CFD Modeling of Near-Field Tornado Flow Using RANS Turbulence Models

The primary objective of this study is to model the velocity distribution in the near-field of a tornado-like flow using computational fluid dynamics (CFD) simulations conducted in ANSYS Fluent 2025 R1. A complementary goal is to develop a computational framework capable of analyzing tornado-induced velocity fields while maintaining a simplified model structure that minimizes computational cost and processing time. The paper begins by introducing the phenomenon of extreme wind events, their classification, and representative cases, with particular emphasis on tornado characteristics and the inherent challenges in accurately capturing their dynamics through numerical modeling. Several Reynolds-Averaged Navier–Stokes (RANS) turbulence models were applied and compared with an analytical Rankine-type vortex profile. The methodological section presents the model development process and the comparative evaluation of simulation results. Various modeling approaches were analyzed to minimize numerical errors, with particular attention paid to mesh refinement and grid sensitivity analysis. The results include a comprehensive investigation of velocity distributions across different model configurations, initial wind speeds, computational domain geometries, and radial distances from the tornado core. Based on these analyses, a representative wind velocity profile was formulated. The findings demonstrate that both the selected RANS turbulence model and the computational domain geometry significantly influence the predicted velocity field in the tornado near-field. The study emphasizes the comparative performance of the Spalart–Allmaras, Realizable k–ε, and SST k–ω models, which showed the closest agreement with the analytical vortex profile.