Laboratory vibration testing has been used for many years as a method to ensure the mechanical and functional integrity of a system or subsystem throughout its expected service environments. However, the vast majority of vibration specifications are developed with the assumption of stationarity. The stationarity requirement results from the shortcomings of traditional vibration signal analysis tools, such as the Power Special Density, to distinguish the time varying features of a nonstationary signal.;Unfortunately, the stationarity requirement of classical vibration specification development procedures prohibits addressing a potentially large number of nonstationary vibration environments. This is especially true of many military vehicles performing tactical maneuvers as demonstrated by the helicopter based example considered in this work.;There are two main objectives in this research effort. First is to explore the application potential of wavelet theory in the analysis of nonstationary vibration data. Specifically, wavelet theory will be researched to establish its potential for decomposing a nonstationary vibration signal while maintaining the signal's time varying power and spectral characteristics. Second is to develop a general concept for a compactly supported nonstationary vibration specification. The resulting specification will be designed to exploit the benefits of motion replication based implementation.;This dissertation illustrates the effectiveness of wavelet based signal analysis techniques on an ensemble of nonstationary vibration data. Combining wavelet based signal decomposition with an already well established set of traditional signal analysis tools, a nonstationary vibration specification methodology is proposed and demonstrated.
展开▼
