IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Texas Railroad Commission to Use AI for Seismic Reviews

The Railroad Commission is turning to artificial intelligence to address increasing seismic activity across West Texas. A machine-learning algorithm has been programmed to process seismic data and reduce backlogs.

Low-angle view of a freight train in motion.
(TNS) — As seismic activity continues to shake up West Texans, the Railroad Commission is turning to increasingly powerful technology to address the issue.

The agency is turning to artificial intelligence to improve the process of conducting seismicity reviews, which are conducted by the Underground Injection Control Department for injection or disposal well permits in areas susceptible to seismicity and in certain geologic zones.

"This is a first for the Railroad Commission," Sean Avitt, manager of the UIC, told the Reporter-Telegram by email.

His department has programmed a machine-learning algorithm to help with the large amount of data to be processed and digested. Tasks performed by the algorithm — along with some other changes — has allowed the UIC to wipe out a backlog of seismic reviews to zero.

Avitt said data quality has increased recently, which he attributed partly to the TexNet Seismic Monitoring Program.

"The TexNet Seismic Monitoring Program has increased seismometer density in Texas, which increases earthquake location accuracy," he wrote. "UIC's permitting seismic monitoring incentive has helped contribute seismometers to TexNet."

The University of Texas at Austin's Bureau of Economic Geology, which oversees the TexNet program, has published several reports — with more on the way — which have yielded a better understanding of seismicity in Texas, including better fault location data, he added. Oil and gas operators also submit fault locations in their UIC permit applications, faults that may not have previously been known, he wrote.

Like the agency's technical analysts who review the applications, the algorithm weighs many factors related to the number, severity and proximity of earthquakes and uses a decision tree to assign a grade to the review. The higher the grade, the more the permit would be allowed to inject. If the algorithm issues a low grade, the technical analyst will consult with the agency's seismologist on whether the application should be denied or allowed a minimal amount of disposal — 10,000 barrels a day. The algorithm has a high accuracy rate but it's the technical analyst who reviews the data and ultimately makes the final decision.

While artificial intelligence in reviewing seismicity reviews is a first for the agency, Avitt wrote that other applications are being evaluated to see if they may be good candidates for AI.

"Analyzing ways to incorporate automation of computing tasks is part of the drive to perform RRC's duties more efficiently," he wrote.

©2022 the Midland Reporter-Telegram, Distributed by Tribune Content Agency, LLC.