Generative deep learning for designing irregular metamaterials with programmable nonlinear mechanical responses

International Journal of Mechanical Sciences, Год журнала: 2025, Номер unknown, С. 110123 - 110123

Опубликована: Март 1, 2025

Язык: Английский

---

Inverse design of 3D cellular materials with physics-guided machine learning

Materials & Design, Год журнала: 2023, Номер 232, С. 112103 - 112103

Опубликована: Июль 4, 2023

This paper investigates the feasibility of data-driven methods in automating engineering design process, specifically studying inverse cellular mechanical metamaterials. Traditional designing materials typically rely on trial and error or iterative optimization, which often leads to limited productivity high computational costs. While approaches have been explored for materials, many these lack robustness fail consider manufacturability generated structures. study aims develop an efficient methodology that accurately generates metamaterial while ensuring predicted To achieve this, we created a comprehensive dataset spans broad range properties by applying rotations cubic structures synthesized from nine symmetries materials. We then employ physics-guided neural network (PGNN) consisting dual networks: generator network, serves as tool, forward acts simulator. The goal is match desired anisotropic stiffness components with unit-cell parameters. results our model are analyzed using three distinct datasets demonstrate efficiency prediction accuracy compared conventional methods.

Язык: Английский

---

Perspective: Machine Learning in Design for 3D/4D Printing

Journal of Applied Mechanics, Год журнала: 2023, Номер 91(3)

Опубликована: Окт. 5, 2023

Abstract 3D/4D printing offers significant flexibility in manufacturing complex structures with a diverse range of mechanical responses, while also posing critical needs tackling challenging inverse design problems. The rapidly developing machine learning (ML) approach new opportunities and has attracted interest the field. In this perspective paper, we highlight recent advancements utilizing ML for designing printed desired responses. First, provide an overview common forward problems, relevant types structures, space responses printing. Second, review works that have employed variety approaches different ranging from structural properties to active shape changes. Finally, briefly discuss main challenges, summarize existing potential approaches, extend discussion broader problems field This paper is expected foundational guides insights into application design.

Язык: Английский

---

Micromechanics-based deep-learning for composites: Challenges and future perspectives

European Journal of Mechanics - A/Solids, Год журнала: 2024, Номер 105, С. 105242 - 105242

Опубликована: Янв. 18, 2024

During the last few decades, industries such as aerospace and wind energy (among others) have been remarkably influenced by introduction of high-performance composites. One challenge, however, for modeling designing composites is lack computational efficiency accurate high-fidelity models. For design purposes, using conventional optimization approaches typically results in cumbersome procedures due to huge dimensions space high expense full-field simulations. In recent years, deep learning techniques found be promising methods increase robustness a variety algorithms multi-scale this perspective paper, short overview developments micromechanics-based machine given. More importantly, existing challenges further model enhancements future perspectives field development are elaborated.

Язык: Английский

---

A novel bio-inspired design method for porous structures: Variable-periodic Voronoi tessellation

Materials & Design, Год журнала: 2024, Номер 243, С. 113055 - 113055

Опубликована: Май 31, 2024

This paper introduces a novel approach, namely Variable-Periodic Voronoi Tessellation (VPVT), for the bio-inspired design of porous structures. The method utilizes distributed points defined by variable-periodic function to generate tessellation patterns, aligning with wide diversity artificial or natural cellular In this VPVT method, truss-based architecture can be fully characterized variables, such as frequency factors, thickness factors. approach enables optimal structures both mechanical performance and functionality. varied, anisotropic cell shapes sizes provide significantly greater flexibility compared typical isotropic addition, not only micro-macro multiscale materials, but is also applicable meso-macro scale structures, constructions, biomedical implants, aircraft frameworks. work employs Surrogate-assisted Differential Evolution (SaDE) perform optimization process. Numerical examples experiments validate that proposed achieves about 51.1% 47.8% improvement in compliance damage strength, respectively, than existing studies.

Язык: Английский

---

A discontinuous Galerkin method based isogeometric analysis framework for flexoelectricity in micro-architected dielectric solids

Computers & Structures, Год журнала: 2025, Номер 308, С. 107641 - 107641

Опубликована: Янв. 8, 2025

Язык: Английский

---

Data-Driven Nonlinear Deformation Design of 3D-Printable Shells

3D Printing and Additive Manufacturing, Год журнала: 2025, Номер unknown

Опубликована: Фев. 6, 2025

Язык: Английский

---

Inverse design of spinodoid structures using Bayesian optimization

Computational Mechanics, Год журнала: 2025, Номер unknown

Опубликована: Фев. 19, 2025

Abstract Tailoring materials to achieve a desired behavior in specific applications is of significant scientific and industrial interest as design key driver innovation. Overcoming the rather slow expertise-bound traditional forward approaches trial error, inverse attracting substantial attention. Targeting property, model proposes candidate structure with property. This concept can be particularly well applied field architected their structures directly tuned. The bone-like spinodoid are class materials. They considerable thanks non-periodicity, smoothness, low-dimensional statistical description. Previous work successfully employed machine learning (ML) models for design. amount data necessary most ML poses severe obstacle broader application, especially context inelasticity. That why we propose an inverse-design approach based on Bayesian optimization operate small-data regime. Necessitating substantially less data, small initial set iteratively augmented by silico generated until targeted properties found. application elastic demonstrates framework’s potential paving way advance

Язык: Английский

---

Inverse Designing Surface Curvatures by Deep Learning

Advanced Intelligent Systems, Год журнала: 2024, Номер 6(6)

Опубликована: Апрель 10, 2024

Smooth and curved microstructural topologies found in nature—from soap films to trabecular bone—have inspired several mimetic design spaces for architected metamaterials bio‐scaffolds. However, the approaches so far are ad hoc, raising challenge: how systematically efficiently inverse such artificial microstructures with targeted topological features? Herein, surface curvature is explored as a modality deep learning framework presented produce as‐desired profiles. The can generalize diverse features tubular, membranous, particulate features. Moreover, successful generalization beyond both data space demonstrated by designing that mimic profile of bone, spinodoid topologies, periodic nodal surfaces application bio‐scaffolds implants. Lastly, mechanics bridged showing be designed promote mechanically beneficial stretching‐dominated deformation over bending‐dominated deformation.

Язык: Английский

---

Inverse design of irregular architected materials with programmable stiffness based on deep learning

Composite Structures, Год журнала: 2024, Номер 340, С. 118210 - 118210

Опубликована: Май 14, 2024

Язык: Английский

---