> No details about the planned solar panel array or its potential power output are known at this time
Taking the 171 acre figure we can compute at least the limits of what it could do.
171 acres is about 700,000 square meters. At a solar incidence of about 1 KW / square meter and an efficiency of 23% that's 161 MW peak power and an average of 4.7 sun hours per day will give about 750 MWh of usable electricity every day.
This calculation ignores shading effects and assumes that the panels will be stationary rather than tracking, you could probably easily deduct another 50% to account for those effects, so say 375 MWh give or take and 80 MW peak. Likely these figures are still very high, think of it as the maximum amount of solar power that you could extract at that location from incident sunlight without further accounting for conversion losses (solar panels produce DC, you'll need inverters if you want AC power, some data center equipment can run on 48V DC but you'd still need a conversion step for that) and so on.
I wonder how big a chunk of the total power requirements for a datacenter that size will be covered by this PV system, to be 100% effective it would have to be sized at least 6 times larger than the total power consumption of the DC, using the grid as a storage system.
WolframAlpha measures an average US household's yearly consumption at 12,000 KWh, or 32.8 KWh per day. The upper bound in the parent comment with the 50% fudge factor, 375 MWh per day, would power (375 MWh/32.8 KWh) = 11,430 households day-for-day, or 22,860 households day-for-day without the fudge factor.
You can feed to, and take from the grid, but it's not a battery.
The Grid is like an open water pipe with holes in it - you can put water back in, or take water out, but you can't store water in it. The power stations are pumping in at one end, and the users are taking out all along the pipe.
I can't imagine any scenario where Greenpeace would actually be happy besides companies never using any sort of energy at all.
To replace coal (7PWh) with solar energy, you'd need the entire state of New Jersey (~22e9 m^2) filled solar panels averaging 20% efficiency and the equivalent of 4 peak hours of sunlight over the year. With the area of Pennsylvania you could start to approach the total energy consumption of the US. Maybe those environmental impacts would be worth it?
You know, I recently rode my bike through some villages in Germany. Just about every house had solar panels on their roof. My parents have solar panels on their roof. They produce more electricity than they use and only need very little oil and some wood for heating in winter. And they are no exception.
Germany currently produces 20% of its energy using regenerating resources (water, sun, wind).
It's not like you have to put all the solar panels in huge parks. Just put them on roofs. My uncle works at a big company and they just put solar panels on their roofs. That company now produces more energy than they need. Solar panels are already efficient enough to usually provide more energy per area than people or businesses consume in the same area.
The next step would be to have batteries in every household, to save on energy losses during 'transportation'. Also, solar panels are getting more efficient and cheap every year.
Perhaps this is a silly question, but do we have any information on available roof space across the United States? I'd be interested to see what kind of power output could be produced if we were able to add solar panels to a large majority of the nation's buildings.
Though I suspect that kind of distribution would significantly increase costs for duplicate components and may fall victim to low efficiency transfers into the grid. Afraid I don't know enough about how today's systems operate.
This would be a great trend, datacenters becoming their own utilities. It would help them better control opex and reliability. I wonder if utilities will try and prevent this sort of self-sufficiency from otherwise potentially lucrative customers.
Sunlight is one of the least reliable forms of power out there. If the grid goes down then... well, you're screwed unless it just happens to be daytime and sunny.
Knowing Apple's recent interest in glass and materials science I can't help but wonder if even their solar farm will be awe inspiring and functionally ahead of the curve in some way.
Taking the 171 acre figure we can compute at least the limits of what it could do.
171 acres is about 700,000 square meters. At a solar incidence of about 1 KW / square meter and an efficiency of 23% that's 161 MW peak power and an average of 4.7 sun hours per day will give about 750 MWh of usable electricity every day.
This calculation ignores shading effects and assumes that the panels will be stationary rather than tracking, you could probably easily deduct another 50% to account for those effects, so say 375 MWh give or take and 80 MW peak. Likely these figures are still very high, think of it as the maximum amount of solar power that you could extract at that location from incident sunlight without further accounting for conversion losses (solar panels produce DC, you'll need inverters if you want AC power, some data center equipment can run on 48V DC but you'd still need a conversion step for that) and so on.
I wonder how big a chunk of the total power requirements for a datacenter that size will be covered by this PV system, to be 100% effective it would have to be sized at least 6 times larger than the total power consumption of the DC, using the grid as a storage system.