Crack detection in the rotor ball bearing system using switching control strategy and Short Time Fourier Transform

In case of simple Jeffcott rotors, bearings lie on edges of the rotors, hence bearing nonlinearities are not properly absorbed in the dynamic response. Real rotors containing multiple disks, flexible bearing supports, centrally located ball bearings and couplings are far more complex in terms of their dynamic response than simple Jeffcott rotors. Nonlinear dynamic modelling of the flexible rotor with centrally located ball bearing has been developed and used for simulation purposes. It has been observed that crack detection techniques based on the steady-state vibration response work well for simple Jeffcott rotors only. The presence of ball bearings and other system nonlinearities inhibit the direct application of steady-state dynamic response based crack detection techniques for real rotor-bearing systems condition monitoring. However, transient response based crack detection techniques can be successfully applied to work in the presence of system nonlinearities. Crack propagation is enhanced during high amplitude vibrations, hence, instead of working in the resonance region, working in pre-resonance regions is advantageous due to lower vibration amplitudes. Switching control strategy based on a combination of active vibration control and Short Time Fourier Transform is proposed in present work. The proposed technique basically operates on the signal in pre-resonance regions (with low vibration amplitude) hence avoids chances of the crack propagation. This technique can be used to detect the presence of even small cracks and works well in the presence of bearing nonlinearities and flexible bearing supports that are inherently present in real rotor-bearing systems and qualitatively change dynamics of the overall system.