Detecting Deformation Mechanisms in Alloys Under Uniaxial Loading via Acoustic Emission and Machine Learning.

Abstract

Deciphering the complex acoustic emission (AE) signals of coexisting deformation mechanisms is critical for advancing the fundamental understanding and predictive health monitoring of metals and alloys. While AE technology is uniquely suited for this dynamic, non-invasive probing, the superposition and intertwining of signals from multiple active mechanisms present a formidable challenge for reliable identification. This review highlights how emerging synergies between physical statistical models of deformation mechanisms and advanced machine learning (ML) techniques are forging new pathways to deconvolute these deformation mechanisms under uniaxial loading. We first establish the foundational correlation between specific AE features and underlying deformation processes under uniaxial loading. We then critically evaluate ML-driven strategies for mechanism discrimination, assess progress in real-time monitoring and failure prognostics, and conclude by outlining key challenges and future opportunities aimed at transitioning this methodology from laboratory insight to engineered application.