Jump to content



Featured Articles

Check out the latest featured articles.

File Library

Check out the latest downloads available in the File Library.

New Article

Product Viscosity vs. Shear

Featured File

Vertical Tank Selection

New Blog Entry

Low Flow in Pipes- posted in Ankur's blog

- - - - -

Water Content Removal From Air Compressor Inter-Stage Cooler


This topic has been archived. This means that you cannot reply to this topic.
6 replies to this topic
Share this topic:
| More

#1 Process_Nerd

Process_Nerd

    Brand New Member

  • Members
  • 8 posts

Posted 05 February 2017 - 08:32 AM

Hi,

 

I am working on an air compressor inter-stage cooler. The cooler is S&T type with TEMA type BEW . Air on the shell side is cooled with cooling water on the tube side.Air contain water content which upon cooling  from 160 deg.C to 45 deg C.  condenses. Air enters the shell side from the top of rear head and leaves the shell from the top side of front head. A liquid outlet is provided at the bottom of shell. I have attached a sketch in an effort to give some clarity to my query. Now, the challenge is to separate the water content from air before it leaves the cooler and is further compressed in the compressor.

 

I would request the experts to please advise how the water content can be removed. It should be noted that the air moves with a velocity of 12 m/sec and I'm afraid the air will sweep out the moisture(water). Certainly, one way could be installing a filter downstream of this cooler, before it enters the 2nd stage compressor. but I am looking for any solution within the cooler. Can installing a longitudinal perforated plate below the bundle and 2) Providing space(compartment) for liquid separation at the shell bottom below the bundle help? I would appreciate if anyone could advise any practical solutions.

 

Process Parameters:

 

Shell Side Fluid: Air

Shell side mass flow: 54095 kg/hr

Moisture (water) content in air: 662 kg/hr. (0.012 mass fraction of total air flow)

Shell side inlet pressure/temp. : 3.7 kg/cm2.g/ 160 deg.C

Shell side outlet pressure/temp. : 3.0 kg/cm2.g/ 45 deg.C

Tube side fluid: Cooling Water

Tube side inlet pressure/temp. : 3.5 kg/cm2.g/ 33 deg.C

Tube side outlet pressure/temp. : 3.0 kg/cm2.g/ 42 deg.C

 

Regards,

Bilal 

Attached Thumbnails

  • Inter-stage Cooler Sketch.jpg


#2 Pilesar

Pilesar

    Gold Member

  • Members
  • 1,347 posts

Posted 09 February 2017 - 01:27 PM

The air velocity in the exchanger will be less than the air velocity in the piping. You have 12 m/s where inside the exchanger? Your velocity at the end of the exchanger will be less than at the beginning of the exchanger due to cooling. You don't have a lot of liquid, so draining the liquid at the bottom of the exchanger should be sufficient if your drain line is large enough. Consider a collection drum for the liquid at the same pressure as the exchanger to allow equipment protection system in case of a high liquid level. Have you modeled this exchanger with design software such as HTRI or HTFS to help in visualizing the internal flow characteristics? Your exchanger baffle design matters a bit. Your thumbnail seems to indicate a large non-tubed region above the bundle which would also help reduce entrainment. If entrained droplets in the vapor are critical, consider using mist eliminator equipment for peace of mind. Mist eliminators are reasonable options for compressor suction vessels. 



#3 Art Montemayor

Art Montemayor

    Gold Member

  • Admin
  • 5,780 posts

Posted 09 February 2017 - 06:18 PM

Bilal:

 

I have installed and operated many compressor intercoolers in the past - some (approximately 25) I have designed and fabricated myself.  My response to your request, based on practical experience, is as follows:

 

  • Get your basic data correct and credible.  You say you are designing a BEW TEMA type exchanger, but yet your attached drawing shows a BEM type.  Were you just being lazy and copied a drawing from somewhere instead of generating your own, specific drawing?
  • For the process conditions you cite, what would be a normal and practical type of TEMA design would be a BEM and NOT a BEW.  I don't believe you have the basis nor justification for selecting a BEW as a compressor intercooler.  A BEW is a straight tube unit with one floating head and one stationary head.  The floating head is generally sealed with an O-Ring.  The thermal conditions on a compressor intercooler don't justify the need for a floating head.  You certainly can't justify a BEW because of the need to clean the shell side.  The BEW will be a much larger unit due to its inefficiency as compared to a BEM, which is more compact.  A fixed tube sheet will do just fine.  If your cooling water is dirty and you need to clean the tubes, this is done by using a bonnet on one end and a channel with cover on the other end.  Removing the cover and rodding out the tubes does the job easily.

