Machine learning for modelling unstructured grid data in computational physics: a review

2025-02-13

AI Summary

To address the challenges posed by unstructured mesh data in computational physics—namely topological irregularity, geometric complexity, and high-dimensional dynamics—that hinder conventional machine learning modeling, this paper proposes the first unified machine learning methodology framework tailored to this domain. Our approach innovatively integrates graph neural networks, spatial-attention Transformers, interpolation-coupled machine learning, mesh-free physics-informed neural networks (PINNs), generative models, and reinforcement learning–driven intelligent mesh generation. This establishes a bidirectional inspiration paradigm: physics-informed insights guide ML innovation, while ML capabilities enhance numerical methods. We systematically survey open datasets and delineate the applicability boundaries of each technique in fluid dynamics and environmental simulation. Notably, we are the first to clarify the technical pathways for generative modeling and adaptive mesh optimization, thereby providing theoretical foundations and practical guidelines for high-fidelity, interpretable, and generalizable physics-integrated modeling.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Humans and AI: Human-in-the-loop Machine Learning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Unstructured grid data are essential for modelling complex geometries and dynamics in computational physics. Yet, their inherent irregularity presents significant challenges for conventional machine learning (ML) techniques. This paper provides a comprehensive review of advanced ML methodologies designed to handle unstructured grid data in high-dimensional dynamical systems. Key approaches discussed include graph neural networks, transformer models with spatial attention mechanisms, interpolation-integrated ML methods, and meshless techniques such as physics-informed neural networks. These methodologies have proven effective across diverse fields, including fluid dynamics and environmental simulations. This review is intended as a guidebook for computational scientists seeking to apply ML approaches to unstructured grid data in their domains, as well as for ML researchers looking to address challenges in computational physics. It places special focus on how ML methods can overcome the inherent limitations of traditional numerical techniques and, conversely, how insights from computational physics can inform ML development. To support benchmarking, this review also provides a summary of open-access datasets of unstructured grid data in computational physics. Finally, emerging directions such as generative models with unstructured data, reinforcement learning for mesh generation, and hybrid physics-data-driven paradigms are discussed to inspire future advancements in this evolving field.

Problem

• Addressing ML challenges with unstructured grid data
• Reviewing advanced ML methods for computational physics
• Exploring ML integration in high-dimensional dynamical systems

Innovation

• Graph neural networks
• Transformer models attention
• Physics-informed neural networks