Simulating how air or water moves around complex objects—from high-speed trains to microscopic particles—requires a difficult trade-off. While the Navier-Stokes equations perfectly describe fluid motion, solving them directly for industrial scales is computationally impossible. Instead, researchers rely on Reynolds-Averaged Navier-Stokes (RANS) methods, which model turbulence rather than resolving every tiny eddy.
As Daoming Liu et al. note in Urbanflow: Designing Comfortable Outdoor Areas, RANS provides the stability and speed necessary for iterative engineering design, avoiding the numerical oscillations that plague more direct solvers. The challenge lies in choosing the right "closure" model to represent the missing turbulent physics.
Blending Scales with SST k-omega
The Shear Stress Transport (SST) $k$-$\omega$ model has emerged as a primary tool for flows where the boundary layer—the thin sliver of fluid touching a surface—dictates the outcome. The intuition behind SST is a "best of both worlds" approach: it uses the $k$-$\omega$ model to handle the sensitive physics near the wall and switches to the $k$-$\epsilon$ model for the simpler, free-stream flow away from the surface.
This hybrid nature is critical for capturing flow separation, where fluid loses its "grip" on a surface and creates turbulent wakes. For example, M. Chávez et al. utilized this model in Numerical Investigation of Galloping Instabilities in Z-Shaped Profiles to accurately predict when architectural shading devices might begin to oscillate dangerously. Similarly, Shifeng Dai et al. found that SST $k$-$\omega$ was essential for mapping the complex pressure gradients on non-streamlined underwater robotic manipulators.
The model's versatility extends to internal flows and rotating machinery. In Evaluating the Transient Energy Dissipation in a Centrifugal Impeller under Rotor-Stator Interaction, Xiaoran Zhao et al. demonstrated that SST $k$-$\omega$ could pinpoint energy losses caused by stall vortices in pumps. It has even been adapted for high-speed aerospace applications; Zhuoran Liu et al. used it to resolve the shock-wave and boundary layer interactions inside supersonic scramjet inlets.
Capturing Strain with RNG k-epsilon
While SST $k$-$\omega$ excels at walls, it can struggle with flows characterized by high strain rates, rapid acceleration, or intense swirling. For these scenarios, researchers often turn to the Renormalization Group (RNG) $k$-$\epsilon$ model.
Standard $k$-$\epsilon$ models assume turbulence is isotropic (uniform in all directions), which fails when a flow is stretched or rotated. The RNG variant, derived using mathematical renormalization group theory, adds an extra term to the turbulence dissipation equation. This term acts as a "correction" for the mean rate of strain.
The authors of Numerical Simulation of Flow Field of DN2400 Hedge Dissipator, Huipan Tan et al., applied this to energy dissipators where high-velocity water is forced through perforated structures, creating complex shear flows. The model's ability to handle streamline curvature also makes it the preferred choice for vehicle aerodynamics. Jian W. Du et al. employed RNG $k$-$\epsilon$ to simulate the violent pressure imbalances caused by broken windows on high-speed trains, while Xiaobo Zheng et al. used it to eliminate vortices in pumping station forebays.
Calibration and Specialized Extensions
Despite their robustness, these models are not "plug-and-play" for every environment. The accuracy of RANS simulations often depends on calibrating empirical constants to match specific physical settings.
In Research on the parameters of turbulence model and modeling of equilibrium atmosphere boundary layer in CWE, Wei Yang et al. showed that the standard industrial value for the turbulence parameter $C_\mu$ (0.09) fails to maintain a stable atmosphere in wind tunnel simulations. By tuning $C_\mu$ to 0.04, they significantly improved pressure predictions around buildings.
Furthermore, some flows require even more specialized logic. Qiqiang Duan et al. utilized a transition-sensitive SST model that incorporates intermittency equations to predict exactly where a flow turns from laminar to turbulent on a turbine blade—a critical factor for engine efficiency. For highly complex aerodynamic fields like those found in coal burners, Xiang Lin et al. bypassed eddy-viscosity assumptions entirely, using a Reynolds Stress Transport (RSM) model to capture the dominant swirl effects at the combustor outlet.
Key Foundational Work
Numerical Investigation of Galloping Instabilities in Z-Shaped Profiles (2014) — contributed 1 concept elements to this direction.
Research on the parameters of turbulence model and modeling of equilibrium atmosphere boundary layer in CWE (2006) — contributed 1 concept elements to this direction.