Open Collections

Zabihi, Mojtaba

University of British Columbia

This dissertation focuses on mitigating airborne disease transmission through two primary strategies: (1) assessing the impact of large-scale indoor airflow patterns and (2) implementing localized airflow control to remove respiratory particles before their dispersion. Large-scale airflow behavior was investigated numerically using a classroom model based on a real university classroom at the University of British Columbia. A new ventilation design, called UFAD-CDR, is proposed. The UFAD-CDR design demonstrated an 85% reduction in airborne particle concentrations compared to baseline ventilation. Furthermore, a new parameter, Horizontality, is introduced to characterize large-scale flow patterns, highlighting their impact and suggesting modifications to ventilation standard codes to account for these effects. The durability of different particle sizes under various ventilation strategies and rates was studied. It was shown that 10 or 15-minute breaks are not sufficient as fallow times unless the ventilation system is specifically designed to mitigate airborne transmission. The effectiveness of localized airflow control was investigated in a consultation room by proposing a novel personalized ventilation (PV) system based on the push-and-pull concept, designed to enhance air quality while preserving occupant comfort. The system was evaluated against conventional PV, baseline ventilation, and offset device positions simulating posture changes during meetings. Results showed a major reduction in infection risk with the new device, lowering probability from 91% (baseline) to 9% after 30 minutes. Conventional PV reduced it to 47.6%, but performed worse than baseline when misaligned. Conversely, the novel system maintained strong performance, similar to an ideally aligned high-flow PV setup (49.8%). A fully transient Eulerian–Lagrangian approach was employed which combines URANS turbulence models with unsteady stochastic tracking of particles to simulate turbulent aerosol dispersion. The numerical model was validated through numerically replicable in-house low-concentration aerosols dispersion experiments. Additionally, among commonly used turbulence models the SST k-ω model showed the best agreement with experiment, due to its more accurate and lower predictions of particle deposition. This was attributed to its superiority near walls, predicting smaller turbulent scales (weaker-shorter-lived eddies). While enabling the turbulent kinetic energy production limiter improved the RNG k-ε accuracy, the SST k-ω remained the most consistent with experimental results.

Text

2025-04-28

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/90838

Mechanical Engineering

University of British Columbia

UBCO

http://creativecommons.org/licenses/by-nc-nd/4.0/