7 replies to this topic

[shadow13]

Hello Everybody,

I am currently sizing a knock-out drum at the suction of a compressor.

Set-up is like this. The fluid exits the first compressor then passes an air cooler then proceeds to a knock-out drum before another compressor.

Simulation data shows no liquid phase occuring at the stream going to the KO drum. How am I to size the vessel when there is no liquid present in the stream? I tried adjusting the temperature to ambient then to the worst minimum temperature in the area but still there is no condensation. I saw a KO drum calculation that just arbitrarily specified 1 gpm but this seems to have no basis. Would you know of any criteria for sizing KO drums where there is a specified assumption of liquid present in the stream for sizing?

Thanks.

[ankur2061]

shadow13,

As per API 12J a reference figure of maximum 0.1 U.S. gallon per MMSCFD of gas is permitted for liquid entrainment in the knock-out drum gas outlet. However, this is just a reference value and you need to check the permissible liquid content in your compressor suction with the compressor vendor/manufacturer. Most compressor vendors/manufacturers will provide this data. The incoming gas to the KO drum as well as the outlet gas composition needs to be specified by you to proceed with your sizing calculations. The Souders-Brown equation is commonly used for the sizing of knock-out drums. API 12J has sample calcualtion for sizing of a KO drum also.

There have been innumerable posts on KO drums on this forum and you need to use the search feature of the forum to access all those posts on KO drums.

Good luck in your sizing exercise.

Regards,
Ankur.

[jprocess]

Good engineering practices provide some minimum values from bottom TL to different liquid levels that can be used when you faced with gas filled vessels.

[shan]

Gas flow velocity and liquid retention time are two factors to size a gas/liquid separator (KO drum is a separator). For KO drum, gas flow velocity is always the governing factor to determine the vessel size because of small amount of liquid flow. Therefore, it doesn’t matter that you presume liquid flow 1 gpm or 0.1 gpm or 10 gpm. Anyway, there is plenty of retention time and you will have the same size vessel. I would say that 1 gpm liquid flow is legitimately assumed rather than arbitrarily specified.

[NGL Licensor]

It is very important to make sure that liquid knock out will not happen. This is not just a simulation.

You need to see the overall picture of the compressor. Where the feed is coming from? What is the expected turn down? Will you experience different operating modes.

May be I can be of a more specific help, if you can tell what is the real service of this compressor (booster gas position or stabilizer OVHD compressor, depletion gas compressor, etc


AL. King

[Technocrat]

Dear Shadow13,

Please give us complete data like: gas name, vol. flow rate, no. of stages, temperature and pressure of gas inlet and outlet for each stage, coolant inlet and outlet temperature for each intercooler, flow diagram. Try to give max. information, don't assume that some information is not required here.

After receipt of this information we can start some fruitful discussion.

Regards.

hell every body .
i was also designing the same type fo knock out drum at suction of the compressor. and simulation results shows no liquid in knock out drum .it was basically a Natural gas compressor and there is no liquid in gas.
i just used as simple equation of V = (K) * (((desntiy of liquid) - ( desnity of gas))/(density of gas))^0.5. i just put the compostion in HYSYS and get the desnties of both gas and liquid.
the velcoity figure guide me to calculate the Diamter and then L/D should be b/w 3 ~ 4 so i took 3 and calculate the length .


DID I DO A GOOD JOB ???

[djack77494]

hell every body .
i was also designing the same type fo knock out drum at suction of the compressor. and simulation results shows no liquid in knock out drum .it was basically a Natural gas compressor and there is no liquid in gas.
i just used as simple equation of V = (K) * (((desntiy of liquid) - ( desnity of gas))/(density of gas))^0.5. i just put the compostion in HYSYS and get the desnties of both gas and liquid.
the velcoity figure guide me to calculate the Diamter and then L/D should be b/w 3 ~ 4 so i took 3 and calculate the length .

DID I DO A GOOD JOB ???

You did great. As indicated throughout this thread, the sizing of the KOD is not likely to be impacted by the presence of the expected very small quantities of liquid. So why are you worrying about it? You must have the drum present to handle any liquid that does drop out or gets entrained by an upstream upset and then is blown into your KOD. Remember that because your simulation does not show liquid to be present does not mean that it will not be there; it just means that you don't know the mechanism, quantity, or composition of liquid you will get. Makes calculating liquid residence times tough, but push through any problems of extreme nurdiness and just accept a reasonable value of (say) 1 gpm of liquid. Now take a bow - job well done!