抄録
This paper describes the estimation technique of an anomaly cause and the interactive explanation technique aiming at mitigating the burden to plant operators in an abnormal situation of an engineering plant. These techniques use partial models of the plant that are constructed previously by considering structural, behavioral, and functional information. The estimation of anomaly cause is need by expanding the search space from the partial model corresponding to the component or system where an anomaly is detected. The estimated results of anomaly causes and the search space are shown by using a simple system diagram showing the connection of plant components, and estimation process is shown by the operator request. Thee techniques are demonstrated by several anomalous scenarios of an oil refinery plant.
