Why does demand knock power plants offline?

[Algr]

https://www.cnn.com/2022/05/14/us/texas-heat-wave-ercot-conserve/index.html

Why would they make plants like this? If demand gets too high, the plant just stops? And potentially knocks out other plants? Wouldn't it make more sense to cut off a few places at a time so that the plant can continue, rather then stopping the plant and having a big regional blackout?

 

[anorlunda]

The article said that the plants were knocked out by high temperatures, not demand.

Why a plant in Texas can't stand high temperatures is still a good question. My guess is lack of maintenance.

ERCOT is supposed to assure that all plants meet the minimum requirements before they are allowed to sell power to the grid. It appears, that ERCOT is failing at that responsibility, both in hot and cold weather. It is likely that political interference is the root cause. If a plant doesn't follow the rules but it does have support of people in state government who boss ERCOT around, interference results. It sounds to me like evidence of a corrupt government.

 

[Algr]

At other times this has happened in winter. And a few years ago, it happened on the east coast.

 

[anorlunda]

And a few years ago, it happened on the east coast.

link to the source please.

 

[Algr]

https://en.m.wikipedia.org/wiki/Northeast_blackout_of_2003#Effects

 

[anorlunda]

https://en.m.wikipedia.org/wiki/Northeast_blackout_of_2003#Effects

No, that was not caused by power plants knocked off by demand. But if your question
relates to blackouts in general, not "power plants knocked off by demand", then it can
be answered.

Nothing can be reliable 100% of the time. Unlimited reliability would cost unlimited money.
In the USA, an organization called NERC sets standards for power companies. Their standard for blackouts says that a blackout affecting two million (or more) people should not happen more than once every 10 years. When they do happen, the trigger causes are varied.

Do they meet that standard? Where I come from, New York, we had big blackouts in 1965, 1977, and 2003. That's 3 times in 39 years. Compare that to 3 times in 30 years from the standard. In all three of those cases, power transmission was closer to the root cause than power plants.

In the recent events in Texas, as I mentioned in post #2, the problem seems to be mismanagement.

If you would like to learn more about the causes of blackouts anywhere, post again.

 

[bob012345]

In the recent events in Texas, as I mentioned in post #2, the problem seems to be mismanagement.

The fact is there have not been brownouts yet in this heatwave. But heat waves are always a concern. It appears the generation capacity quickly came back if I read this correctly.

https://www.ercot.com/news/release?id=8b772e9e-51d0-4c3c-e653-1e5079f28e89

Here are the current grid conditions in Texas;

https://www.ercot.com/gridmktinfo/dashboards

 

[Algr]

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

Yes, the question is a general one, not specific to this heat wave or power loss. Shouldn't cutting off demand or brownouts happen before power generation is forced to actually shut down?

 

[russ_watters]

The article said that the plants were knocked out by high temperatures, not demand.

Unless I missed something, all I'm seeing on that is that it can be inferred from ambiguous grammar. Can quote what you are seeing that lead you to say that?

No, that was not caused by power plants knocked off by demand. But if your question
relates to blackouts in general, not "power plants knocked off by demand", then it can
be answered.

I'd like to know the answer to the question implied by the above though; can excessive demand itself knock power plants offline? It's my understanding that it can and that that's what causes cascading blackouts.
[edit]
You posted seconds after I did and gave a partial answer. It mostly answers about the effect of high demand on the grid. But what about individual power plants?

 

[anorlunda]

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

Yes, the question is a general one, not specific to this heat wave or power loss. Shouldn't cutting off demand or brownouts happen before power generation is forced to actually shut down?

It does. Or at least it is supposed to.

The #1 defense is reserve capacity. Each area purchases about 20% excess power plant capacity specifically to handle situations when weather or other problems cause loss of the regular supplies. Things like industrial gas turbines are expensive to run, but they make very good backup reserves.

The #2 defense is to reduce demand by cutting voltage. Sometimes called brownouts.

#3 is voluntary or contractual load reduction by consumers consuming less. This is mostly done at the commercial level, not residential homeowners. For example, in the NW US mining companies can use 20-30% of all electric power produced. But they can also afford to easily shut down mining during periods of high electric demand.

