Understanding rotor modal excitations is crucial for high performance centrifugal compressors and other rotating machines. Assuring low vibration levels of such machines at operating conditions before delivery is important both for original equipment manufacturers (OEMs) and end users. In this paper, transient simulations of a full scale test rig rotor subject to sine sweep excitations are performed to investigate the forward and backward rotor whirling response. The applied sine sweep excitations are circular forward, circular backward, and elliptical forward, respectively. The effects of excitation force amplitude are also investigated to determine the minimum force required to accurately identify the rotor system modal parameters.
The transient simulation results are then used to investigate a forward and backward mode system identification method for rotating machinery stability based on sine-sweep excitations. Both simulations and experimental testing on a full size rotor with an electromagnetic actuator were performed to verify and validate the method. The traditional Multiple Input Multiple Output (MIMO) Frequency Response Function (FRF) is transformed into a directional Frequency Response Function (dFRF) form. This transformation recasts the real number field into complex number field via a transformation matrix. This transformation separates the MIMO FRFs into forward and backward components, which improves the accuracy of the identified results. This method is used to identify the first forward bending modal parameters to estimate rotor stability. The rational polynomial method is used to fit and identify both the dFRFs. Excellent correlation was obtained between simulation results and the identification experiments. The results of this paper provide new insights for avoidance of rotor instability in centrifugal compressors.
Local polynomial method for frequency response function identification