Optimization of Fatigue Test Signal Compression Using the Wavelet Transform

Traditional approaches to accelerated fatigue testing rely on heuristic methods with thresholds based mostly on experience and engineering judgment. These methods generally do not apply to the multiaxial dynamic loading situations characteristic of most aerospace applications and often result in uncharacteristic fatigue damage and failure modes during testing. To overcome the limitations of traditional spectrum compression techniques, ATA Engineering, in collaboration with Sandia National Laboratories, proposes to develop a wavelet-based signal editing tool capable of creating optimally compressed fatigue test input signals while maintaining damage equivalence and other constraints. The tool will identify underlying signal features in the time-scale domain using multilevel wavelet decomposition and remove nondamaging features. Following signal reconstruction, successive time-domain signal editing will be performed using automated truncation techniques to maximize signal compression. Consideration of the transfer function between test inputs and stress responses enables fatigue damage to be matched at multiple response locations while maintaining load sequencing and allowing recovery of the full stress tensor, thus enabling support for multiaxial fatigue. In Phase I, the team will focus on methods development and will demonstrate feasibility on relevant single-input examples. In Phase II, experimental validation will be conducted, and the method will be extended to multiple-input applications.