AI and machine studying (ML) are used as a instrument within the oilfield to optimize hydrocarbon output and predict failures earlier than they occur. However the business nonetheless has some kinks to work out.
Engineers nonetheless should make changes by way of trial and error to seek out the candy spot between know-how and human intervention, business consultants mentioned on the Society of Professional Engineers’ (SPE) Artificial Lift Conference and Exhibition on Aug 20.
Superior information analytics present the required info to design synthetic raise methods. However the complicated information additionally exhibits the business is much from being fully hands-off, Courtney Richardson, an engineer at Occidental Petroleum, mentioned on the convention.
“The explanation that I say that’s in the event you’ve checked out any high-resolution surveys for wells these days, we try to pump by way of some very complicated wellbore geometries and sucker rod pumping that creates a singular problem,” Richardson mentioned.
Producing a number of wells from a single pad complicates the issue.
“If you happen to’re multi-well pads, these 4 nicely trajectories are very totally different nicely by nicely, so we take these case by case,” she mentioned.
Whereas software program and associated algorithms are useful, there may be nonetheless no substitute for human enter, Richardson mentioned.
“I feel that the predictive design software program that we use is nice and it provides us a elementary base for designing our wells,” she mentioned. Nonetheless, there’s a giant quantity of enter inside the predictive design software program that the business doesn’t absolutely perceive, she added.
On the design stage, a hands-on strategy is critical to know the logic and construct round attainable eventualities. However that doesn’t imply Occidental isn’t benefiting from AI and ML the place it may well—corresponding to in surveillance and optimization, Richardson mentioned.
Though high-level information analytics, ML and modeling know-how are economically possible for a lot of wells, there’s a cost-based restrict on its software, particularly for these managed by a small group of engineers.
“We have 8,000 sucker rod pumped wells and the overwhelming majority of these are very low marginal producers,” mentioned Chevron engineer Amine Zejli. From an financial standpoint, it doesn’t make sense for Chevron to spend cash to optimize each nicely, Zejli mentioned. Dependable automation is crucial.
The mix of applied sciences revealed beforehand unrecognized nicely dynamics, enabling the business to be extra aggressive with failure mitigation efforts.
In the end, the objective is to construct capabilities to handle particular discipline wants and engineering issues as they come up.
Occidental can also be attempting to optimize energy utilization by making use of synchronous electrical motors—the place the rotor rotates on the identical pace because the magnetic discipline within the stator—of their beam pumping models.
“So we have put in a few these everlasting magnet motors on pumping models, and we’re actually early in these trials, however we all know that it has been wildly profitable within the ESP [electric submersible pumps] world,” Richardson mentioned.
Artificial raise has come a good distance from the times of pump-jack pushed sucker-rod pumps. The trail for future generations can be paved by digital applied sciences and superior surveillance strategies to adapt to an ever-changing panorama.
“Artificial raise wants digital twin know-how to develop,” mentioned Lawrence Camilleri, CEO of Camilleri & Associates. Digital reproductions have to be physics-based in order that the basics are captured. The worth of this strategy lies in evaluating potential to precise manufacturing because the roadmap for optimizing hydrocarbon output, he mentioned.
The business additionally must embrace a holistic strategy.
“We cannot simply be synthetic raise engineers, we cannot be manufacturing engineers, we cannot be reservoir engineers, we are going to act as one engineer that sees no limits to the appliance of math and physics to resolve the business’s most difficult issues,” Camilleri mentioned.