Improving the Vector Basis Neural Network for RANS Equations Using Separate Trainings

TL;DR

A new data-driven turbulence model for Reynolds-averaged Navier-Stokes equations called $\nu_t$-Vector Basis Neural Network, which predicts separately the turbulent viscosity $\nu_t$ and the contribution of the Reynolds force vector that is not already accounted in $\nu_t$.

Abstract

We present a new data-driven turbulence model for Reynolds-averaged Navier-Stokes equations called $ν_t$-Vector Basis Neural Network. This new model, grounded on the already existing Vector Basis Neural Network, predicts separately the turbulent viscosity $ν_t$ and the contribution of the Reynolds force vector that is not already accounted in $ν_t$. Numerical experiments on the flow in a Square Duct show the better accuracy of the new model compared to the reference one.

Figures (4)

• Figure 1: Square duct domain and square $yz$-section.
• Figure 2: Intensity of the secondary motion $\| (u_y, u_z)^T \|_2 / U_b$.
• Figure 3: Intensity of the streamwise velocity $u_x/U_b$.
• Figure 4: Velocity profiles across the red lines in Figure \ref{['fig:square_duct']}.