Is artificial intelligence the future of E&P?

AI companies are lining up to work with the oil-and-gas sector, but challenges remain.

Artificial intelligence expert Michael Simioni admits AI makes some people nervous, particularly in the oil and gas industry.

“Subject-matter expertise is highly regarded in oil and gas,” says Simioni, who is the APAC regional director for AI startup Endila. “People build careers around subject-matter expertise – they build decision-making capability around it and they build authority in the industry around it. And there’s this perception AI will take away from that subject-matter expertise.”

A large part of Simioni and his team’s job is reassuring oil and gas stakeholders that AI won’t erode mid-to-senior-level roles. “In some industries, AI does remove a workforce in some capacity,” he concedes. “A lot of data preparation, for example, can be almost fully automated with artificial intelligence screening and sorting techniques.

“However, in oil and gas, AI is actually a complementary component of your workflow. There are plenty of very routine, very trivial, very data-intensive tasks that oil and gas staff do in their day-to-day jobs that could be done by AI so that they can concentrate on the work that adds more value.”

He provides an example: “We have clients that are looking at hundreds of well logs for downhole drilling. They’re looking at 600, 700 wells and trying to find the 10 that have the amount of data they need to go to the next step of analysis. They can either wade through these 600 or more datasets, or an AI platform can nominate candidate wells that they can then consider for further analysis. It’s not removing experts from the workflow – it’s about adding efficiency.”

Numerous integration opportunities

Tech companies and oil and gas decision-makers seem to agree that AI has a role to play in the sector’s future. According to a 2019 survey by EY, 92 per cent of oil-and-gas companies are either investing in AI or plan to in the next two years.

But Simioni says many oil and gas companies have yet to incorporate AI in a meaningful way. He also points out too many AI developers participating in the E&P sector are applying generic platforms and models rather than building AI solutions specifically tailored to oil and gas.

Endila, he argues, is different because it considers oil and gas to be one of its core markets. The company, which is headquartered in the UK and has an office in Melbourne, assists oil and gas stakeholders with AI strategy and also develops POCs for artificial intelligence applications within oil and gas businesses.

When an oil and gas company engages with Endila or one of its peers, it finds the opportunities for AI integration in E&P are numerous.

Says Simioni: “One compelling example is natural-language processing, which is a sub-discipline within machine learning where the system is able to understand the sentiment of terms that are related to a specific context.

“For example, in a lot of drilling software, there’s a free text field where the driller or rig manager will write their notes regarding their productivity in the well, what went wrong and any other relevant observations. NLP allows you to look across all of your wells and see what was said when certain events or incidents occurred.”

Simioni says NLP has two main benefits for a company involved in drilling. “The first is that you’re able to do a retrospective analysis of your drilling workflow and identify where you can optimise,” he says. “The second, which is a bit more experimental, is predictive real-time analytics to avoid problems down the line.

“Say you, the operator of the drilling platform, are writing in the free text box. You will say: ‘We started drilling an offset well in a certain direction and received under certain conditions’, and the system will identify that you’re leading yourself into a loss zone or are about to find yourself in a difficult situation.”

Other AI programs for E&P are designed with one simple goal: to save people time. “When an operator gets all their petrophysical data from downhole, there are a lot of inconsistent attributes that need to be removed,” says Simioni. “An AI platform can do that very quickly and across thousands of wells, and then it can present its findings to a petrophysicist, who can either say: ‘Yes, that’s a good analysis, I can build on that’, or: ‘This isn’t quite right, what is wrong here?’”

Maintenance is another area Simioni says could be tightened up with the correct application of AI.

“You can have a pump that is pumping away, with a vibration detector on it,” he says by way of example. “What the system can do is learn from the normal running of the pump over its previous history. When there is a perturbation in the vibration – and that could be very subtle, a slight change in frequency or oscillatory motion – it can notify a maintenance worker that that pump needs to be serviced or it can trigger the SCADA system and shut down the pump before a catastrophic event occurs.”

These are just three examples of hundreds contained within AIForesight, an Endila database that aggregates different E&P use cases published in the public domain. Endila uses the database to demonstrate to clients the breadth of AI applications that are now possible.

Reshaping the future

It’s not just Simioni and his colleagues who see AI playing a major role in the oil-and-gas sector’s future. Some commentators believe AI has the potential to fundamentally reshape the industry and help address the ‘brain drain’ afflicting the sector.

“AI has the potential to address the oil and gas industry’s core talent challenges in three ways,” say Jeff Williams and Keith Strier from EY. “First, by making existing oil and gas operations more efficient, so fewer resources are needed to maintain them. Second, by codifying some of the knowledge of retiring engineers into virtual agents. Thirdly, by making the industry more attractive to younger hires.”

It’s this third point that Williams and Strier suggest is the most significant. AI, which promises to make E&P safer and cleaner, could help convince many environmentally conscious young professionals to take up jobs in oil and gas.

But even AI’s biggest supporters, such as Simioni, believe there’s a limit to how widespread the use of the technology will become in the near term.

“There’s been a lot of AI research by external parties and academics that sits in reservoir engineering, production and drilling, and those are very sensitive parts of the industry where the E&P operators differentiate themselves,” he says.

“It’s unlikely these discipline areas will be the first to throw themselves comprehensively into AI, except maybe for the largest operators who can afford to employ armies of data scientists and have the historical data to work with. We believe greater uptake of AI in the oil-and-gas industry sits instead in things like health and safety, planning and predictive maintenance – things that help everybody equally and could be de-risked through collaboration.”

He adds: “Anything that can affect production and reserve estimation is ‘Do Not Touch’ as far as we’re concerned. The industry would rather trust a human to do that for the time being, even if AI could potentially provide a benefit.”

AOG 2021 will run 10-12 March at the Perth Convention and Exhibition Centre.

To stay up to date, subscribe here.

  • Stay up to date with the latest news, industry insights and AOG Energy updates.
  • Subscribe
Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now