Document Type

Article

Publication Date

Summer 5-16-2025

Abstract

To enhance energy harvesting efficiency, this paper explores the optimization of a cantilever-based piezoelectric energy harvester by integrating advanced machine learning (ML) methodologies. Leveraging a meticulously trained model on data sourced from a sophisticated two-dimension (2D) COMSOL Multiphysics numerical simulation, the study focuses on the critical input parameters, particularly the dimensions of the piezoelectric thin film. Through extensive simulations, the analysis delves into power density extraction and resonance frequency for various configurations. The culmination of rigorous simulations and analysis has led to the identification of an optimal design configuration for the cantilever piezoelectric energy harvester, characterized by a length of 13 mm and a thickness of 0.5 mm, to exhibit remarkable resonance characteristics, resonating at a frequency of 52.7 Hz, and a peak power density of 8.68 mW/mm2. The study aims to address real-world challenges and maximize energy conversion efficiency through a convergence of cutting-edge technology and thorough analysis by investigating the system’s impulse response and response to pseudo-random excitation patterns.

Download

Share

COinS