Beschreibung
The dissertation focuses on the probabilistic investigations of turbomachinery, evaluating the impact of modelling uncertainties on the machine. The geometrical variability is of particular interest when investigating axial compressor aerofoils. Deviations of these components’ geometries from the nominal cyclic symmetry affect both the structural dynamics of the components and the flow field.
The geometrical variability introduced during the manufacturing process is investigated as the source of uncertainty for the aeroelastic studies. Optical surface scans are used to measure the aerofoils’ real geometries. A digital twin is defined to stochastically describe the geometrical deviations. The quantities of interest are the aeroelastic forces on a rotor blade integrated disk. The impact of the geometrical deviations on excitation forces and aerodynamic damping is studied numerically.
Different methods are developed for the evaluation of the aerodynamic damping and the modal forces, while limiting the geometrical uncertainty space. The results indicate the necessity of including multi-passage effects in the evaluation of the aeroelastic forces. The optimized stochastic modelling of the geometry, combined with the investigation of its impact on sector aerodynamic influence coefficients, allowed to evaluate the aerodynamic damping for a mistuned rotor. Considering the modal forces, the full spectrum of the mistuned excitation could be reconstructed as a linear superposition of disturbances from individual aerofoils’ geometrical deviations, allowing the stochastic investigation of the problem. The methods are compatible with aero-mechanic reduced order models to include also the structural mistuning in the investigation of the forced vibration amplitudes.