Your conditions call for cooling 26,000 Scfm of compressed air and dropping out 3 gpm of condensed water out of the same stream.  This is a hefty capacity and probably a centrifugal compressor and you certainly don't want to incorporate an intercooler that is inherently bigger than most due to its inefficiency.  You have to be practical - as you state - and using a horizontal BEM with multiple condensate nozzles in the bottom of the unit, in-between baffles is what I've done in the past.  The nozzles lead to a common header that goes to a level pot with a level instrument that keeps a positive water level seal at the bottom of the exchanger shell side.  The exchanger baffles are, of course notched in the bottom to allow all water formed in each baffle compartment to inter-communicate.

 

I prefer vertical cut baffles in this type of application, but you can also use horizontal type.  I space my baffles in increasing distances starting from the inlet of the hot air.  The baffles are not equally spaced because I want more turbulence (better heat transfer and water condensation at the first baffle passes and I want lower velocity (with less heat transfer efficiency) at the last baffle pass to allow for easier and better air-water phase separation with little or no entrainment.  Both the air inlet and outlet are at the top of the shell ends.

The unit would be single pass on the shell side and 2-pass on the cooling water tube side.

For your benefit, I would hope that our expert member, Dale Gulley, jumps into this thread with his valuable advice and experience.  His input would definitely put a master's touch on a design you can count on.  His experience level is well above mine.

 

A downstream suction drum is always recommended after a compressor intercooler.  This drum is there as a snubber and a separator to catch any liquids that may collect or be entrained from the upstream cooler.  You certainly don't need a "filter".  You can't "filter" out entrained water from an air stream.  You have to separate it out, and the suction drum does that.



#4 Process_Nerd

Process_Nerd

    Brand New Member

  • Members
  • 8 posts

Posted 10 February 2017 - 08:53 AM

Thank you very much gentlemen for your valuable input and recommendations.

 

Dear Pilesar,

I have sized the exchanger using HTRI and the flow monitoring results for hydraulics and thermodynamics of the exchanger are attached for your review. The air velocity does decrease through the shell length and is the lowest at the outlet but apparently not low enough to separate out the water content via gravity. The thumbnail I attached earlier was not the actual depiction of my case and was just for flow orientation. I am afraid Installing a mist eliminator won't be very effective with such high velocities.

 

Dear Art,.

I have attached a proper sketch now as per your recommendations.  Please forgive my negligence.  I should have presented my case more professionally.

 

I have provided a sump for liquid drainage and accumulation. The liquid level will be such that enough water column is maintained to avoid air bypassing.  The air velocity at the shell exit is around 8 m/sec.  Please provide your opinion and valuable recommendation.

 

Dear Mr. Gulley,

Please help me in designing this exchanger and provide your expert opinion.

 

Regrads

Attached Files



#5 Pilesar

Pilesar

    Gold Member

  • Members
  • 1,347 posts

Posted 10 February 2017 - 10:34 AM

The pressure drop across the exchanger is very large at 12.5 psi. Your compressor is using energy to overcome the pressure drop and you may be able to justify a new exchanger based on energy savings alone. You could reduce the pressure drop by using double segmental baffles or an exchanger shell type that divides the flow path through the shell. Finned tubes might be an option to reduce exchanger physical size. What is your total pressure drop between the compressor stages? Since your velocity in the exchanger is so high, I would certainly also check whether the interstage piping is large enough.

   Baffle-type mist eliminators depend on high velocity to be effective. But if you are going to put in extra equipment, compare the costs. For expert help, some of the mist eliminator vendors can do a decent job of calculating knockout vessel size with their recommended internals. Equipment space can be an issue around compressors so you really need someone to consider you specific situation.



#6 Art Montemayor

Art Montemayor

    Gold Member

  • Admin
  • 5,780 posts

Posted 10 February 2017 - 02:21 PM

Bilal:

 

Pilesar, in my opinion, has hit the nail on the head.  The important negative points coming out from the HTRI results (assuming that the data inputs were correct) show that this application is getting very expensive due to the large energy loss imposed by the large pressure drop between your 1st and 2nd stage.  Losing essentially one atmosphere of pressure is causing the compressor driver to consume more energy.  I recommend that you expand your project Scope of Work to include the negative effect this application can impose on your compressor and driver if you design for a high pressure drop between stages.  What I normally would expect as an acceptable pressure drop between stages would be 3-5 psi.  12.5 psi, in my opinion, is not acceptable and the price to pay is expensive in driver horsepower and operating energy consumption.

