Alex Nania

CFD Validation of Indoor CO₂ Dispersion

I replicated Gilham et al.’s (1997) enclosed-space CO₂ dispersion study using transient CFD and validated the results against the published experimental data. I developed a locally refined, 16,197-cell mesh and compared standard k–ε and k–ω SST turbulence models using concentration histories at four sensor heights. The simulations captured CO₂ accumulation near the floor and vertical stratification, while revealing differences in predicted mixing between the models. This project combined mesh development, species transport modelling, and experimental validation to assess CFD predictions of indoor gas dispersion.

Centrifugal Pump

2D MRF Frozen Rotor + RPM Sweep

Modeled a 2D impeller with rotating and stationary regions in an 80,725-cell mesh. At 1,000 RPM, the model produced a 31.9 kPa pressure rise, 9.57 N·m blade torque, and 85.9% hydraulic efficiency. Re-running the sweep at dynamically similar flow recovered the affinity laws: pressure scaled with N², shaft power with N³, and efficiency stayed between 85% and 92% from 500 to 1,600 RPM.

Humid Air Mixing

Species Transport + Buoyancy

Coupled species transport, incompressible ideal-gas density, and gravity for equal-speed streams of cold dry air and warm humid air. The staged solution closed the water-species and overall mass balances, predicting a mixed outlet at 31.6 °C and 27.1% relative humidity for a ventilation-relevant view of psychrometric mixing and stratification.

Mixing Elbow

Mesh Independence + Discretization

Built a poly-hexcore model of conjugate hot-and-cold water mixing with k-ω SST and second-order upwind discretization. An eight-case adaptive refinement sweep made outlet temperature and heat transfer mesh-independent, while a matched first-versus-second-order study exposed the hot-inlet transition as the right target for local sizing.

Exhaust Manifold

Conjugate Heat Transfer

Simulated a cast-iron manifold with roughly 275,000 polyhedral cells, three 925 K inlets, and external convection to a 300 K ambient. Mass and energy balances closed to machine precision, with a 917.5 K outlet and 32.2 m/s outlet velocity. A second SIMPLE solution matched the outlet temperature to 917.533 K, demonstrating solver-scheme independence.

2D Airfoil

Fan Blade Section + AoA Sweep

Modeled a fan-blade cross-section at 15 m/s with local airfoil and wake refinement. A staged first- to second-order solution reduced net mass imbalance from 4.9×10⁻⁶ to 6.9×10⁻⁸ kg/s. I extended the model into an angle-of-attack sweep, applying a −2.3° inlet correction to zero the geometric angle and resolving lift and drag coefficients.