Ask an Engineer

Electrical substation in Houston, Texas, February 20, 2021. (Go Nakamura/Reuters)

Seven questions, and more than seven answers, about the crisis in Texas.

Sign in here to read more.

Seven questions, and more than seven answers, about the crisis in Texas

NRPLUS MEMBER ARTICLE W ould things have worked out better in Texas if it weren’t on an electrical grid that is separate from the rest of the country?

Probably not, says Professor Wei-jen Lee of the University of Texas at Arlington, who answered questions with several of his colleagues in a panel discussion organized by the Institute of Electrical and Electronics Engineers (IEEE). Lee, the director of the university’s highly regarded Energy Systems Research Center, notes that there was not enough capacity in the surrounding grids to have prevented the Texas blackouts even if the grids had been connected.

Doug Houseman, a grid-modernization expert at engineering-and-construction firm Burns & McDonnell, agreed. “Even if it had been connected, there would have been blackouts, because SPP [the Southwest Power Pool] and MISO [the Midcontinent Independent System Operator] had no excess [capacity], and Southwest [which is distinct from SPP] had only four to six gigawatts, but we were turning off ten-and-a-half gigawatts at that point.” Far from having spare power to share, adjacent grids were experiencing rolling blackouts of their own.

“Officially, there are three U.S. systems, but, in Texas, we like to say that there are two: The Texas system and the non-Texas system,” Lee says. “In reality, most such systems are in effect independent systems, regardless of whether they are connected or unconnected, from a macro-systems point of view. The size of the system in Texas is about 85 gigawatts, which is larger than a lot of countries in system capacity.”

In the 1990s, Texas undertook a study of what it would need to do to connect to the neighboring grids and what the effects of doing so were likely to be. And here’s a counterintuitive finding: That connection might very well have made the overall network less resilient. Because Texas produces a great deal of energy at relatively low cost, connection probably would have led to the early retirement of some generating capacity in the adjacent systems, which most of the time would have been able to buy power from Texas. If Texas were connected to the neighboring grids, it might very well have meant more widespread dependency on the very Texas systems that failed.

How big a problem was the interruption in wind and solar power?

By one estimate, Texas’s wind-power output fell to 2 percent of its installed capacity. Even so, that problem was probably not a very big one, according to an engineer with a background in grid operations. “Wind doesn’t work in these conditions,” he says, “but they don’t expect it to work.” As he reads the numbers, electricity production from thermal sources (meaning natural gas, coal, and nuclear) ramped up at the beginning of the crisis to just about double its previous level, but then fell by about 25 percent as the weather interrupted generation in different ways in different facilities. When it is running at full output, wind can provide as much as 60 percent of Texas’s electricity needs, but the grid managers don’t count on that in a blizzard.

Exactly how did the cold weather interrupt electricity generation?

In a variety of ways that are still being reported on and analyzed. In one instance, a sensor that malfunctioned in a nuclear plant necessitated taking some power offline for safety reasons. In previous cold snaps, such as the catastrophic one that paralyzed southern Louisiana in 1989, steam systems froze up, as did water pipes and some small connections. It’s likely the same kinds of failures are behind much of Texas’s recent trouble. It’s also likely that there was a kind of vicious cycle at play: In an effort to improve their environmental profiles, oil and gas operators have spent years replacing gas-driven compressors and other parts with electric equipment. Without electricity, that equipment fails and the gas stops flowing to the electrical plants, reducing their ability to keep up with demand elsewhere. Blackouts can cause more blackouts by taking electrical plants’ fuel-delivery systems offline.

Even when there is additional generating capacity available, bringing new power online is more complicated than flipping a switch. “There are two times when things really go wrong,” one engineer says. “When you turn something off, and when you turn it on. That’s the time when you are most likely to see equipment failure in plants.”

Professor Massoud Amin of the University of Minnesota expects that frozen well heads will end up being the main factor reducing the natural-gas supply. There are heaters and dryers that can help, he says, but these are insufficient given the extreme cold Texas experienced.

Is the Electric Reliability Council of Texas (ERCOT) simply incompetent?

Amin doesn’t think so. He’s the former chairman of the Texas Reliability Entity, which (the bureaucracy here gets complicated) oversees ERCOT under the authority of the North American Electric Reliability Corporation (NERC). “ERCOT has some of the best people I’ve ever had the privilege of working with,” he says.

