|M.Sc Student||Baruch Eyal|
|Subject||Fault Detection and Wave-Control of Imperfect Cyclic|
|Department||Department of Mechanical Engineering||Supervisors||Professor Izhak Bucher|
|Dr. Harel Plat|
|Full Thesis text|
The goal of this research is propagation of traveling waves and identification of location and severity of isolated defects in cyclic symmetric systems. These structures are common in the industry. Some notable examples are turbine bladed disks, ultrasonic motors, and toothed gear. Cyclic symmetric systems experience doublet mode shapes, which have a common natural frequency, and can be exploited to propagate traveling waves at resonance. However, a phenomenon called frequency split, which can occur as a result of small imperfections, can drastically decrease the efficiency of the propagation of traveling waves. Therefore, a form of excitation which overcomes the imperfections in the system and propagates pure traveling waves at resonance is developed. This method applies Autoresonance excitation to automatically excite the system at the resonance frequency. Modal filtering is used in order to excite certain wavelengths, and a phase shift between the sine and cosine components is introduced to automatically propagate traveling waves. Extremum seeking adaptive control is incorporated to optimize two parameters of the excitation, ensuring efficient excitation of pure traveling waves. Adaptive control of the parameters allows the tracking of optimal excitation in a system where the dynamics vary over time. An alternative form for obtaining the optimal frequency of excitation is proposed, using an offline optimum search.
Deviations from cyclic symmetry result in loss of efficiency and localized strains due to elevated vibration levels. Therefore, it is of interest to identify the defects in the system, which cause these deviations. Detection of these perturbations via classical system identification approaches is time-consuming, indirect, and exhibits low sensitivity to defects and are affected by measurement noise. The present work utilizes low-level forces that automatically lock onto a weighted rotating projection of the system modes to enhance the detectability of small structural imperfections. The spatial localization in the mode shapes of the perturbed system due to the defects is exploited to identify multiple defects' locations. The defects' severities are estimated based on the deviation from the circular structure's analytical mode shapes. Fast and enhanced precision of defect identification is obtained by employing the modal filtered Autoresonance technique.
To validate the presented methods for propagation of traveling waves and fault detection, an experimental system consisting of a ring of coupled Helmholtz acoustic resonators was developed. Excitation was accomplished using loudspeakers, and sensing was executed with microphones. Experimental results are compared to simulations, verifying the ability to automatically propagate pure traveling waves in perturbed systems, and to identify the location of defects which cause deviations from cyclic symmetry.