Torsional vibration feature extraction method from lateral vibrations based on decomposed forward and backward whirl motions

The high-speed turbomachinery rotor-bearing system faces potential torsional vibration problems due to its large span, flexible structure, and the wildly use of variable frequency drive systems. In this paper, a method based on the Hilbert technique to extract weak torsional vibration signals from measured lateral vibration signals is proposed, and the forward whirl signal is constructed to make full use of the information contained in lateral vibration signals. Furthermore, it was found that the forward whirl signal has better stability and stronger anti-noise ability after comparing the utilization of horizontal, vertical, forward, and backward vibration signals applied to the proposed method to extract the torsional vibration signal from simulated vibration responses of a rotor system. Numerical simulation and experimental results indicate that the method can detect torsional vibration signals accurately, with a high resolution in frequency estimation. This method enables additional monitoring of torsional vibrations for machines without dedicated torsional vibration testers and provides a convenient way to measure torsional vibrations in engineering applications.