Steam drum area of disengagement 5

[curve3104]

A certain area of water surface is required in a boiler from which to release the steam at an acceptable velocity with minimal carryover. I am looking for design standards or calculation methods for determining steam drum diameters and lenghts to ensure that the area of the normal water level would be acceptable for a given steam load.
Anyone know where I could find these answers or does anyone have experience with this topic?
Thanks!

 

[davefitz]

For low pressure drums below about 60 psig, this method is used. See the B+W book "steam" , and Coulter,E (1989)"Moisture seperation and and steam washing" In Water Technology for Thermal Power Systems , P Cohen (ed) pp10.1-10.43, ASME NY

For higher pressures and higher steam loading, moisture sperators are used, usually cyclone separators plys demisters or mesh pads. See Carsen W R and Williams H K (1980) in Epri report NP-1607 and also Carter H R and Prueter W P (1980) in Proc symp.polyphase flow and transport tech., ASME NY

 

[Montemayor]

curve3104:

Proper disengagement, when considering 2-phase separation, is more related to the volume on top of a liquid surface - not to the area of the liquid surface. As an example, consider the difference in disengagement effect when dealing with a vertical cylindrical vessel as opposed to a horizontal cylindrical vessel. How much disengagement height is required (even prior to entering a demister pad) is a matter of the fluids in question, the pressures, and the physical configuration inside the vessel. This height value is normally an empirical one, dependent on the experience of the designer. There are no hard-and-fast rules here. Prior success and results deem what is more appropriate and recommended.

I believe that by basing yourself on a liquid surface area you are making a mistake. The answer is much more involved and complex than that. The Souders-Brown relationship is sometimes used for this application and in others, the settling velocity of droplets is used to size the separation vessel (particularly horizontal vessels). Most text books, Perry's Handbook, and the GPSA deal with this subject.

I hope this helps you out.

Art Montemayor
Spring, TX

 

[athomas236]

In the steam drums of water tube boilers, separation almost always takes place in two stages. The first stage is usually cyclone separators that use centrifugal force to separate the steam from the steam and water mixture entering the drum from the evaporator circuits.

This first stage is not 100% efficieny hence the need for a second stage. The second stage is usually demister pads that separate any remaining water from the steam.

With this approach the drum size is then determined to allow this separation equipment to be installed in the drum with access for maintenance. Some times there is a requirement for a minimum water storage quantity in the drum to cover the case when a feed pump is tripped and the standby pump started.

Each major boiler supplier has his own proprietary design of separation equipment which then leads to each supplier have different drum sizes for the same boiler operating conditions. I have kept records of drum sizes offered to us since about 1980 and there are very wide variations in steam drum loadings.

athomas236

 

[curve3104]

After giving this some additional thought, I am in agreement with you Montemayor. The height of the water level to the top of the drum or demister pad is the critical variable to determine steam quality. The overall length of the drum can be considered to be fixed based on other design criteria such as the required number of tubes in the boiler and heat release rates.
The question now becomes, is there a relationship between steaming capacity, the quality requirement of the steam, and the height to the water level that can be estimated to calculate the drum diameter?
Thoughts anyone?

 

[Montemayor]

curve3104:

I have designed liquid vapor separators (both vertical and horizontal orientation) using the Souders-Brown equation as well as the gravity separation method which applies the terminal velocity of a particle settling under the action of gravity. I have preferred the Souders-Brown because it hasn’t failed me yet – with or without internal baffles, plates, or wire mesh.

Mott Souders and George Granger Brown developed their relationship over 65 years ago and it was originally for the purpose of designing the disengagement space that is found in the spacing between trays in distillation towers. It was later applied to mechanical separators used in-line and as separate vessels to separate a 2-phase mixture. The literature contains a lot of information on the application and the empirical value of "K".

