Accurate RANS modeling of turbulent dispersion from arbitrary sources

• Limitations of most common RANS modeling of turbulent dispersion shown and discussed.
• New RANS turbulent dispersion model formulated.
• Proposed model validated against theoretical and experimental data.
• Radical improvement of the accuracy when using the novel model demonstrated.

A RANS (Reynolds averaged Navier–Stokes) model of turbulent dispersion is developed, which can reproduce both the initial and the asymptotic regimes of dispersion (as they were described by Taylor (1921) [1]). The dispersion process of a single transported scalar is modeled as the dispersion of a discrete number of scalars, each of which is characterized by a turbulent diffusion coefficient. The proposed model is expected to provide improved accuracy in flows where the large scales of turbulence are comparable to the scales of the flow. In the present work, the model is validated against exact data and experimental data and the model results are also compared to two earlier models. The validation shows that the proposed model provides a significant improvement compared to the earlier models.