Search Results

Noise due to the flow over an automobile A-pillar rain gutter in isolation was computed using a two step procedure. Initially the flow solution was obtained by solving the Reynolds Averaged Navier Stokes (RANS) equations. Acoustical Sources were extracted from the flow solution and propagated to the far-field using the Lighthill-Curle equation. Experiments were conducted to evaluate the computations. Compared results include steady pressures, time dependent pressures, and sound intensity levels. Computed results and experimental data were reduced in a similar way to ensure a one to one comparison. Computed results are in good agreement with the experimental values. A-weighted noise levels are predicted reasonably well.

Effects of boundary conditions on the computational simulation of time dependent flows is studied. In particular, the effect of various boundary conditions for the flow over a half circular cylinder which is known to exhibit periodic shedding under certain conditions is investigated. A type of convection boundary condition called the radiation boundary condition is demonstrated to eliminate the secondary frequency which contaminates the solution due to the partial reflection of the fluid structures at the exit plane. However, the boundary condition implementation comes at the additional cost of storing results corresponding to three time levels.

Reynolds Averaged Navier-Stokes equations (RANS) were solved for three cases, namely, flow over a half cylinder in a uniform stream, near a stationary wall and near a moving wall. Time accurate, incompressible computations were performed for a Reynolds number of half-a-million. Computed time histories of force coefficients exhibit a strong tonal nature representative of regular periodic shedding. The effect of the ground on the periodic shedding was studied and the computational results were compared with experiments for the free stream and stationary wall cases. Strouhal number computed from the computational simulations agrees well with the experimental results. Time averaged pressure distribution on the body and wall show reasonable agreement with experimental data.