Experimental Validation of Friction Models for Rotor-Rubbing
Abstract The work presented here is a continuation of a set of experiments that were designed for predicting friction forces during rotor rubbing. The experimental set up consisted of a rotor rubbing a fixed surface. The surface had two force sensors, one aligned with the tangential force, and the other aligned in the radial direction. This set up allowed us to measure the friction component and the normal force. The measurements were complemented with a couple of accelerometers mounted on the bearings, and the accelerations and the friction force were measured simultaneously. All the data were analyzed using the Continuous Wavelet Transform (CWT) and the cross-correlation function. The CWT produces a spectrogram that is useful for identifying the nonlinear behavior of the phenomenon. The cross-correlation is used to measure the similarities between the friction force and the acceleration measurements. At low friction levels, experimental results show a sub-synchronous vibration at half of the rotating speed. This pattern is always present regardless of the friction source, but it is impossible to reproduce this effect using current friction models. The experimental results were compared to numerical results, these results were computed with the Muszynska’s model that is based on the physics of the phenomenon, but their predictions differ significantly from the experimental results. One of the reasons for these discrepancies lay on the fact that rotor-rubbing models consider the friction as an external force instead of modifying the system parameters.