Depressurization risk assessment in multi-unit residential buildings using pressurized corridor ventilation systems

Pressurized corridor (PC) ventilation systems in multi-unit residential buildings (MURBs) are prone to instances of depressurization driven by occupant behaviour and weather conditions. Depressurizing events are not monitored in most buildings but can drive the transfer of odour and contaminants into common spaces, negatively impacting residents.

This paper describes the development of a data-driven XGBoost classification model to assess depressurization risk across various conditions. Trained on data from three units in a 17-storey MURB, the model performed best on units with well-balanced datasets, validating its diagnostic potential given quality training data.