A novel rotordynamic experimental testing via controllable lubrication techniques and operational modal analysis

Flexible rotors – compressors, turbines, pumps, rotary atomizers, etc – supported on fluid-film bearings have their operational conditions limited by the lack of damping and lateral vibration instabilities. The instability threshold is strongly influenced by the fluid-film forces coming from the thin layer of the lubricant as well as seal forces. The changes in natural frequencies and damping ratios of flexible rotors are strongly depending on the rotor operational conditions, such as the angular velocity and loading. Consequently, there is a growing demand among industrial monitoring systems for the ability to continuously and real-time monitor the damping ratios of rotating machines. For this purpose, Operational Modal Analysis (OMA) can be employed which in comparison to Experimental Modal Analysis (EMA) does not require knowledge of the excitation forces. A general challenge in the OMA framework is to ensure that the excitation sources are Gaussian and, thinking of rotordynamic applications, that unimportant harmonic components are properly eliminated from vibration response. The paper gives an original contribution to the problem of rotordynamic experimental testing, investigating the possibility of combining controllable lubrication techniques with automated OMA techniques. A tilting-pad journal bearing is used as a shaker to dynamically perturb the rotor-bearing system without altering the rotor-bearing equilibrium position, allowing for non-invasive testing. The non-invasive fluid-film forces are generated via radial oil injection controlled by two servo valves. The non-invasiveness of such fluid-film forces is thoroughly investigated and the advantages of using such a novel technique are elucidated and discussed for industrial applications. Natural frequencies, damping ratios, and mode shapes will be determined for different rotor operational conditions using semi-automated OMA techniques and compared to EMA estimates using a mechanical shaker.