When it comes to event detection, a common question we receive is, what kind of events can AI detect? We sub-categorize these events of interest into four different types: 1) High-Risk Events, 2) Sub-Optimal States, 3) Events Needing More Lead Time, and 4) Silent Killers.
We will explain these event detection types below, but please realize this is not an exhaustive list.
- High Risk Events: Catastrophic events which cause cessation of production immediately, and requires replacement of equipment and rig work. A good example of this will be broken shaft on an ESP, the operator will have to schedule a work-over and replace the pump.
- Sub-optimal events: If you are working in a place, say Bakken, you might see freezing on gas chokes quite frequently. This might not lead to an immediate cessation of production but over time it leads to a drop on oil production which accumulates to a significant volume over time.
- Events Needing More Lead Time: Events that are detectable by field personnel, but AI can help you detect them earlier. A hole in tubing on a gas lift well may cause the production to plummet. However, using AI we may detect the hole a couple of days or a week before the production drop is too apparent.
- Silent Killers: These are events which we call the ‘Silent Killers’. We recently heard from an operator in the Permian that they generate 6000 dynacards on their rod pumps in a month. There are so many wells and so few engineers to analyze these. Chances are that some of the vital information is missed which causes a failure at a future stage, and the dynacards are evaluated much later after realizing that the production was down for days.
If any of these events are occurring in your oilfield, we would love to speak to you about how we can help with event detection and optimizing your oilfields through the use of our AI solutions. To learn more about our solutions, view our Production Intelligence Solutions page or Contact Us directly.