Prediction and analysis of broadband interaction noise using synthetic turbulence

Broadband interaction noise is a major source of noise in turbofan engines and will become more dominant with the increase of the bypass ratio. Generated by the interaction of turbulent fan wakes with outlet guide vanes, it can be numerically predicted by restricting the configuration to the stator only and introducing turbulent fan wakes with stochastic methods. In this work, a new method to inject turbulence using multiple sources of vorticity is proposed. It can generate either one-component or two-component frozen turbulence while being easy to implement and has no influence on the parallelization of an already existing solver. This method is successfully applied to an isolated 2D aerofoil and compared to experiments. In complex 2D geometries, the distortion of turbulence upstream of an aerofoil plays an important role in interaction noise, yet little is known regarding the mechanisms involved. Thus, the second part of this work focuses on turbulence distortion near the leading edge. To provide more physical insights in these mechanisms, a numerical vorticity approach in the frequency domain is developed. It allows the decomposition of the vorticity field into the incoming vorticity which is distorted close to the leading edge, and the bound vortices at the solid surfaces and wakes which represents the vorticity response of the aerofoil to respect a non-penetration condition at its boundaries. The size of the stagnation region at the leading edge is found to be a key factor in understanding turbulence distortion. Indeed, the larger this region, the more skewed the incoming vorticity field is. It results in attenuated transverse velocities retrieved from this vorticity field as well as reduced velocity perturbations from the bound vortices, which decreases the incoming turbulence levels, and therefore leads to noise reductions. However, at low wavenumbers, as the wavelength is large with respect to the stagnation region, no turbulence alteration is observed. To maximize the effect on the distortion of turbulence, the geometric changes need to be located as close as possible to the leading edge, as it will have more influence on the size of the stagnation region. Indeed, in this work, the geometry after the maximum thickness is found to have no effect on turbulence distortion