Dynamic analysis and design optimization of hemispherical dielectric elastomer generator for efficient energy harvesting

Abstract

Dielectric elastomer energy generators (DEG), serving as efficient vibration energy harvesters, hold great potential in energy reuse and environmental energy exploitation. This study proposes a hemispherical DEG that employs pre-stretching and out-of-plane expansion deformation mode to achieve high-efficiency energy harvesting. For the proposed DEG, considering the viscoelasticity and electrostriction of DE materials, the study establishes a dynamic model of the hemispherical DEG based on the fractional viscoelastic model and the deformation-dependent electrostriction model. The model is then verified through experiments. Moreover, the influence of excitation parameters and structural parameters on the dynamic response of the proposed system is systematically investigated to better select the appropriate range of design parameters. Based on the established theoretical foundation, this paper leverages deep neural networks (DNN) combined with genetic algorithms (GA) to circumvent the complex dynamic solving process, and establishes a mapping relationship between multiple parameters and the energy output of the hemispherical DEG and seeks design solutions that maximize output power under different energy harvesting environmental conditions. This work provides new insights into the design and modeling of high-performance DEGs, which aids in the design and optimization of DEGs in low-frequency vibration environments.