The Control Room Assistant (CRA) has been developed to give control room operators real-time support by detecting abnormal situations and identify root-cause, predict future consequences, and suggest mitigation plans.
The CRA will always be on alert and immediately present the root-cause with its consequence behind an abnormal situation and recommended counteraction steps. The objective for CRA is to make sure the operator is assisted with cause/consequence information and counteraction suggestions on relevant observations during operation.
The CRA uses a stream of real-time sensor data from the control and safety systems as input to a functional deterministic model or digital twin. The model starts with the explicit plant design parameters and combines human tacit knowledge and experience to improve the results. Core technologies is MFM (Multilevel Flow Model), rule based explainable AI with propagation analysis.