#4 is called load shedding, either automatic or manual. That refers to involuntary termination of power supply to some customers. Sometimes that is called rolling blackouts. But load shedding that is unplanned and not announced in advance is something that power companies hate to do because of the bad PR. The 1977 blackout in NYC was such a case. The NY State operator ordered the NYC operators to shed big fractions of the NYC load. The operators were afraid to do that. As a result, the whole state was pulled down.

Cascading failures tend to happen in two cases. First, dynamic oscillations in the grid caused by design flaws in the controls. Se cond, are limitations on how much power can be shipped from one area to another. If there are 3 routes to get power from Idaho to Southern California, and one of them gets overloaded and trips, the remaining 2 need to handle all the power themselves. That may overload a second line, and it trips. Then the third trips and we have a cascade. The limitations there are more often in the transmission system, than in the power plants. Ontario Canada may have surplus power that could help Texas, but moving it all that distance and across borders doesn't work. Texas is special, in that it has no transmission connections to the rest of North America.

But think back to what I said before. No matter how things are designed, no matter how much money we spend, they can not be expected to never fail. Power grids are no exception.

How often is acceptable? The acceptable limit was set at once per ten years for major blackouts. But blackouts that might affect only one customer at a time, happen several thousand times every day. Between those extremes is a continuum of intermediate cases. State regulators set performance goals for mean time between outages, mean time to repair, and other quality metrics.

But politics and public reactions can be emotional. Every time there is a major outage event, people get on TV and claim "I'll make sure this will never happen again." Never say never.

Edit: @russ_watters , we crossed posts. See if this answers your question. Short form, excess demand strains the transmission system. So it is transmission that cascades, not the power plants per se. Things like load shedding prevent the loads from becoming more than what the generating capacity can handle. Of course, involuntary load shedding can be described as a deliberate blackout, but we don't use that terminology. Figuratively, the PJM region could say "Dump Pennsylvania to save the PJM region." They would be correct in PJM's view, but PA customers would not see it that way.

 

[Algr]

But think back to what I said before. No matter how things are designed, no matter how much money we spend, they can not be expected to never fail. Power grids are no exception.

It's not that things shouldn't fail, it is that failures should not cascade. They know how much a given power line can handle, so send that and nothing more. If that isn't enough, let the destination deal with the shortfall, rather then shutting down one system after another. That would confine blackouts to the area near the failure, rather than having a regional colapse.

 

[anorlunda]

Sorry, out of time tonight and I'm traveling tomorrow. I'll be back to PF Tuesday and I'll write better answers. But I'm not the only one here that can answer power system questions. Maybe before Tuesday, others can pitch in..

 

[anorlunda]

OK, now I have some time. I like your questions. They are good ones.

It's not that things shouldn't fail, it is that failures should not cascade.

Consider two areas A and B. The probability of an outage in some unit of elapsed time is 0.01. P_A=P_B=10^-2 That's the reference level of reliability for the customers. The probability of both A and B being out at the same time is P_A*P_B=10^-4, because they are statistically independent.

Now suppose we create a super area AB. A backs up B and B backs up A. Now the probability of an outage is P_AB=10^-3. Better service for both areas! But if B does fail, it cascades to A. If A fails it cascaded to B. Therefore, the probability of both being out is 10^-3 instead of the old 10^-4. Should the customers complain? "Failures should not cascade. That would confine blackouts to the area near the failure, rather than having a regional collapse. " The areas are either independent or they aren't. You can't have it both ways.

They know how much a given power line can handle, so send that and nothing more.

Think of the grid as a circuit with many branches. Source power plants, and loads are not considered part of the circuit, but external boundary conditions. Within the circuit, voltages and currents follow Ohm's law and Kirchoff's Laws, just like any other circuit. We could put a fuse, or even a "smart device" in series with every component. I think that would make things worse, not better.

