How is it possible to indicate root cases in a process, based on a stream of process sensor data?
The plant design (as documented in P&IDs) are modelled in a functional language where AI is used to build the relations between root cause and effect using mass and energy balances.
The result are knowledge graphs, representing the propagation failures in the plant where sensors, actuators, operational tasks, and measures are documented together with the consequence of the failure.
When the built plant model is exposed to live sensor values, it is possible to identify root causes and forecast consequences in real-time.
Since the model combines thermodynamics and operational experience and procedures, Kairos has established a standardised functional model library where our clients can get and re-use the captured learnings among their projects and plants.