Locating and quantifying local defects in transmission power systems using wavelet -based signal processing techniques.

摘要

The two-fold objective of this thesis is to: (i) develop an efficient signal processing strategy for locating defects, and, (ii) demonstrate that a mathematical algorithm can be used to assess the health of rotating machinery.;To carry out those objectives a combined analytical/experimental investigation was undertaken. For the experimental phase, a test rig, consisting of a motor, meshing gears, and a belt driven generator, was constructed. Defects were provided for by introducing cracks of different lengths at the root fillet of one of the pinion teeth. A mathematical model of the system, which accounts for the effect of crack size and the time-varying stiffness of the gear pair, was developed. A Newmark-Beta numerical scheme was used to solve the system of non-linear coupled equations and simulated accelerations of the pinion were determined. Comparison of the predicted and experimental obtained accelerometer signals, processed in the Wavelet domain, showed remarkably good agreement.;It has been demonstrated that processing data in the wavelet domain is extremely effective for detection and quantification of cracks in spur gears. It is thus suggested that the results generated by mathematical algorithms can successfully be used as standards for assessing the health of rotating machinery.