Mechano-electrically integrated variable-reluctance gear for self-powered wireless sensing and intelligent diagnosis

Gear condition monitoring is essential for predictive maintenance, yet conventional external sensors are susceptible to long transmission paths, noise interference, and power-supply constraints in sealed enclosures. This study proposes a mechano-electrically integrated variable-reluctance gear (ME-VRG) that combines energy harvesting, self-sensing, wireless transmission, and intelligent diagnosis. An electromagnetic unit is embedded in the driven gear, and periodic changes in magnetic reluctance caused by gear rotation convert mechanical energy into electricity without internal frictional contact. The ME-VRG is intended for low-load auxiliary gear stages outside the primary power-transmission path. Its feasibility and performance were evaluated through theoretical modeling, finite-element simulation, and experiments. The optimized two-coil series configuration achieved a maximum RMS output power of 5.21 mW. The harvested energy successfully powered an integrated wireless module and enabled intermittent signal transmission at a stable interval of approximately 14 s after initial charging. For fault diagnosis, the self-induced voltage signals were converted into time–frequency representations using the short-time Fourier transform and subsequently classified using a CNN–BiLSTM model. Healthy gears and gears with different faults were distinguished with an accuracy of 95.94%. These results demonstrate that the proposed ME-VRG provides a compact and self-sufficient solution for wireless sensing and intelligent condition monitoring in sealed gear transmission systems.