Paper Conference

2018 Building Performance Analysis Conference and SimBuild co-organized by ASHRAE and IBPSA-USA


Combined Ontology-Driven and Machine Learning Approach to Monitoring of Building Energy Consumption

Parastoo Delgoshaei, Mohammad Heidarinejad, Mark Austin
National Institute of Standards and Technology, Gaithersburg, MD
Illinois Institute of Technology, Chicago, IL
University of Maryland, College Park, MD

Abstract: A number of studies within the building domain have independently considered knowledge representation and reasoning (KRR) and machine learning (ML). This paper explores opportunities for how these artificial intelligence (AI) technologies can be combined to provide synergic assistance in intelligent building systems. A case study is used to demonstrate the construction of semantic knowledge from weather and utility ontologies. That knowledge is used as a semantically annotated training set in clustering algorithms that identify consumption patterns, and the results of clustering are stored in ontologies for further inference. This paper presents a roadmap to intelligent building monitoring techniques that uses both historical data and the underlying semantic knowledge.
Pages: 667 - 674