10 replies to this topic

[ravion143]

I have been doing a flash separation calculation using a pressure drop as the energy input prior to a distillation column and have come across issues in the sizing of the flash vessel. I have used the Rachford Rice Equation to determine the stream compositions and flow rates and am satisfied with the results as it is completely what i expected.

 

The result are in the attached spreadsheet and it can be seen there is a very low flow rate in general.

 

When sizing the vessel (using a similar design approach as in http://www.chemicalf...00ab&topic=8913) for these low flow rates i have ended up with a diameter of the vessel of 0.026m and a calculated liquid height of 14m with a hold up time of 5 minutes.

 

The values seem way out of proportion for the design specifications for any pressure vessel and it would be very impractical if this were used for the final design.

 

My question is would it be possible to increase the diameter of the vessel to reduce the required height whilst keeping the desired volume of the vessel (approximately twice the volume of the normal liquid level) constant and then using the solver function to minimized the difference in volumes from the required volume and that calculated using a set L/D ratio. I know this shouldn't be a design specification or a criteria for the design of the vessel but i am unsure of how else to obtain a reasonable vessel size. And are there any  errors within my understanding or calculations of this unit operation.

 

Any help would be greatly appreciated.

 

p.s. I have completed an aspen plus simulation of the entire process however I am now trying to understand the  mechanisms involved and justify the use of certain parameters within the distillation column and do a sizing of the flash vessel for costing purposes then move on to size and cost calculations for the distillation column.

Attached Files

•  vessel sizing.xlsx   20.92KB   24 downloads

Edited by ravion143, 06 April 2013 - 07:07 AM.

[Bobby Strain]

Google my name to find my website. You can use software there (Liquid Drum) to get a proper size, maybe.

 

Bobby Strain

[ravion143]

Thank you for the advice, I will check our the applications. May I ask what equations and principles it uses to calculate the various design specs?

[Pilesar]

Your liquid flow rate is 0.1 m3/h and your vapor flow is 0.8 m3/h. This is really too small to design based solely on classical separator calculations. Figure out how you are going to control the level and design your vessel around that. I would try to use a vertical section of NPS 6 or NPS 8 pipe for the separation. This is plenty wide for the vapor rate and a section 1 or 2 feet long will hold 5 minutes of your liquid.

 

If this were an industrial separator drum, you would first calculate the vessel minimum diameter based on the vapor flow rate. The vapor velocity has to be slow enough that the vapor flowing force pulling liquid droplets upward is less than the gravity force on the liquid droplets. The larger the diameter, the slower the upward velocity. But you do not have to use the minimum diameter for your drum! Larger diameter will just give vapor velocity lower than the maximum. One of the questions you want to answer is "What size drum is cheapest to build that will satisfy both the liquid holdup requirement and the maximum vapor velocity?" For a vertical drum, the height to diameter ratio should usually be between 2:1 and 5:1. At high pressures (500 psi or so), drums near the 5:1 ratio are usually cheaper. As you approach atmospheric pressure design, 3:1 drums are better. (The cost differences are not based solely on total weight of steel -- the cost for vessel heads increase with pressure at a greater rate than the cost of straight sided shell.) So the final dimensions are the result of an iterative sizing process. Choose a drum diameter... determine the low liquid level and high liquid level...check the liquid holdup. For small drums, the diameter will be based on nominal pipe sizes. For larger drums, diameter will be to the nearest foot or half-foot. Check that the height to diameter ratio is reasonable. After a few iterations, you can round up to the next standard larger diameter that satisfies all your criteria. When you have a lot of liquid and just a little vapor, horizontal drums might be cheaper. For your "drum" you will probably use pipe caps or blind flanges instead of formed heads and the standard rules of thumb do not apply. The cost of your pipe vessel will be determined by factors other than the cost of the steel.

[ravion143]

Thankyou for the reply it was very useful and I will look into different piping to use for this application.

 

Are there any special considerations to the pipe inlet and positioning (apart from above the high liquid level) and various other aspects with the design that may also need to be considered. I think from the flow rates that will be produced there would be no need for a mesh to aid condensation?

[Pilesar]

You posted in the student forum and I wonder how much you understand what you are doing. Why are you flashing the input to a distillation column in a separate vessel? Why not feed the mixed phase stream directly to the distillation column? Perhaps you would be better off designing the distillation column to separate your feed into useful products. You are asking equipment design questions when you may really need to be looking at the larger process system. Do you have a process flow diagram completed? Do you have a final heat and material balance for your process? You should complete these steps before beginning your detailed equipment design. You said you had a complete Aspen simulation. But getting a simulation to converge is not the same thing as developing a process flow diagram. You can do lots of things in Aspen that are not practical in the real world. I did some impossible things in my senior design project and did not really understand why they would not work until after I graduated.

