Evaluating Window Operation Behaviour and its Effect on IEQ and Energy Consumption in Retrofitted Dormitories

Buildings consume a considerable amount of total energy during their lifetime. Building retrofits are a common method for improving energy performance and indoor environmental quality (IEQ) to achieve sustainability goals in the building sector. However, the efficiency of such interventions depends not only on technical interventions, but also on the behaviour and preferences of the occupant. This study performed a detailed analysis of one-year high-resolution pre-and post-retrofit data gathered from a university dormitory in the USA to investigate the impact of window usage on IEQ. Three important IEQ factors were studied: thermal comfort, indoor air quality (IAQ), and energy use associated with Heating, Ventilation, and Air Conditioning (HVAC). Thermal comfort was assessed according to the requirements of The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) Standard 55, and the indoor CO2 concentration was used as an indicator of IAQ. Random Forest (RF), extreme gradient boosting (XGBoost) classifiers, and logistic regression models were used to identify the best models for predicting window operation using indoor and outdoor environmental variables as inputs. The results revealed that the ensemble models consistently outperformed the other models in predicting occupant-driven window operation. The most significant features that affected window opening were outdoor temperature, indoor CO2 level, and the time of day. The opening and closing of windows also revealed a decrease in HVAC energy use of 80.20% and 79.50%, respectively; however, the reduction occurred at the cost of occupants experiencing thermal discomfort 67% of the time. Despite these findings, this study was limited to specific buildings, and a broader study is required for generalizability.