Examination of human interaction on indoor environmental quality variables: A case study of libraries at the University of Alberta

• Published in: Building and environment Vol. 207; p. 108476
• Main Authors: Shum, Caitlyn, Alipouri, Yousef, Zhong, Lexuan
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.01.2022; Elsevier BV
• Subjects: Air conditioning; Air quality; Classification; Control systems; Energy consumption; Environmental quality; Feedback control; Field study; Gender; Gender impact; Human relations; HVAC; HVAC equipment; Indoor environmental quality; Indoor environments; Libraries; Local weighted support vector machine; Noise levels; Occupancy; Social sciences; Support vector machines; HVAC; Local weighted support vector machine; Gender impact; Indoor environmental quality; Occupancy
• ISSN: 0360-1323; 1873-684X
• DOI: 10.1016/j.buildenv.2021.108476

It is well-known that indoor environmental quality (IEQ) can be affected by the occupants and their activities in indoor environments. What is essential to be investigated is the impact of occupant gender on indoor environmental variables. In this report, a field study was conducted over the span of three months to measure the various components of IEQ – including thermal, acoustic, air quality, and lighting – within the Cameron Library for Sciences, Engineering, and Business and the Rutherford Humanities & Social Sciences Library at the University of Alberta in Edmonton, Canada. The purpose of the study was to investigate the impacts of occupant gender ratios on the IEQ. The selected spaces had varying noise level restrictions to facilitate occupant activities within the library spaces. Through the field study, it was found that occupant gender distribution had an impact on the IEQ of the selected spaces. In addition, in this paper, a novel support vector machine (SVM) model was proposed to predict the impact of the female to male ratio on IEQ. The IEQ variables are highly nonlinear and hard to be classified by conventional methods. Hence, in this paper, a locally weighted SVM was proposed in which the classification is focused on the neighborhood and region around the new sample. This way, the classification model could achieve higher accuracy. This work provides further insight into the development of heating, ventilation, and air conditioning (HVAC) feedback control systems to optimize both IEQ and the energy consumption of indoor spaces. [Display omitted] •Selected library indoor spaces generally met ASHRAE standards regarding thermal comfort and ventilation rate.•Occupant gender distribution has an impact on indoor environmental quality (IEQ).•IEQ variables are highly non-linear and cannot be classified by conventional machine learning methods.•Indoor CO2 levels can be predicted using the proposed support vector machine (SVM) model.•The locally weighted SVM model obtained the highest prediction accuracy.