Computational Science and Engineering at TU Wien
supported by the Institute for Microelectronics

openKB4BMS – A Semantic BMS Server

openKB4BMS is an open source knowledge base (KB) acting as a building management application with Apache Jena as its core. The overall system architecture is shown in the figure above. openKB4BMS sits on top of an integration layer established by oBIX or OPC UA. Alternatively, it can directly connect to IoT devices via a Webservice interface. A Web ontology language (OWL) framework is used for representing the static information about a building and its technical equipment (geometry, building physics, network topologies and devices) as well as the current dynamic state. This includes temperature and humidity values, presence states and blind positions but also the state of building automation systems. This way, openKB4BMS represents a holistic knowledge base of a building and its components.

Analogously to SQL for databases, there exists a query language for knowledge bases called SPARQL. Via this interface, semantic queries on a building management system can be executed. Examples for possible queries to openKB4BMS include but are clearly not limited to:

  • Are every switching and dimming actuators of a distinct floor in “off” state?

  • Which lamps in a building have exceeded a distinct operating time?

  • Which lamps in a building are broken?

  • How many rooms of a building are occupied?

  • Which unoccupied rooms at the southern or western front of the building have their shutters in “down” position?

The SPARQL interface answers with boolean values or with the respective positive instances. On the other hand, it is also possible to update property values in the openKB4BMS by SPARQL statements:

  • Perform a “central full open” to all shades of a distinct building part.

  • Reset all cycle counters exceeding a distinct value.

  • Activate or deactivate, respectively, proper electricity consumers for demand side management or flexibility trading purposes.

The changes performed on the openKB4BMS knowledge base are propagated to the underlying automation networks in order to update actual runtime datapoints.

Members: Andreas Fernbach, Thomas Frühwirth