Damping ratio identification in rotordynamic system with inverse directional Fourier transform, multiple-tests-multiple-outputs and clustering techniques

Abstract The damping ratio of the first forward mode in rotating machinery such as gas/steam turbine and centrifugal compressor is a key indicator to evaluate the rotordynamic stability and predict the stability margin. Aiming to improve the accuracy, stability and confidence of estimated damping ratio and to simplify the actuator equipment, a method which combines the inverse directional Fourier transform, multiple-tests-multiple-outputs and clustering techniques (idFT-MTMO-CT) is developed. By using the inverse directional Fourier transform, the response vibrations related to the forward and backward whirls can be separately constructed when exciting in the horizontal or vertical direction. In the numerical analysis, the introduction of multiple-tests-multiple-outputs and clustering techniques helps to obtain stable statistical estimates with normal distributions, as well as to accurately identify the first forward modal parameters even with serious noises. Experimental result shows the estimates are very stable and the blocking excitation force can be reduced when using the combined idFT-MTMO-CT approach. From the consideration of operation safety, a small excitation force should be recommended. This paper mainly contributes to simplify the complex excitation approaches, reduce the excitation force (as well as reduce the size of actuator equipment), and provide stable and confident estimates based on the statistical clustering technique.