An ontology-based approach for building automation data analysis


  • Eunju Park Fraunhofer Institute for Building Physics IBP
  • Sumee Park Fraunhofer Institute for Building Physics IBP
  • Sebastian Stratbücker Fraunhofer Institute for Building Physics IBP



Ontology, Semantic model, Anomalies detection, Data mining


For the efficient and sustainable operation of building automation systems, it is critical to consider various aspects such as users’ comfort requirements and energy consumption. The successful application is associated with the integration of multiple and heterogeneous data sources. However, the high complexity of the data poses a challenge. To address this problem, various ontologies have become popular with many applications for data modelling, management and analysis through harmonizing different data sources, as well as efficient querying. In this work, the design, implementation and usage of semantic approaches is investigated to exploit building automation data for customized room automation. As a main contribution, a building automation ontology focusing on room automation is proposed, which is represented in the Resource Description Framework (RDF). Furthermore, several scenarios with the proposed model are demonstrated in-situ, showing easier access to various data sources using a query language like SPARQL. Based on the ontology in RDF format, building data from different sources such as commercial building automation system (e.g. KNX), weather station and room monitoring sensors (e.g. temperature, humidity) are considered for multiple scenarios: (a) anomaly detection of shading automation systems, (b) monitoring user’s shading controls in automatic and manual mode, (c) identifying influential factors affecting user’s preference. The ontology-based approaches have benefits especially in multiple and heterogeneous data environments using a standardized common and controlled vocabulary. It allows engineers and researchers to enrich and interlink with various databases. Additionally, it explicitly describes the relationships between variables that make data understandable for both humans and machines.




How to Cite

Park, E., Park, S., & Stratbücker, S. (2022). An ontology-based approach for building automation data analysis. CLIMA 2022 Conference.

Conference Proceedings Volume