkala]

Yes, flashing LPG creates low temperatures and potential outside freezing. It is evaporated to some part, but rest remains as liquid. Liquid is not certain to be vaporized on its way to flare. Check data sheets of relevant PSVs to see whether liquid is specified in their discharge.
I assume a (low point) drain pot (if any) can collect condensates. If there is a water seal drum, heavier hydrocarbons could be traced in its water. There may be other methods of detection. Frozen or cold header up to flare during a big discharge can be also a strong indication for inside liquid.

Edited by kkala, 08 October 2011 - 12:15 PM.

[smalawi]

Hi,

yes, this was an LPG flare, but I dont have much doc to share. the design is very simple, just to have capacity to hold any liquids for say 5-15 min (depending on the process). these likley be from vessel liquid overfilling

as for seal drum in LPG service, you better use glycol fluid rather than water. the latter freezes and can plug the flare line especially if submerged in it.

regards,
sm

[R NESAMANI]

Thank you all for your contribution,

Kindly add or correct the below statements,
Purpose of the KO drum is
1) To seperate the LPG liq.particle size of more than 300 microns.
2) To prevent the liq. spray from hot or cold flare
3) To have a complete combustion of the LPG vapors and prevent excessive smoking, burning rain and flame out.

For LPG KO drum applications, we dont need any pump to take out the liq. in bottom of the KODrum. As this Liq. slowly vaporise and exits in the flare. Am i right..?

Is KO drum is useful for cold vent also...?

So we will install the knock out drum just before the hot flare stack.

[fallah]

For LPG KO drum applications, we dont need any pump to take out the liq. in bottom of the KODrum. As this Liq. slowly vaporise and exits in the flare. Am i right..?

Nesamani,

Care should be taken to apply the above conclusion for all cases!

Actually, need to a pump to take out the LPG from relevant KO drum depends on size of drum, maximum inventory once to be released to drum, frequency of receiving such inventories,...., and as far as i know in many applications even with having submerged electrical heater to assist LPG vaporization pumps have been installed to take out the LPG from the drum.

Fallah

[R NESAMANI]

Hi Fallah, We will install the KOdrum small, so that only vapor exit from flare tip. This will prevent the Liquid rain and safe venting.
We dont want the liq.LPG to be pumped back. Let it take its time to become vapor and exits.
In normal maintenance, we vent max of 1.5 MT of LPG at a time.

Does KO drum has any dis-advantages...?

[Qalander (Chem)]

Dear, Nesamani, as you should realize that

• each vessel or container be it a KOD(knock out Drum)
• has a limited capacity and without any pump-out arrangement.
• It is bound to fill(eventually) topped up with liquid(slightly depends on ambient temperatures)
• As such Nobody can assure No liquid fire rain from flare tip if excessive slug finds it way up to flare-tip.

Hope this proves helpful

[S.AHMAD]

1. If you have decided not to install pump, you must ensure that the liquid hold-up volume is big enough so that no liquid overfill the drum during an emergency.

2. Having liquid overfilling and carried over to flare stack is even more dangerous than having a mixture of vapor and liquid.

3. Have you done any flash calculation to determine the amount of liquid?

4. API RP 521 recommends 20-30 minutes hold-up volume.

[Qalander (Chem)]

I will still have this question where this 20-30 minutes liquid will be safely disposed-off finally?

• One has to be very clear about the conceptual safety in design
• to avoid "pitfalls" and "disasters,catastrophes" from being repeated time and again.

Sorry if too harsh,but Texas refinery's recent past event from atmospheric vent must not be forgotten.

Hope this helps/ everything will be duly taken care off.

[S.AHMAD]

1. Good question by Qalander.

2. Where to dispose the hold-up volume?

3. 3 feasible options

1. Natural evaporation - this will take sometime. Waste of valuable product

2. Install steam coil - faster than 1st option. Add cost but waste of valuable product

3. Pump-out - recovery of valuable product. This add cost but with economic return

4. Whatever it is - "Safety First"

Edited by S.AHMAD, 18 October 2011 - 07:00 PM.

[Qalander (Chem)]

Dears, I do not know if any point was deserved/earned by me or not for previous post!

However just to further clarify

• the normally existing KOD's (physical point) pressure conditions are most usually a very low (much lower than) other upstream system segments and components
• to permit UN-obstructed maximum De-pressurization possible within minimum possible time
• to ensure all upstream system components staying safe and healthy
• with simultaneously ensuring that this Knock out vessel/drum (if available) is never subjected to very high pressures.

All this necessitates a pump-out system in-place

• to remove any or all condensed liquids of any kind
• be it the LPG,Naphtha like H-C's or any other material traveling to the sky-flare
• for disposal in emergency or near emergency conditions of plant's upstream.

Hope this proves helpful&guiding in the right direction!

[kkala]

I would agree with Qalander (Chem) that removing collected liquids from the knock out drum is critical. This is apparent in http://www.cheresour...-flare-ko-drum/, post by paulhorth. Critical function of pumps removing collected liquid in the drum is apparent in http://www.cheresour...h__1#entry44049.
Steam coil is another option instead of pumps. Does it have wide application in flare knock out drums? Advice on it would be welcomed.
Suppose that the knock out drum is being filled with droplets of liquid. As far as I see, it is not easy to predict total volume of liquid to be collected, how can we be sure that this will not be repeated before emptying the drum? So we had better not let liquid in the drum for long time.
Drum capacity covers the 20 min time for the operator to find the reason of liquid entrainment and eliminate it, but then it has been (almost) full. It can remain almost full for long (natural evaporation is assumed slow), in case that collected liquid is not forced to go out.