SMART STRUCTURES AND SYSTEMS, cilt.36, sa.6, ss.293-308, 2025 (SCI-Expanded, Scopus)
The present study exhibits the idea, mathematical formulation, and confirmation with machine learning of a smart spherical harvester that is able to harvest efficiently ultra-low-frequency marine vibrational energy. The harvester which is made for ocean applications is spherical in shape which helps it to interact more actively with the low-frequency vibrations that are usually found in the ocean environment. By combining the piezoelectric and electromagnetic methods of harvesting, the harvester is provided with the capacity to fully exploit the energy out of the mechanical vibrations which are very difficult to capture because of their small amplitude and low frequency. During the development of theoretical models, the focus was on aligning the resonant frequency with the environmental marine vibrations in predicting the harvester's efficiency in converting energy. These models take into account the forces caused by the water movement, the properties of the material making up the spherical shell, and the interactions occurring between the components of energy harvesting and the surrounding marine medium. Theoretical predictions are validated through experimental tests performed in a controlled marine environment to assess practical performance. Moreover, deep neural networks (DNNs) are applied to the experimental results verification, which in turn, enhances the accuracy and stability of the performance analysis. The outcomes reveal that the smart spherical harvester can successfully catch and convert very low-frequency vibrations into power, thus producing reasonably high power even in real marine conditions. This proves the harvester's position as a green energy source for very off-grid marine applications, such as sensor and monitoring systems, and even underwater drones. Working smartly on materials and energy harvesting technologies for different functions, this work is helping the incorporation of such devices into marine energy applications.