New Data Center New Data Center - Page 12 - Talk of The Villages Florida

New Data Center

Reply
Thread Tools Search Thread
  #166  
Old Today, 03:38 PM
ton80 ton80 is online now
Silver Member
Join Date: Oct 2012
Posts: 189
Thanks: 39
Thanked 168 Times in 78 Posts
Default Some Basic Principles and energy/utility demands for a 1 GW Data Center

Quote:
Originally Posted by OrangeBlossomBaby View Post
In an actual closed-loop cooling system, the water evaporates into pipes - which then channel the vapor back into the system. That's part of what makes it a closed-loop system.

However, these kinds of systems require MUCH more energy, which means a) more drain on the public grid, OR b) more and noisier generators and the required fuel to run them, OR c) some combination of the two.

There's an interview with David Cote in the October/November issue of Fortune Magazine. Cote runs Vertiv, the most successful manufacturer of data center cooling systems in the country. I recommend it to anyone who has access to the article. I have a paper subscription to Fortune, and don't create accounts online to anything I have a paper subscription for. So you'd have to google it yourself and see if it's available to you. In the paper issue, it starts on page 76.
OBB From a basic principles standpoint I do not see how the water in the closed loop gets hot enough to vaporize. Boiling point at atmospheric condition is 212F. How can that be created in the cooling system?

More to the big points:
1. Utility demands such as power and water makeup can occur at a remote site not in Sumter County. Power Gereation can be elsewhere and power deliverred to new Dta Center.
2. The new Data center designs are very different than the older smaller Centers as pointed out in earlier posts.
3. I tried to get typical cooling water supply and return temps but there are no consistent designs. The electronic equipment has varying needs. It far beyond my comprehension level to try some categorization.

Some Basic Principles for consideration.
1. Closed loop does not mean self contained. If the heat removed from the electronics and transferred to the cooling loop liquids, does not get rejected, the data Center will Not function. I believe OBB agrees with this statement.
2. Recirculating Cooling Water system cooled over a cooling tower, my example, and OBB hybrid system are essentially the same in principle. This is the cheapest equipment cost but requires a lot of water
3. The recirclated water in the closed loop is highly treated with chemicals to avoid fouling and corrosion. The liquid has a lot of glychol which means it can be used for low temps below 32F if desirable for the electronics.
4 Using outside chillers saves water at the plant , but this additional power requires water at the power generation site.requires more equipment, and power consumption. It generally more expensive.

Now for numbers to be used for understanding the size of the impacts for a 1GW Dtat Center.

a. Power Generation Modern Plant Combined Cycle
1.64 GW (2 at .82GW) cost 1.3 Billion$ in 2019 Crystal River Station.
b. Power Station combined cycle efficiency is about 55 to 60% but there are distribution losses. So need to buy about 2GW worth of NG rough number
c. Power delivered to Data Center is mostly turned into waste heat that must be removed. So need to extract and dispose of 1GW heat from the recirc water.

b. Power sent to Data Center eventually needs to be removed from the recirculating water. Water is enclosed, heat is NOT.

So my bottom line is that a 1GW Data Center is a big impact some at site and other impacts at power gen and water supply.

There are a number of vendor/designers for these cooling systems with various schemes to reduce cost. But the Devil is in the details. I am signing out.:
Reply

Tags
water, center, proposed, data, trash


You are viewing a new design of the TOTV site. Click here to revert to the old version.

All times are GMT -5. The time now is 03:39 PM.