Passive Directional Airflow Control Using Tesla Valve Geometry
Peer-reviewed paper at ATI 2026: CFD and full-scale experiments on a Tesla-valve double-skin facade
Published in the proceedings of ATI 2026, Yaşar University, Izmir, 7 to 9 October 2026
A double-skin facade cavity is prone to flow reversal, recirculation and pressure swings as the wind changes, and the usual answer is motorized vents and controls. This paper asks whether a fixed shape can do the job instead: a Tesla valve profile built into the cavity, which lets air move upward easily and resists the reverse direction with no moving parts. We combined CFD with a full-scale prototype in a wind tunnel, measured velocity, temperature and CO₂ decay, and checked the model against the measurements. The work grew out of the spring 2024 studio research linked in the project details, where I built the first prototype and ran the first CFD.
Morishita, Y., Li, J., Wang, H. and Kim, J. (2026). Passive Directional Airflow Control in Building Façades Using Tesla Valve Geometry: CFD Simulation and Experimental Validation. Proceedings of the Fifth International Conference on Architecture, Technology, and Innovation (ATI 2026), Yaşar University, Izmir, 7 to 9 October 2026.
Abstract. This study investigates the integration of Tesla valve geometries into double-skin façades (DSFs) as a passive strategy for enhancing natural ventilation and mitigating flow instability. The system leverages the Tesla valve’s asymmetric, fixed-geometry design to promote unidirectional airflow and restrict reverse flow without mechanical actuation. A hybrid methodology was employed, combining computational fluid dynamics (CFD) simulations and full-scale physical experiments to evaluate the aerodynamic and thermodynamic behavior of the system under wind-dominated and buoyancy-driven conditions. Results demonstrate that under wind-dominated scenarios (3.0 m/s inlet), the valve geometry significantly accelerates upward flow, achieving top-vent velocities of approximately 4.8 m/s, a 1.6-fold amplification, while maintaining negligible flow at the bottom vent (below 0.5 m/s). Under buoyancy-driven conditions, the system effectively sustained passive airflow via the thermal stack effect, with peak velocities reaching 0.26 m/s. Experimental validation using a full-scale prototype confirmed these directional trends and evaluated ventilation performance through CO₂ decay monitoring. The integration of a slit partition recorded approximately 60% reduction in pollutant concentration within 100 seconds under wind-driven conditions, consistent with displacement ventilation effects. Quantitative validation using root mean square error (RMSE) further confirmed the model’s reliability in predicting momentum and pollutant flushing dynamics, despite observed thermal discrepancies under buoyancy. These findings suggest that Tesla valve integration provides a robust, low-maintenance mechanism for passive airflow regulation in high-performance building envelopes, offering a resilient alternative to complex automated façade systems.
| Condition | Measure | Result |
|---|---|---|
| Wind-dominated, 3.0 m/s inlet | Top-vent velocity | About 4.8 m/s, a 1.6-fold amplification |
| Wind-dominated, 3.0 m/s inlet | Bottom-vent velocity | Below 0.5 m/s, reverse flow restricted |
| Wind-dominated, slit partition | CO₂ flushing | About 60% reduction within 100 s |
| Buoyancy-driven, no wind | Top-vent velocity | Peak 0.26 m/s from the stack effect |
| Model accuracy | RMSE, wind cases | Velocity 0.22 m/s, temperature 0.28 and 0.08 °C |
| Model accuracy | RMSE, buoyancy cases | Temperature 3.35 and 7.10 °C, localized radiative-convective heat transfer not fully replicated |
The physical experiments use a 1:1 section of the cavity with a CNC-machined high-density urethane frame, cast concrete thermal mass and an acrylic outer skin, tested in a custom wind tunnel with velocity, temperature and CO₂ sensors, so the CFD could be checked against measured data.
Download the paper (PDF, 1.7 MB)
Conference presentation
Manuscript
