Superheated Gas / Flash Liquid Mixing

[HumbleMan]

I’m a bit out of my comfort zone/ knowledge base and need assistance (maybe to just be educated enough to define the problem completely…)

What is a good way to mix throttled liquid/ flash gas and hot gas in a low side of refrigeration system to achieve a uniform superheated vapor in a minimum amount of space? The initial design has the feed streams injected into the top of separate sides of a plate heat exchanger. At the bottom of the heat exchanger, the two streams are connected and then allowed to run through about 20' of tubing with 4 180 degree return bends. There is a liquid stream rolling along the bottom of the pipe showing incomplete evaporation is evident.

After initially seeing the evaporation was not complete and that there was a significant temperature differential across the heat exchanger, inlet piping to the heat exchanger was connected together. This visually looks a little better but there still is a liquid stream flowing along the bottom of the pipe.

The proportions of flow are 64% superheated gas and 36% flashed refrigerant liquid which is at about 0.13 quality at the beginning of the mixing. Flow velocity is about 2400 ft/min in the 20' section of the tubing.

There is insufficient room to add much additional piping. Would static mixers which I believe are typically used for liquid blending be of any benefit? Are there some other “chemical engineering tricks of the trade” that are typically used for this type of application?

TIA,
Mike

 

[dcasto]

not enough info, but i just guessed at some values on a typical refrigeration and ended up with about 10% still liquid.

You have curious process there. Why not let the exchanger vaporize the refrigerant and let the hot gases mix

 

[HumbleMan]

I inherited the project, the undersized heat exchanger was already installed as was the rest of the piping on this initial commissioning of this test loop. It was initially designed for one refrigerant but will now be used with a different one that is about twice as dense.

Definitley a curious "process", actually it's a compressor test loop to provide superheated vapor for a compressor.

What additional info do you need?

 

[dcasto]

well, we can simulate the process and see how much liquid there should be. what are the temperatures and pressures in the loop along with refrigerant.

As for the curious, you said it, its a test loop to try to artificially load a compressor.

 

[Montemayor]

Humble:

I've done this numerous times. Most of the applications were to cool superheated gas prior to recompression for refrigeration applications. All it is, is simple adiabatic mixing and I've done it with an "engineered" in-line mixer as well as in a pipe itself.

There is no Chemical Engineering trick of the trade to it at all. Mechanical Engineers are just as prepared and adept at resolving it as Chem Es. All that is required up front is a good, accurate heat and mass balance around the mixing pot or point. You basically have two equations:

1) A mass balance; and,
2) A heat balance (done with the fluid enthalpies).

There certainly is no need for a simulation program. I can come up with the correct, accurate results using nothing more than refrigerant tables and a slide rule - but I'll settle for a hand-held calculator. With the accurate and free thermo properties available at the NIST website

http://webbook.nist.gov/chemistry/fluid/

You can make the heat balance using the liquid and vapor enthalpies found there. You know the feed stream; all you have to solve for is the liquid flash stream - which is an adiabatic flash product at constant enthalpy.

Once you calculate the size of the feed and resultant streams, you can easily design the related piping.

At 2,400 ft/min fluid velocity (which I doubt) you are really whistling the stuff through the piping.

 

[MikeHalloran]

I have a little practical experience mixing liquid and gas. Doing it >in_minimum_space< can be a challenge.

That plate heat exchanger will never be as effective as direct contact, as you have proven.

There is a trick to making direct contact work; you have to maximize the surface area of the liquid stream. E.g., force it through an array of pin nozzles that convert it to a sheet, and blast the gas through, normal to the sheet.




Mike Halloran
Pembroke Pines, FL, USA

 

[dcasto]

Cut some slack here Mr M., Chem E's use to use slide rules too.
I was verifying that he doesn't have enough hot gas by-pass, next after iterating by slide rule you'll find the proper split ratio. Next you'll need to size the valve that makes the split so that you have a controllable test and not build up heat in the loop.

Then I assume he'd want to do some thermo work on the test loop to calculate effiencies, man that gets my eyes strained looking at tables for that.

