ARTIFICAL INTELLIGENCE in
Oil and Gas Industry

The case is a deployment of a reliable application using AI for a petrochemical plant. The plant had two key challenges that was aimed to address with reliability application. The first was to improve the reliable operation both of the equipment in the plant and the processes in the plant. And the second challenge doing that at scale. That meant doing it at scale across the entire plant and all of the equipment in that plant and then also being able to replicate that for all of those 200 plants. An initial analysis was to look at the unplanned maintenance costs at this particular plant which translated to 100 million dollars on the table that the reliability solution could address. 

The goal of the application was to adjust any available data that included maintenance data, asset data and tag data. So sensor data can predict when things might go wrong. And that included events both at the equipment level and at the process level. This deployment covered five unique data sources and configured machine learning models for over a thousand individual piece of equipment. This was done by training and evaluating tens of thousands of individual models and identified issues on those thousand pieces of equipment both stationary equipment like heat exchangers or distillation columns and rotating assets like compression units and turbines.