The US is using so much solar power it'll have to prepare for the August eclipse | Hacker News Reader

The US is using so much solar power it'll have to prepare for the August eclipse

qz.com

113 pointsobi1kenobi40 comments

Open articleView on HN

This occurred a few years ago in Europe - I graphed what happened with my solar panels.

https://shkspr.mobi/blog/2015/03/the-eclipses-effect-on-sola...

For UK / Europe the effect was broadly similar to normal cloud cover. Something that California rarely experiences ;-)

An eclipse is more predictable than the weather - so other sources of energy can be used to offset any disruption.

Yup! Grid operators will have this planned in well ahead of the date.

Yes indeed, we are mandated by NERC to have reserved for such events...it won't be an issue at all.

QZ has become a spam and advertising cesspool. I couldn't find the story the first two times I tried until I realized by default it only shows ads and garbage until you click the small "show story" button.

>QZ has become a spam and advertising cesspool.

These companies keep popping up promising to "disrupt" the media world (I'll throw Medium in there as well), and end up right back at the basic model, spamming readers with advertising.

It's such an interesting market; it's literally delivering text, and we have such a convenient way to consume it. Sure some text is (far) better than others, but like many products, the average person doesn't care at best, and can't tell the difference between crap and good at the worst.

People are going to have to get used to paying for good content. I hope it happens sooner than later.

It appears that this article is a summary of one in the Financial Times, which requires payment. The fact that the summary from qz.com got to the front page and the original from ft.com didn't shows you what people actually prefer, regardless of what they might say.

I don't see the cesspool that the grandparent comment describes, though. Even in incognito mode with my adblocker disabled, I see the article immediately. There's a gigantic annoying video banner ad at the top, but the story begins right below it, and after that there's one more fairly reasonable ad in the middle of the article. Maybe they're doing some analytics to decide how much they can annoy you, and the other person got incorrectly pegged as someone who can take a lot.

I was wondering about his experience too… if viewed on a shorter screen like an 11" Macbook, it's 100% ad.

I read Quartz on RSS, so my knee-jerk is "what are you talking about, they do exemplary journalism." The web version is indeed butt.

I clicked over to the linked site using Firefox running NoScript in default deny JS mode.

The whole text of the article appeared, ready to read. Granted it is a very short article, but it appeared just fine.

Us NoScripters experience a different Internet which many surprisingly call intolerable.

Same with uBlock with Privacy Badger.

This happend a couple years ago in Germany. The actual challenge for the local utility company (one of the biggest 4) was not to compensate the missing energy output from solar, but in calculating the optimal backup energy plan to buy on the future/spot market. If you buy too much or not enough upfront, it just gets expensive. But the whole problem is keeping the cost low. Not a black out.

Friend of mine was in the dev-team for the spot trading software. (Sorry for my english)

You know, darkness isn't something that just happens every day... wait...

This is during peak air conditioning load, and air conditioning is used most when when... wait... the sun is out.

Yeah but this article is silly. Solar isn't a given, you can have a cloudy day. Or days. So the grid already has to be prepared to deal with this.

It's a known problem, happens every night. The company that solves storage in a meaningful way is gonna get rich. By meaningful I mean at scale and portable. I know you can pump water up to a reservoir and let it back out to generate power but that's not portable. Anyone know if someone has an answer other than Musk's powerwall (which is kinda cool in that it is the definition of portable, completely distributed, but also sort of limited).

Seems like a fun problem to work on.

No, seriously, during summer, peak residential and commercial load correlates with solar heating. Unless you have residential load-shedding installed in your footprint, coordinating with industrial loads is your best bet to reduce peak demand.

The article isn't silly.

It's a known problem, but utilities and operators move at a glacial pace due to their scale and regulatory issues. This can make it difficult for startups to get sales and stay afloat.

Lots of people are trying to get rich in this space and it turns out to be a non-trivial set of problems.

And yes, it is a fun problem to work on. I would encourage anyone who is interested to get into the storage / smart grid management space.

It's not really a problem at all...the WECC utility companies have ample reserves, CA will just buy MW from APS, BPA, and PGE on the EIM (run by CAISO)...I'm not sure how this process is affected by utility company and operator "pace" in the slightest, in fact the ability to buy energy and transfer it from one BA to another happens every second...how would you envision this industry being "disrupted"?

So the problem that I (and I think others) see is a time where solar is providing more than 50% of our power. Or 100% of our power if that makes it more clear.

Then where do you pick up the slack?

Maybe your point is that we get it long distance from places that are not covered in clouds?

