Using geothermal to cool pipes

[cookyb]

I wanted to get some help on setting up a cooling system for a saltwater aquarium. The tank water is 82F and needs to be held constant by the use of heaters and some type of cooling system. They make chillers for this use but they are upwards of $500 and I figured I could make something for cheaper than that.

My proposal is to pump the water from the tank through a pipe that is X' long and use the crawlspace under the house to cool it, either by just laying the pipe in the crawlspace or burying it. The crawlspace temp ranges from 60F-65F and I'll need a flow rate of approx 6 gallons per minute of 70F water to cool the tank. My question is, how long (or how do I calculate) the length of pipe I will need to cool the 82F water to 70F before it returns to the tank?

I was initially hoping to use copper tubing but I've heard stories of it leaching copper into the water which can kill the fish so I may have to use PEX, which I know is much less efficient at transferring heat through the walls.

Any thoughts or suggestions would be helpful. I think the time it will take for the water to cool may keep this idea from working but I thought I'd check into it first.

Thanks,
Steven

 

[ivymike]

regardless of how you design the cooling loop, I would suggest that you limit the depth of water which can be siphoned from the tank in the event of leakage from the system (that way the tank doesn't dry itself out suddenly one day when you're out).

 

[jmiles]

Hi Steven,

I jsut tossed some numbers into a spread sheet i have, and well alot fo thigns are different, but using a 2" carbon steel pipe with your given flow range and temperatures, and keeping in mind that CS will conduct heat away better than plastic will, and assuming you just lay the pipe out on the floor, i figure you will need about 1400 meters of pipe, so yeah go buy the cooler its much cheaper.

The thing is your working with a very small delta T here and water holds alot of heat that you want to dissipate. free convection tot he air from a pipe is a poor way to do this. and i can easily see any other methods quickly going well past 500 dollars.

Plus im not sure how constant your crawlspace temperature will really be i mean do you have a sampling form all year round?

 

[cookyb]

Haha! Yeah, that's a lot of pipe. I don't have a temp sample from year around, but I do have one from the summer which would be worst case (and probably the only time the cooler would be in use). Can I use your spreadsheet one more time? How about if we used 1/2" pipe in 40 degree water?

Thanks a bunch for looking into this,
Steven

 

[mcgyvr]

What size tank?

 

[cookyb]

180 gallon.

 

[jmiles]

Hi Cookyb,

I am assuming you meant putting the pipe through a tank of water held at 40°F?

If thats the case you would need about 20m of pipe (my spread sheet is set up for 1" as the smallest so i went 1" sch 160 to approximate the flow volume of 1/2")

So this is a much better scenario to consider, but your water sink will warm up over time unless your constantly changing it, and if your going to that kind of trouble why not just bring in cold make upw ater, and build a little mixer? unless the water is treated with chemicals and such... Just a thought either way.

 

[cookyb]

That is exactly what I meant. I'm thinking of a small fridge unit to keep the water cold and run the piping through it. Was your spreadsheet using the thermal conductance of plastic or steel for the second run?

 

[jmiles]

It was using steel as well for the second one, but i will dig into my sheet a bit more and see if i cant change the thernam conductivity to that of plastic. (Currently the sheet is automated to use steel and it varies with temperature through some calculations)

Ouch.... I plugged ina thermal resistance of .19 W/mK from

http://www.engineeringtoolbox.com/thermal-conductivity-d_429.html

and it jumped up to 90m of pipe.

With steel at 54 W/mK i guess this is to be expected, sorry i should ahve dug itno the spread sheet a bit more first. you really will need to go to a metal pipe for this application, or just introduce colder make up water.

FYI i used 1" sch 10 for the thickness, of the PVC.

 

[cookyb]

Good to know. Thanks for all of your info!

 

[jmiles]

No problem, glad i could help a little.

 

[Zapster]

cookyb,

Perhaps I do not understand the basis for the question. It sounds like you are trying to circulating 6 gpm of water through the tank with a 12 degF temperature drop. How are you generating 36,000 BTU/hr in a 180 gallon tank? Have you considered providing some type of shade for the tank?

 

[cookyb]

The thing with keeping reef fish and corals is that you have to mimic (as close as possible) the light and intensity of the sun in order for the corals to grow. In doing so you end up with some pretty powerful lights that need to run about 6 hours per day. Not to mention the circulation pumps, filters etc. All of this builds up heat in the tank.

I only need to lower the temperature of the tank about 5 degrees. The system I was describing before wouldn't always be running; only when the temperature was getting too high and it would kick on.

 

[jmiles]

Yeah i think given that it does not need to run all the time just bringing in cold makeup water with perhaps a mixer would be better.

 

[moltenmetal]

Figure out the wattage of the lights, pumps etc. Assume all their input energy ends up in the water- that's a worst case assumption.

That will give you the duty. Cp x m dot x delta T will give you the delta T you really need to keep up with this heat load.

Chances are, once you're more realistic about the duty, radiative loss alone out of a reasonable length of PEX tubing will give you the delta T you really need. Once we know the duty we can do the calc for you, or at least teach you how to do it yourself.

Be careful about the end connectors used with PEX tubing, though, as they're often brass (both Zn and Cu are phytotoxic if I remember correctly). Wirsbo do sell end connectors that are non-metallic, but only limited types etc. For a recirc situation like what you're talking about, non-metallic external compression fittings (Jaco etc.) will also do the job.

 

[cookyb]

I figure I'm looking at a maximum of about 600 watts at the peak (about 6 hours a day), 150 watts for about 6 additional hours and 100 watts during the night. I have some fans that help vent the heat due to the lights (which account for about 450-500 watts) but a worst case scenario would be a good way to go in case something happens and the fans die.

Being a structural guy I don't get into these kind of calcs much (read: ever). I would like to learn it, though.

Thanks!

 

[Zapster]

600 watts is around 2050 BTU/hr. With a 6 gpm flow you now need less than 0.7 degF drop in temperature. This still may be high if the temperature in the tank is higher than the room temperature. There may be some cooling from the surfaces of the tank.

 

[cookyb]

Is this the correct equation to determine heat loss through a pipe? q=2*PI*k*(to-ti)/(ln(ro/ri)) where to=outiside temp, ti=water temp, ro=outside diam, ri=inside diam and k=thermal conductivity of the pex?

I think I read that k for pex is around .26 Btu/(ft hr F). Can anyone confirm this? If so, my q would be about 80 Btu/hr resulting in only needing 25' of pipe to balance the heat gain due to the 600 watts. This seems low to me. Any ideas?

 

[Zapster]

You have neglected the film coefficients on both the inside and outside of the pipe.

 

[danberry]

How about using an automotive radiator and an electric fan