AI Helps Researchers Find Lost Oil and Gas Wells
The oil and gas industry is increasingly embracing artificial intelligence (AI) to drive innovation and improve efficiency. With the promise of purpose-built AI tools, companies aim to streamline operations and control costs with unprecedented precision.
Just a few years ago, achieving this level of optimization and granular control seemed out of reach, but advancements in AI are making it a reality.
When we think of oil and gas exploration, the first that comes to mind is the immense economic opportunities and the potential for striking it rich. Oil is known as the "black gold" in reference to its significant economic value. However, this narrative often overlooks the environmental and climate challenges. Undocumented orphaned wells (UOWs) pose serious safety hazards. They leak methane, contaminate water, and release toxic chemicals, harming the environment.
Experts believe that there are hundreds of thousands of oil and gas wells in the United States that remain undocumented and unowned.
(Zia-Liu/Shutterstock)
Researchers use a variety of modern tools to help them find and plug UOWs, including laser imaging, drones, and sensors. However, with over 3 million square miles to cover in the contiguous United States, these methods can be inefficient. This is where AI steps in. It can enhance the process and make it more effective across such a vast area.
Scientists at Lawrence Berkeley National Laboratory and other institutions have developed an AI system to analyze 45 years of historical U.S. Geological Survey maps, identifying 1,301 potential UOWs in California and Oklahoma.
“While AI is a contemporary and rapidly evolving technology, it should not be exclusively associated with modern data sources,” said Fabio Ciulla, a postdoctoral fellow at the Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab) and lead author of a case study on using AI to find UOWs.
“AI can enhance our understanding of the past by extracting information from historical data on a scale that was unattainable just a few years ago. The more we go into the future, the more you can also use the past.”
Starting in 2011, the United States Geological Survey began uploading 190,000 digitized topographic maps dating from 1884 to 2006. These maps are geotagged, which allows each pixel to be linked to specific geographic coordinates for easy reference. Geotagging has been crucial, as it links each map pixel to specific geographic coordinates, enabling the AI system to detect and analyze potential orphaned wells accurately.The flagged UOWs are verified using satellite images, field surveys, and sensors like magnetometers.
The Berkeley Lab research team trained the AI models to identify the correct points on the maps and avoid false positives. The model also has to work with maps of varying terrain, colors, and conditions. Some of the old maps were torn and stained, making it challenging to work with.
“For a human being, looking at this circle and recognizing it is extremely easy,” Ciulla said. “Until recently, this was the only available method to extract information from these maps – but that strategy does not scale well if we want to apply it to thousands of maps. This is where artificial intelligence comes into play.”
While the AI tool developed by the Berkley team relies on AI, it does allow manual adjustment. The researchers can vet the findings to ensure the AI is correctly interpreting the symbols on the maps. The researchers have verified 29 of the UOWs discovered by the AI models using satellite imagery, and another 15 through field surveys.
So far, the researchers have taken a conservative approach, prioritizing fewer false positives over false negatives. While the results serve as a proof of concept, given the scale of undocumented orphaned wells across the United States, the model will need to scale rapidly.
Looking ahead, the researchers aim to combine AI’s predictive ability with modern tools like drones. Preprogrammed with set fly routes, the drones equipped with magnetometers could access areas where aerial detection is challenging. The drones could also be equipped with sensors or LIDAR-equipped planes could be used to check the air for leaks. This combination of AI and advanced technology could pave the way for better speed and accuracy in locating and plugging UOWs.
