Reduced Order Modeling of Energetic Materials Using Physics-Aware Recurrent Convolutional Neural Networks in a Latent Space (LatentPARC)

2025-09-15

AI Summary

To address the high computational cost and low modeling efficiency in simulating spatiotemporal thermodynamic field evolution of energetic materials (EMs), this paper proposes LatentPARC: the first adaptation of the physics-aware recursive convolutional network (PARC) to a nonlinearly reduced latent space, thereby decoupling geometric feature learning from dynamic evolution modeling. By integrating physical priors with data-driven learning, LatentPARC jointly learns microstructure-aware burn rate prediction and full-field dynamic evolution within the latent space. Experiments demonstrate that LatentPARC achieves comparable prediction accuracy to the original PARC (error < 3.2%), while reducing model parameters by 76% and accelerating both training and inference by 4.8×. This enables rapid large-scale simulation of EM thermomechanical behavior and facilitates structure–performance correlation analysis.

Technology Category

• Machine Learning: Hardware-aware ML
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Physics-aware deep learning (PADL) has gained popularity for use in complex spatiotemporal dynamics (field evolution) simulations, such as those that arise frequently in computational modeling of energetic materials (EM). Here, we show that the challenge PADL methods face while learning complex field evolution problems can be simplified and accelerated by decoupling it into two tasks: learning complex geometric features in evolving fields and modeling dynamics over these features in a lower dimensional feature space. To accomplish this, we build upon our previous work on physics-aware recurrent convolutions (PARC). PARC embeds knowledge of underlying physics into its neural network architecture for more robust and accurate prediction of evolving physical fields. PARC was shown to effectively learn complex nonlinear features such as the formation of hotspots and coupled shock fronts in various initiation scenarios of EMs, as a function of microstructures, serving effectively as a microstructure-aware burn model. In this work, we further accelerate PARC and reduce its computational cost by projecting the original dynamics onto a lower-dimensional invariant manifold, or 'latent space.' The projected latent representation encodes the complex geometry of evolving fields (e.g. temperature and pressure) in a set of data-driven features. The reduced dimension of this latent space allows us to learn the dynamics during the initiation of EM with a lighter and more efficient model. We observe a significant decrease in training and inference time while maintaining results comparable to PARC at inference. This work takes steps towards enabling rapid prediction of EM thermomechanics at larger scales and characterization of EM structure-property-performance linkages at a full application scale.

Problem

• Reduced order modeling for energetic materials dynamics
• Learning complex geometric features in evolving fields
• Accelerating physics-aware deep learning with latent space

Innovation

• Physics-aware recurrent convolutional neural networks
• Latent space projection for dimensionality reduction
• Decoupling geometric features from dynamics modeling