Piping Thermosiphon-Oil Cooler

[gwapa]

Dear Friends
I appreciate your comments regarding the size of the liquid a return ammonia pipes from thermosiphon to oil cooler and viceversa

ASHRAE give us some diameter as a funtion of the load (Tr) but does not give us the criteria.

Thanks for your comments
Regards
Gwapa

[Dacosta]

I generally design my thermosiphon lines with the typical 2 PSI loss per 100ft (liquid) and 1 PSI loss per 100ft (suction) plus one size larger for wet suction. The thermosiphon is like a flooded system, flow rate is typically 3:1 up to 10:1 or more, but generally I stick to 5:1.

You can convert your oil cooling heat of rejection (OCHR) to tons and use the ASHRAE tables at 40ºF suction and 5:1 liquid flow rate from your liquid lines.

Or you can do it the long way and calculate your refrigerant flow rate and calculate your head loss to select your pipe size. If you want to go this way, I would recommend you download “System SYZER” from Bell and Gossett. Although you will have to convert your flow rate to Gallons per minuet (GPM) and look up Viscosity and Specific Gravity (SG) for the refrigerant. This typically gives me the same values from the tables, but a good exercise none the less.

Dan Acosta

PS. I’m too new so I can link anything.

[US Iceman]

There was a very good book written some time called "Thermal Hydraulics". It was bound in two small volumes. The reason I bring this up is the risers are a complex model. What works at full load might not work properly at part load. As the oil cooling load decreases the volume of gas is reduced. This affects the drag force ability of the vapor flowing up through the riser to carry the liquid up.

The pressure losses I use are a lot lower than 2 psi/100 ft. I generally use a 3:1 rate. Overfeeding more liquid does not improve the heat transfer or make the system work better. However, a 5:1 rate should not hurt anything either.

At full load I believe a larger diameter riser (one pipe size larger than liquid feed pipe) is OK. However, at part load it might be better to use a smaller pipe size to keep the flow regime in annular flow. If the velocity drops to slug flow the cooler can quite working.

[gwapa]

Thanks Friends
The topic looks to me very complex.
In the Oil cooler we can figure out:

Q=m(hl-hg)
Q= OCHR (known)
m=flow rate(unknown)
hl= liquid condensation(known)
hg= entahalpy gas outlet (unnown)

so the flow rate(m) is a funtion of the gas entalpie outlet (hg) .This is a hyperbolic funtion, that is biger m less hg
If we fix m with a recirculation rate of 5:1 we automatically fix hg and all the ammonia propierties.
Are there any literature about the rate of circulation in this tipe of applycation?
Thanks
Gwapa

[US Iceman]

I think what your are looking for is the void fraction. This is the mixture of the liquid and vapor if you assume it to be homogenous (as in a single uniform two-phase mixture).

The void fraction is a ratio of exiting liquid and vapor mass from the oil cooler. Using this, you can determine the mixture enthalpies from a set of rules called the lever laws (or something like that). The reason it is called the lever law is because this process is like a simple lever. If you move the fulcrum, the lever is easier to move or will move more mass. As the void fraction changes (the fulcrum) the enthaplies change more or less..

Then you use the new enthalpies (in two-phase mixtures) to determine the slip velocity for the volume of gas moving through the riser in relation to the volume of liquid being dragged up the pipe.

Ideally you want the flow regime to exist in an annular flow. Gas inside the core of the pipe with liquid flowing along the inside pipe wall.

One way to look at this is like the system is a gravity flooded air unit for air cooling. The surge drum is the pilot receiver and the air unit is the oil cooler. In this application both pipes are typically the same size. I think the bigger issue is making sure you can get liquid to the oil cooler myself.

The last oil cooler is always a problem if you miss anything!;)

[Dacosta]

The flooded system occurs at constant pressure, therefore you know the vapor pressure and therefore you know hg (saturated vapor)

Dan Acosta

[US Iceman]

The flooded system occurs at constant pressure, therefore you know the vapor pressure and therefore you know hg (saturated vapor)


So does a thermosiphon for all practical considerations. The only difference is the temperature and where the liquid comes from.