[ravion143]

The design in brief is the super critical transesterification of an oil (triolein) containing a small amount  (3%) of free fatty acids(FFA, modeled as oleic acid) to form biodiesel (FAME, modeled as methyl oleate) and glycerol prodcuts.a

 

1 mole of triglyeride (triolein) produces 1 mole of glycerol and 3 moles of FAME.

 

Each mole of FFA produces 1 mole of water and 1 mole of FAME.

 

The reaction pressure is 126atm and a temperature of 280C with large excess of methanol (24:1 methanol/oil ratio)l and propane as a co-solvent( 20:1 methanol/propane). The design conditions have been taken from literature and separation of the methanol, propane and residual hexane (from the initial oil extraction) is necessary to obtain a mixture biodiesel and glycerol products which are then separated using the imiscibility and density difference of the two materials.

 

Methanol and propane is recycled and to prevent the build up of hexane and water (by product from the free fatty acid reaction in the feed oil) within the system a purge is also employed. My thinking was to use the flash separator to make sure of the high pressure to separate a fraction of the feed using a simple and cheap piece of equipment and then separate the remaining contaminants using a distillation column. Lowing the total flow being handled within the distillation column, reducing costs. Using the flash vessel also lowered the quantity of material that needs to be compressed back up to 5 bar compared to when a distillation column was the only unit operation used.

 

I have attached the flow sheet for the aspen simulation and the PFD (the heat input and removal will be heat integrated at a later date and the temperature decreases are to prevent degradation of the products and side reactions after the transesterification step). Mass balances were done by hand prior to simulation and similar results were obtained.

Attached Files

•  PFD.jpg   30.88KB   18 downloads
•  mass balances.xlsx   15.42KB   15 downloads

Edited by ravion143, 06 April 2013 - 09:26 AM.

[Pilesar]

You have an interesting process to work on. Your KO drum does reduce your feed to your recycle compressor from 20 cum/hr to 17 cum/hr, but is that worth another piece of equipment with a level controller and possibly a relief valve? Will your compressor equipment cost or column cost be significantly different? 

Some other questions come to mind: What will you do with your methanol-hexane vapor purge stream? What are you trying to remove from your recycle and why did you set your purge rate where you did? Pumps are generally preferred over compressors when possible. Do you need to compress your recycle vapor to condense it before pumping? Or can you condense it without compressing it first? 

[ravion143]

Thank you for your continued advice with this process, your input has been very valuable in helping me consider alternatives and asking myself different questions.

 

I think I will attempt to do the separation with a single unit operation also and compare the two cases to see how it compares at the low flow rates and also investigate the implications if a higher flow rate is used as the next stage of the development would be to do a possible capacity increase to investigate the economic feasibility at different capacities so I would expect it to have a greater significance in that case.

 

The methanol/hexane purge  hasn't been addressed yet (still slightly unsure on how to process it) however the purge rate is decided on keeping the hexane content within the system below a certain level due to at these values the fact that it doesn't have any negative impact on the reaction extent and actually provides a slight advantage in reducing operating temperature and pressure but to a much lesser extent than the propane.

 

The recycle was compressed due to the propane solvent being soluble at 5 atm and 40C allowing for the pump P-303 to be used to increase the pressure to 126 atm for the reaction instead of equipment that could handle both liquid and vapor which would be more expensive.

Edited by ravion143, 06 April 2013 - 08:07 PM.

[Pilesar]

What components would not condense in your recycle stream at 25C and 1 atm? Only the propane? You have 0.6 kg/hr propane in your recycle compressor stream. You can buy propane in liquid form for about $1/kg. What is the cost of energy to operate the compressor? This stream comes off the top of your distillation tower. Why not take a liquid distillate stream instead of a vapor distillate and use a pump? If you have some non-condensibles, then you may be able to vent them to a safe location. You know your system requirements better than I do, but I would jump through some hoops to avoid using a compressor.

[ravion143]

The main concern was the propane, also the additional purchasing and operating cost of the compressor would be much greater than the pump which i think would be easily be offset by having a liquid distillate. The non condensibles would be mostly propane which i could probably involved in another part of the process, probably by burning it off to produce heat energy for different parts of the process such as the reboiler. 

 

Thanks for the note, I didnt consider a vent/two phase collection after the distillation column removing the propane.