 

[DickRussell]

If I understad the description of your configuration, your original design uses the plate HX to pass heat from the superheated vapor to the liquid to vaporize it. Given the sensible heat of the vapor vs. the latent heat of the liquid, having a 64/36 vapor/liquid mix tells me (and the simple heat balance should confirm the visual evidence) that you don't have enough heat in the system to end up all vapor. If this is the case, you don't have a mixing problem at all, but one of providing heat to the mix. Why not put both streams, or at least the liquid, into one side of that HX and some heating fluid into the other side to provide the heat you need? Do we understand the heat balance side of your problem correctly?

 

[HumbleMan]

Thanks to all for the quick responses, sorry to not get back to you sooner but I took a half day vacation as these are the last days before the cold weather sets into the northern plains, probably until March or April.

Montemayor,
For gas velocities, 2400 f/m is correct for the test case described and there will be conditions where this will exceed 4000. Yes, it does go sonic and whistles at times, overall noise level is usually below 85 dB as sound attenuation is used inside the facility.

Mike Halloran,
Yes, this is the usual engineering dilemma: 10 # of stuff and a 5# bag. The “Pin Nozzle” is concept intriguing: Are you thinking something like a smaller tube with small holes (~#60 drill?) that is suspended inside a larger tube where the hot gas flows?

DickRussell,
I have already repiped once and currently we are mixing the liquid and hot gas together prior to entering both inlets of the heat exchanger. The configuration is the liquid comes in horizontally and the hot gas blasts down vertically from above prior to entering the heat exchanger horizontally and the mixture emerges at the bottom of the heat exchanger. This orientation is probably not best since it forces the liquid against the bottom of the line but it was the easiest way to pipe with the existing piping configuration.

Using mass balance and first law shows that I’m slightly wet but the way that the facility is setup, I should be able to provide plenty of hot gas. Looks like it may be an instrumentation problem (2 separate RTDs were misapplied), will be correcting and see if this helps; will let everyone know what I find.

Haven’t used a slide rule in two weeks since giving a demo for some Boy Scouts while teaching the Engineering merit badge. They now have an appreciation of its power and are thankful they only had to correctly punch numbers into a calculator. They also learned about accuracy and not take more than 2 or 3 digits of a number seriously.

Also posted at another site, their response about static mixers was:

http://www.chemicalprocessing.com/experts/answers/2006/064.html

Will check back in tomorrow, thanks again for the responses.

Mike

 

[MikeHalloran]

No. The beauty of a pin nozzle is that it doesn't need small holes or a large pressure drop.

Imagine, say, a 1/4" diameter simple nozzle with sharp edges. The exiting stream should be fairly well defined. Now, imagine a right circular cylinder, say 5/16" diameter, axially aligned with the nozzle, and held say 1/4" downstream from it. The liquid stream strikes the flat end of that cylinder and spreads into a flat circular sheet, interrupted only by whatever structure holds the pin relative to the nozzle. The end of the pin needs to be flat, abrasion resistant, sharp-edged, and plain. Flutes or other surface features (e.g. like you see on fire sprinkler heads) will make it less effective for the extant purpose.





Mike Halloran
Pembroke Pines, FL, USA

 

[unclesyd]

Here a mixing device we used for mixing a superheated vapor with liquid/catalyst to minimize the mixing time to prevent secondary reactions.

http://www.komax.com/det-gasliquidcontacting.html

 

[HumbleMan]

Mike Halloran,
So if I got this concept correct, the liquid flow is moving in the same direction as the gas flow and it hits a circular target that should atomize the liquid and force it to flow normal to the gas until it hits the pipe wall or is carried downstream by the gas flow. This should cause the liquid surface area to increase so that there is a larger area for heat transfer to occur.

I'm going to sketch something up this morning with a 1/2" liquid tube and a 2" gas line, should be fairly easy to construct by using a tee.

unclesyd,
The Komax mixer looks like it should also work, I will pursue this if my homemade pin nozzle does not work. Time is running short to fix this so I need to try something fairly quickly.

Thanks for all the ideas, I'll post back if they do not work.

 

[dcasto]

Mix all you want but if the enthapy balance says you must have liquid, you will get liquid.

 

[HumbleMan]

dcastro,

I agree, I'm a "law abiding engineer" and would never try to violate any physical law, especially the first law of thermo. In this case, there is enough energy for the heat transfer, just not enough mixing time as there is less than 0.4 seconds between where the flows join and when they enter the test unit with the current piping configuration.

Mike