Optimizing energy capacity, and vibration control performance of multi-layer smart silicon solar cells using mathematical simulation and deep neural networks

This study focuses on optimizing the energy capacity and vibration control performance of a multi-layer silicon solar cell reinforced with graphene oxide powder (GOP) and equipped with sensor and actuator layers. Mechanical properties such as Young's modulus, Poisson's ratio, and density are determined using the Halpin-Tsai method and the law of mixing, ensuring accurate modeling of material behavior. The formulation is developed by employing the von Kármán geometric nonlinearity assumptions and Hamilton's variational principle while incorporating the constitutive relationships for the sensor and actuator layers to capture their electromechanical coupling effects. The computational framework is based on isogeometric analysis (IGA), leveraging B-Spline and NURBS basis functions for higher-order continuity and precise geometric representation. The proposed model is validated against published studies, demonstrating its accuracy and applicability for advanced engineering problems. To enhance efficiency, a deep neural network (DNN) is introduced, providing low computational cost while preserving predictive accuracy for mechanical behavior and dynamic responses under complex simulations. The DNN framework is designed with optimized hyperparameters, ensuring robust performance and paving the way for future studies. This hybrid approach offers an innovative and reliable solution for analyzing the energy and vibration characteristics of smart solar cells, integrating advanced numerical methods with machine learning techniques. The results highlight the potential of this methodology for the design and optimization of high-performance energy systems, contributing to sustainable engineering and renewable energy advancements.