Numerical Surface Roughness Influence on the Aerodamping of an Axial Transonic Compressor at Nominal Speed and Part-Speed

In a turbomachine, with time, wear and depositions modify the surface from smooth to rougher characteristics. These effects are also prevalent in emerging manufacturing technologies such as additive manufacturing (AM). Surface roughness characterization is based on a correlation between a physical length scale or surface statistic moments, and a non-physical equivalent sand-grain roughness (ks). Depending on the flow characteristics and ks, three wall regimes can be considered: hydraulically smooth, transitional or fully rough. In compressors, an increase in surface roughness translates into a reduction in efficiency and pressure ratio. While steady-state roughness effects over airfoils and stage performance are well documented, its consequences over the aerodynamic damping are not. This paper aims to investigate numerically the effect of surface roughness on the aero-damping for the three wall regimes. The test geometry is the first stage of an open-source transonic axial compressor. Operating points considered are peak efficiency and near-stall conditions at nominal speed (N100) and part-speed (N70). The presented numerical simulations are obtained using the commercial software Ansys CFX with SST as the turbulence model with a reattachment modification. At near-stall part-speed, there are non-physical separation regions that are mesh independent but assumed to come from the numerical roughness implementation. Amplitude fluctuations in the aerodynamic damping per unit area, ζ*, are driven by the tip gap presence as well as normal shocks at N100 whereas the presence of separated flow regions appear to have a negligible effect at N70. The phase between the pressure and blade motion appears to remain almost constant regardless of the wall regimes, implying that only the amplitude distribution drives the aerodynamic damping stability. This numerical observation is aimed to be experimentally tested at the transonic linear cascade at KTH Royal Institute of Technology.