Sizing a Knockout Drum 3

[venividivici]

I am trying to verify the size of a knockout drum, but I am not sure what all considerations must be made. I have tried to list the process conditions and the assumptions I have made.

Conditions:
Flow: 18,000 BPH Dry Crude Oil Vapor and Air (Concentration varies from 0%-70% vapor)
Operating Pressure: 0.1-0.4 PSIG
Vapor Temp: ~70degF
Ambient Temp: ~90degF

This is a code vessel rated for internal pressure of 50PSIG, so I am not worried about over-pressure. My main concern is how do you size a knockout drum for 1) proper volume to collect condensates and 2) proper diameter to avoid erosion.

Thank you in advance for any help.

 

[LittleInch]

Run a HYSYS calculation.

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

 

[venividivici]

Thank you for the reply. Could you please expound on what you mean?

 

[LittleInch]

RHysys our something similar will give you some idea of the liquid content of your vapour. Size of the ko pot is dependent on a number of issues. These include velocity of gas entering, size of droplets, percent liquid you want to catch, outflow rate from ko pot.

Designs of ko pots can vary and the difference between those and scrubbers is sometimes rather opaque. Keep gas velocities below 20m/sec and avoid direct flow of gas with liquid in onto tees or flat surfaces.

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

 

[eduardoeh]

I am sizing a horizontal knockout drum for the flare line based on API 521 for a transmission line booster station. The incoming gas is very dry and we assume that any liquid will be mostly water.
Data
Vapour rate Q'v = 64 MMSCFD
Vapour rate Qv = 27222 ACFM
Liquid rate Q'l = 10.3 bbl/d
Operating pressure P = 24.0 psia
Operating temperature T = 520 Rankine
Gas compressibility Z = 0.9700
Gas viscosity µv = 0.0114 cP
Smallest liquid droplet size to be removed D = 500 μm
Contingency accumulation time t = 30 minutes
Slope and drain volume VL1 = 36.25 ft3
Gas operating density ρ'v = 0.00886 lb/ft3
Liquid / oil operating density ρ'l = 64.3 lb/ft3
Gas operating density ρv = 0.00886 lb/ft3
Liquid / oil operating density ρl = 64.3 lb/ft3
Calculation
Drag coefficient Cd = 1.80
Dropout velocity uc = 16.78 ft/s

Having given this information, I am not sure in a few points.

1. Being that it is horizontal, will the droplets be falling at the dropout velocity?
2. If the above is true, then what will be my determining gas velocity across the vessel having satisfied the required minimum length? I ask this because I can get many dimensions (std sizes) and still satisfy the minimum length and lets assume they're all the same as for fabrication cost.

 

[bimr]

Milt Beychok outlined the procedure.

 

[eduardoeh]

bimr,

Thank you for the reference but that explanation in the link is for a vertical vessel containing a mesh pad which determines the flow conditions inside. My question is based on a flare header knockout drum with no internals but a baffle at the inlet.

 

[bimr]

Use Chemworks outline:

http://chemwork.org/PDF/board/API-521 Flare KOD Design and Even More _Part 1.pdf

 

[georgeverghese]

Most large oil companies have inhouse standards to be followed for equipment sizing, which includes flare KODs'. If such standards are not applicable, you can use the method in API 521, which relies on the estimation of the Stokes Law terminal velocity of a 300-600micron oil droplet. If the incoming gas is generally dry, you could use the upper limit of 600microns droplet size.

Since the gas is dry, level heights will default to the minimum permissible level gaps for the various operating, alarm and shutdown level alarms. But you should confirm that there are no other relief that may result in a 2phase feed stream.

Generally speaking, use an L/D ratio ranging 3-6; the larger the ratio, the thinner walled will be the vessel, and the easier fabrication will be.

Dont forget to install the inlet feed elbow and the shellside wear plate. If the gas rate is high, a dual split feed with single central exit will result in further vessel dia reduction.

 

[eduardoeh]

@georgevergheses
Would you agree that when the vapor enters the vessel and then flow at a much higher rate than the dropout velocity, the drops will have a downward vertical velocity equal to the dropout? This is what I understand when in API RP 521 the dropout time is calculated by dividing the available liquid height by the dropout velocity.

 

[georgeverghese]

From memory, I recall the superficial vapor velocity ( based on the available vapor side cross sectional area) should not exceed the liquid droplet terminal velocity.

 

[eduardoeh]

That is true, but it applies for vertical vessels due to vertical forces: gravity, buoyancy and flow velocity. In a horizontal vessels the droplet is only subjected to gravity and buoyancy forces, thus gravity being the large of both. So, the droplet will be falling as it flows through the horizontal vessel.

 

[georgeverghese]

Yes, you are right - my last response is valid only for the case of a vertical KOD. Thanks for the refresher.