Page 19: of Offshore Engineer Magazine (Jan/Feb 2025)

18 20

You can’t just place an algorithm on top of a mass of data, adding use cases for machine-learning, at least with OKEA.

Foss-Sjulstad says. “It’s about contextualizing the data and Yet, it’s their well-test app that has convincingly won over a triggering algorithms in a correct way,” so having data infra- client list that’s the Who’s Who of hydrocarbons. structure in place is key for both producer and AI supplier. “The well-test app … is fairly off-the-shelf, standardized and works from the get-go,” Foss-Sjustad says, adding that

SEEING HUMAN ERROR deduction testing has also won over clients because “its

Gathering and interpreting client data sets are how Foss- crucial to know when you’re not getting enough informa-

Sjulstad and her overwhelmingly under-35 team of data tion during the additional time when you’re shutting in a scientists ?nd solutions to challenges that range from oil- well. Being able to see the uncertainties and how they con- water remedial work to models that don’t re?ect well tests. verge and be able to say, “Okay, now I can actually start up

Another challenge lies in showing the limits of physics- my well”.” Beyond the algorithm, visualization, can now based production modelling. A data-driven offering ap- end a test earlier than once thought correct, and at least pears to offer easy data control and model maintenance. one client is on the record saying they avoided deferred

Engineers’ fears should be eased by knowing Solution production worth about 6 million kroner (USD 600,000)

Seeker products and tech have since 2013 been developed per well test (they average 12 tests a year)! “It’s fairly sim- in close collaboration with production experts. ple. It creates value from the get-go.”

Virtual ?ow metering is a “tricky and complex” ?eld of Solution Seeker’s recently released Neural Compass technology. The key info is the well test, which Solution VFM app is used as both backup or as sole evaluator of

Seeker de?nes a good when “you’re actually approving tests ?ow. Some oil companies might also use a mechanistic and you have a very standardized work?ow” while being MPFM which would have to be “fused” to it to add “in- able to easily import the well test data, other standard in- formation value”, making setting up “out-of-the-box” dif- strument data like pressure, temperature, choke sizes etc., ?cult and highlighting an industry with no “off-the-shelf” when you’re “good to go”. Solution Seeker say they see hu- app to relay on for production. It doesn’t help that “Data man mistakes during well tests, as their tech autodetects infrastructure isn’t standardized across producing ?elds, well tests and provide info on “uncertainties” for different (digital) platforms and different companies”. Custom set- test results. “When we look at previous well tests, we see ups, the ?ne-tuning of existing apps and the adding of new that there has been … a lot of human error. So, if it isn’t features “in-stride” appears to be the rule. very intuitive, then you might not enter a certain time that perhaps you should have looked at to recognize that per- TRAINING AI AND STAFF haps this wasn’t a really good well test,” Foss-Sjustad says. Asked whether operator employees or the AI itself had

When talking to Solution Seeker, it is dif?cult not to to be trained, Foss-Sjulstad says, “It depends.” ?ash back to a time — not long ago — when Increased Oil “VFM (Virtual Field Manager) is a service we deliver

Recovery and Enhanced Oil Recovery were all the rage, mostly from our side, training and updating the model as before cost-cutting became the thing. Solution Seeker, de- well as deployment. What the client sees is an updated signal, spite its software start-up vibe (it is an NTNU university like any other tag. Training enters the picture on products spinoff), did, in fact, emerge out of that era (in 2013) and that are work?ow based. For the well test app, you need to was created to answer all those needs. Their young vibe have a tutorial with the client…The same applies to a prod- is real, however, although anchored in solid, grey-haired uct we have (Neural Compass), where you create actions, and reservoir-?ow and well-integrity experience. you can cooperate with the offshore engineer or the offshore operator ?rst to discuss an action and then to communicate

DEFERRED PRODUCTION back when done. That’s when you need to train engineers.”

The staggeringly wealth-generative and tech-heavy Nor- As everywhere, selling AI isn’t all smooth sailing. Gone wegian oil industry is well onboard with Solution Seeker’s are the fundraising rounds, the Norwegian Research approach. Only Equinor and Svall are understood to not Council funding and the help from known innovators. be using Solution Seeker tech among Norway’s operators. It’s hard, too, to convince proud engineers using trusted

While their current AI offering streamlines work?ows physics-based models to open up to data-driven produc- and backs up physics-based production modelling, Solu- tion management, however sound. tion Seeker are nevertheless upping their AI game and also That, too, “depends on the use case”.

january/february 2025 OFFSHORE ENGINEER 19