PDA

View Full Version : Ammonia Refrigeration Knock Out Vessel Design







Pdag
10-04-2013, 02:03 AM
Hi all I am a student doing chemical engineering and I am in the process of completing my final year design project. I am currently stuck on my detailed design and any guidance would be greatly appreciated.

Overview of unit:
The unit that I am designing is an existing refrigeration tower used in industry to flash liquid ammonia from inlet feed conditions of 8.3kPa, 19C to AtmP, -33.3C over 5 stages. This allows the ammonia to be cooled to it's bubble point without adding any energy to the system other than into the compressor which is used to re-compress and recycle vapour flash streams to inlet. For the purpose of the design I have treated each stage as an individual flash or knock out vessel for simplicity.

These 5 stages are each incrementally lower in pressure than the previous until atmospheric pressure is reached which provides the flow driving force between stages. I have also done calculations for 1,3,and 7 stage system to compare my design with.

The main reason for the stages is to reduce the duty on the compressor(s), which re-compress the flashed ammonia vapour back to inlet conditions so they can then be recycled.

As a result the compressor is compressing smaller vapour streams which are already at higher pressures and temperatures as opposed compressing the whole vapour stream from atmP back to 8.3kPa which is what would be required if a single knock out vessel where used.

Where I am now stuck:
I have performed an energy balance and calculated the Liquid and vapour flow-rates through the system using thermodynamic data on ammonia that I got from NIST database. I am now trying to size the vessel(s).

My problem now is that I donít know how to determine the required height and volume of the column. I assumed that I could just set a diameter and determine the height by dividing the volumetric flow-rate by it. That way I get a velocity but this does not shed any light on scale because of the time component. For example if I were to do the calculations in seconds I would get a much smaller height than if I did it using hours which has left me thinking I am missing something, probably obvious. I thought that it would be necessary to incorporate acceleration due to gravity and have been playing around trying to get the desired units to no avail.

I have since used the Souders-Brown equation with a a k value of 0.03m/s however this gives me a diameter of 0.52m and height of 2.8m using a h/d ratio of 4 which seems very small as the unit is processing a total Ammonia flow-rate of 3.5tonnes/hour.

Any Information on designing a flash vessel would be greatly appreciated and thank you for taking the time to read this if I haven't bored you already!!!

ps. I have tried attaching an excel spreadsheet of what I have done so far however the uploader doesn't seem to recognize the file so if you have any guidance on how to upload it I will be happy to share my working. Thanks again!!!!

Magoo
10-04-2013, 03:38 AM
Designing and selecting a 'knock drum/ surge vessel " is a fine art. All to do with operating levels, surge volumes and high limits, demister sheets and sparge headers internally. Add velocity and density volumes, operating conditions and compressor loading rates., drop leg heights overfead rates or static heads and a multitude of other critical criteria. Basically no one system is the same, and you can only rely of manufacturers published data to be totally confident of designing a stable reliable surge vessel.

Pdag
10-04-2013, 04:07 AM
Hi thanks for the reply Magoo I have since assumed a hold-up time of 1/2 an hour and divided by volumetric flow (7291m^3/hr) by it to get the volume of the vessel then by varying the diameter to height ratio making sure its between 3-5 (this is the range most sources sources I see have it at being) I came up with a height of 38.4m and a diameter of 11m. This seems like a huge vessel so I dropped the hold-up time to 12mins which gives me a volume of height of 28.3m and diameter of 8.1m do these values seem reasonable?

Magoo
11-04-2013, 03:23 AM
You need to look at a petrochem cracking tower concept with controlled pressure reducers between each section, height will be the criteria. Bit off a trick question being ammonia related. Pressure reduction between sections will reduce diameter calculations.

lowtempman
11-04-2013, 03:29 PM
Assuming there is no need to hold an inventory of refrigerant in this vessel the design basically reduces to a matter of sizing the vessel to assure separation of the gas formed from flashing and the liquid that remains. The reason is that liquid carryover to the compressor must be avoided. A book written by Professor Will Stoecker of the University of Illinois will give you the theoretical equations that are involved. From a practical standpoint, we designers use a very simple equation involving gas and liquid density and a factor that depends on how turbulent the mixture of gas and liquid enters the vessel. Research is still ongoing on this subject. A gas velocity of 100 feet per minute for the inside velocity in a vertical vessel will be a good number. For example if you have 500 CFM of ammonia gas leaving the vessel an inside area of 5 square feet will be OK. A vertcal distance of 24 inches or more is recommended from the top gas outlet of the vessel to the entrance level of the entering flow. This is a very basic answer and there are other considerations to the final detailed design.

sterl
09-05-2013, 02:56 PM
I'm fascinated by what you are trying to accomplish.

The premise of dropping pressure on warm liquid in stages has merit but the implication in a normal compression arrangement is that you have a corresponding suction pressure for each stage of the flashing process. This can be accomplished with a fixed displacement screw or scroll compressor at a fixed discharge pressure but the plumbing and control side is pretty exotic; and the internal compressor pressure is not fixed for a specific size of port so local flow reversal and internal bypass are both possibilities.

Is that what you are doing with the flash vapor?