Similarly, in Toronto, cold water from Lake Ontario is used for building air conditioning.
And after the water is used for cooling, it’s “pumped to the city’s water filtration plant for treatment and distribution to residential and commercial users.”
When we lived in Arizona, years ago, the water supply was warm enough that my parents didn’t use the water heater. Tap water was already warm enough for showering or doing laundry or dishes. They started using heated water when they bought the dishwasher.
An interesting video that gives some perspective on the issue…
Care to tell any of us what the 20 minute video says is the problem, if it’s not water?
Sure, I apologize for not doing so earlier.
It’s not that water isn’t a problem. It’s just not as big of an issue as it seems…overall. Sure, in some areas, it IS a big problem, like where water sources are already strained. But don’t let those few very real stories cloud you’re overall judgement of data center water usage, which is miniscule compared to other various industries. Also, power usage is the larger concern, etc.
“From a distance, it would probably look like an immense can of Red Bull.”
That would explain the hundreds and hundreds of 14 year-old boys playing Counter-Strike 24 hours a day…
What I struggle to understand is why these immense buildings aren’t all being built 40 feet underground to START with. Geothermal constant and all.
The water would be at 50º f. The building would be as well. Is the Earth itself much less efficient at cooling water than evaporation?
In the longer term yes.
Thermal conductivity of rock poor. Specific heat is OK. Most rock specific heat is about one fifth of water at ordinary temperatures. Which isn’t bad. Thermal conductivity of solid rock varies a lot, but in the range 2-5 W/m/K. Contrast that to copper at about 400. Dry sand is 0.6 because of all the pore space. Wet ground is better but not great. This is part of the problem.
You can use the ground around you to provide a useful thermal sink for a dwelling or ordinary building. During the summer you can dump heat into it, and in the winter pull heat out. Usually with a heat pump system. The coefficient of efficiency obtained is really good. But you can’t exceed a fairly modest energy flow . Which is why an annual cycle of driving heat in and out works. Continual injection of heat doesn’t work.
Whilst the volume of earth defines your heat sink capacity, its surface area defines its ability to conduct heat away. Which is why it doesn’t scale. Nor can you go deep as the temperature of the earth increases again past a couple of hundred meters depth. This also places a bound on where your injected heat can go. So things don’t even scale with surface area, but eventually only scale with length.
Tl;dr. You can’t dump megawatts of energy into the ground. It will just heat up vastly faster than it can conduct away.
Re: data center cooling. I just read this article about elastocaloric cooling.
Either that just acts like a heat pipe (which we already have, and which work quite well), or it results in needing to dump even more heat into the environment. In neither case does it help the actual problem with datacenter cooling.