.png){kind=icon}
.png){kind=icon}
Return
Collaborative representation of acoustic emission signals for automatic characterisation of tensile processes of aluminium alloy sheet
.png){kind=icon}
PRE
AI
J
Z
Y
T
DOI：10.1080/10589759.2026.2686344.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The stress-strain state is a critical parameter for evaluating quality and stability in material forming. Nonetheless, its real-time, online measurement during the actual forming operation remains a significant challenge. In this study, a block collaborative representation classification (BCRC) framework is proposed that treats acoustic emission (AE) waveforms as observable fingerprints of the underlying micro-mechanical response. AE signals were acquired during tensile tests of AA6082 aluminium alloy sheets and processed via Fast Fourier Transform (FFT) and Short-Time Fourier Transform (STFT) to construct spectral and time-frequency datasets. Key parameters influencing the BCRC performance, including dictionary dimension and training sample size, were optimised. Experimental results demonstrate that the proposed method can effectively identify four distinct tensile stages: elastoplastic, yield plateau, hardening, and necking/fracture, with higher accuracy for the elastoplastic and necking/fracture stages. The FFT-based dataset yielded superior identification rates compared to the STFT-based approach, attributed to its global spectral sparsity. The proposed method achieves favourable cross-specimen generalisation for necking and fracture stages, whereas its identification performance in micro-deformation stages remains sensitive to microstructural variability. The BCRC method offers a computationally efficient and reliable tool for real-time monitoring and automatic classification of material deformation stages using AE waveform signals.
Keywords:
Acoustic emission (AE) signal
tensile process
stress-strain state
block collaborative representation classification (BCRC)
automatic classification
Journal
N
IF:
4.2
Papers:
1.7K
Citations:
2.1K
Organization
Cited Papers