 

In setting the physical dimensions of the cooler, did you consider the pressure drop?  The HTRI printout shows a "standard" tube length of 20 feet (6 m).  For fabrication economy, the longer tube length gives lower fabrication cost - but it starts to create tradeoffs such as the pressure drop and the higher velocity in the shell side (with the related entrainment).  Practical engineering, as Pilesar has suggested, may call for use of finned tubes as one means to decrease the pressure drop while contributing more heat transfer area on the most difficult side (the humid air).  I would follow this suggestion and try to find out if this technique works to reduce the pressure drop and ensure complete condensation of the water moisture.  Here, be mindful of the clue that I gave in my previous post: "You certainly can't justify a BEW because of the need to clean the shell side."  In other words, your scope of work should recognize that the compressed air, by common sense, is not a fouling or troublesome fluid.  It comes from a compressor stage that is designed to handle only clean, filtered air and as such will not need any means for future cleanout.  The shellside of this unit does not need any means for access or cleanout - with the exception of inspection ports.  Therefore, you are free to make this large exchanger as compact as you can.  The use of finned tubes should not be a source of maintenance concern or complaints.

 

Another way to reduce the pressure drop and related velocity might be to increase the shell diameter (your sketch does not identify this) and reduce the tube length.  This will probably force you into multiple tube passes - which shouldn't be a problem, other than increasing the partitions employed in each head.  Here, the available space for the unit might be affected, but you can add a lot of tube heat transfer area with relatively small diameter increases.  Further HTRI runs are called for to investigate these potential options and their results.

Accept the outlet velocity that satisfies the required total heat transfer and the complete condensation or the associated water without any entrainment.  I would not use outlet velocity as an HTRI input, but would let the program calculate it.  I stress the need for a 2nd stage inlet separator as an additional vessel that should be included in your scope.  This is normal practice and serves to protect the compressor and the integrity of intercooler.  It is at this separator that you attach such items as PSVs, surge protection connections, added streams, purge facilities, drains, pressure and temperature indicators, etc.  This vessel adds to the design and operability of the compressor unit.

 

I have used the type of condensate drain facility that you show in your cooler-condenser sketch.  I have applied this design in about 10 cases in the past.  But be very careful in applying this type of condensate collection and drainage because it's final physical design is subject to the operating pressure drop across the cooler-condenser.  This is where Pilesar's experienced concern for the large pressure drop comes into play on a practical basis.  You cannot expect to make the drain connections you show without detailed hydraulic calculations that show that the vertical legs of each drop will remain sealed during operation.  Common, practical sense should tell you that if you are expecting a pressure drop of an atmosphere between the first condensate drain leg and the last one, then the difference in condensate height between both legs will equate to the pressure drop.  This could be quite considerable and call for a cooler-condenser installation height that is ridiculously too high for economic, safety, and practical reasons.  Pilesar's advice calling for attention to reducing the pressure drop is more than warranted for this additional reason.  The ten cases where I have used this type of design all were as cooler-condensers for the overheads streams from stripper towers - most of which were using MEA solution for CO2 removal.  The units were quite large - approx 16 ft long x 30-36 ft diameter, done using a slide rule and no HTRI.  The average pressure drop I allowed in these units was approx. 1-2 psi and I allowed for sufficient elevation to allow multiple drains on the shell while maintaining a water seal in each.  If you don't maintain a positive seal on each drain, you will obviously create a convenient gas bypass from the high pressure shell side to the low pressure side, and not have any cooling of this bypass, humid gas.  You can't tolerate this type of operation.

 

Again, follow Pilesar's recommendations closely and I am crossing my fingers, hoping that we also hear from SrFish (Dale Gulley) on this thread.  Keep us posted with your return comments and missing data - like the diameter and additional related equipment you need like the 2nd stage suction drum.

 



#7 srfish

srfish

    Gold Member

  • ChE Plus Subscriber
  • 408 posts

Posted 11 February 2017 - 04:24 PM

It is common practice in calculating the velocity to use in vapor liquid separation is to use an entrainment coefficient of 15 or less. In using this coefficient of 15., the present velocities are more than 10 times faster than that needed for good separation.






Similar Topics