[morelisp]

The size (not always duration) of outage in your second link would be considered routine in many American cities. To take SF as an example:

- 2020, well, here you go

- 2018/2019 should also need no overview, CA is probably the worst power situation in OECD countries in the 21st century

- 2017, 88k without power https://sanfrancisco.cbslocal.com/2017/04/21/power-outage-sh...

- 2016, 23k https://sanfrancisco.cbslocal.com/2016/10/14/pge-power-outag...

- 2015, 45k https://sanfrancisco.cbslocal.com/2015/06/08/thousands-of-ba...

I'm only picking the first Google result for each year. I could keep going but I hope the point is made: Berlin got a blackout once in a decade like most American cities get multiple times a year. It reacted comparatively slowly because it's so infrequent there was no recovery plan. The political impact of the outage was huge, the country's attitude was that everything is collapsing - and by "collapsing", they mean "like America is all the time."

[tripletao]

> Berlin got a blackout once in a decade like most American cities get multiple times a year.

The Berlin power outage linked above was over thirty hours, and the SF outages you linked were a few hours. To bury that in "(not always duration)" seems misleading to me. That said, the 2019 power cuts were indeed approaching third-world duration and frequency, to the point that homeowners who could afford it applied the usual third-world workarounds (gas generators, etc.). That was mostly outside the biggest cities, but huge numbers of people were affected.

And how did we get from SF to "most American cities"? California's grid reliability is notoriously and distinctively bad, and a huge political topic here too (though with little progress after many years). The World Bank[1] puts the overall USA's quality of electricity supply solidly in the middle of high-income countries though, slightly ahead of Germany.

We also seem to be mixing load shedding (i.e., the utility realizes they can't safely supply all their customers and therefore deliberately cuts power to some) with accidental outages. They look the same to the customer, but perhaps imply different kinds of bad planning by the utility.

1. https://tcdata360.worldbank.org/indicators/ha7db856d?country...

[morelisp]

> We also seem to be mixing load shedding... with accidental outages.

Not 'we'; I'm well-aware of the difference but the person I was responding to gave examples of accidental outages.

The US has considerably more of both than Europe, especially Germany (SAIDI measured in hours vs. 15-20 minutes), regardless of what investment-focused metrics the World Bank is giving.

[mpweiher]

> how did we get from SF to "most American cities"

2020

https://www.forbes.com/sites/lisettevoytko/2020/08/07/blacko...

https://www.theblackoutreport.co.uk/2020/08/12/chicago-black...

https://www.nola.com/news/article_a64debc0-c3dc-11ea-8d5a-57...

https://eu.usatoday.com/story/news/nation/2020/06/11/detroit...

https://eu.detroitnews.com/story/news/local/michigan/2020/07...

...

2019

https://en.wikipedia.org/wiki/Manhattan_blackout_of_July_201...

https://eu.freep.com/story/news/local/michigan/2019/07/22/dt...

https://www.rt.com/usa/385646-blackouts-hit-la-ny-sf/

etc.

https://www.utilitydive.com/news/detroit-hit-by-major-power-...

https://en.wikipedia.org/wiki/Northeast_blackout_of_2003

https://en.wikipedia.org/wiki/New_York_City_blackout_of_1977

Note that the Berlin blackout affected 30000 people and was caused by a construction company cutting through a cable. All of these affected many more (Chicago: 800K), occur rather frequently and seem to be caused by regular phenomena like weather.

One major reason is that so much of power-distribution is above-ground, on these poles you see everywhere. Storms will knock these over or have falling trees take them down.

>We also seem to be mixing load shedding ... with accidental outages.

Yes. The claim was that "most 1st world countries" experience rolling blackouts. That's not true.

[tripletao]

To be clear, I agree that most developed countries don't routinely experience rolling blackouts like the one that just happened in California. This includes the USA, unless we count twice in twenty years in the state with the worst grid as "routinely". I don't think it's useful to compare anecdotes, though I'd note again that most of the outages linked above are far shorter than the Berlin outage. Anyone who believes that most American cities experience multiple thirty-hour, 30k-customer outages per year (as the comment I first replied to implied, despite the hedge with "not always duration") is mistaken.

morelisp's later comment linked to SAIDI numbers for the USA and Germany, and by that standard[1] Germany is indeed far better than the USA, about 4.5 nines to our 3.5. Singapore[2] is yet better, about 6 nines. The USA is far worse than the EU average, somewhere around Italian[3] standards. That's an undeniable difference, though I'm not sure all of that is attributable to incompetent American operators (though certainly some is, as PG&E has repeatedly shown).

The grid can't target perfect reliability, since the cost for that would exceed the benefit in almost all applications. It's cheaper to provide backup power for the few exceptions (data centers, medical, etc.). The cost per customer to achieve a certain reliability will be higher with lower population density. The cost for backup power is fixed. So the economically efficient grid reliability should be worse with lower population density, which seems to be what we see empirically--Australia is bad too, for example. I'd guess that much of the difference is simply because the USA has decided that Germany-level reliability isn't cost-effective, in the same way that Germany has decided that Singapore-level reliability isn't cost-effective. I'd guess the World Bank ranking I cited somehow tries to capture that fact and therefore ranks all three countries about the same, though I can't find any details of their methodology.