But there's a better way to do it. We allocate shares of the total demand to power plants in smart ways. So we can find an allocation of shares such that no circuit branch is overloaded. Let's start with no overloads. Now, we consider contingencies. Every branch could fail, every power plant could trip, every load could trip. Each contingency causes voltages and currents in the circuit to adjust. Now suppose we find an allocation of shares that guarantee no overload after any of those hypothetical contingencies. Better still, we can consider all combinations of contingencies taken two at a time or three at a time and find an allocation that guarantees no overloads even after the contingencies. That is the type of software called "security constrained economic dispatch." It is a big linear programming optimization problem to find the solutions. So central software can act to forestall cascades in advance.

But central software can never be available 100% of the time itself. We can use redundant computers, but there is no 100% guarantee nevertheless. What do we do if the software is temporarily out of service? Shut of the power to all customers as a precaution? Certainly not. We keep operating without that added safeguard.

In the 2003 blackout, MSIO, the operator in Ohio, was supposed to have such software but it was out of service. Unavoidable? But in their case, the software had never worked and they operated for years without it. In my opinion, that crossed the line from unavoidable to negligent.

So despite our technology, the fact remains that one blackout per 10 years affecting 2 or more million people is still a realistic goal. Never is not realistic.

 

[Drakkith]

In the recent events in Texas, as I mentioned in post #2, the problem seems to be mismanagement.

Indeed. It's a product of a deregulated industry and simple greed. It costs money to upgrade the equipment used to produce and move both electricity and fuels, money that the power companies don't want to spend. Even bringing extra capacity online prior to a possible grid strain costs money, hence why the Texas grid has very little reserve capacity at anyone time (according to wiki, but I can't access the NY Times article linked as a ref). Also, many regulations regarding extreme temperatures are voluntary, not mandatory. From a Texas Tribune article:

Woodfin, of ERCOT, acknowledged that there’s no requirement to prepare power infrastructure for such extremely low temperatures. “Those are not mandatory, it’s a voluntary guideline to decide to do those things,” he said. “There are financial incentives to stay online, but there is no regulation at this point.”

According to Michael Webber, an energy resources professor at the University of Texas, Austin and an expert on Texas’ unique grid, Texas is the 3rd largest natural gas producer in the world after Russia and the rest of the U.S., so to lose half of its production due to cold weather is telling. For ref, see this link and go to about the 4:00 mark on the audio file.

 

[Algr]

I'm still not convinced that a billion dollar industry understands the issue better than me. :/

 

[gmax137]

I'm still not convinced that a billion dollar industry understands the issue better than me. :/

I am. ;)

I don't know where you are located, but see if you can find your regional independent system operator "ISO" or power pool. Some of them have pretty good websites. And some used to offer tours, if you happen to be nearby.

 

[anorlunda]

Indeed. It's a product of a deregulated industry and simple greed. ... Also, many regulations regarding extreme temperatures are voluntary, not mandatory.

It gets political from this point out. ERCOT's rules, and ERCOT's structure and ERCOT's governance are set by the Texas Legislature, not by the industry.

In other ISOs, an important mechanism is called ICAP (Installed CAPacity). That makes payments to power plants to be ready and able to bid into the energy market and to produce power. ICAP payments can be substantial, as much as 1/3 of a power plant's gross income for being ready compared to 2/3 for making power. The utilities that buy ICAP and make those payments demand to get what they paid for. Therefore, they force the ISO to set eligibility rules for getting ICAP payments, rules like winterization. And the ISO inspects the plants to confirm compliance before signing the ICAP payment check. Customers have the power to force sane and effective rules because the ISO is governed by participants, not by an appointed board of directors. Also because in every state except Texas, regulation is Federal and the state government has no jurisdiction. The Texas laws that created ERCOT rejected all that stuff.

I understand the deeply rooted issues with distrust of industry, versus distrust of government. I have my own biases. But in this case, we can compare regulated and deregulated regions of the country. Some work better than others.

By the way, they are still doing it; micromanaging by legislative statute things that ought to be decided by engineers.
https://www.spglobal.com/commodityi...-changes-for-ercot-in-post-feb-14-storm-bills