Why experts say a solar storm could cause trillions of dollars worth of damage

[Marco]

I'm sceptical. AFAICS just the modeling of generator response to bias current is very recent, with models for regional bias current prediction from solar observation at the conceptual stage. When I read stuff like "There is no commercial product dedicated to GIC measurement." in a 2021 paper I'm going to say the industry isn't quite ready. Certainly not to the point of responding to space weather forecasts with a blackout.

You need good estimates for bias current limits for every generator on the grid, you need good bias current measurement systems everywhere to calibrate a model with and then maybe if you're lucky and the observations gives you high reliability model to predict the currents you can use it to black out regions. I don't think they are there.

PS. I suspect the models will have too poor predictive value to be used for a blackout decision.

[Andy Chee]

I'm going to say the industry isn't quite ready. Certainly not to the point of responding to space weather forecasts with a blackout.

You need good estimates for bias current limits for every generator on the grid, you need good bias current measurement systems everywhere to calibrate a model with and then maybe if you're lucky and the observations gives you high reliability model to predict the currents you can use it to black out regions. I don't think they are there.

PS. I suspect the models will have too poor predictive value to be used for a blackout decision.

You're correct about long range forecast of solar weather, the modelling is nowhere near ready for a planned shutdown (e.g. analogous to the nightly news 7-day weather forecast).

However, the observatory located at L1 can be used to give advance warning for an emergency shutdown (in the order of a few minutes).  As I said, since the L1 observatory was launched in 1996, there hasn't been a strong enough storm to test the fast reflexes of grid operators.

For example, a fast CME takes 15 hours to reach earth.  Sun to earth distance is 147 million kms.  L1 to earth distance is 1.5 million kms.  When a massive CME is detected at L1, that means it will be 100% guaranteed to hit earth in 9 minutes time.

What can grid operators do in 9 minutes?  Not much I would've thought!  Fortunately, most CME usually takes days to reach earth, not hours, which gives grid operators even more time to react to L1 detection.

[Marco]

If it's truly catastropic, they might be liable to prepare to shutdown (though I doubt they will actually do it till the first grid shuts down without manual intervention).

If it's only slightly superior to March 1989, everyone is just going to wait and see.

[Psi]

It's very hard speculate on the damage, or mitigation needed, without deciding on the exact level of solar storm we're talking about.
The problem with that is the really big solar storms happened too long ago to have much useful technical info recorded about them.

We know people were getting arcs and shocks off metal objects in the street, like light poles etc, due to voltage induced in overhead lines, but we don't know if those arcs were like 3mm or 30mm long. 
Modern electronics can handle some surges and small HV arcs but only too a point and not continuously. Solar storms can last for hours.

If you had enough warning to flip your main house breaker off and unplug any copper phone lines you would at least be somewhat protected, so long as the voltage couldn't arc over the breaker, but I'd expect all that induced power to take the easier path of going through every other house that didn't have their main breaker tripped. Rather than getting high enough to arc over your breaker.

[Marco]

It won't generally induce significant voltage on the mains. The cables from distribution transformer to homes are too short (or in my case, buried) and the distribution transformer doesn't pass DC unless things go catastrophically wrong. I'd say that if the distribution transformer does pass DC due to a solar storm you probably need to be in a bunker.

[coppercone2]

if you get a shock off a metal object on the street why on earth wont the long wire be energized? I assume that means people found that mailboxes and signs were arcing (if they had metal signs)

[johansen]

Allegedly most of the high voltage long distance transformers are delta and will not be affected by 10 volts per mile of dc on either end


The rest of the grid with grounded Y transformers, the issue is the electrical metering equipment is not setup to measure dc current, so the transformer saturates and the grid does not even notice the extra load, and what happens is the transformer catches fire internally. Not due to failure of conductors due to overcurrent, but by the saturated transformer overheating steel components located near the core, which normally have very minimal magnetic flux flowing around them.

Us congress investigated this extensively for many years and you can read the reports that concluded that about 50 million bucks for the 1,000 largest substations is all it would cost to cut the neutral, install a capacitor and lightning spark gap.

[coppercone2]

does anyone have the first hand anecdotes about voltage? because if metal objects in the 1800's are arcing, thats pretty small stuff. 10 v per mile does not sound like it could cause citizens to see random arcs

[johansen]

does anyone have the first hand anecdotes about voltage? because if metal objects in the 1800's are arcing, thats pretty small stuff. 10 v per mile does not sound like it could cause citizens to see random arcs

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

10 volts per mile also depends on how high the line is off the ground and what direction its run, that's the highest number i've seen anyone suggest is possible.

10 volts per mile is not enough to melt anything. it will not destroy anything.. the issue has always been transformers saturating because they are grounded on both ends. cut the ground, problem solved.

[NiHaoMike]

Now the question is how to get the utilities to do it. Remember how PG&E neglected to do maintenance which led to some forest fires?

[Psi]

does anyone have the first hand anecdotes about voltage? because if metal objects in the 1800's are arcing, thats pretty small stuff. 10 v per mile does not sound like it could cause citizens to see random arcs

There was conflicting and odd reports with no real way to know what is true vs not.

So, as with most things like this it is looked at, after the fact, with a filter on of...
"what actually makes sense given what we know about the subject" and any reports that go against that tend to be ignored or treated as false.
Which may be correct, but is not a certainty.

