Research on occupants’ window opening behavior in residential buildings based on the survival model

• Published in: Sustainable cities and society Vol. 60; p. 102217
• Main Authors: Chen, Wanyue, Ding, Yan, Bai, Li, Sun, Yuexia
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.09.2020
• Subjects: environmental indicators; prediction accuracy; survival models; window opening probability; window status; window opening probability; survival models; prediction accuracy; window status; environmental indicators
• ISSN: 2210-6707; 2210-6715
• DOI: 10.1016/j.scs.2020.102217

•A new survival model was proposed to predict window opening behavior.•Window opening duration was introduced as an independent variable into the model.•Survival model outperforms Logistic model on dynamic description of window switching.•Model prediction accuracy can be improved by considering occupant living habits. Natural ventilation can reduce a building’s energy consumption while improving indoor thermal comfort and air quality. However, window opening behavior is prominently random and highly dependent on meteorological conditions. It is of considerable significance to explore indicators influencing window opening behavior and establish a model to predict the window opening probability. In this study, four environmental indicators significantly influencing window opening behavior (indoor air temperature, the concentration of CO2, outdoor air temperature, and the concentration of PM2.5) were screened out from the measurement. With the selected indicators as input variables, the Cox model for survival analysis was established. For comparison, the logistic model was also established, coupled with the Cox model, to simulate the window opening probability under the interaction of environmental indicators. According to the comparison results, although slightly less accurate than the existing logistic model in predicting the window opening probability throughout whole survival lifetime, the accuracy of the Cox model was 1.3% more than that of the logistic model, which will provide a foundation for further revision of the human behavior model.