The Winds of Change: Wind Energy PPAs Are the Future for Data Centers

But now a few tech companies are using green energy to solve the power supply problem.

No they’re not.  Not one company, tech or otherwise, is relying on “green” (wind and solar) energy to actually run their operations.  It’s all greenwashing.  Here’s how you tell:

1. Does the tech operation, such as a server farm, operate on intermittent and interruptible power flows?  (Of course not; they all require 24/7 power.)
2. If the answer to 1 is “no”, do fossil fuels provide the backup for the “green” energy supplies?  (Unless the backup is hydro, the answer is “yes”.)  This is proof positive of greenwashing.

Companies in any business can purchase “renewable energy credits” to cover their consumption from the grid, but this is a paperwork dodge; they are not actually using all “green” power, and probably couldn’t keep their operations going if they did.  Further, the fossil plants that have to ramp up and down (but can’t stop operating) to follow the waxing and waning of fitful wind and fickle sun run less efficiently and burn more fuel than if they were run closer to their optimum.  Thus, “green” energy on the grid does not translate directly to fuel not burned and carbon not emitted.  It is all an elaborate fraud, pushed and even mandated by policy and subsidized by the taxpayer.

The wind farm that HP is drawing from is 300 MW. If they get first dibs on generation,it’s not out of the question that 95 percent of the time it will generate more than 112 MW which would be a 37 percent CF. If a 150 MW SMR was planted central to the data centers to limit transmission (one of the theoretical advantages of SMR’s), I doubt it could hit the same 95 percent CO2 free power generation for HP over a 20 year contract period. So if I had to make the decision for reliable carbon free power, I would lean towards wind,solar,lots of transmission, with as little hydro and gas back up as possible rather than depending on one SMR.

he wind farm that HP is drawing from is 300 MW. If they get first dibs on generation….

Consider the rest of the grid for a minute.  Is anyone ELSE operating on intermittent and interruptible power flows which are only the surplus from the “probably on” fraction that the data center has first dibs on?  Or is everyone and everything expecting “always-on” power and relying on fossil-fired generators to make up the difference?  Of course they are.

So if I had to make the decision for reliable carbon free power, I would lean towards wind,solar

If you actually had to make it work, you’d either pay through the nose or fail.  Nuclear and hydro are the only options today, and nuclear is the only one that scales.

Texas wind, in 2013, generated 36 million megawatthours?!

The figure represents a scant .88% of U.S. consumption for the year – a waste of money and resources about which even Facebook can’t get me excited.

Bob, you may or may not be correct about the extra cost, but the consumer decides what is worth the extra cost, not policy makers, economics or even the science (though science should be most important) behind it all. There are many, many consumer choices that are a waste of money to one person but not another.

the consumer decides what is worth the extra cost, not policy makers

Apparently you haven’t heard that the policy makers have subsidized and even MANDATED the purchase of unreliable power regardless of cost or actual usefulness.  It’s called “production tax credits” and “renewable portfolio standards”.

I’m strongly tempted to tell you to go play with the other children; that’s the level of ignorance you’re working on.  Or are you outright pushing disinformation?  It’s hard to tell sometimes.

    ‘…it’s not out of the question that 95 percent of the time it will generate more than 112 MW which would be a 37 percent CF’.

That’s not how wind works. The Vestas 100 2.5 makes 2600kw at 14 m/sec wind; at 7 m/sec ( 15 mph )  it only makes 700kw – wind power scales with the cube of the wind speed. At 3m/sec it makes nothing. Here’s a graph of texas wind output –

http://www.eia.gov/todayinenergy/images/2014.06.23/main.png

In 2014 it was occasionally peaking at 10 Gigawatts, but often dropping below 2, and that’s for the whole state, not one wind farm.

I decided to analyze your assertion and took as my test case the windy month of March in the windy state of Texas.  Short answer:  you’re wrong.

http://ergosphere.blogspot.com/2015/10/how-unreliable-is-wind.html

“If a 150 MW SMR was planted central to the data centers to limit transmission … rather than depending on one SMR.”

The point of SMRs is not to avoid grids; grids are a good thing.  A more helpful example would be a city with 500 MWatts of baseload demand.  One option would be to share a 1000 MWatt nuke with a neighboring city, and install a 500 MWatt transmission line, and 500 MWatts of spinning reserve (baseload portion only).  

Instead, we can use ten 50 MWatt SMRs and only 50 MWatts of spinning reserves.  The transmission link to the rest of the grid is still a good idea, as it allows sharing of the reserves, but it can be smaller, and the average loss can be lower.

In the case of the data center that wants reliable and clean power, the reliability comes from the grid in both cases.  The important difference is that in the nuclear case, after the data center takes their portion of the plant’s power, the remaining portion is still baseload with optional load following.   In the windfarm case the remaining power is extremely variable, with even lower capacity factor than the raw windfarm output (as EP has shown in his excellent analysis).

Thus because the windfarm does not produce power in the form the datacenter needs, and the datacenter did not buy power smoothing, for example, from a pumped-hydro facility, the datacenters “100% clean energy purchase” is essentially fraudulent: they have guaranteed that an amount electricity will be produced by fossil fuel which actually exceeds the amount they used, in order to balance their “clean energy” purchase.

Great point regardnig SMRs Nathan.  As you describe it, the cost savings between big nuclear plus spinning reserve and the SMR solution approaches 50% at the grid level, something operators must be keen to realize.