Design and Optimisation of Piezoelectric Energy Harvesting Using Evolutionary Methods

• Grant number: 22/09350-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: October 01, 2022
• End date: December 31, 2024
• Field of knowledge: Engineering - Mechanical Engineering - Mechanics of Solids
• Principal Investigator: Paulo Sergio Varoto
• Grantee: Catarina Luz Moura Ibiapina Barros
• Host Institution: Escola de Engenharia de São Carlos (EESC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated scholarship(s): 23/15931-1 - Bayesian Calibration and Physics-Informed Models for Enhanced Piezoelectric Energy Harvesting on Fan-Folded Structures., BE.EP.IC

Abstract

The main goal of the present proposal concerns the application of evolutionary algorithms in the design and optimization of multiple degrees of freedom piezoelectric energy harvesters as well as their applications to low-frequency structural vibration signals. Particularly, the scope of the project falls within the well-known area of Vibration Piezoelectric Energy Harvesting, which seeks to design and optimize electromechanical devices using piezoelectric materials that are capable of converting structural vibration signals from a given dynamic environment into usable electrical voltage. Of special interest are the cases where the main spectral components of the vibrating signals usually rely on the frequency range corresponding to 0 100 Hz, which is the case of most of the vibratory phenomena encountered in nature (environmental vibration signals). The well-known cantilever beam partially covered with a thin layer of piezoelectric monolithic ceramic is one of the most commonly used models in the design of piezoelectric power generators. Although very useful, this model has limitations regarding the operating frequency range in applications that impose restrictions on the maximum volume occupied by the generator. In these cases, the device generally has a reduced equivalent mass and a high stiffness value, thus resulting in a high value for the natural frequency corresponding to the fundamental mode of vibration, that can be above the usable frequency range. Another limitation of the cantilever model is the fact that it presents satisfactory results only for the first natural frequency, thus not offering the desirable adaptability to other frequencies present in the disturbance signal and which, in principle, could improve the harvester's performance. Improvements in the performance of a given harvesting system are possible through optimization methods, which aim to determine a set of geometric and electrical parameters that, ultimately will maximize the electromechanical conversion process. Another alternative coping with the need to obtain natural frequencies in a certain range where the main spectral components of the external disturbance occur is to combine several structural elements of the same type, forming a structure of higher geometrical complexity than the cantilever beam. Depending on the way these structural elements are combined, it is possible to obtain harvesting device configurations that present more than one of their natural frequencies in a certain frequency range of interest. In this context, the present research project offers as its main goal the application of evolutionary algorithms in the design and optimization of piezoelectric energy harvesting systems. Particularly, we seek to apply genetic algorithms and artificial neural networks in the formulation of evolutionary models that allow alternative ways of optimizing a given harvesting device using several of its design parameters. Initially, a structural distributed parameter model of the energy harvesting system with the associated electrical circuit will be formulated using established beam theories. This modeling step will commence with the cantilever model with a single beam with mass at the free end. This model will be used in numerical simulations and after the construction of a physical prototype, experimental results will be obtained to verify the formulated model. Once the theoretical model has been verified, a methodology will be formulated to optimize its parameters using a combination of genetic algorithms and artificial neural networks. From this methodology, the parameters for maximizing the generated energy will be obtained. The project also foresees the study of more complex geometric configurations from the combination of beam-like structural elements in the form of fan-folded structures