ERCOT's AI Planning Reduces Texas' Storm-Related Grid Outages

---

Written by Paul Wiseman for IIR News Intelligence (Sugar Land, Texas)

---

Utilities Learn to Weather the Storms

Winter Storm Uri, which struck Texas from February 13 to 17, 2021, led to the deaths of 246 Texans and caused more than $300 billion in damage. The Electric Reliability Council of Texas (ERCOT) failed on the "reliability" part, with 33% of its capacity offline during that event. Thousands of Texans faced rolling blackouts in deeply sub-freezing conditions.

In a webinar hosted by power industry media platform Energy Central, titled "From Reactive to Predictive: The Case for a Weather Intelligent Grid," three panelists, including one from ERCOT, discussed how planning and preparedness can, and has, improved those numbers. The discussion also covered AI's role in data monitoring and analysis.

According to Industrial Info Resources data, 4,252 power plants are under ERCOT's purview.

Panelists included Renny Vandewege. General Manager of Weather and Climate Intelligence at weather risk management firm DTN; Dustin McKen, Director of Asset Management at Black Hills Energy, which serves parts of South Dakota, Wyoming and Montana; and Raihan Khondker of ERCOT, who oversees their generation and transmission weatherization and inspection efforts.

While ERCOT has its own weather team in-house, Black Hills partners with DTN for their weather-monitoring and storm-preparation input.

Measurable Improvement

If anything good came from Uri, it was that ERCOT and other utilities were forced to acknowledge their unpreparedness and ineffective forecasting models. Khondker said ERCOT has employed programs both predictive and preventive in the past five years, and that improvement has been measurable.

"During Winter Storm Uri, at the worst point of the event, the number was approximately 33%, which means a third of the dispatchable generation fleets were unavailable when we needed them most."

Khondker pointed to Winter Storm Fern, the second-most significant storm following Uri, where the outage numbers showed improvement. "Despite the level of stress on the system, the peak force outage ratio dropped to approximately 7%."

By the Numbers

• 33% down to 7%: Amount of power ERCOT lost in Uri in 2021 vs. Fern in 2026, due to better preparedness
• 4,252: Number of ERCOT plants Industrial Info Resources is tracking

Using AI and Other Weather Tools to Move the Needle

Weather events used to be a scapegoat for utilities to blame for poor performance, said Vandewege, recalling reports he gave during his previous career as a weatherman. Especially, "we didn't see this storm coming." Now, he continued, "That's unacceptable in today's world."

Weather forecasting models today have greatly improved, all panelists agreed. While they are still not 100% accurate, utilities are taking the decision-making process more seriously, regarding deploying personnel and preparing resources such as emergency fuel supplies.

Khondker said storm preparation starts very early: "Coordination starts well before the event arrives. The structure becomes progressively more focused as forecast confidence increases," and "decisions are made based on forecast confidence and priorities."

AI's ability to process and send alerts on massive amounts of data is one of the keys, McKen said. Black Hills is looking into installing a quantity of early detection cameras, comparing that to a cadre of forest fire watchtowers, with humans watching for smoke. "There are opportunities to take cameras and put them on those peaks, and you've got AI running in the background monitoring for smoke."

Once detecting smoke, AI would alert humans for verification and appropriate action.

AI offers scale that would be impossible with direct human oversight, McKen said. It also allows for modeling based on experience from previous events. McKen spoke of "risk-based prioritization that definitely helps us by combining historical inspection data, performance trends during different weather events, and also facility-specific patterns."

For ERCOT, Khondker said they are working toward a model to pinpoint areas needing more focus, "so that we can identify those vulnerabilities ahead and select ways to improve the grid reliability."

Saying Black Hills also starts early, McKen added: "It's taken a lot of conversation and decision making internal to Black Hills to determine how we're going to do that. I don't think there's a silver bullet to this."

Next, "In partnership with DTN, we've had to build a matrix around confidence levels days out and build our action plan in that manner."

No More Silos

An integrated view of readiness adds real value; while data may be siloed, an event in one area of the grid can domino through others, unless the system is anticipated and prepared.

AI helps manage the utility's huge spread, Khondker said. "With ERCOT, we're talking about 1,460 facilities and even more, and it's growing each day, and we have over 6,700 substations, so the population is huge."

Irony of Using Energy-Intense AI to Inform the Grid

Using a grid-straining system like AI data centers to inform utilities of how to manage that grid in an emergency is an irony that's top of mind to the panelists.

Said Vandewege, "One of the unique things about AI is how much energy consumption it takes to run all of this, and where you're seeing that impact the utilities."

McKen added, "It's a different kind of problem that I hadn't experienced early in my career, as we think about taking loads the size of entire metro areas and trying to plug them into the grid."

Key Takeaways

• Uri being a wakeup call to ERCOT and others, residents and business are getting better reliability during more storms.
• Use of AI is helping improve modeling and planning, while still creating a drain on the grid.

---

About Industrial Info Resources
Industrial Info Resources (IIR) is the leading provider of industrial market intelligence. Since 1983, IIR has provided comprehensive research, news, and analysis on the industrial process, manufacturing, and energy-related industries. IIR's Global Market Intelligence (GMI) helps companies identify and pursue trends across multiple markets with access to real, qualified, and verified plant and project opportunities. Across the world, Industrial Info Resources is tracking over 250,000 current and future projects worth $30.2 trillion (USD).