Ásgeir Daniel Hallgrímsson
Decision making and control in present-day operation of power plants and other industrial systems is hampered with floods of alarms and other plant information in abnormal operating situations. This makes it difficult for operators to identify critical alarms, assess abnormal situations, and make adequate control decisions; hence industry needs methods and tools, which can be used to extract system descriptive and decision relevant information from sensor data.
Multilevel Flow Modelling (MFM) is a methodology for functional modelling of complex industrial systems that provides the means for representing and reasoning about the knowledge of the system that cannot be done by quantitative approaches based on first principles. MFM expresses the modelled system in terms of a means-end structure; goals and objectives are met by elementary functions that describe the mass, energy, and control flow in the system. The relation between functions and objectives facilitates MFM reasoning which is the analysis of their influence on each other. The reasoning relies on deviating the states of the modelled components away from nominal behaviour.
The current state of MFM theory lacks an understanding of what constitutes as a deviation of model states away from their nominal behaviour.
The purpose of the project is to develop methods and tools for integration of signal processing techniques for sensor fusion and statistical event detection with evaluation of process system states on a symbolic level.