After a slow start, the oil and gas industry seems to be eager to adopt all sorts of digital technology as they help companies keep costs lower while boosting efficiency. Artificial intelligence is being flaunted as the answer to all problems or, at least, a better answer to many problems than older approaches. And it is maturing.
First of all, it needs to be said that a lot of the people who talk about AI enthusiastically do not mean artificial intelligence literally, as in an autonomous system capable of making decisions on its own. What they most commonly mean is predictive and analytic algorithms, and the process that allows for the deployment in a huge variety of tasks in the upstream industry: machine learning.
While the hype is considerable, it’s not all without merit. The co-founder and chief executive of oiltech statup OilX, Florian Thaler, told Oilprice that “amidst the general hype, there is indeed an epochal shift: the current exponential growth in oil data from sensors and satellite is unprecedented and is not showing any signs of slowing down.”
But this data has to be clean and high-quality, Thaler explains. Once the quality and reliability of data collected are good enough, the data can be used to create a whole platform based on machine learning that effectively works as a digital oil analyst. And this is just one application of machine learning, often inaccurately referred to as AI.
Some believe 2019 is the year that will mark AI’s advance to practice from theory. One of these optimistic experts is Jan Ren, the chief executive of software developer Atomiton.
Ren told Forbes’ Mark Venables that “The biggest thing that will impact you as a company is that AI is going to go from theory to practice. So far it has been mostly theoretical, but people don’t understand how to do it and what it can do. I think now people understand its potential and more projects are being fulfilled and implemented. The infrastructure industry will start to push AI from the how, the technical, to the what, which means AI will be recognized by the problems it has solved instead of what data it collects.”
This is quite general as far as forecasts go, but here’s something rather more specific: machine learning can be deployed in more than one area of the industry. In the field, for example, it can help oil and gas producers see how a well’s yield will change over time.
“We have a lot of data points and we assume that if the well has produced long enough it will behave as other wells have performed in the past. … A neural network is a way that allows us to do this in some way,” Rystad Energy senior analyst on shale Alexandre Ramos-Peon recently told EPMag’s Velda Addison. He added, however, that “This technique only seems to work for wells that have a sufficiently long production forecast.”
This makes perfect sense, of course: the more historical data you have, the more accurate predictions for future well performance you could make, especially if you have the help of algorithms. And these algorithms are now so sophisticated, according to Ramos-Peon, that “You just throw all of this data to the computer and it will train itself somehow to guess the best value so the accuracy is as high as possible.”
The AI market in oil and gas has been estimated to reach US$2.85 billion by 2022, growing by a compound annual growth rate of 12.66 percent. No wonder, given how quickly the technology evolves and how many applications it could have in oil and gas.
Ren believes next year will demystify the concept of artificial intelligence as adoption expands and trust towards the concept grows along with understanding of how it works. Of course, this development is by no means risk-free: cybersecurity is just as much a problem in oil and gas—of not more—as in any other industry.
