Numerical investigation of the unsteady loading of a model rotor in Hover

The paper describes a numerical study performed using the NASA OVERFLOW code to investigate the unsteady loading components of a rotor in hover. The nominal conditions investigated followed the model rotor test data of Caradonna and Tung, (Ref. 1). Variations to the baseline case included a single-bladed rotor and with and without a central hub. An additional study simulated the experimental setup of Tangier, Wohlfield, and Miley. (Ref. 2) Multiple simulations of the Caradonna and Tung model demonstrated that higher order differencing, together with increasing wake grid resolution, resulted in marked increase in the magnitude of the unsteady components in the rotor thrust convergence history. This was accompananied by a marked increase in the detail of resolution of the rotor wake structure and the appearance of tip vortex instabilities. The simulation of test configurations from the report by Tangier et al. were used to help validate that the higher fidelity model were accurately resolving the nature of the wake structure, especially the presence of helical vortex instabilities. It is found that under the given simulation parameters the tip vortex instability is present but not consistent with experiment. Fourier analysis showed two dominant frequency components of this unsteady loading for cases using adaptive mesh refinement and one dominant frequency in cases using fixed grid systems. Findings suggest that the lower frequency component is due to unsteady wake dynamics near the rotor hub and are unrelated to unsteady vortex dynamics while the higher frequency component is spurious and is demonstrated to be due to the adaptive mesh refinement routine used by the flow solver. Additional numerical experiments were completed to investigate the observed unsteady aerodynamics of a single bladed rotor case which may also be compared to the instability modes predicted for this helical vortex system by Widnall. (Ref. 3) It is postulated that under certain conditions this spurious frequency component due to adaptive mesh refinement may cause an nonphysicial aerodynamic forcing on the rotor system which can potentially pollute numerical simulations of any unsteady phenomenon such as those involving blade dynamics, deformation (Ref. 4), or result in inaccurate noise predictions (Ref. 5) (Ref. 6) when coupling results with aeroacoustic codes such as WOPWOR (Ref. 7) There is an additional subharmonic signature in the Cj convergence history, the source of which is not yet identified. Hanning windowing was used in data processing in attempt to remove this from the single revolution FFT.