Journal Article

Advancements in Physics-Informed Neural Networks for Laminated Composites: A Comprehensive Review

Yazdani, Muhammad Haris, Khalid, Salman, Elahi, Muhammad Umar, Azad, Muhammad Muzammil, Kim, Heung Soo, Raouf, Izaz

2025

Overview

Physics-Informed Neural Networks (PINNs) integrate physics principles with machine learning, offering innovative solutions for complex modeling challenges. Laminated composites, characterized by their anisotropic behavior, multi-layered structures, and intricate interlayer interactions, pose significant challenges for traditional computational methods. PINNs address these issues by embedding governing physical laws directly into neural network architectures, enabling efficient and accurate modeling. This review provides a comprehensive overview of PINNs applied to laminated composites, highlighting advanced methodologies such as hybrid PINNs, k-space PINNs, Theory-Constrained PINNs, optimal PINNs, and disjointed PINNs. Key applications, including structural health monitoring (SHM), structural analysis, stress-strain and failure analysis, and multi-scale modeling, are explored to illustrate how PINNs optimize material configurations and enhance structural reliability. Additionally, this review examines the challenges associated with deploying PINNs and identifies future directions to further advance their capabilities. By bridging the gap between classical physics-based models and data-driven techniques, this review advances the understanding of PINN methodologies for laminated composites and underscores their transformative role in addressing modeling complexities and solving real-world problems.

• Publisher: MDPI AG
• Subject: Accuracy / Adaptability / Analysis / Behavior / Boundary conditions / composite material optimization / Composite materials / Decomposition / Efficiency / Energy consumption / Failure analysis / Fluid dynamics / Interlayers / Laminar composites / laminated composites / Laminated materials / Laws, regulations and rules / Learning strategies / Machine learning / Methods / multi-scale modeling / Multilayers / Neural networks / Optimization / Optimization techniques / Ordinary differential equations / Partial differential equations / Physics / physics-informed neural networks / Strain analysis / Structural analysis / Structural health monitoring / Structural reliability