![AI in the oilfield](/sites/default/files/styles/hart_news_article_image_640/public/image/2024/04/ai-oilfield.jpg?itok=gv6Ej9ww)
While some operators might be hesitant to implement AI in the oilfield, some experts are saying its living up to the hype.(Source: Shutterstock)
In 2024, all things technology seem to touch on artificial intelligence (AI)— both as specter and savior for various industries, including oil and gas. And while some operators might be hesitant to implement AI, experts say its living up to the hype.
“AI has enabled us to integrate all of our data and reduce our capex in seismic processing, interpretation and reservoir characterization, and we also have increased uptime,” Tanima Dutta, geophysical adviser for ONGC Videsh, said during an April 9 panel at the 2024 AI in Oil & Gas Conference in Houston. “In field development, AI has also increased workflow and efficiency, and on the production side, we have used AI machine learning models to get answers and actionable results and insights.”
AI is part of a broader spectrum of machine learning that has been used across various value streams in the oil and gas industry for years, said Adam Phan, technical delivery supervisor for Chevron. However, its recent takeoff can be attributed to a concerted effort to upscale its use while lowering computing costs.
“With the semiconductor industry going off the charts and all the different investments coming into that… there’s a lot of capital actually coming into [artificial intelligence], and that's improving the cost of computing, which held us back,” Sandeep Mukherjee, geoscience adviser for Callon Petroleum, said during the panel.
Niche AI
![Srimoyee Bhattacharya, team lead for Americas Portfolio at Shell. (Source: Energy Conference Network)](/sites/default/files/inline-images/Srimoyee-Bhattacharya.jpg)
While made up of large language models, AI in the oilfield oftentimes serves niche purposes. However, Shell’s Srimoyee Bhattacharya, team lead for Americas portfolio, doesn’t see that as a negative, but a boon.
“AI helps us, because we are able to find issues small and fix them... We don't rely on maintenance that has to be done every three years or two years. We are now predicting beforehand whether or not our machine needs maintenance and proactively managing,” she said. “That has saved a huge amount of time, increased efficiency and reduced waste overall.”
AI offers a multitude of uses in the oilfield: it can comb through older files and interpret data faster; predict maintenance needs; reduce the number of workers needed on a given operation; and assist in sustainable operations by helping reduce emissions.
“If you add photography, satellite photos and even drones, you start capturing emissions that you weren't even aware [of]. They're leaking from your pipeline and are so small that you weren't aware of it, but yet you can still measure that,” said Agustin Diz, an adviser for Blockchain for Energy. “Now if you start using AI to warn you of where there is a failure, you are able to start working on it.”
AI may have niche uses, but the technology’s use is just getting started.
![Chris Hanton, director of data solution at Ikon Science. (Source: Ikon Science)](/sites/default/files/inline-images/Ikon_ChrisHanton.jpg)
“Our approach to [machine learning] has been quite niche, quite focused on individual tasks,” said Chris Hanton, director of data solutions at Ikon Science. We see a lot of work in terms of predictions and subsurface volumes. Those models, at the time, were very powerful, but they're also very limited in terms of their scalability and how far we can start spreading them across different assets on the subsurface.”
Hanton said the potential for an “AI explosion is going is to be able to upscale it, being able to come up with solutions that can now be applied to multiple different concerns across the space.”
Despite all the benefits AI provides in the oilfield, it is not a magic bullet, Hanton said. AI needs to meet certain stipulations to properly collect and interpret data.
“Whether it be traditional or generative, AI starts with data, clean data, curated data,” said Rajeev Aluru, director of innovation and analytics at Smartbridge. “That’s so critical for us right now, and in generative AI it's even more important because the results come fast, and if the result is not validated and not done …using curated data, it is dead on arrival.”
Humans plus AI
AI has its detractors, some of whom worry only about the safety of the data the software is fed, but also the security of their jobs.
Such fears couldn’t be further from the truth, Raoushan Ummulwara, data and analytics manager for Chevron Phillips Chemical Co. said.
“It's not like AI is going to replace humans, but it's always going to be humans plus AI … when it comes to process improvement or any of those projects, there will always be some form of human intervention.”
Stephen Johnson, vice president of software consulting firm Improving, agreed.
![Stephen Johnson, vice president of sales at Improving. (Source: LinkedIn)](/sites/default/files/inline-images/SJohnsonImproving.jpg)
“It's the easy way out to say that doing something with AI is going to make us 20% more efficient, therefore we can cut people,” Johnson said. “You got to take the extra step … If you’re saying we want to be 20% more efficient, because by doing so the same amount of people can drill an extra well which is going to make us $20 million or better, then you keep the people because they're going after another well.”
While AI isn’t foolproof and plenty of trial and error will occur when implementing new solutions in the oilfield, the technology is here for the long haul, Hanton said. The days of niche solutions will be gone, and AI will be prevalent throughout the entire production process from upstream to downstream, he said.
“I think AI is going to become less of a niche toolkit and it's going to find itself more into products and applications that are used on the day-to-day basis.”
Recommended Reading
Artificial Lift Meets Artificial Intelligence, with Hit and Miss Results
2024-08-23 - Oxy and Chevron are among companies trying to find the sweet spot between data analytics and machine learning with human intervention to optimize production and predict failures before they happen.
AI & Generative AI Now Standard in Oil & Gas Solutions
2024-07-25 - From predictive maintenance to production optimization, AI is ushering in a new era for oil and gas.
ProFrac, IWS Taking the Garbage Out of Oilfield Data Transfer
2024-07-16 - ProFrac and Intelligent Wellhead Systems’ MQTT protocol promises to speed up communications at the frac site, not only by saving costs but laying the foundation for future technological innovations and efficiencies in the field, the companies tell Hart Energy.
Give Us a Signal: Tech Firm Ups E&P Coverage in Remote Plays
2024-07-02 - As E&Ps struggle to transmit information over public networks in out-of-the-way oil and gas basins, tech firm Digi is working to improve its reception.
Cracking the Fracking Code: Efficient Approaches to Optimize Wellbores
2024-06-30 - Technology and process innovations improve operational efficiencies even as companies scramble for greener fracking solutions.