Unless you’re an oilman, one doesn’t always think of oil and gas when prompted on industries
that will flourish in the coming decades. On the contrary, artificial intelligence will almost
always make that list.
This makes AI in oil and gas a rather unexpected marriage.
Nobody wants to be a bummer at the wedding, clapping unwillingly in an awkward stance,
especially if there’s actually a rather interesting picture to this partnership. Consider this:
Brian Williams—a professor at MIT who designed AI software powering the Curiosity rover on
Mars — partnered with ExxonMobil back in 2016 to design robots for deepwater exploration.
The logic seems sound. Hunting for traces of oil in terrestrial waters must seem like child’s
play if you have designed software for a rover to select scientifically significant rocks on another
planet.
ExxonMobil aims to apply AI to enhance it’s natural seep detection capabilities.
Natural seep is a phenomenon where oil escapes subsurface rock under the ocean floor and migrates
upwards. Half of the methane eventually escapes into the atmosphere, but the other half dissolves
into the water along with the heavier oil particles. These are presumably the trace elements
detected by the robots, while the AI models predict the position of the oil reservoir itself
based on flow patterns. Of course, since ExxonMobil has not reported the deployment of any such
robots in public, we can only surmise how they actually work.
Can yesterday’s data ensure oil and gas has a tomorrow?
There are other companies attempting a more straightforward method — using historical data. By utilizing data on previous drill attempts, companies such as Extract AI are now able to predict accurately which locations will produce the most crude oil and raw natural gas at the most efficient cost for an upstream producer. Considering the fact that oil and gas companies used to spend millions of dollars (or waste, as epitomized by this incident in the Beaufort Sea) simply on seismic and geological studies to determine where to drill, the aforementioned technology can definitely be a game changer in terms of oil exploration.
Having a better idea of where to drill isn’t the only facet of AI’s impact on the oil and gas industry upstream. In fact, AI is also changing how companies drill for oil. Shell for example, has publicised the role AI plays in its own drilling operations. Algorithms are used to guide drills as they move through the Earth’s subsurface, making multiple decisions every second based on historical data from exploration efforts. Such data encompasses both mechanical input from the drill bit and seismic data like temperature and pressure. Shell claims that the use of AI results in more efficient drilling operations with less wear and tear on the expensive machinery.
The recent boom in US oil production does little to disguise the fact that oil exploration
is essentially a race of what’s left, as oil companies are forced to search for resources in
increasingly challenging environments. This means the Arctic circle, the yawning depths of the
Pacific and Atlantic oceans, and war-torn regions of Africa. Existing frameworks and technologies
for oil and gas exploration will be tested to their limits, and the use of artificial intelligence
will prove to be one of the best tools to expand those limits.