1. https://spectrum.ieee.org/energywise/energy/the-smarter-grid...

2. https://www.ema.gov.sg/cmsmedia/Publications_and_Statistics/...

3. https://www.occto.or.jp/en/information_disclosure/miscellane... , Tables 45 and 46

[mpweiher]

What do you mean "twice in 20 years"?

2019:

"California’s largest utility is under severe scrutiny by state regulators and customers over its rolling blackouts that have left millions of people without power.

Just two weeks after a massive power shut-off that the company acknowledged it mishandled, Pacific Gas & Electric announced on Wednesday that it will cut electricity to 179,000 customers in California in the face of a new wildfire threat. Southern California Edison, the state’s second-largest investor-owned utility, also warned on Wednesday of power outages to 308,000 customers."

https://www.cnbc.com/2019/10/23/pge-rebuked-over-imposing-bl...

That's at least 1 and possibly 3 power cutoffs (not outages, deliberate cuts) just in 2019.

Several days so far in 2020, with millions affected.

https://www.vox.com/2020/8/15/21370128/california-blackouts-...

And the year is still young.

And of course the energy crisis in 2000-2001 caused many, many blackouts over the course of two years, it's not the single incident you are implying.

https://en.wikipedia.org/wiki/California_electricity_crisis

> far shorter than the Berlin outage

Shifting goal posts. This was a single outage affecting 30K people. Just the one Chicago outage affected 800K people, that's 26x more people. This isn't comparing Apples and Oranges, this is comparing Apples and Aircraft Carriers.

And most of the others are over 100K people. And there are multiple ones per city. Per year. And I stopped looking after it was clear that many cities were affected by much larger outages multiple times a year. My guess is that outages affecting only 30K people just didn't make the news.

> The grid can't target perfect reliability

Yes, being resilient against construction workers cutting the cable is difficult. However, being resilient against storms, storms that occur just about every year should be something you can expect, and in non-third-world countries is something that you actually can expect. That doesn't mean there never is storm damage, just that power isn't routinely cut when there's a storm as seems to be the case. Or when the weather is hot.

Austria has much worse snow storms and ice than what we saw that winter in Detroit, every year. And avalanches. Yet I've never seen or experienced anything like that power outage. And it doesn't look like anything substantial has changed. The US has been under-investing in infrastructure for many decades, and it shows.

If you don't believe it, visit other places.

> reliability will be higher with lower population density

New York has a low population density? Chicago? The Bay Area?

[tripletao]

The article under discussion is titled "California issues first rolling blackouts since 2001". For that to be true, the author appears to be distinguishing between (a) outages due to load shedding due to insufficient generation capacity and (b) outages due to deliberate shutdowns of inadequately-maintained transmission lines in weather with high fire risk. The 2019 outages--which in an earlier comment I called "approaching third-world duration and frequency"--were (b).

That distinction is a little subtle, since they're both deliberate cuts. But I chose to follow the author, and I believe the "rolling" reasonably conveys "just (a)", since outages due to (b) don't actually "roll" (despite your CNBC article using that phrase). The power there stays off permanently until the fire danger passes and the line is re-inspected. This is actually worse, since customers don't get brief opportunities to re-cool their freezers and such, but it's different.

I don't see why you think the duration of the outages is irrelevant. For better or for worse, the USA accepts occasional outages of a few hours as "normal". People are inconvenienced, but not hugely--your freezer doesn't melt, water tanks in (older, at least) tall buildings don't empty, cell sites and other infrastructure with backup power doesn't run out of battery or diesel, and so on. I thought the Berlin blackout was particularly significant because its long duration meant those stopgaps were exhausted; but perhaps the two countries' overall attitude to power outages is sufficiently different that I'm missing something.

I mentioned country-level population density because I was comparing country-level SAIDI numbers. American cities still do worse than Germany, but by less. For example, my Singapore report above put the SAIDI for NYC at 20.53 minutes, around Germany's overall average.

In any case, I've lived many places in the USA and other countries, and California's grid is uniquely terrible. The rest of the USA's grid seems fine to me though, just one that makes different tradeoffs from Germany's. If I understand correctly, Germany's electricity prices are roughly double[1] those of the USA. So it seems like Germany just chose a higher-quality, more expensive approach, and USA chose quick and dirty, all very much in line with national stereotypes.

1. https://energytransition.org/2015/05/german-power-bills-low-...

[dmode]

I don’t know what you are comparing. A power outage due to an accident, affecting a limited population is very different than blackouts. Such emergency outages happen every day

[morelisp]

> I don’t know what you are comparing

I'm comparing what the parent asked to compare, even though it's not load shedding.

> Such emergency outages happen every day

In America. Not Germany. Like, yes, literally some level of outage occurs daily, but it's an order of magnitude difference.

US: https://www.eia.gov/todayinenergy/detail.php?id=37652

Germany: https://www.bundesnetzagentur.de/EN/Areas/Energy/Companies/S...