On the measurement of nonlinear aeroacoustics characteristics of solar cell structures using a novel metrological framework validated by a machine learning algorithm

Abstract

This study created a new metrological framework for the modeling of nonlinear phase velocity, propagation, and aeroacoustic evaluation of multilayer silicon solar cell structures with graphene platelet (GPL)-metal layers. The solar cells were evaluated under the combined external sound radiation effects and airflow pressure subjected to coupled vibrational acoustic responses. A corrected structural modeling method using higher-order shear deformation theory (HSDT), where transverse shear stresses continuously vary through the thickness, and modified couple stress theory (MCST) to account for size-dependent phenomena at the microscale. The governing partial differential equations using the variational energy method were constructed, and an analytical harmonic-based approach is used to solve the governing equations. The Newmark’s transition is used to compute dynamic vibration response in order to accurately define transient outcomes. To ensure both reliability and predictive ability, the framework is validated with a hybrid deep neural network model (HDNNM) that uses machine learning to relate input parameters—GPL weight fraction, frequency stemmed from the excitation, and airflow velocity—to general nonlinear response performance metrics of sound pressure level and phase velocity changes. The contribution of the proposed mechanics-based modeling with machine learning represents a reliable method to evaluate metrics measured from metrology specific to solar cell structures subjected to aerodynamic acoustic conditions generally within the environmental regime. The findings further document the contributions of GPLs as strengthening or reinforcement of a structure’s stiffness and damping characteristics, which similarly enhances a structure’s energy conversion stability and acoustic properties. Lastly, these divergent thinking methodologies and experimental evidence will promote new insights into structural health monitoring, noise mitigation, and the evaluation of the next generation of solar energy devices.