Vibration signature analysis for fault detection in gear and bearing components

Gears and bearings are the most universally used machine elements. With higher requirements in operating speed, applied load, and lighter weight, there is a general increase in premature failures in bearing and gear components due to excessive wear and material fatigue. Such premature failures often result in substantial financial losses, and sometimes may even lead to catastrophic consequences.; In this dissertation, both experimental investigations and numerical simulations were performed to identify and quantify damage or wear in rolling element bearings and gear tooth surface profiles in a gear transmission system. The experimental results used in this study were obtained from the high-speed ball bearing test rig and the spur gear test rig in the vibration test facilities at the University of Akron. During these tests, vibration signatures due to a variety of bearing failures (including inner race and ball element failures) and gear damage/wear (including single and multi-tooth damage conditions) were acquired for identification and quantification studies. The numerical study was performed by solving the system equations of motion including the effects of changes of gear mesh and bearing support stiffnesses. A modal procedure was applied to reduce the degrees of freedom of the system for efficiency in generating transient and steady state solution of the system. Gear mesh stiffness changes were evaluated based on change of gear surface profile due to wear which usually result in significant shifts in phase of the mesh stiffness during changes from single to multiple gear tooth contacts.; Vibration signatures due to various damage or wear and operating conditions were examined in both time and frequency domains for identification purposes. Joint time-frequency analysis such as the Wiper-Ville Distribution and the Wavelet Transform were also used extensively in detecting and identifying various types of gear and bearing damage. The modified Poincare Maps based on chaotic vibrations were also successfully applied in identifying and quantifying wear and damage in ball rolling element bearings.; From the results of this work, considerable success has been achieved in generating a comprehensive database of the vibration signals in gearboxes and rotor-bearing systems. It will be valuable to the development of a machine health monitoring system as well as for the prognostication of machine component life.