Abstract :
[en] This paper proposes a novel analog computing framework for bearing fault diagnosis using piezoelectric energy harvesters (PEHs). The system leverages the harvest vibrational energy to compute health indicators (HIs) directly in the analog domain, thereby reducing the need for high-frequency data acquisition and digital post-processing. A numerical model based on isogeometric analysis is employed to describe the electromechanical behavior of the PEH, while the Intelligent Maintenance Systems (IMS) dataset is used for validation. Results show that the energy accumulated in a capacitor by the PEH can serve as a reliable HI, enabling the identification of critical degradation milestones such as the early fault point, elbow point, and remaining useful life (RUL). Compared with conventional approaches, the proposed method achieves savings in sensing energy and wireless transmission, paving the way for self-powered, ultra-low-power smart bearing.
