Smart materials are bidirectional transducers responding to physical fields (e.g. electrical, magnetic, and thermal fields). They can be used as sensors, actuators, and generators. Easy and multiscale access to material data and models enables efficient research and development with regard to the selection of the materials and their optimization towards specific applications. However, different working principles, measurement and analysis methods, as well as data storage approaches lead to heterogeneous and partly inconsistent datasets. The ontology based data access (OBDA) approach is a suitable method to access such heterogeneous datasets easily and quickly, while material models can transform material data across certain scales for different applications. 

In order to combine both capabilities, an extended OBDA approach has been developed within the research project SmaDi. OBDMA (ontology-based data and model access) combines data access with model-based working steps. Within the joint project, the approach considers four exemplary smart material classes: thermal and magnetic shape memory alloys, piezo-electric ceramics, and dielectric elastomers.

By rewriting and unfolding a SPARQL-query, data can be elevated to ontology level (without being actually realized) using mappings. Storing knowledge not only in the ontology, but also in the database and mappings leads to a lightweight ontology. This distributed approach increases the interchangeability and enables variable datasets, which is essential, especially for dynamic research fields like smart materials. In addition, the user-defined functions (UDF) stored in the database link data access to on-the-fly data analysis, resulting in a flexible and powerful tool for material development and optimization. In this contribution, the developed OBDMA system is presented for one specific use case in detail, followed by an overview of the multitude of implemented use cases for the material classes.