Towards end-to-end optimization in multimaterial 3D printing

2026-07-14

AI Summary

This work addresses the challenge of jointly optimizing material distribution and structural topology in multi-material 3D printing, which is hindered by high-dimensional design spaces and complex constitutive modeling. The authors propose an end-to-end computational framework that integrates a composition-aware physics-informed neural network with FEniCSx-based adjoint topology optimization. This approach learns explicit hyperelastic constitutive relationships directly from experimental data and employs symbolic differentiation to compute exact gradients. For the first time, it seamlessly unifies interpretable machine learning with continuum topology optimization, enabling co-design of highly anisotropic contact responses under non-failure tensile constraints. Demonstrated on a soft robotic gripper, the method successfully achieves integrated material-structure optimization, showcasing its efficiency and practicality for complex multi-material fabrication.

Abstract

Multimaterial 3D printing enables the fabrication of functionally graded components, but optimizing their spatial material distribution alongside structural topology remains a formidable challenge due to high-dimensional design spaces and complex constitutive modeling. This paper presents an end-to-end computational framework integrating sparsified physics-augmented neural networks with finite-element-based topology optimization. By extracting closed-form, composition-aware hyperelastic constitutive laws from experimental data, this approach facilitates exact symbolic differentiation via the adjoint state method implemented with FEniCSx, efficiently circumventing the bottlenecks of applying neural network constitutive models. This pipeline is deployed on soft robotic gripper applications, demonstrating continuous composition optimization for highly anisotropic contact responses, and the concurrent optimization of macroscopic topology and material distribution under non-failure stretch constraints. This methodology could replace laborious empirical prototyping, establishing interpretable machine-learning models as practical, robust design primitives for advanced multimaterial additive manufacturing.

Problem

• multimaterial 3D printing
• topology optimization
• material distribution
• constitutive modeling
• design space

Innovation

• multimaterial 3D printing
• physics-augmented neural networks
• topology optimization
• constitutive modeling
• adjoint state method