Mesh2shell neural operator improves inverse shape optimization for physics driven design

Mesh2Shell Neural Operator for Inverse Shape Optimization

Machine Learning

Summary

Designing shapes that perform well in physical tasks can be hard, especially when the shapes don't fit into known categories. The authors introduce a system called Mesh-to-Shell Neural Operator (M2S-NO) that turns a detailed surface mesh into a format better suited for physics calculations. This approach lets them optimize shapes more freely without being stuck to specific shape families. They tested it on things like jet-engine brackets and neutron moderators and found improvements in performance compared to existing designs.

What this means in practice

  • For mechanical design engineers: Optimize structural components like brackets with flexible shape modifications that improve performance beyond preset design families.
  • For nuclear engineers: Design neutron moderators with enhanced neutron-sensor currents using physics-based shape optimization across diverse geometries.

Authors

Xinyi Li, Xi Deng, Yubo Cao, Hudson Liu, Anima Anandkumar

Abstract

Physics-driven design seeks high-performing geometries, which need not conform to a learned shape category. Parametric and latent-space approaches facilitate surrogate-based optimization, but restrict deformations to prescribed shape families or learned decoders. We introduce the Mesh-to-Shell Neural Operator (M2S-NO) as the mesh-to-shell component of a category-agnostic inverse shape optimization framework. We separate the explicit mesh used for design from the volumetric representation used for physics prediction. M2S-NO maps oriented surface samples to a surface-centered shell representation, such as TruncatedSDF or HeatSDF. A separate problem-specific PDE neural operator (PDE-NO) maps this representation to physical fields. Objective gradients update mesh vertices through both operators, without a latent shape decoder. Preconditioning and remeshing support free-form deformation, while fixed interfaces are controlled directly on the mesh. Optional shell-space proposals enable topology changes. We evaluate geometric matching, jet-engine brackets, and neutron moderators, with independent FEM and OpenMC evaluation. Relative to the best training designs, M2S-NO reduces maximum bracket displacement by 14.3% and increases neutron-sensor current by 13.8%. Optimization also succeeds from an out-of-distribution bracket initialization, demonstrating the framework's geometric flexibility.