Elastodynamic Damage Chronography for cyclic fatigue evolution in mechanical components using distributed piezoelectric transducers and ensemble transfer learning

Fatigue-induced degradation is one of the most common failure mechanisms of mechanical components under cyclic loading, and advanced monitoring techniques are required to detect the damage in advance. A framework of continuous fatigue monitoring using distributed piezoelectric transducers and elastodynamic wave propagation is proposed, called Elastodynamic Damage Chronography (EDC). It allows real-time crack initiation detection and monitoring of the fatigue evolution, quantification of stiffness degradation, and estimation of remaining structural integrity. A temporal damage chronograph of structural degradation is obtained by combining the guided-wave excitation, the electromechanical impedance measurement, and finite element simulations. The Ensemble Transfer Learning model is used for the classification of fatigue damage states, performed on multi-domain piezoelectric signatures, from different operating conditions. Experiments are performed on metal samples under cyclic fatigue tests. The results show better crack detection, damage localization, characterization of fatigue evolution, and prediction of the degradation of structures when compared to traditional piezoelectric monitoring techniques. The proposed structure offers a smart and scalable solution for fatigue monitoring, predictive maintenance, and enhanced structural reliability for the fields of aerospace, automotive, civil, and industrial structures.