Key Objectives of Bringing Expert-Guided Machine Learning to the Digital Oilfield
Artificial intelligence (AI) and machine learning (ML) promise to drastically improve the efficiency and output of today’s oilfield. Here are just some of the objectives we are striving for:
AI models can monitor each well 7 days a week for 24 hours a day, automatically detecting and diagnosing any sub-optimal operating conditions across every field as soon as the problem starts. Operators and engineers don’t have to spend hours a day reviewing the operations of each well.
Instead, they can let automated systems handle this tedious, repetitive monitoring task and feel confident they will be notified of any potential problems immediately. This allows operators and engineers to focus their expertise on fixing the underlying issues and optimizing performance before losing production or incurring unnecessary wear and tear on the pumps.
Computer systems are much better at watching other machines than humans are, and these automated systems analyze each and every well constantly. This should free up an engineer’s time where their wisdom is required, such as designing, planning, and field management.
Today’s control systems can monitor pressures, temperatures, and other signals, and they can make simple changes like shutting down the pump when some threshold is exceeded. However, these simple threshold ranges cannot handle the combinations of patterns across different signals and the underlying petroleum engineering models that reveal what is happening at the well.
Today, operators and engineers must actually watch these signals and make judgments based on their experience and different calculations and models. This monitoring can be very tedious and consume a significant amount of time from this skilled personnel.
The real value comes from correcting the problem and improving the operation, not watching and waiting. Automated systems can handle the monitoring and point operators precisely to the priority problems that need to be addressed quickly.
To hear about these objectives and OspreyData’s Expert Guided Machine Learning, access the webcast replay featuring Venkat Putcha, Ph.D. (Data Scientist and Petroleum Engineer) and Mike Pennell. Access this webcast replay entitled “Is it really Man vs. Machine? Expert-Guided Machine Learning & Your Digital Oilfield”.: