An Ontological Approach for Modelling Evolutionary Knowledge of Prognostic Method Selection

The selection of appropriate predictive maintenance methods based on the current state of a manufacturing system, its machines, and its components is not an easy task due to the multitude of physical and virtual resources available. Moreover, the value of the prognostic information provided by a prognostic method, when applied to a given machine, depends on the system structure and the production and maintenance planning process. Therefore, it is necessary to consider the impact of such information on the system’s key performance indicators to assess the real benefits of each prognostic method. Based on knowledge from predictive maintenance approaches, manufacturing system simulation, and production and maintenance planning, an appropriate semantic model allows establishing a shared vocabulary and understanding among all these fields, along with a description of their relationship. Thus, this paper proposes an ontology of domain termed Ontological MetaMaintain (OntoMM), which specifies the description of concepts and their relationships to provide evolutionary knowledge on domain structure which further specializes in information flows. It is a novel semantic architecture in an ontology network with three modules that address automated prognostic method selection in an industrial environment.