Semantic Technologies for the Data Model of the Digital Twin

Abstract

During the lifecycle of an aircraft a huge amount of data is produced. Data is generated directly by the systems of the aircraft, e.g. through sensors and computers. Additionally, offline information is produced such as product documentation and records of maintenance, repair and overhaul (MRO) activities. This data can be valuable for different stakeholders to plan and optimize their processes. During repair processes the extent of damage as well as previous repairs of the affected part are important information to the maintenance and repair organization. However, this kind of information is often not known to the organization until a damaged component arrives at the facility. In case of historical information, it is sometimes not available at all. Furthermore, the heterogeneity of data sources and types as well as the necessity of interaction between different data bases pose a problem to building an efficient data network, which would allow access to data to all stakeholders in the MRO service chain. In recent years, the development of virtual representations of physical products and its environment, socalled Digital Twins, have gained increasing attention in the industry. One of the key roles of this technology is the implementation of product data management systems. Semantic web approaches have been demonstrated to be methods for building networks, which enable efficient data management. In this work, key contributors of the component lifecycle data network are defined and modeled in ontologies, which are then transferred into a graph data base. This includes event, damage and process ontologies, as well as a lifecycle ontology which links the information over time. This approach of using semantic technologies establishes database interoperability and, by logically linking the data, allows analyzing relationships as well as inferring new knowledge, e.g. additional sources of repair input data for more efficient planning of maintenance activities. The advantages of an integrated data network such as improved maintenance capabilities are demonstrated, analyzing the processes surrounding the visual assistance system for manual scarf repairs developed in a mutual research project with an industrial partner. The model allows to take advantage of already existing data sources and identifies potential gaps in the data network, in order to improve MRO processes. Also, the data model can serve as a basis to further develop other Digital Twin functionalities such as predictive maintenance applications.