Validation of engineering designs represented in OWL according to ISO 15926
Intervenant⋅e⋅s
Résumé
ISO 15926 is a series of standards for the integration of life cycle data for process plants. The scope of the standards includes all life cycle phases include conceptual and detailed design, procurement construction and commissioning, and operation and maintenance.
The standards have a formal basis with a top-level ontology that uses a 4D approach to change over times, and a possible worlds approach to design. The ontology was defined initially using the EXPRESS information modelling language, but a representation in OWL has been standardised as part 12 of the series. The top-level ontology is used alongside a taxonomy of activities and physical objects relevant to process plants and their properties.
During the design of a process plant the functional requirements for a facility within the plant are defined, and then the type of equipment item to be installed is selected for procurement. The validity of a select equipment item for its intended duty can be validated automatically by inferencing engines. If it is not valid, then the reason why can be reported.
The rendering of the 4D possible worlds approach in OWL DL to support inferencing presents challenges, because using this approach much property and design data involves relationships between user defined instances of classes. The way in which this can be hidden from OWL to produce a useful and computationally tractable view is discussed.
Auteurs/Autrices
David Leal has worked on formal representations of engineering data for many years. He was a developer and user of finite element analysis systems, and produced early representations of analysis data as relational data bases. Subsequently he worked within ISO committee TC 184/SC4 “Industrial data” on EXPRESS data models in international standards for engineering analysis data and for life-cycle data for process plants. Recently he has been involved in transforming the data models in these standards into OWL ontologies.