Arrival time detection of AE signals using multi-threshold wavelet denoising and improved AIC for composites damage localization

Acoustic emission (AE) technology, with its exceptional high-frequency, high-sensitivity, and real-time monitoring capabilities, is increasingly recognized as a promising solution for structural health monitoring of composites. However, accurately locating damage sources in composite structures under high-noise conditions remains a significant challenge, primarily due to the difficulty in precisely detecting the time of arrival (ToA) of low signal-to-noise-ratio (SNR) signals. To address this issue, this study proposes an enhanced AE signal ToA picking method that ingeniously integrates the advantages of multi-threshold wavelet denoising and the Akaike Information Criterion (AIC). The performance of the proposed method is systematically benchmarked against state-of-the-art methods based on Manhattan distance statistics, higher-order statistical kurtosis measures, and characteristic function surrogate entropy variance model. Compared with the other methods, the proposed method demonstrates superior accuracy while maintaining high computational efficiency. Localization experiments conducted on carbon fiber reinforced polymer (CFRP) plate show that, compared with ToA detection based on manual identification, the new method not only maintains a higher localization success rate but also exhibits a similar error distribution. The results indicate that this method provides significant advantages in detecting the ToA of low-SNR AE signals propagating within composites, and offering a more reliable and efficient technological support for structural health monitoring, thereby advancing the technological development and practical application of this field.