Modeling occupancy distribution in large building spaces for HVAC energy efficiency

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journalpeer-review

2 Scopus Citations
View graph of relations



Original languageEnglish
Pages (from-to)1230-1235
Journal / PublicationEnergy Procedia
Publication statusPublished - Oct 2018


TitleCCUS2018-Applied Energy Symposium and Forum 2018: Carbon Capture, Utilization and Storage
Period27 - 29 June 2018



In large spaces, co-operative HVAC terminals are usually installed to provide services for different virtual thermal zones. The lack of high-resolution occupancy distribution in large spaces is often perceived as one of the main causes of underperformed HAVC systems. Current studies usually considered occupancy information of the whole space or room, such as occupancy count level, other than the zone-level occupancy distribution. Although the count of total occupants in space might stay constant, the actual occupancy distribution might be different, which will bring with different operations for each HVAC terminal. Therefore, to find out one high-resolution occupancy level, this research proposed the idea of integrating k-Means clustering and k Nearest Neighbors (kNN) classification algorithm to detect the occupancy distribution via the dual Bluetooth Low Energy (BLE) and Wi-Fi signal technology networks. One experiment place was conducted in one indoor area of the typical office room at the City University of Hong Kong for measuring the signal distribution of BLE and Wi-Fi. In this study, the occupancy preference cluster could be mapped into the indoor thermal zones and three case studies are chosen for validation of occupant number confirmation in each thermal zone. Finally, zone occupancy based energy performance analysis was presented with the assistance of wireless sensors nodes to compare energy saving potential under actual occupancy distribution and detected occupancy distribution and the importance of zone occupancy information for the demand-driven control mechanism is stressed.

Research Area(s)

  • Cluster-classification algorithm, HVAC energy efficiency, Occupancy distribution

Download Statistics

No data available