Automated Point Mapping for Building Control Systems: Recent Advances and Future Research Needs

Journal Article · Mon Jan 01 04:00:00 EST 2018 · Automation in Construction

DOI:https://doi.org/10.1016/j.autcon.2017.09.013· OSTI ID:1495345

Wang, Weimin; Brambley, Michael R.; Kim, Woohyun; Somasundaram, Sriram; Stevens, Andrew J.

This paper presents a review of recent research and development on methodologies relevant to automating mapping of points in building control systems and between building control systems and external or replacement software and hardware. Manual point mapping is labor intensive and costly, presenting a major impediment to innovations in building control (e.g., automated fault detection and diagnostics, self-healing, and automated commissioning for existing building control systems). The methods reviewed focus on classifying building control system points, especially sensor classifications by sensor type. Fewer publications address other important aspects of the point mapping problem, such as discovering spatial and functional relationships among points, relationships between control system points, physical systems, and equipment, and between various equipment and the systems of which they are part, and discovering metadata, normalizing it to a common namespace, and assigning the metadata to control system points. To motivate further development of new automated point mapping approaches, we identify many research questions organized into three key technical needs: 1) a complete solution and underlying problem formulation, 2) alignment of methods with the actual point mapping problem, and 3) test cases, data sets for testing, explicit test procedures, and consistent performance metrics for reporting testing and evaluation results.

• Research Organization: Pacific Northwest National Laboratory (PNNL), Richland, WA (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-76RL01830
• OSTI ID: 1495345
• Report Number(s): PNNL-SA-123329
• Journal Information: Automation in Construction, Journal Name: Automation in Construction Journal Issue: C Vol. 85; ISSN 0926-5805
• Country of Publication: United States
• Language: English

Related Subjects

Building automation systems
building control
machine learning
point mapping
metadata
semantics