Digitaler Zwilling für die Zustandsüberwachung und vorausschauende Instandhaltung im ATO

Predictive maintenance is a key requirement for autonomous systems, particularly in the rail sector where it enables autonomous rail operation. In order to perform effective condition monitoring and predictive maintenance the integration of operational data, engineering models, and hybrid predictive models is necessary and enables a centralised access to all necessary data. The digital twin (DT) concept fulfils this requirements and is technically realised in the Eclipse BaSyx Asset Administration Shell (AAS) environment as a central hub for capturing models, measuring data and maintenance information. The implementation makes use of Industrial Digital Twin Association-Standards (IDTA) and models the train components in a hierarchical structure. However, in the current BaSyx implementation manually creating the hierarchical DT structures is error-prone and difficult to standardize. This paper presents an adapter that creates and maintains hierarchical DT structures based on predefined concepts. The adapter facilitates the annotation of heterogeneous data and models, their mapping to a desired DT hierarchy, and their deployment into a standardized BaSyx-based AAS infrastructure by providing an easy-to-use user interface. The hierarchical DT structure resembles the hierarchical structure of the system and its models and allows the exchange of parts or components physically and virtually. The development of the adapter accompanied the development of physical, data-driven and hybrid models to enable predictive maintenance for rubber-metal springs in rail bogies. The application is illustratively for predictive maintenance in the rail sector and yielded practical insights for engineering cyber-physical systems. By addressing current user interface limitations, this work provides a scalable, standard-compliant approach for integrating different data sources, creating the initial DT structure and continuously operating a DT in rail maintenance management.