I took a long break from the nyah-nyahing over the Texas grid failure in February. Today's WSJ carries a fascinating
piece about the vulnerability of the "black start" capacity of the grid, which not only sends chills down my spine about how bad things could have gotten if ERCOT had delayed even a few more minutes before cutting off a huge fraction of Texas customers in the middle of the night, but also explains more than I'd read before about what happens if a grid shuts completely down and has to start back up. The article is behind the usual paywall, but you can get there by
Googling.
How is it that we keep reading about these disasters in which the back-up systems turn out to be vulnerable to the same conditions that cause us to need a back-up? I call that anti-resilience.
When a grid has rolling blackouts or even partial long-term service interruptions, a crucial core of the grid stays active. "Crucial" means not only things we'd really rather not shut down, like hospitals, but the power plants themselves. Power plants shut down not only because they can't get fuel and electric power, but also because a grid with too low a frequency can damage their workings. Lack of fuel is a temporary problem, but disconnection from the grid or staying on a low-frequency grid are long-term bad news: a damaged plant will take time to repair, and restarting a plant and reconnecting it to a dead grid is tricky.
I guess I always assumed that a power plant generated its own internal power as a matter of course, but apparently that's not so. If the grid shuts down, or even the part of the grid that's attached to the power plant, the power plant doesn't hum along on its own power. It can't: the power has to go to a load. So when the local grid area shuts down, the plant shuts down, too. It needs a special "black start" local generating unit to get it going again.
Even if all the black-start units operate perfectly, it's a wildly delicate operation to start the whole grid up again from scratch. If 1/2 or 3/4 of the grid is down, it's easier to add new sections gradually, though no picnic, with delicate attention to balancing the new power and the new load it can serve. When the whole grid goes down, it can take anything from days to the unthinkable months to black-start it.
In this case, ERCOT didn't have to do a black-start, which is a good thing, because about half of our black-start resources evidently were iffy. If they're to be reliable at all, they need a large standby fuel source. Gas that's got to come through vulnerable pipelines won't cut it. Nuclear is nice, as is hydro; failing those, giant oil or gas tanks would be good, or huge mountains of coal. The just-in-time inventory style has made those less common, and we're not giving power generators the right financial incentives to keep large expensive emergency fuel inventories handy.
This issue came up when ERCOT had a near-miss emergency in 2011. Predictably, we addressed the issue by ordaining committees to study the issue and work together to improve yada yada, the usual word salad. We won't have an actual solution until we figure out what it costs to have reliable standby power and reach a consensus on how to pay for it with real money from real electric power customers who have decided it's worth the price. What I'm reading is arguments about whether the free market or regulation is the panacea. The Texas system, while somewhat less regulated than some others, is hardly a free market, though it has a strong emphasis on market signals in some areas, generally in an attempt to force efficiency and keep prices down. Nevertheless, it's not all about efficiency, unless you include adequate backup resources for extreme emergencies in the concept of efficiency. Extra security costs money. We're going to have to get past the thinking that either a market or a regulation can change the cold equation telling us that something valuable has to be paid for by someone. "Someone" is going to be be (1) users or (2) people donating to users.