But incentives matter. Another engineer with a more critical view points to the economic incentives at work. “The commercial models are driving behaviors,” he says, “and there’s no real penalty if they don’t show up with the power. People get paid for what they deliver — people don’t get paid for providing guaranteed capacity. In addition to actual power, there’s reserve capacity, and that’s not really covered in the ERCOT model. If you want the capacity there for an emergency, then someone has to provide it. And, unless you are willing to pay for it through the PUC (Public Utility Commission) or ERCOT, people are not going to do it from the goodness of their hearts — it’s too much money.”

Just as the economic incentives do not support the maintenance of a lot of reserve capacity, they do not provide much support for the storage of natural gas and other fuels. It’s a just-in-time world for energy, and that creates both efficiencies and risks.

Another engineer, writing at Judith Curry’s “Climate Etc.”, observes

Providing extra generation capacity, ensuring committed (firm) deliveries of gas during the winter, [and] upgrading transmission facilities are all expensive endeavors. Premiums are paid to ensure gas delivery and backup power and there is no refund if it’s not used. Such actions increased the annual budget and impact rates significantly for something that is not likely to occur most years, even if the extreme weather projections are appropriate.

Was this just a “black swan” event that nobody could reasonably plan for?

Maybe, maybe not. “We need to rethink how we approach extreme phenomena,” says Panos Moutis, a systems scientist at the Scott Institute for Energy Innovation at Carnegie Mellon University. “We need to consider whether these extreme phenomena are going to be appearing more and more as we go further down the line.” There’s a lot of math involved in answering that question, but there are also non-technical factors in play. “At the end of the electrical grid are customers, including families with children, old people, sick people, hospitals — these should be at the head of every concern about the electrical grid.”

Lee concurs:

From the customer point of view, the biggest complaint we have is that ERCOT’s rotating outages was supposed to be a planned outage. We expect that if it’s a planned outage, then we should be informed about when power will be out and when power can come back on, so that people will be able to prepare. If I can prepare, then I’m willing to work with the utility system, because this is an unusual situation. But if you leave people in the dark. . . . I had a friend who lost power for 50 hours — that’s not a rotating outage.

The narrative, as usual, is way out ahead of the hard data. “An awful lot of information doesn’t exist yet for us to understand what happened,” Houseman says. “There are technical, regulatory, and legislative reasons it happened — and there are technical, regulatory, and legislative things that need to happen to fix it. There’s no silver bullet.” His advice: “Be skeptical of what you hear, especially on places like Twitter.”

“This was a rare but high-impact black-swan event,” says Pete Wung, chairman of the IEEE Smart Grid Program. But rare is not the same thing as unforeseeable or unthinkable: The 1989 deep freeze that hit Texas and Louisiana is one precedent, and unusual winter storms as recently as 2011 have strained power systems.

In any case, though, the main challenge for Texas and the surrounding states going forward probably is not going to be blizzards that happen once every ten years, but summers that are hot every year, with bigger populations living in bigger houses making bigger demands on the system during the air-conditioning season, which is to say most of the year.

What can be done, and what will it cost?

Substantially improving the North American grid would mean adding about 9 percent to the existing stock of high-voltage lines at about $2 million a mile, or a total cost of $84 billion, Amin says. For that, we’d get a “stronger, smarter, interconnected backbone.” Reorganizing sections of the grid into semi-autonomous microgrids, as has been done at facilities such as universities and industrial parks, is another possible improvement. States in New England have successfully experimented with AI-based systems that predict failures and demand spikes in their systems, something that could be replicated relatively easily elsewhere.

There’s always the go-to solution of simply building more capacity, too. But that isn’t straightforward, because building enough capacity to weather a storm like the one that paralyzed Texas would be an extraordinarily expensive undertaking. “How are utilities going to get approval to spend the kind of money we’d have to spend for a black-swan event?” asks Steve Collier, the vice president for business development at Conexon. “The number of gigawatts of capacity that would have to be built, the strengthening of the utility systems, would not be worth it for one event.”

Is this going to happen again?

It very well might. This wasn’t one failure but a complex of interrelated failures, and fixing one or two of the issues won’t fix the overall problem.

“Everything went wrong,” Amin says.

For perspective, Houseman points out that Texas was obliged to try to produce enough power to cover nearly a month of ordinary usage in four days.

The unmet demand, Moutis says, amounted to about 1 terrawatt-hour, equal to the consumption of a small country. “The rolling blackouts were not as rolling as we’d wanted. We have to understand how severe it was, how beyond expectation it was, beyond our planning,” he says. “One failure leads to cascading failures, and that takes us further from what we can guarantee in terms of system reliability. This was beyond what our system could guarantee.”