Embedded Files
Data centres are now some of the heaviest users of electricity on Earth.
Row upon row of processors and memory cards each generating substantial amounts of heat as they search for, store, and allocate information.
This heat has to be regulated and controlled.
This needs other circuits. Liquid cooling circuits, and one isn't enough.
Cooling circuit one is removing the heat on the circuit boards by using liquid that is pumped around its own 'liquid' circuit. The heat goes into heat exchangers then back to cool the chips and circuit boards again.
Cooling circuit two: the heat exchangers themselves need cooling. This is sometimes fans but can be another (secondary) liquid cooling circuit.
Cooling circuit three. The pumps that push the cooling liquid around need cooling. So do the fans.
And so it goes on. Forever. An endless need for cooling, all of it electrically powered. More data storage, more heat extraction needed. More cooling. More electricity.
Eventually all this waste heat ends up the atmosphere. It takes a lot of energy to keep data centres operating properly.
We have another solution.
Data centres and riverside brownfield land
If we look again at our tidal ideas they are well matched to data centres.
If, instead of building houses on the roof, we turn this space into a base for data centres we can offer the data centre not only round the clock electricity - megawatts of reliable, clean power free of supraharmonics, - but also unlimited cooling. Instead of consuming vast amounts of electricity for fans and pumps, with cooling equipment needing cooling itself, we use the cold water of the tide as it ebbs and flows.Cooling to liquid offers an efficiency improvement of around 1000 times over cooling to air.
For most basins we have been considering we have water 'turnover' of around 20,000 tons an hour. The water is cold but never freezes.
This is is the actual temperature taken over a year on the river Tyne near Newcastle:
The cooling capacity of cold water
It's worthwhile doing a simple calculation just to show how effective such a huge cooling facility can be.
Let's say we want to remove 10 megawatts of heat. This is just an arbitrary number to start with, and this amount is produced constantly. We have 20,000 tons of cold water available.
If we want to remove all this heat using the cold water, how much would the water warm up?
We have that it takes 4,200 joules of energy to heat 1 kilogram of water up by 1 degree C. We can call this number 4,200 our "unit" for now.
10 megawatts is a lot of energy. But 20,000 tons is lot of water.
A joule of energy is just 1 watt of power per second. So this arbitrary 10 megawatts represents (about) 2380 of our "units" (keeping to reasonably round numbers).
If 1 "unit" of our energy can heat 1 kilogram of water by 1 degree Centigrade and we have 20 million kilograms then this 2,380 "units" is frankly going to have an absolutely negligible effect on the temperature of the reservoir:
Temperature rise per second: 20 million divided by 2,380 which is (close to) 0.00012. Over an hour (3,600 seconds) the rise is 0.43 degrees C.
The turnover of cold water is of the order of the number above: 20,000 tons an hour.
The cooling capacity is virtually infinite. The water that has been 'heated up' by less than half a degree disappears and a fresh amount arrives. And at the same time as the water is absorbing this heat the flow is also generating electricity.
***************************************
We can also use heat interchangers; two circuits interacting through thermal transfer, but not actually intermixing. This would allow the 'waste heat' from the data centre to be used locally, for example in heating circuits for housing estates, food production, or other industrial purposes - such as hydrogen production - shown in the 'overall capabilities' block diagram on page 1.
Report abuse