Building-Specific Comfort Models
Occupant Feedback in Automated Fault Detection and Diagnosis of HVAC Systems
Martín Mosteiro-Romero (TU Delft - Civil Engineering & Geosciences)
Nitant Upasani (Eindhoven University of Technology)
Laure Itard (TU Delft - Architecture and the Built Environment)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Automated fault detection and diagnosis (FDD) of systems can allow component and control faults to be identified and rectified early to minimize their impacts. However, occupant behavior and feedback are typically neglected in FDD processes. In this paper, we develop building-specific comfort models for whole building HVAC system FDD. The models are based on random forest classification with two target variables: thermal preference and indoor air quality. The methodology is applied on a seven-floor office building in Delft, the Netherlands. The model is trained on subjective feedback provided by 15 study participants through a smartphone application as well as data from distributed sensors deployed in their offices during the period September–October 2024. The model is used to represent a “typical” occupant’s comfort perception in the case study building using the limited information available in the BMS data, allowing for long-term FDD in HVAC systems without the need for recurrent feedback.
Files
File under embargo until 18-12-2026