For the maximum vapor velocity (which will set a vertical vessel's diameter), use this equation:

Vmax = (k) [ (dL - dV) / dV ] ^ 0.5
where:
Vmax = maximum vapor velocity, ft/sec
dL = liquid density, lb/ft3
dV = vapor density, lb/ft3
k = 0.35 (when the drum includes a demisting section; I have used 0.45 w/o demister)

You can also design applying the GPSA design criteria for Horizontal separators (without mist extractors) using gravity as the sole mechanism for separating the liquid and gas phases. This method requires a particle diameter to be set.

In the Brown-Souders relationship you have to empirically set a vapor disengagement height, since the relationship only gives you the vessel diameter. Depending on the application and the fluids, I have used 1-3 feet in the past; the higher, the better - but of course, the cost increases on the vessel.

I hope this short description helps you out.



Art Montemayor
Spring, TX

 

[athomas236]

Sorry guys but as far as I know gravity is not a major consieration in the design of separation equipment in steam drums of modern water tube boilers

athomas236

 

[25362]

To curve3104, aren't you by chance confusing a steam drum with a steam accumulator ?

 

[25362]

Some thoughts as aked by curve3104 follow.

A steam drum associated with (natural or forced circulation) tube boilers, with its millipede-like configuration can hardly be called an efficient liquid-vapor KO drum. Rapid upward moving liquid disengages from steam on one side, and returns on the other side through relatively cool downcomers to the bottom of the heater to renew its pass upward. Some tubes are too near the steam outlet nozzle short-circuiting the needed steam path length for efficient V/L separation.

One of the features of these drums is the variability of their level and their inverse response characteristics. A sudden increase in demand brings in more feedwater that provokes condensation and bubble collapse reducing the rate of boiling. As a consequence, an increase in feedwater flow may actually cause the liquid level to fall momentarily before the increasing inventory begins to raise it again.

Depending on the operating pressures (and the concomitant vapor densities) pressure changes can produce transient swelling and shrinking of the liquid level. In particular level may swell when a sudden increase in demand drops the drum pressure, even when the feedwater inflow didn't yet compensate for the increased steam withdrawal !

All these factors are brought into consideration in designing steam drums. Since entrainment is somewhat "tolerated" and treated more efficiently downstream, from the viewpoint of vapor-liquid separation efficiency, one could call steam drums just precleaners.

General (published) ROT for horizontal V/L disengaging spaces are: minimum 12 inches for drums of 5-ft. ID and smaller, and 20% of the diameter for larger drums. When optimizing sizes economic criteria are always included by the fabricators.

One reasonable site that can be helpful as a tutorial is

http://www.spiraxsarco.com/learn/default.asp?redirect=home.asp

 

[curve3104]

Montemayor,

I believe we are on the right track with the Souders-Brown relationship. I would guess to say that this would be valid for a horizontal application as well?
As far as the values for K, you say you use 0.35 for any application that uses a demister pad, correct? Is this a typical value given by the seperator manufacturers? In my involvement with them, they will say that for a given capacity, you must allow anywhere from 7-10 inches for the vapor disengagement height (again I am speaking in terms of a horizontal design in a range of 10,000-200,000 lb/hr of steam, with drum diameters in the 30-54" range).
Then, based on the design velocity and given capacity, I would be able to estimate the required surface area of disengagement. Based on that information, determine required disengagement height set by the seperator manufacurer and that will give me the diameter and length required for my vessel.
I think I am right here, but would like some feedback on my thinking.
Thanks again!

 

[25362]

Horizontal gravity separators need too large cross-sectional areas for handling large flowing streams. They are used mostly for the separation of drops greater than about 250 micrometers.

Horizontal travelling distances play a role in the selection of the Sounders Brown K factors and, in general, the horizontal vapor flow is preferably laminar.

Thus, horizontal steam drums on water-tube boilers can't be efficient V/L separators, since the separation takes place also within the liquid/vapor mix entering through the tubes along the longitudinal axis. Pressures play an important role in the sense that they markedly affect the vapor densities.

Anyhow, if you are interested in sizing steam accumulators, kindly see

http://www.spiraxsarco.com/learn/default.asp?redirect=html/3_22_01.htm

and look for "sizing a steam accumulator" where you'll find a worked out example.

