AI Research

A Physics-Informed Neural Network Framework for Ela... | AI Research

Key Takeaways

A Physics-Informed Neural Network Framework for Elastodynamic Wave Propagation in Bimaterial Systems This research introduces a new computational framework t...
Physics-informed neural networks (PINNs) provide a promising framework for solving partial differential equations while embedding the underlying physical laws directly into the learning process.
This study presents a PINN-based framework for modeling transient elastodynamic wave propagation in bimaterial systems governed by the axisymmetric equations of linear elasticity.
High-fidelity finite-element simulations performed using ANSYS Workbench Explicit Dynamics are used for validation and as supplementary data constraints during training.
Mesh-sensitivity studies confirm numerical robustness, while additional material combinations demonstrate the generality of the proposed methodology.

Paper AbstractExpand

Physics-informed neural networks (PINNs) provide a promising framework for solving partial differential equations while embedding the underlying physical laws directly into the learning process. This study presents a PINN-based framework for modeling transient elastodynamic wave propagation in bimaterial systems governed by the axisymmetric equations of linear elasticity. A steel-aluminum specimen representative of a Split Hopkinson Pressure Bar configuration is considered, and the governing elastodynamic equations, together with the corresponding initial, boundary, and interface conditions, are incorporated directly into the network through a physics-informed loss function. High-fidelity finite-element simulations performed using ANSYS Workbench Explicit Dynamics are used for validation and as supplementary data constraints during training. The proposed framework accurately predicts wave transmission and reflection across the bimaterial interface and reproduces axial and radial displacement histories, face-averaged responses, and the dominant stress and strain evolution with close agreement to the finite-element solutions. The trained network further demonstrates the ability to predict wave responses at previously unseen time instants and for modified material properties without requiring additional finite-element simulations, providing a continuous surrogate model for elastodynamic analysis. Mesh-sensitivity studies confirm numerical robustness, while additional material combinations demonstrate the generality of the proposed methodology. The results show that integrating physics-informed neural networks with explicit finite-element analysis provides an accurate and computationally efficient framework for elastodynamic wave propagation in heterogeneous solids, offering an effective surrogate modeling approach for high-rate solid mechanics and impact engineering applications.

A Physics-Informed Neural Network Framework for Elastodynamic Wave Propagation in Bimaterial Systems
This research introduces a new computational framework that uses Physics-Informed Neural Networks (PINNs) to model how stress waves travel through heterogeneous materials. By combining the governing laws of physics with data from high-fidelity simulations, the researchers created a model capable of predicting how waves move, reflect, and transmit across the interface of two different materials—specifically steel and aluminum—in a setup similar to a Split Hopkinson Pressure Bar.

Bridging Physics and Machine Learning

Traditional numerical methods, such as finite-element analysis, are highly accurate but can be computationally expensive when researchers need to run many simulations for optimization or material testing. This study addresses that challenge by embedding the Navier–Lamé equations—the fundamental laws of linear elasticity—directly into the neural network's training process. By doing this, the network does not just learn from data; it learns to obey the physical laws that govern wave propagation, such as displacement continuity and stress balance at material interfaces.

How the Framework Works

The researchers designed a fully connected neural network that takes spatial and temporal coordinates as inputs and outputs the resulting radial and axial displacements. To ensure the model is accurate, they used a "physics-informed loss function." This function penalizes the network if its predictions violate the governing equations or the specific boundary and interface conditions of the experiment. Additionally, the team incorporated displacement data from ANSYS Explicit Dynamics simulations as "soft constraints," which helps anchor the network’s predictions to high-fidelity reference solutions while maintaining physical consistency.

Key Findings and Capabilities

The proposed framework successfully reproduced complex wave behaviors, including axial and radial displacement histories and the evolution of stress and strain, with close agreement to finite-element results. A significant advantage of this approach is its flexibility: once trained, the network can predict wave responses at time points that were not part of the original training data and can adapt to modified material properties without needing to perform new, time-consuming finite-element simulations.

Implications for Engineering

This methodology offers a robust and efficient surrogate model for high-rate solid mechanics and impact engineering. By demonstrating that the framework remains numerically stable and accurate across different material combinations, the authors show that this approach is a viable alternative for analyzing heterogeneous solids. It provides a way to perform rapid parametric studies and inverse analyses in fields like aerospace engineering and the design of impact-resistant, multilayer structures, where understanding how waves interact with material boundaries is critical.

Comments (0)

No comments yet

Be the first to share your thoughts!