Impact of ventilation stacks on airflow in naturally ventilated buildings: Simulation vs. measurements

Modeling of mass and energy flows in naturally ventilated buildings enables the prediction of indoor thermal conditions and the evaluation of ventilation performance under different operational scenarios. However, the accuracy of simulation results strongly depends on the quality of the airflow model, which must be validated against experimental data. The greatest discrepancies typically arise from difficulties in capturing rapidly changing outdoor conditions and unstable airflow patterns. Model validation is therefore essential, as it improves the reliability of predictions and provides a basis for optimizing ventilation strategies, reducing heat losses, and enhancing the overall energy efficiency of buildings. This study investigates a mixed-use office and educational space located on the second floor of a five-story building at the Silesian University of Technology, Poland. The rooms are naturally ventilated, with outdoor air entering through infiltration and being exhausted via grilles connected to vertical stack ventilation ducts. The building is of reinforced concrete frame construction with ceramic brick infill, and its envelope has been thermally retrofitted with insulation. Three simulation approaches were compared: (1) co-simulation between EnergyPlus and CONTAM, where mass flows are calculated in CONTAM and exchanged with EnergyPlus at every timestep, (2) the Airflow Network model in EnergyPlus, and (3) the simplified Leakage Area and Wind and Stack Open Area airflow model in EnergyPlus. The models were validated using measurements of CO2 concentration in room, airflow rates through the ventilation grille, and local meteorological data from the nearest weather station. Both calibration of input parameters and validation of simulation outcomes were carried out for each modeling approach. The results demonstrate that simplified infiltration models can provide acceptable agreement only under limited and highly specific conditions, while the Airflow Network model fails to realistically represent airflow in the presence of vertical ventilation chimneys. In contrast, the EnergyPlus-CONTAM co-simulation achieved good agreement with measurements across different window operating scenarios.