Face Velocity in Air Coolers. 2

[Peregrino7]

Face velocity in Air Cooler is usually not greater than 600 ft/min. But I don't find any information that explains why this top value, and would happened if I have velocities that are bigger than 600 ft/min. I will appreciate your help if you know where I can find information on this issue (website, paper, etc)
Thanks...!

 

[quark]

This is to reduce water(moisture) carryover along with the air stream. Actually it should be between 500 and 600 fpm.

Regards,

 

[25362]

Air face velocities in air-cooled heat exchangers vary inversely to the depth of -usually finned- tube rows. Ludwig Vol. 3, gives for 1" OD tubes and 2^3/8", triangular spacing, the following typical standard face velocities (FV):


Depth, tube rows 4 6 8 10 12
FV, fpm 595 540 490 445 405

These FVs have apparently been fixed by economic considerations taking into account the HT coefficients, the fan energy and the overall equipment weight and cost.

The Buffalo Forge Company in its Fan Engineering Hadbook gives a wider range of FVs.

More often than not the FV is expressed as air mass velocity, lb/[(h)(ft² face area)]. For example, 1,400 to 3,600 lb/[(h)(ft² face area)].

 

[friartuck]

The velocities can be higher if you use an eliminator plate, but keeping the velocities down is also good practice as it keeps the pressure drops down to a reasonable level.

PD is proportional to the velocity squared.

Friar Tuck of Sherwood

 

[25362]

It seems to me friartuck and quark are referring to fans on cooling towers or otherwise air fans accompanied by water evaporation. In these cases drift eliminators are provided with 4 to 5 velocity heads to keep the carryover of water droplets down.

With "dry" air cooling services, as used in the chemical and petrochemical industry, a face velocity of ~600 fpm is quite common, as Peregrino7 says.

 

[quark]

25362!

By "dry air cooling" if you mean cooling air by circulating chilled fluid in coils (indirect cooling), the face velocity limit remains same. Beyond this velocity there is a chance for air to pick up the condensate formed on the fins and coil tubes.

If your air is totally dry or if you are cooling the air above its dew point then the velocity limit can be flexible.

Regards,

 

[speco]

Peregrino7,

There really is no right answer to this one. Like most other engineering problems, the air velocity in a cooler is a question of what is most appropriate for the application. The "typical" values as given by Mr. 25362 are somewhat reasonable, but only as "typical". Probably 99% of all air coolers (that is, finned tube air-cooled heat exchangers) have between four and six rows of tubes. It is a rare and weird application that would use more than this, but it does happen.

Let me throw in my two cents on the what is typical. Most cooler programs actually work with air-side mass velocity through the net free area of the cooler bundle. Since tube pitches, tube diameters, and fin configurations vary all over the place, rules of thumb are good as a starting place or for estimating purposes. That said, the most typical configuration you are likely to see in process coolers would have a 1" tube, 2-3/8" triangular tube pitch, and aluminum fins about .016" thick spaced at 10 fins/inch. In this configuration, the actual net free area is almost exactly 50% of the face area. Most coolers have an air-side mass velocity in the range of 5000 to 6000 #/hr-ft^2. (or 2500-3000 #/hr-ft^2 of face area).

Some manufacturers like to use a lot of fan horsepower to keep the capital cost down, and may use velocities of over 7000. However, the capital cost savings of such designs are at the expense of great operating cost. However, this may be necessary where there are severe limitations on plot area, say on offshore platforms.

In some noise-sensitive applications, the opposite is true. The mass velocities may need to be cut to half the typical values in order to slow the fans(s) down enough to keep them really quiet. Such coolers have much more capital cost, but the operating cost is extremely low.

It's a tradeoff that has to be considered for each application. As in any heat transfer problem, there is an infinite number of solutions. Some are much better than others. The trick is to figure out what works "best" for each one.

Regards,

Speco

http://www.stoneprocess.com

 

[25362]

To quark, by "dry" air coolers I meant heat exchangers on which external air is the cooling medium, such as in steam or process condensers and coolers.

I think Speco has given a thorough and practical answer to Peregrino7's query, and merits a star for that.

 

[quark]

I have to convert the mass velocities to volumetric velocities before I make any comment on the previous posts. However, the velocity limitations are well specified in black and white and followed world over, particularly for air conditioning applications.

ASHRAE HVAC Systems and Equipment Handbook suggests a strict 400-500 fpm velocity for cooling and dehumidifying coils and evaporative air coolers.

Principles or Refrigeration and Air Conditioning by Shan K Wang suggests a wide range of 400-600 fpm in the absence of drift eliminators.

For heating applications, the velocity ranges in between 400-800 fpm.

Regards,

 

[25362]

For a range of face velocities, and other considerations, on air conditioning, air-cooled heat exchangers, various other fan systems and applications, Fan Engineering of the Buffalo Forge Company may serve as an excellent reference engineer's handbook. I've seen the 8th Ed. of 1983, I assume new editions may be available.

 

[imok2]

Air must travel across the face of the coil at velocities that can't be too low or too high. If the air is traveling too fast across the coil, then you get inefficient
heat transfer and high air pressure drops. If the air is traveling too slowly, then you get practically no heat transfer, because there is no air turbulence.

 

[Peregrino7]

Dears Sirs,
Thanks to all of you for your answers.
Your discussion has helped me to see this issue in many points of view.
I have to say that I work with HTRI software, and they do not limit this value, but the Reynold number. So I have designs of Air Cooler Heat Exchangers with 10 tube rows and a pressure drop in air side not bigger that 0,7 in H2O and face velocities not bigger than 600 ft/min.
I know that there is a lot of reference that limit this value, but few people knows really why (scientifically rather than empiricist) and this is why I put this question in this forum.
I think that Speco’s answer give a line: that is a trade off between motor power and noise production, and energy save. Nevertheless, I find designs with Face velocities of 800 ft/min, and when I rate the motor and fan (for example with Moore software) it shows 85dB at 3 ft (which is according to API).
I wonder why there is no Paper published on this issue? And Why there is so many different point of view of the reason of something that almost all of us agreed to be a good practice.
Sincerely,
Peregrino7


G

 

[speco]

Peregrino7,

In my previous posting to this question, I mentioned that most air-cooled exchangers for process applications have between 4 and 6 rows of tubes. There are two very practical reasons for this. Both really relate to the overall cost of the exchanger.

1. In a very deep tube bundle (that is, one with several rows of tubes), the cooling air tends to heat much more than it one with fewer tube rows. This reduces the LMTD (Log-mean temperature difference), thereby requiring much more surface to do the same amount of cooling. The HTRI program is pretty good. I don't think it has any way of comparing equipment cost for different designs. Of course, manufacturers always have to look at this, since every job is a competitive situation.

2. Process coolers are also pressure vessels. They almost always have rectangular box headers at each end of the tube bundle. As the number of rows of tubes increases, the thickness of the tubesheet and plugsheet tends to increase as well. Rectangular header boxes are designed per ASME Appendix 13. When these thickness increase, the material cost goes up. There is also a significant increase in the welding cost. In general, the higher the design pressure, the more significant these costs are.

It may be worthwhile to work with a manufacturer (or get competitive bids from a few manufacturers) to design the cooler for your application. You might find that the overall cost can be reduced considerably compared to your ten-row design. You can then check their designs with the HTRI program to confirm that they actually work. I believe that was the intent of HTRI in developing it.

By the way, the Moore Fan program is very good. It is an excellent tool to use in conjunction with a cooler design program.

Regards,

Speco

http://www.stoneprocess.com