That's the whole point of an energy market and a grid system...if Cali can't cover their own energy needs (which they don't do already btw), they'll purchase however many MWs are needed from AZ, OR, WA, NV, or UT

On a really dark cloudy day, you're still generating something like 30-40% of the energy as you would in bright sunlight.

Eclipses are much tougher since even on cloudy days, there's enough variance over a region that your output would rarely drop more than 50% (there's a mitigating benefit to clouds where they reflect a lot of light back to the earth). During an eclipse though, something like 5,000 square miles of land drop to 0%.

But again, the sun is not out during an eclipse.

The temperature change outdoors is detectable, but brief. Also, buildings even the effect out, due to mass and retained heat.

https://sunearthday.nasa.gov/2006/faq.php

"Typically, how big a temperature drop do you get during a total solar eclipse? It would probably be equal to the typical daytime minus nighttime temperature difference at that time of year and location on the Earth. It would be modified a bit by the fact that it only lasts a few minutes, which means the environment would not have had much time to thermally respond to its lowest temperature, so it would probably only be 3/4 or 1/2 the maximum day-night temperature difference."

Not at 11AM it isn't.

dont these solar companies have to deal with rains, thunderstorm and other stuff too??? not every 11AM is a sunny day

Weather moves much more slowly.

Any individual power company would have to deal with this, but the graphic implies that the entire US will be inside the eclipse for a while. The usual strategy of buying from someone else on the grid who has a surplus won't be viable, they'll be relying entirely on standby generators.

I love the GIF in the article and how it was copyrighted in 2000 about an event that would happen in 2017. I'm a reader/follower of N. N. Taleb and agree with him about how difficult it is to predict future events. And meanwhile, the conditions in space and the laws of physics allow for a model like this to be made with what seems to be a lot of precision. Can anyone speak to how accurate/precise that GIF model _could've been_ in 2000?

You mean the gif of the eclipse path? We've had that capability for a while, it's just orbital mechanics. We know exactly when and where every eclipse is going to happen for the next several thousand years at least.

The ability to predict the locations of celestial objects to high accuracy is one of the pinnacles of modern science. The approach, called "ephemeris", is pretty straightforward- the entire solar system, with all of its components, is modelled using straightforward physics and math (Newton's law, with some relativistic corrections). We can accurately predict the location of all major celestial objects to hundreds of years in the future (barring some unknown external influence).

Every single time we send probes to other planets and they collect more data (such as the non-spherical shape of a planet or moon), that data is incorporated into the ephemeris calculations and they become that much more accurate.

Now, were a rare, unpredictable event to occur- say, a rogue star entering the solar system with low albedo so we couldn't see it coming- that would introduce a change in the dynamics of the system that we couldn't easily predict. it;s not exactly a "black swan" (to use taleb's nomenclature) because we know rogue stars exist.

The predictions aren't perfectly accurate. We don't have total knowledge of every element in the solar system (tiny asteroids affect the motion of the sun, in a very long term way) so the predictions eventually fall off in accuracy.

It's still a long term discussion whether the solar system is a truly chaotic system, but we'll leave that to JPL and Minsky.

I think this can be handled with a little of preparation just like it was handled elsewhere in the past.

If you look at the chart in the PDF, at the point of maximum impact the net load will be 23.5GW instead of 16.5GW

We're tracking live electricity generation from solar in the US here: https://www.electricitymap.org Not sure we will be able to see the effect talked about here though.

Good thing they have 6000MW of dispatchable generation they can crank up when they need it. I wonder how much spare capacity they have and how much the people that own it get paid to have it on standby.

CAISO will do it's job and CA residents won't notice a thing...APS, BPA, and PAGE will make a good profit that day...utility planners/schedulers will minimize line work (especially on the BES), and everything will be normal

The article doesn't mention a decrease in demand from air conditioning. Has that been factored in?

Are you implying that people will turn off their air conditioners for the brief <5 minutes of darkness? I really doubt this.

The partial eclipse will last about 2.5 hours. I don't know how much air conditioning would be impacted, but it's not just the totality at work.

Well air conditioners are set to keep houses at a certain temperature. If the five minutes of darkness will have a measurable affect on temperature then air conditioners won't have to work as hard to keep houses cool.

In a very near total solar eclipse in I believe the summer of 1991, it was a blistering hot day in Arizona and during the eclipse the temperature dropped substantially. I don't know the numbers, but I remember being surprised at how nice and comfortable the temperature drop was, which lasted for much longer than 5 minutes.

How much of a lag was there between the eclipse itself and the drop in temperature? Even if the drop is immediate I don't expect the thermostats to turn off right away, as it will take some time for the cold to migrate into buildings.

It's negligible

Eclipse for some, bright day for others.

-- Elon Musk

Reply on news.ycombinator.com