Locating fatigue damage using temporal signal features of nonlinear Lamb waves

• Use of temporal features processing of nonlinear Lamb waves to locate small-scale fatigue damage.
• Combination of sensitivity of nonlinear waves and efficacy of temporal processing in locating damage.
• Use of embeddable PZT sensors to form a sparse sensor network, conducive to in situ SHM.
• Quantitative comparison between imaging qualities using linear and nonlinear features of Lamb waves.
• A synthesized monitoring strategy for continuous monitoring of growing cracks under varying ambient conditions.

The temporal signal features of linear guided waves, as typified by the time-of-flight (ToF), have been exploited intensively for identifying damage, with proven effectiveness in locating gross damage in particular. Upon re-visiting the conventional, ToF-based detection philosophy, the present study extends the use of temporal signal processing to the realm of nonlinear Lamb waves, so as to reap the high sensitivity of nonlinear Lamb waves to small-scale damage (e.g., fatigue cracks), and the efficacy of temporal signal processing in locating damage. Nonlinear wave features (i.e., higher-order harmonics) are extracted using networked, miniaturized piezoelectric wafers, and reverted to the time domain for damage localization. The proposed approach circumvents the deficiencies of using Lamb wave features for evaluating undersized damage, which are either undiscernible in time-series analysis or lacking in temporal information in spectral analysis. A probabilistic imaging algorithm is introduced to supplement the approach, facilitating the presentation of identification results in an intuitive manner. Through numerical simulation and then experimental validation, two damage indices (DIs) are comparatively constructed, based, respectively, on linear and nonlinear temporal features of Lamb waves, and used to locate fatigue damage near a rivet hole of an aluminum plate. Results corroborate the feasibility and effectiveness of using temporal signal features of nonlinear Lamb waves to locate small-scale fatigue damage, with enhanced accuracy compared with linear ToF-based detection. Taking a step further, a synthesized detection strategy is formulated by amalgamating the two DIs, targeting continuous and adaptive monitoring of damage from its onset to macroscopic formation.