A new graph-based surrogate model for rapid prediction of crashworthiness performance of vehicle panel components

2025-03-16

AI Summary

To address the high computational cost of finite element (FE) simulations and the trade-off between accuracy and efficiency in existing graph neural network (GNN) surrogate models for crashworthiness assessment of automotive components (e.g., B-pillars), this paper proposes ReGUNet—a novel GNN-based surrogate model. ReGUNet innovatively integrates U-Net–style multi-scale graph up/down-sampling with a recurrent temporal propagation mechanism, enabling simultaneous capture of fine-grained local geometric features and stable multi-timestep dynamic responses. Trained exclusively on FE simulation data, ReGUNet achieves a relative error of only 0.74% in predicting peak intrusion depth for B-pillar side-impact cases. It reduces average deformation prediction error by over 51% compared to baseline GNN models and significantly accelerates inference—thereby fulfilling industrial requirements for rapid, high-fidelity crashworthiness evaluation.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Natural Language Processing: Safety and Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines

Abstract

During the design cycle of safety critical vehicle components such as B-pillars, crashworthiness performance is a key metric for passenger protection assessment in vehicle accidents. Traditional finite element simulations for crashworthiness analysis involve complex modelling, leading to an increased computational demand. Although a few machine learning-based surrogate models have been developed for rapid predictions for crashworthiness analysis, they exhibit limitations in detailed representation of complex 3D components. Graph Neural Networks (GNNs) have emerged as a promising solution for processing data with complex structures. However, existing GNN models often lack sufficient accuracy and computational efficiency to meet industrial demands. This paper proposes Recurrent Graph U-Net (ReGUNet), a new graph-based surrogate model for crashworthiness analysis of vehicle panel components. ReGUNet adoptes a U-Net architecture with multiple graph downsampling and upsampling layers, which improves the model's computational efficiency and accuracy; the introduction of recurrence enhances the accuracy and stability of temporal predictions over multiple time steps. ReGUNet is evaluated through a case study of side crash testing of a B-pillar component with variation in geometric design. The trained model demonstrates great accuracy in predicting the dynamic behaviour of previously unseen component designs within a relative error of 0.74% for the maximum B-pillar intrusion. Compared to the baseline models, ReGUNet can reduce the averaged mean prediction error of the component's deformation by more than 51% with significant improvement in computational efficiency. Provided enhanced accuracy and efficiency, ReGUNet shows greater potential in accurate predictions of large and complex graphs compared to existing models.

Problem

• Predict crashworthiness of vehicle panels quickly
• Overcome limitations of current surrogate models
• Improve accuracy and efficiency in crash analysis

Innovation

• Uses Graph Neural Networks for crashworthiness prediction
• Introduces Recurrent Graph U-Net (ReGUNet) architecture
• Improves accuracy and computational efficiency significantly