A Foundation Model for Material Fracture Prediction

2025-07-30

AI Summary

Predicting material fracture faces three key challenges: scarcity of labeled data, poor model generalizability across materials and configurations, and prohibitive computational cost of high-fidelity physics-based simulations. To address these, we propose the first multimodal foundation model specifically designed for fracture modeling. It jointly encodes textual inputs—such as material properties and boundary conditions—with both structured and unstructured finite element meshes using large language model embeddings and a Transformer-based architecture. This enables zero-shot transfer and single-shot fine-tuning across diverse materials, geometries, and loading conditions. Evaluated on five material systems—including plastic-bonded explosives, steel, aluminum, shale, and tungsten—the model achieves high accuracy in predicting crack evolution and failure time. Remarkably, it generalizes to unseen materials (e.g., titanium alloy and concrete) with only one calibration sample, drastically reducing data dependency. Our approach consistently outperforms state-of-the-art domain-specific machine learning methods.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Accurately predicting when and how materials fail is critical to designing safe, reliable structures, mechanical systems, and engineered components that operate under stress. Yet, fracture behavior remains difficult to model across the diversity of materials, geometries, and loading conditions in real-world applications. While machine learning (ML) methods show promise, most models are trained on narrow datasets, lack robustness, and struggle to generalize. Meanwhile, physics-based simulators offer high-fidelity predictions but are fragmented across specialized methods and require substantial high-performance computing resources to explore the input space. To address these limitations, we present a data-driven foundation model for fracture prediction, a transformer-based architecture that operates across simulators, a wide range of materials (including plastic-bonded explosives, steel, aluminum, shale, and tungsten), and diverse loading conditions. The model supports both structured and unstructured meshes, combining them with large language model embeddings of textual input decks specifying material properties, boundary conditions, and solver settings. This multimodal input design enables flexible adaptation across simulation scenarios without changes to the model architecture. The trained model can be fine-tuned with minimal data on diverse downstream tasks, including time-to-failure estimation, modeling fracture evolution, and adapting to combined finite-discrete element method simulations. It also generalizes to unseen materials such as titanium and concrete, requiring as few as a single sample, dramatically reducing data needs compared to standard ML. Our results show that fracture prediction can be unified under a single model architecture, offering a scalable, extensible alternative to simulator-specific workflows.

Problem

• Predicting material fracture across diverse conditions and materials
• Overcoming limitations of narrow ML models and fragmented physics simulators
• Enabling flexible, scalable fracture prediction with minimal data requirements

Innovation

• Transformer-based model for diverse fracture prediction
• Multimodal input with structured and unstructured meshes
• Generalizes to unseen materials with minimal data