Energy-Adaptive Micro-Mechanical Systems Driven by Intelligent Materials: Experimental and Simulation Study Based on the Piezoelectric Effect

The miniaturization of mechanical and electronic systems has driven demand for self-powered micro-electro-mechanical systems (MEMS) in robotics, biomedical implants, and distributed IoT networks. Piezoelectric materials offer direct mechanical-to-electrical energy conversion, making them promising candidates for energy-autonomous microsystems. Existing studies largely optimize either material performance or system-level design independently, neglecting the integration of microstructural electromechanical behavior with adaptive energy regulation. Moreover, nonlinear piezoelectric responses under micro-scale cyclic loading remain insufficiently quantified, leading to 15-30% discrepancies between simulations and experiments. This work develops a hybrid PZT-PVDF composite thin film integrated into a cantilever MEMS with a closed-loop adaptive feedback controller. A coupled multi-physics finite element model links stress-strain fields to voltage output, while real-time stiffness adjustment maximizes energy harvesting efficiency under variable vibration conditions. Experimental validation spans 150 samples with cyclic loading across 50-500 Hz. The proposed adaptive MEMS achieves a peak voltage of 5.8 ± 0.2 V, output power density of 2.45 ± 0.05 mW·cm - 3 , and conversion efficiency of 77.1 ± 1.1%, representing a 28% improvement over non-adaptive hybrid systems. Voltage degradation after 10 5 cycles is limited to 3.2 ± 0.4%, and the system converges to steady-state power within 12 ± 2 control iterations. By integrating material-level physics, multi-physics modeling, and adaptive control, this framework enhances energy conversion reliability and provides a reproducible approach for designing energy-autonomous MEMS suitable for micro-robots, implantable sensors, and distributed IoT devices.