Flash Vessel Sizing

[hadams417]

Im a mechanical guy, but alot of the threads I've found on similar topics were dominated by you chemical guys, so I though I'd post this here hoping some of you may be able to help.

Please advise me on how to properly size a flash vessel. Here is my application:

Condensate inlet flowrate = 51, 348 lb/hr
Inlet pressure = 600 psig
Internal tank pressure = 160 psig

Which corresponds to 15.1% flash steam. Given this amount of flash steam for the given condensate flow, this corresponds to:


Steam Flow (from tank) = 51348 * .151 = 7753.6 lb/hr

Liquie Flow (from tank) = 51348 * (1-.151) = 43595 lb/hr

Now, using the specific volume data of the vapor and liquid (2.6 ft^3/lb and 53.6 ft^3/lb, respectively) and target velocities of 10 fps for the steam and 50 fps for the water exiting the tank, my calculations for the tank diamter are as follows:


Steam: Dia = [((7753.6*2.6)*4) / (3.14 * 36000)]^(1/2) = .844 ft = 10.1 in (dia. necessary to provide steam flow of 10fps)

Water: Dia = [((43595*53.6)*4) / (3.14 * 180000)]^1/2 = 4.07 ft = 49 in (dia. necessary to provide water flow of 50 fps)

Clearly these are substancially different diameters. I do not want to undersize my vessel and experience overflow as a result. However, I do not want to over size either. I am stuck at this point. How do I determine proper tank heigt? Ive read places that state "use a 3:4 ratio for dia. to height" and others that say "tank height = 3*dia or minimum of 36in). Please advise me on how to complete the sizing procedure.

 

[katmar]

Flash vessel sizing was/is discussed about 10 threads down from your query. I'm sure a search on "flash vessel" will reveal plenty more.

 

[Montemayor]

hadams417:

Katmar is correct and giving you helpful advice. You've given us some detailed basic data and also shown some of your own effort in this application, so I'll give you my input. One very experienced contributor to this forum is Milton Beychok and he may also throw in his worthy opinions.

The design criteria used is a variation of the famous Brown-Souders equation:

Va = K [(dL - dG)/dG]^0.5

where,
Va = Maximum allowable superficial vapor velocity, ft/sec
dL = Liquid density, lb/ft3
dG = Gas density, lb/ft3
K = a constant, based on the application; usually varies between 0.25 and 0.40

Wire mesh is commonly used to aid in separating the two phases. It is located in the upper, vapor phase area of the vertical vessel, at least 12 inches above the highest liquid level and preferably 18 inches.

The relationship between velocity and the vessel cross-sectional area is:

Superficial Velocity = (vapor mass rate)(specific volume)/(cross-sectional area of vessel)

Stokes law is incorporated in the Brown-Souders relationship and the time of residence in the separation zone is important - it allows for the liquid droplets to separate and drop out by gravity. The length of the vapor trajectory (which sets the residence time) is critical to separate out the liquid droplets.

The "correct" K to use for the new operating conditions depends on the system, the fluids, and your expertise and experience in this field. The K is a constant that is empirically obtained and subject to a lot of variables that can't be defined in writing.

Try to build in some liquid residence time in the sump of your flash vessel so that you can reasonably control the level and maintain it well below your vapor section. I usually design for at least 5 minutes liquid residence time. You may need more. A reasonable ratio of height:diameter is in the order of 3 to 2. For difficult flashes I've seen as much as 5:1 - that is, for a 12" diameter vessel, a height of 60" was used. This is rare; a ratio closed to 3:1 is common.

I hope this helps in guiding you.

 

[katmar]

There are many interesting correlations in engineering and I think I have just discovered a new one. An engineer's patience is in direct proportion to the number of grand children he/she has. Art, you beat me by a mile.

 

[Montemayor]

Harvey:

Thanks for the build-up, but I believe that when it comes to patience and understanding you have proven yourself worthy of being called a Champ. Like you, I admire and take pleasure in responding to a well-written, detailed, and concise query. It indicates interest, work effort, and resourcefulness – characteristics that are indicative of good engineers. I also maintain a personal database of all past threads I’ve responded to and can quickly cut-‘n-paste a response to a frequently-asked question. Besides that, I have a lull in my daily action routine right now: my two granddaughters are taking a cat nap after their swimming lessons this morning and prior to my taking them to see “Madagascar”, the latest and greatest kid’s movie for this summer. Grandparenting is the greatest, and I’ve earned my stripes.

 

[hadams417]

To all -

I want to thank you both, as well as Milton Beychok, for your timely responses to my inquiry. Mr. Beychok contacted me was more than happy to answer my questions and set me on the right track. Im not able to relate, however, to your most recent discoveries correlating grandchildren and patience. I am merely an intern brand new to the professional engineering field. I would like to say that this website and everybody associated with it has certianlly warmed me up to the idea of getting out into the working world. I really appreciate that members of the engineering commumity are so overly willing to offer help to young people who are just entering the industry. Thanks again.

 

[katmar]

hadams147,

I would not have guessed that you were young from your post, and the quality of the post obviously impressed Art Montemayor because although I complimented him on his patience, he has been known to explain to some posters in no uncertain terms that they have not set out their query properly!

I'm certain you will do well in engineering, and please stick around here as well - we need guys like you. I apologise for the private joke to Art. When I compared my first reply and his the stark contrast was just so obvious. Art's reply shows that he is not only patient, he is also highly organized. We can all learn from him.

 

[quark]

I generally check initial data by the OP in any post and if found to be ok, take the other values given in the original post. Enthalpies and flash steam percentage seemed to be OK, so I didn't bother about the other stuff.

The original post baffled me for sometime and I wondered how such a flash vessel could be constructed with the given liquid and vapor flowrates. Then I found out that the OP used density instead of specific volume in his liquid calculation. Actually it should be 0.018241cu.ft/lb x43595lb/hr = 795.21cu.ft/hr

Further, 50fps for a liquid line is something I avoid at any cost. With 795.21 cu.ft/hr, it can be as low as 10fps which gives us 0.167ft or 2" pipe diameter.

Regards,