By Clifford Louis

Artificial Intelligence (AI) powered operations provide decision support to mitigate risks as well as actionable insights to eliminate unplanned downtime, bottlenecks in process/workflow, and reduce safety incidents. However, the decision support process in the industry needs to align with current physics-based simulation models and algorithms enabling explainable AI for key decisions.

The promising emergence of the differential programming paradigm will enable the industry to augment data-driven AI with physics, while domain-driven models will enable better decisions and insights. However, the differential programming paradigm is not mainstream yet and the current machine learning frameworks and toolkits do not provide adequate support for the differential programming paradigm.

What Next?

As the world emerges from the current turbulent times, it is quite likely the oil and gas industry will tread into a new world order of lower demand and unlikely alliances. This new normal will push the industry to efficient ways of operations with integrated teams breaking the existing cross-functional silos in both data and governance.

The new normal of lower demand and crude prices will need drilling efficacy to improve multifold with a larger set of integrated processes and data flow. This could include integrated reservoir models which can be rerun based on the real-time drilling data and production data from existing wells, or geological and geophysical interpretation models which can be rerun based on the real-time data from the drilling process and learnings from the previous drilling reports.

The industry will continue to invest in AI technologies, but the demand and expectation from the technology will be scaled. Successful operators in the new normal will be defined by their ability to blend in the current physics and engineering technologies with data-driven analytics, machine learning, and automation.