Session: 01-05-02: Applications of Artificial Intelligence/Machine Learning for Aerospace Structures 2

Paper Number: 161676

161676 - Physics-Informed Neural Networks for Transient Heat Transfer Analysis in Cfrp Composites With Thermal Barrier Coatings 

Physics-informed neural networks (PINNs) combine the representational power of deep neural networks (DNNs) with the physical principles embedded in governing equations of multiphysics systems (e.g., thermodynamics, electromagnetics, and fluid dynamics). By learning from both experimental data and physical laws, PINNs improve prediction accuracy while avoiding overfitting, even with limited or noisy data. This hybrid approach efficiently captures complex material behavior while ensuring compliance with key conservation laws and boundary conditions.

In this work, a PINN model will be developed for three-dimensional (3D) transient heat transfer analysis of carbon fiber-reinforced polymer (CFRP) composite with thermal barrier coating (TBC). The simulated baseline CFRP and TBC/CFRP composite will be subjected to a moving heat source with various Gaussian distributions to assess their thermal resistance. Threshold TBC thickness to mitigate thermal damage in underlying CFRP composites will be determined for each applied heat source (500~700 °C). The PINN model will be validated against corresponding in-house experimental results and finite element (FE) models with a discrete static heat source; both are already in the open literature. This work will provide valuable insights into how PINN can be utilized to computationally efficiently optimize TBC design that can extend the performance limits of CFRP composites in high-temperature applications.

Presenting Author: Juhyeong Lee Utah State University

Presenting Author Biography: Juhyeong Lee is an assistant professor in the Department of Mechanical and Aerospace Engineering at the Utah State University. He received his B.S. in Material Science and Engineering from Chonnam National University (South Korea) in 2004, and earned his M.S. and Ph.D. degrees in Aerospace Engineering from the Mississippi State University in 2010 and 2017, respectively. Prior to joining USU, he was a postdoctoral researcher in the Center for Advanced Vehicular Systems (CAVS) at the Mississippi State University (2017-2018) and a research associate in the Bristol Composite Institute at the University of Bristol in UK (2018-2019). Along with academic experience, he has four years industrial experience in carbon fiber and composite manufacturing and design for thermal and automotive applications. His research interests broadly cover extreme mechanics for advanced composite materials/structures through the combination of experiments, finite element modeling, and machine learning