Like there were reports of people getting shocks from door knobs.  Which really makes zero sense given what we know about CMEs and induced voltages, given how small a door knob is.  Did that actually happen often, or was it just one door knob at a telegraph office somewhere, or is there some mechanism at play that we're not aware of.

[SeanB]

Most high voltage power lines have ground fault detection, generally provided by having a Peterson transformer connected to provide a Y connection to ground, and a great big inductor in series with the tap, so that any fault to ground will cause a voltage drop across this coil. AC or DC plus AC, not really any difference, as the current flowing in the coil as a result of imbalance will pull in a solenoid, and close a contact to the SCADA system of the line, telling of a ground fault. Might take 5 seconds to respond, so as to ignore transients, as the voltage will kind of self clear the general fault, by flash cooking the big bird or errant tree branch, or the metal kite string somebody flew into the power line. But a prolonged fault will trip out the line, and, so long as the sending end and receiving end still have power available, they can be commanded remotely, using the fibre networks most utilities have now, run along the middle cable of the long transmission lines as they have no need then to do any EIA report on installation, and also it is easy to install, as they have the equipment and pulleys to draw along the line, and droop down by pylons to follow the existing cable, and then tie off to the existing as well. Long lines will have issues, unless they are HVDC, which already have DC fault blocking and detection, and long AC lines will merely need a few minutes notice to cut power, island the power plants to local generation supply only, and thus not need them to go offline fully, but keep a low spinning load, and run with a largish circulating current instead, excess supplying local loads. Otherwise you need to black start them, and most power plants cannot do this, they need power to run.

By me not an issue, 15 years of rotational load shedding means there is a pretty robust system of alternative power already, thanks to this. USA and Europe has never really had this except short term over relatively small areas, but by me being without power a quarter of the time is somewhat accepted to be normal.

[coppercone2]

I recall reading somewhere of arcing from a scaffolding on a building. no idea wtf that was from. i think they said it had something to do with stars or bombs

maybe it was some survival manual saying to stay away from large metal objects like buildings under construction. or some news story.

[Marco]

Us congress investigated this extensively for many years and you can read the reports that concluded that about 50 million bucks for the 1,000 largest substations is all it would cost to cut the neutral, install a capacitor and lightning spark gap.

I don't think spark gap works. Once it starts arcing it's a short and then what's point? Can't they just handle it by programming a TCSC a little differently? Let bias voltage build up on the capacitor and use the thyristors to bleed off some voltage when necessary (or rather, one of the two).

[coppercone2]

I think the confusion is with HEMP. I am reading that its alot worse. So the carington event was like a little hemp.

[EPAIII]

Thunder? Do you mean lightning? Thunder is just a lot of noise.

Without a large blackout I don't see how it should affect internet.

It could affect modems and other equipment when the blast is strong enough. It is not just about voltage induced in cables. In early days the telegraph system suffered from it, and it could disrupt a lot more with nowadays electronics.

Think of thunder, when it hits close enough to your house it can destroy every electronic device in your house even when they are not connected to a power outlet or a cable for TV or internet.

[Marco]

Even in countries with overhead phone lines, the last mile is mostly very short relative to telegraph lines of old and are twisted pair. Open voltage and short circuit current won't be any where near what they were for telegraphs and are for the grid. Next to nothing. With buried cable it will be even more next to nothing.

[johansen]

Us congress investigated this extensively for many years and you can read the reports that concluded that about 50 million bucks for the 1,000 largest substations is all it would cost to cut the neutral, install a capacitor and lightning spark gap.

I don't think spark gap works. Once it starts arcing it's a short and then what's point? Can't they just handle it by programming a TCSC a little differently? Let bias voltage build up on the capacitor and use the thyristors to bleed off some voltage when necessary (or rather, one of the two).

Spark gap is to discharge lightning transients

[coppercone2]

what is this telephone line. you mean a shielded coaxial cable or  fiber line?

[Psi]

Even in countries with overhead phone lines, the last mile is mostly very short relative to telegraph lines of old and are twisted pair. Open voltage and short circuit current won't be any where near what they were for telegraphs and are for the grid. Next to nothing. With buried cable it will be even more next to nothing.

That is a fair point.

[Marco]

what is this telephone line. you mean a shielded coaxial cable or  fiber line?

Pcprogrammer mentioned modems, so I assumed he meant ADSL. Guess there's cable modem too, either way not going to be much induced current. Earth return telegraph makes a nice big loop for the magnetic field to play with, twisted pair and coax not so much.

[Psi]

I wonder if anything else might cause a problem that we don't really think of as electrical.
Miles and Miles of underground metal water pipes for example.

[Dan123456]

I wonder if anything else might cause a problem that we don't really think of as electrical.
Miles and Miles of underground metal water pipes for example.

That’s a really good point actually!

I’m just imagining the poor souls on a train if it were to happen as they could be sitting on 1000’s of kms of track 

[NiHaoMike]

I’m just imagining the poor souls on a train if it were to happen as they could be sitting on 1000’s of kms of track 

Nothing would happen inside the train, it's an enclosed metal box. Things might get more interesting when loading or unloading the train but I would expect there to already be a system to detect a large voltage between the (grounded) rail and ground which might arise from an electrical fault. (Besides the many systems that use a rail for power, I assume most overhead systems use the rail for return? Having an additional overhead line adds maintenance costs.)

[Marco]

The tracks are earthed, the only real problem is transformer saturation same as grid.