 

[25362]

To curve3104: since you have decided to try and design a steam drum by yourself, rather than by relying on the boiler manufacturers' experience consider the following:

1. Try to accommodate the risers at one end and the demister, water downcomer and the feed water nozzle at the other end to provide maximum travelling distance.

2. For a water tube boiler steam drum there is no need of a high liquid residence time or liquid inventory, so most of the space (cross-section) would be used for disengagement.

3. There is an article (I have an old copy) titled Horizontal Vapor-Liquid Separators by Adolph Scheiman in the Hydrocarbon Processing magazine of May 1964, Vol. 43, No.5 that may be of help. If you give me a fax number I could send it to you asap.

4. The site I mentioned above gives data on free surface of steam accumulators that you can compare with your own estimates on the steam drum.

5. There are two more articles which I would be unable to send you but may also be of help:

a. Selecting gas-liquid separators by P.G. Talavera, Hyd. Proc., June 1990;

b. Optimum design of horizontal liquid-vapor separators by Claudio Purarelli appearing in Chem. Eng., Nov. 15, 1982.

Good luck.

 

[curve3104]

25632:

Thanks for the info. Just to let you know, I'm not trying to design a boiler steam drum, per say. I am simply doing some research to determine why and how boiler manufacturers determine the size of their drums. Typically, diameters of steam drums are supplied as 30, 36, 42, 48, and 54". I am trying to determine, why not 34, or 40, or 45?? Since the steel plate can be rolled into any diameter, why are these the industry standard? I am trying to find a relationship of steam capacity, steam disengagement area, and liquid-vapor separation properties to determine, if in fact, that for a boiler of X capacity, one could use a 35" drum instead of a 36" drum. Most of the design information is proprietary of the boiler manufacturer, or it's based on the fact that "it's been done like this for years, so we must follow tradition." Or, is it a matter of being able to fit all of the steam separation equipment into the drum, and the actual diameter is irrelevant to the capacity of the boiler??
These are the questions I am trying to answer. I appreciate everyone's input. Maybe I have been unclear on my intended goals here. My name is Chris and my fax is 804-226-9321 if you wish to send any of the articles as you have mentioned above.

 

[davefitz]

As explained on another cross-post, the diameters chosen are related to the standcard dies the foundy has in stock , and also related to standard size spun heads available. Since we now have a global economy, one can probabably get any diameter today from overseas foundries.

 

[25362]

Chris, I've sent the faxes a few minutes ago. Hope you received them in a readable form. As for the diameters, could this have something to do with the head-forming machinery available ? Good luck.

 

[curve3104]

25632,
I have this post in another forum, Boiler & Pressure Vessel engineering. According to davefitz above, this was a limitation some years ago, with machinary capable of producing only standard sizes. With new technology, drums and heads can now be made in any size with the new machinary.
Yes, I got your fax. Thanks a bunch. It is a little difficult to read, but shouldn't be a problem. Haven't had a chance to read through, but looks like some good stuff.
Thanks again!

 

[curve3104]

25632,
Got a chance to look at the faxes you sent. Contains a lot of good info, just doesn't really help me all that much. Mainly focuses on sizing seperators for a liquid product. I'm looking more for steam boiler type of information.
Basically looking to re-invent the wheel (more like optimize)
Should be able to find something, somewhere for boiler design???
Thanks to all for the help!

 

[rmw]

See my comment in the other forum regarding optimization vs shop practice and utilization.

Not a good idea to double post questions.

rmw

 

[25362]

Anyway, when sizing a steam drum, the volume dedicated to the liquid holdup should be estimated to cover for, say, 10 minutes of BFW pump failure. Designers follow differing rules on this subject.

 

[curve3104]

So, in essence, the size of a steam drum for a watertube boiler is determined to maintain a suitable water level in the drum between two operating points in the case that the feedwater pumps are temporarily off line?
Are there standard drum hold up times published or are they set on a case by case basis?