|M.Sc Student||Ofer Shomer|
|Subject||Active Diagnostics of Rotating Structures|
|Department||Department of Mechanical Engineering||Supervisor||Full Professor Bucher Izhak|
This work presents a new method to detect minor faults in a rotating system by means of a special probing force and a model-based signal processing technique. Unlike most fault detection methods, the present approach combines a physical model, sophisticated excitation and signal processing, and advanced rotordynamics to enhance the detectability of minute imperfections. Both the active excitation and the physical model of the rotating system are crucial for the success of this approach. As a result, the proposed method is able to detect defects that are several orders of magnitude smaller than what can be found by detection techniques that only use response measurements. The dedicated, real-time diagnostic procedure of the rotating structure is required to monitor the machine, as an aid to schedule maintenance shut-downs, only when necessary.
Mathematically, asymmetric rotating systems are characterized by periodic coefficients which appear in the mass, damping or stiffness matrices. These periodic terms, can cause parametric vibrations that appear in frequencies other than the frequency of excitation, frequency of rotation and their integer multiples. The appearance of these unique spectral lines is directly related to faults and provides highly sensitive detection means. In theory, many faults do produce special features in the response, in the form of additional harmonics or modulation, but in practice, due to the small magnitude of these signal components compared with the unbalance and normal dynamical response, faults are only detected when they are severe and exceed dangerous operating levels, unless the suggested diagnostic methods is employed.
The research presents a novel approach in fault detection of minute, developing faults by utilizing an external excitation device (e.g. Active Magnetic Bearing) as a non-synchronous force exciter. The ability to actively detect asymmetry is studied analytically, numerically and experimentally via increasingly detailed mathematical models of the fault related response.