SoylentNews is people
SoylentNews
As technology within the energy space evolves, solar farms have become a key production source across the world. For example, solar farms in the United States produce energy and a whole lot more, and solar farms in Britain have expanded to deliver impressive amounts of energy in their own right. Unfortunately, one solar farm — the largest in the region — and the energy it produces have supposedly proven too much for what the area's power grid can safely handle. Thus, the North Devon-based Derril Water solar park has been temporarily shut down until September, without any warning or input from those behind it, in the hope of keeping the local grid intact.
Per The Guardian, the order to shut down the Derril Water solar park came from the National Energy System Operator, which ordered National Grid to pull the plug for the remainder of the summer. Concerns of a possible thermal grid overload, involving the movement of excessive energy through a system that generates too much heat and can damage key components, backed the order. This comes after issues with the North Devon power grid were highlighted back in 2023, though the necessary improvements scheduled for installation at the end of 2025 have been pushed out to this September at the earliest.
As reasonable as the rationale for temporarily shutting down the Derril Water community solar farm may seem, it comes with major consequences. Energy production and the livelihoods of those involved monetarily with the solar farm are at stake.
Even though the reason for the Derril Water solar farm's shutdown is tied to grid issues, the consequences of its stoppage fall on those with no responsibility for the grid itself. The first issue the shutdown presents is a lack of solar energy production. Not only is this a negative due to a loss of clean energy accumulation, but this translates to a serious dip in revenue, with estimates placing the lost funds at around two million pounds.
With that amount of money lost, those invested in the solar farm will feel the effects of the shutdown in their wallets. Roughly 9,500 individuals will likely lose out on dividends paid to them for their financial investment into the solar farm in the near future. The Derril Water volunteer board of directors told The Guardian that most of those individuals are understanding of the situation and know that the issue lies with the power grid and not the farm itself. Still, there is plenty of frustration going around, especially since grid operators and insurance companies aren't expected to pitch in over the lost revenue.
As effective as they can be, solar farm shutdowns of some kind are occasionally deemed necessary. Sometimes it comes down to solar energy technology, like in the case of California's $2 billion solar plant shutdown, or it can be a connected grid infrastructure issue. In this situation, the North Devon grid seemingly couldn't handle what the Derril Water solar farm could provide.
(Score: 2) by Rich on Tuesday August 04, @05:13PM (4 children)
Well, data centers (aside from possibly higher cooling requirements in summer) continuously pull the same load. So Deeplearning Datacenters, Inc. writes a tender for supply, and Abundant Atomics Corp. and Pervasive Photovoltaics, Inc. submit (assuming post-crossover point battery prices). AA quotes 12 ct/kWh all year round. PP calculates with 6 ct/kWh for March-October, 24 ct/kWh for November-February. In the first round, both insist on full-year contracts.
On that calculation alone, PP would have already won. So DD asks a second time "what can you do about your quote?". AA replies: "Nothing, because if you don't fully commit, we don't get the financing, and, btw, we'll only be online 10 years from now". PP replies: "Well, we can quote you 5 ct for Summer, but then you're on the spot market in winter.". DD looks at their turbine backup (datacenters are a special case, because they need 100% backup), which produces at 8ct Natgas * 3 efficiency factor = 24ct/kWh, too. Deal closed. Eventually PP might also undercut DD's turbines, and they would switch.
Or if you look at this this way (overnight price calculation): 40 years of seasonal storage = 40 charge cycles. If the battery costs $10, a stored kWh comes at 25ct.
(Score: 1) by khallow on Wednesday August 05, @03:27AM (3 children)
(Score: 2) by Rich on Wednesday August 05, @02:01PM (2 children)
Next to their power plant? Well, they'd be online in 10 years when the power plant is done at 12 ct cost. Meanwhile PP got FOMO, joined the game, got online in 3 years, covers their winter gap with turbines, is already cheaper and binds the customers. After 10 years they have scaled up their batteries and sell their turbines. Complete loss for AA even if the battery price crossover point is not reached and PP resort to fossil gap filling. Which really isn't a CapEx, because in the case of data centers, they need 100% backup anyway. After the crossover point PP would go straight to batteries and save the turbine cost.
Data centers are really the worst examples here, because the NPP has to go down refuel and they need turbines, whereas the battery solution is modular and if big enough can bear the entire backup requirements alone. It would only be interesting before the crossover point, and for a baseload application that doesn't need 100% backup and that can live with a day or two downtime. Or if PP are extremist greenies refusing to add gas turbines even before the crossover, and plan for going online at the same time as AA. Then it would be a gamble if the batteries are cheap enough in 10 years or not.
(Score: 1) by khallow on Thursday August 06, @11:52AM (1 child)
That would be the plan. IMHO now you're talking about the real problem. The enormous capital investment and lead time presently required for a fission power installation. Cut both of those considerably and its back in the game. Not saying that will happen because we wouldn't be in the present situation with an 80 year old technology if it were trending that way.
I still think demand shaping is in the game with data centers. Sure, a lot of stuff needs to be always on. But crypto and AI both can use computation power that can be turned on and off with little to no warning.
(Score: 3, Interesting) by Rich on Thursday August 06, @02:41PM
It's said that SMR are some kind of salvation here, but their pressure vessels aren't much different from what a VVER 440 already was. I don't see much of a change, at least as long as the requirements for containment stay the same. Also, I think any of this molten salt stuff is right out, because that's a huge mess of corrosive chemistry vs. a plain water kettle, and a magnitude harder to handle. So, very unlikely.
That's a very important point I didn't consider at all. Build four data centers, say in Germany, South Africa, California, and Chile, powered entirely on PV with just a few days of battery storage, switch between them to wherever the sun shines, and you get ~2/3 availability with electricity for free. (Or even just two, one in Quito, Ecuador, and the other in Pekanbaru, Indonesia, for 100% availability.). An LLM doesn't care where it calculates, and the effort for data shuffling to there pales in comparison to the compute needed.