A comprehensive comparison of neural operators for 3D industry-scale engineering designs

2025-10-07

AI Summary

Neural operators for 3D industrial engineering design lack fair, systematic evaluation. Method: We construct and open-source six standardized 3D engineering benchmark datasets covering thermodynamics, elasticity, plasticity, and fluid dynamics; uniformly evaluate four representative architectures—DeepONet (branch-trunk), graph neural networks, Fourier neural operators, and PointNet—across accuracy, inference latency, memory footprint, and deployment complexity; and propose an engineering-oriented preprocessing standardization pipeline and practical model adaptation strategies. Contribution/Results: This work establishes the first reproducible, comparable, and deployable comprehensive evaluation framework for neural operators on industrial-scale 3D physical problems. It enables rigorous, cross-model and cross-task benchmarking, providing empirical evidence and practical guidelines for engineering AI model selection and architecture design.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: 3D Computer Vision
• 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
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Neural operators have emerged as powerful tools for learning nonlinear mappings between function spaces, enabling real-time prediction of complex dynamics in diverse scientific and engineering applications. With their growing adoption in engineering design evaluation, a wide range of neural operator architectures have been proposed for various problem settings. However, model selection remains challenging due to the absence of fair and comprehensive comparisons. To address this, we propose and standardize six representative 3D industry-scale engineering design datasets spanning thermal analysis, linear elasticity, elasto-plasticity, time-dependent plastic problems, and computational fluid dynamics. All datasets include fully preprocessed inputs and outputs for model training, making them directly usable across diverse neural operator architectures. Using these datasets, we conduct a systematic comparison of four types of neural operator variants, including Branch-Trunk-based Neural Operators inspired by DeepONet, Graph-based Neural Operators inspired by Graph Neural Networks, Grid-based Neural Operators inspired by Fourier Neural Operators, and Point-based Neural Operators inspired by PointNet. We further introduce practical enhancements to adapt these models to different engineering settings, improving the fairness of the comparison. Our benchmarking study evaluates each model strengths and limitations in terms of predictive performance, computational efficiency, memory usage, and deployment complexity. The findings provide actionable insights to guide future neural operator development.

Problem

• Standardizing neural operator comparisons for 3D engineering designs
• Evaluating model performance across six industrial-scale physical problems
• Benchmarking computational efficiency and deployment complexity of operators

Innovation

• Standardized six 3D industry-scale engineering datasets
• Systematically compared four neural operator architectures
• Introduced practical enhancements for fair comparison