The oil and gas industry is now leveraging generative artificial intelligence (AI) to streamline field operations, ensure regulatory compliance, and enhance carbon footprint management.
Before the advent of ChatGPT, an interface generating content in response to natural language queries, the sector was already a keen adopter of traditional AI, primarily used by programmers.
At the CERAWeek conference in March, Chevron CEO Mike Wirth highlighted the use of AI in subsurface imaging and production planning in the Permian Basin, the largest shale oil and gas field in the United States.
Overcoming Data Challenges
However, extracting the vast amounts of data generated by drilling has traditionally been challenging for major industry players, noted Tim Hafke from AlphaSense, a data analysis firm.
These data sets come from various sources, making them difficult to synthesize into a digestible format. This is where generative AI comes into play, he explained.
This more accessible version of AI, usable via a natural language interface, also allows for exploring digital replicas of refineries and LNG (liquefied natural gas) terminals, known as digital twins, which are becoming more widespread in the industry.
Digital twins help solve operational problems encountered by physical installations.
Predictive AI Models
The energy sector also relies on predictive AI models to forecast equipment wear and enable replacements before failure occurs.
For Matthew Kerner, head of cloud computing in the industry at Microsoft, this is one of the entry points for generative AI, which provides detailed responses, images, or sounds from a written query in natural language.
Some clients use generative AI to understand why models make specific predictions, Kerner said during a CERAWeek roundtable. Generative AI acts like a connector linking various AI interfaces.
Providing Context
In the field, an employee can ask the generative AI about the temperature, pressure, and humidity of problematic equipment, which aids in diagnostics, added Rob McGreevy from Aveva, an industrial software specialist.
Generative AI provides context to people who need to make decisions, even outside their expertise, emphasized Matthew Babin from Palantir, a leading data analysis firm.
In an industry that often requires total or partial shutdowns for inspections and maintenance, an interface gives access to maintenance guides to consult on procedures, he added.
Moreover, digital modeling of infrastructure offers a broad perspective to answer practical questions.
If you are undertaking maintenance, you need to determine, for example, if there is space for a ladder here or scaffolding there, illustrated Rob McGreevy.
Enhanced Efficiency and Carbon Footprint
Faster and more suitable responses, preservation and optimization of equipment through historical data analysis, and more efficient exploration and extraction contribute to a better carbon footprint.
However, these gains are tempered by the fact that generative AI consumes a lot of energy to run data center servers.
In a highly regulated environment, the industry can also rely on generative AI to ensure that equipment and personnel comply with current regulations by integrating the latest updates into its models.
Broadening Access to AI
The interesting aspect of generative AI is that it is usable by individuals other than those who typically interact with traditional AI, such as IT professionals, analysts, and executives, explained Dan Bennett, chief technology officer at S&P Global Commodity Insights.
Now, field staff and administrative employees, who were not previously involved, can benefit from these tools.
These tools can help the next generation of workers adapt, argued Rob McGreevy. There is a way to significantly shorten the time needed to master the operation of installations.