Abstract
Modern buildings are data-rich environments that can contain thousands of IoT devices. However, most of this data is not analyzed in order to reduce energy consumption and improve occupants' comforts. This is often due to the required large manual effort for integrating the data into analytic systems. Semantic models allow to model the required meta-data and to arrive at an automated integration process. This is demonstrated for the new Brick ontology, that comprehensively models meta-data in buildings. It is extended by model concepts enabling to address challenges pertaining to physics and thermal comfort. Moreover, this Brick ontology is further extended by reasoning approaches in order to better exploit knowledge. As an example, the proposed approach is used to compute and diagnose virtual sensors so as to assess thermal comfort in a real building.
Zusammenfassung
Moderne Gebäude sind datenreiche Umgebungen, die Tausende von IoT-Geräten enthalten können. Allerdings werden die meisten dieser Daten nicht analysiert, obwohl sie erlauben den Energieverbrauch zu senken und den Komfort der Bewohner zu verbessern. Dies ist oft auf den erforderlichen großen manuellen Aufwand zur Integration der Daten in analytische Systeme zurückzuführen. Semantische Modelle erlauben es, die erforderlichen Metadaten zu modellieren und einen automatisierten Integrationsprozess zu realisieren. Dies wird für die neue Brick-Ontologie demonstriert, welche die Metadaten von Gebäuden umfassend modellieren kann. Zur Automation des Integrationsprozesses wird sie um Modellkonzepte erweitert, die es ermöglichen, physikalische Prozesse wie dem thermischen Komfort zu beschreiben. Darüber hinaus wird die Brick-Ontologie durch Reasoning-Ansätze erweitert, die das physikalische Wissen automatisch anwenden. Der Ansatz wird beispielhaft verwendet, um virtuelle Sensoren zu berechnen und zu diagnostizieren, die den thermischen Komfort in einem realen Gebäude beurteilen.
About the authors
Joern Ploennigs (Dr.-Ing.) leads the group on Cognitive IoT for Buildings and Environment. He was previously heading a research group at TU Dresden and at University College Cork, as Feodor-Lynen fellow of the Humboldt-Foundation. He is chair of the IEEE IES TC on Building Automation and board member of the IEEE IoT Initiative.
IBM Research – Ireland, Damastown Ind. Est., Mullhadard, Dublin 15, Ireland
Amadou Ba is research scientist in machine learning and control theory at IBM Research Ireland. He received his Ph.D. in control theory and signal processing from the Ecole Nationale Superieure des Arts et Metiers de Paris (France) in 2010. From 2006 to 2011 he was a control system researcher/engineer at Schlumberger.
IBM Research – Ireland, Damastown Ind. Est., Mullhadard, Dublin 15, Ireland
Paulito Palmes is research scientist in machine learning at IBM Research. He holds a Ph.D. from the Toyohashi University of Technology, Japan, and was a Postdoctoral Fellow at the National University of Singapore. He also held an Asst. Professorship at the University of the Philippines and Ateneo de Manila University.
IBM Research – Ireland, Damastown Ind. Est., Mullhadard, Dublin 15, Ireland
Acknowledgement
The authors wish to acknowledge the support of the European Union Horizon 2020 research and innovation program under the grant agreement No. 676760 (TOPAs) and No. 608790 (Tribute) in part funding the work reported in this paper.
©2017 Walter de Gruyter Berlin/Boston