One of the big differences is how everyone considers the two processes to work. In basic terms they both operate on the same principle. Void fraction and all other technical jargon is the same. However, for many years manufacturers designed flooded air units and they have worked quite well, if you keep the oil drained out of them and the liquid supply is adequate.

[gwapa]

it is very interesting thanks

[US Iceman]

I forgot to include some other remarks in my last post...

One of the big differences between gravity flooded & what we call thermosiphon is this:

In a gravity flooded system the vapor is pulled out of the surge drum by the compressor. Therefore pressure drop in piping is not a huge concern.

In a thermosiphon the vapor flows back to the condenser by pressure difference alone. This makes it critically important for the vent line form the pilot receiver to be large enough (greater diameter) to support very low pressure drop. If the pressure loss at full load is too high you change the boiling temperature of the liquid in the oil cooler. That can raise the oil injection temperature and also decreases the LMTD of the exchanger!

Both a gravity flooded and thermosiphon will operate qoute well if the piping is designed correctly. Generally though, if the piping is not designed properly you can see problems arise.

[gwapa]

Thanks US ICEMAN
You know when we are dealing with two fase flow it is very hard to figure out any pressure drop ,specially when the pipe is slopping. All the two fase flow equation are for horizontal or vertical flow (Martinelly)
Some autor recomend for sucction or return to the termosiphon 2xD of normal suction gas.
For the liquid I will use a velocity as low as 0,3 m/s

But even selecting this diameter there are no way to figure out the pressure drop acurely.

Best Regards
Gwapa

[US Iceman]

Hi Gwapa,

When I designed a large thermosiphon some years ago I did a lot of research on the subject. One of the papers I reviewed said not to install slopped piping, and of course that is what the owner wanted to do (on a 45° angle). It seemed to work OK during the complete load range so I was happy with it.

I think the safest way to treat the risers is on superficial gas velocity (gas only). Trying to find the pressure drop takes more time than it is worth in my opinion. And besides, after doing all of the calculations you do not know how close your guess is!

The liquid piping is more critical I think. You need to ensure the flows are balanced and the piping is symmetrical between all units. Otherwise you can get preferential flow to some units. I normally use very low pressure loss allowances for the liquid and make the riser the same size as the liquid supply.

One thing people sometime forget is this is a gravity drain system, not a pressurized line where you can loose a bit a pressure without too many negative effects.

[CanadianIce]

Hi US Iceman

After reviewing the paper regarding not using slopped lines. do you now prefer slopped or not slopped? I've heard of both methods being used but only slopping the return line. Some say slope the retiunr to allow the liquid to run down to the next riser where it will collect and the vapor will push it up, but wouldn't this be cosidered slug flow? Other schools of thought suggest allowing the vapor to carry the liquid along the horizontal lines without slope. Would I be correct in saying that the size of these horizontal portions of the return line would be critical in order to maintain this flow and as you stated earlier, this flow rate would change during lower loads so it may be difficult to size this line correctly. I have never designed a thermosyphon system but I find the topic interesting because I have had to make changes to some existing systems that were not performing well.

Cheers

Canadian Ice

[gwapa]

I prefer slop both pipe the liquid and the gas. Both are in saturated state so any very small pressure drop will produce a bag of gas and it will plug the flow. When you slop the pipes the bag of gas has the chance to return to the termosiphon vessel
Gwapa

[US Iceman]

Let me try to be more clear. When I say slopped I meant inclined.

Slope would mean a small change in elevation from Point A to Point B. This would be mm per meter or inches per foot.

Inclined would be a drastic change in elevation from Point A to Point B. The paper described large changes such as 45° angles. The riser seemed to work OK, but it went against everything the paper stated! As long as it worked I was OK.

Slopping lines is always a good thing. Let gravity help you. Liquid supply piping always has bottom take-offs for branch lines. Return lines always have top of pipe entrances to allow the liquid to drain by gravity. These two things will prevent a lot of problems.

If you wanted to be really particular on the horizontal return piping you could treat it as sewer flow. A liquid puddle running along the bottom of the pipe with vapor on top of the liquid. Just like a two-phase wet suction line!

[gwapa]

Thanks US ICEMAN
I agree with you
If you find the paper you talk about please send me a copy.I really appreciate it
Regards
Gwapa