Today’s artificial lift professionals may rely on tools like predictive analytics, remote monitoring, and automation, but success depends on a strong foundation in artificial lift fundamentals.
During the “What Does a Modern Artificial Lifter Look Like?” session at SPE Artificial Lift Conference and Exhibition on 27 August, panelists said technology and tools can facilitate work, but a firm grasp of the fundamentals is necessary to prevent the phenomenon of garbage in, garbage out.
Kreg Flowers, production engineer at Occidental, said base skills for artificial lift professionals haven’t changed over the years. “The foundational components of what made a good developer, what made a good engineer, make the same quality of artificial lifter today. And I would reduce it down to logic and chain of thought, intuition.”
Access to tools like artificial intelligence (AI) can be helpful but are not a panacea. “AI has made it possible to be wildly cross-disciplined and expand your skill set in very many different areas. So, in traditional development, you have a project manager and then a product owner and a developer. And they’re all different people with different intrinsic core skill sets. AI has made the developer think, ‘I can do the project management and the product ownership.’ And same with the product owner. Now the product owner thinks, ‘Well, I can be the developer now. All it needs is a couple lines of code,’” he said.
Flowers said it’s likely that people with deep skill sets have seen people using AI who think they’ve become an expert in a certain skill, though it’s obvious they’re not. “You can see there’s sort of a lack of depth. Maybe they don’t know how to extend that solution when something unknown comes their way. AI has a plan until it gets punched in the face with reality.”
Maggy Burns, operations supervisor at produced water at ExxonMobil, said the modern artificial lift professional needs to be willing to try new things. “I don’t know if I feel like there’s any specific necessarily skill set more than that behavioral skill of being willing, open to change and open for constant development.”
Beyond that, she said, there’s a basic level of understanding that is necessary in order to make good recommendations and decisions, and it requires starting with the right inputs. “Garbage in, garbage out. That hasn’t changed from the last 70 years of computers. If you’re putting bad data in, you’re getting bad information out. And so, I still think there’s a level of technical, both from physics-based, mathematical, operational, domain knowledge that you’re going to have to always have.”
And given the rabbit hole people can fall into because of the vast quantities of data now available, Burns said it’s important to balance that data with knowledge and common sense.
Araceli Rivera Mandujao, lead sales engineer at SLB, said that while productivity tools are assisting artificial lift professionals, it is important for them to develop and remain in touch with their expertise. “You need to really touch on an expertise because right now we have so much information. We have so much out there, so much noise. I call it noise. And it’s not easy to find that there is an error” served up by software, she said. They need to be able “to look at whatever’s in front of them and see, ‘This is a mistake, this is wrong, and how are we going to fix it?’”
Flowers said fundamentals don’t just apply to professionals; they need to be core in tools that are used, particularly any involved in making decisions. “If you’re creating a tool that’s going to be even assisting in making decisions, much less actually making that decision and implementing the solution, it needs to be very well grounded in the fundamentals.”
And as the oil patch increases its reliance on autonomous tools, there are times when humans must be in the loop, Burns said. “The thing I’m not willing to give over is anything that has an effect on safety. But I think everything else is up for negotiation. When I think reliability, financial loss, deferred production, short-term shutdown, we have to own that risk.”
The number of people with access to data and able to make decisions about artificial lift operations can complicate things, she cautioned. An artificial lift workflow may be watched by people within the operating company as well as those at service companies, plus lease operators, which means all these people could be making changes at the same time, as could closed-loop automation systems, she said.
“Who is owning the responsibility?” she asked. “If you see a person in the field make a change, we built a lot of bandwidth that says you let them make it because they’re seeing something everyone else isn’t, and learn from it.”
