LOCALISED THERMAL COMFORT IN SHARED ROOMS OF EXISTING OFFICE BUILDINGS

Abstract

Cooling accounts for almost 60% of end-use electricity consumption in Singapore buildings. Despite this, occupants still report thermal discomfort and dissatisfaction. Recently, the thermal comfort paradigm is shifting from traditional ‘one-fit-for-all’ approach to ‘personal comfort models’ where additional personal comfort devices like fans, are used alongside the air-conditioning system to enhance the thermal comfort of occupants and improve the energy efficiency of buildings. This thesis addresses challenges in adopting this approach for operational shared office rooms. Firstly, a localised thermal comfort framework is proposed to facilitate the data collection and deployment of personal thermal comfort models. Secondly, online presence detection methods are developed that detect individual user presence with accuracy over 80% using non-intrusive plugload meters and PIR sensors. Lastly, preference learning methods are developed that predict user desk fan usage preferences with 95% test accuracy using environmental and user presence-related factors.