Representing causality and reasoning about controllability of multi-level flow-systems

Safe operation of complex processes requires that operators maintain situational-awareness even in highly automated environments. Automatic reasoning can support operators as well as the automation system itself to react effectively and appropriately to disturbances. However, knowledge-based reasoning about control situations remains a challenge due to the entanglement of process and control systems that co-establish the intended causal structure of a process. Due to this entanglement, reasoning about such systems depends on a coherent representation of control and process. This paper explains modeling of controlled processes with multilevel-flow models and proposes a new framework for modeling causal influence in multilevel flow models on the basis of a flow/potential analogy. The results are illustrated on examples from the domain of electric power systems.