Akademik Veri Yönetim Sistemi
MECHANICS RESEARCH COMMUNICATIONS, cilt.155, 2026 (SCI-Expanded, Scopus)
Utilising artificial intelligence to establish a multi-physical optimisation framework for improving the electromechanical and thermal functionality of graphene platelet (GPL) reinforced polygonal lithium-ion (Li-ion) batteries. The development of the battery model included an innovative multilayer polygonal shell composed of anode, cathode, and separator layers. The outermost layer is composed of a cathode, which is designed as a functionally graded composite with spatially varying distributions of GPL. By examining three distinct functionally graded distribution patterns of GPL (Uniform, Outer-rich & Inner-rich), rigorous comparisons were performed in order to find an optimal spatial arrangement to enhance electrochemical efficiency, mechanical stability, and heat management. A quantifiable way to compare functions of the different GPL distributions will be to evaluate the effect of each GPL on the natural frequency (mechanical stability), thermal conductivity in the through-thickness direction (thermal management), strain distribution (structural integrity), and electrochemical performance for each battery configuration will be made to determine which will be best for that specific application. The governing equations are based on the third-order shear deformation theory (TSDT) and Hamilton's principle, using the Green-Gauss theorem to reduce the complexity of the formulation. A hybrid deep learning approach utilizing both deep neural network (DNN) and Random Forest (RF) algorithms will be used to predict and optimise multiple performance characteristics based on simulation data and experimental confirmation. The findings indicate that the use of GPL increases both the natural frequency and electro-mechanical coupling of batteries significantly and enhances the battery's thermal stability. In addition, this methodology employs artificial intelligence to enable real-time design optimisation and predictive maintenance for energy storage systems. Such applications include electric vehicles, unmanned aerial vehicles (UAVs), and renewable energy storage systems. Furthermore, these results demonstrate how combining physics-based modelling techniques with machine learning provides a novel, data-driven approach to developing future energy systems intelligently.