Translating MFM into FOL: towards plant operation planning

Motoura, Shota, Yamamoto, Kazeto, Kubosawa, Shumpei, Onishi, Takashi

arXiv.org Artificial Intelligence 

A plant is operated on the basis of its manual usually; however, it is not realistic that a manual contains instructions for all cases, especially regarding abnormal ones. For obtaining appropriate operation procedures for a wide variety of cases, multilevel flow modeling (MFM) has been studied ([1]-[3]). MFM is a functional modeling framework, in which a plant structure is expressed as a directed graph. The framework also has a set of influence propagation rules, which consists of if-then rules regarding the states of related components. If the state of a component has changed, the resulting state of the other components can be obtained by applying the rules in the forward direction. Conversely, given a desired state of a component, we can obtain the states of other components to be satisfied for achieving the desired state by tracing back the propagation rules. This leads an action to a desired state. Our contributions are as follows: 1) We propose a method to translate MFM into an FOL. This enables the application of techniques used in the FOL to MFM, such as inference engines and abductive reasoners [6].

Duplicate Docs Excel Report

Title
None found

Similar Docs  Excel Report  more

TitleSimilaritySource
None found