BAKU, Azerbaijan, June 1. The oil and gas
sector aims to predict risks in advance, Machine Learning Manager
at Caspian AI Institute, Ulvi Zamanbayov, said at a panel
discussion within the framework of the 31st International Caspian
Oil and Gas Exhibition in Baku today, Trend reports.
He noted that the main priority in the application of artificial
intelligence (AI) in the oil and gas industry is safety, and
therefore, the transition to fully automated, closed-loop systems
should be carried out in stages.
According to him, since the oil and gas sector is a high-risk
area, the cost of mistakes here is very high, and safety should be
at the center of all decisions.
"We aren't trying to fully automate everything right away. First
of all, it's important to ensure the security of systems, maintain
human control, and ensure maximum participation of the business
side in the processes. Only then can we move to broader
automation," he explained.
Zamanbayov pointed out that the main goal in the industry is to
move from a reactive approach to a predictive and recommendatory
approach.
"Instead of solving problems after they occur, we are working to
identify risks in advance, predict production, and propose the most
optimal action options," he emphasized.
According to him, the Caspian AI Institute, established at the
initiative of SOCAR, is working on the development of artificial
intelligence solutions by bringing together specialists in various
fields.
"Our team includes geologists, chemical engineers, software and
data science specialists. This approach allows us to better
understand business needs and industry-specific features. It is
also important to involve user representatives in the process as
product owners during project development," said Zamanbayov.
He added that the institute prioritizes the creation of scalable
products that can be widely used across the company rather than
solving individual problems.
One of the projects presented during the panel was the "Virtual
Flow Meter" system, which is currently being developed jointly with
SOCAR.
According to Zamanbayov, this project is based on "soft sensor"
technology for more accurate assessment of production in wells.
"Previously, in many cases, well production was measured once a
month, and it was assumed that this indicator remained unchanged
throughout the month. However, production is constantly changing.
The Virtual Flow Meter system allows you to predict production on
an hourly or daily basis using high-frequency data such as
pressure, temperature, and control equipment indicators," he
noted.
According to him, this approach allows for more accurate
monitoring of production processes, improved decision-making, and
increased operational efficiency.
Stay up-to-date with more news on Trend News
Agency's WhatsApp
channel
