Implementing the inverse design and vibration isolation applications of piezoelectric acoustic black hole beams by machine learning

Thin-Walled Structures, Год журнала: 2025, Номер unknown, С. 113074 - 113074

Опубликована: Фев. 1, 2025

Язык: Английский

---

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

---

Design strategy of curved-beam based metamaterial with unprecedented lateral deformation mode transition

International Journal of Mechanical Sciences, Год журнала: 2025, Номер unknown, С. 110136 - 110136

Опубликована: Март 1, 2025

Язык: Английский

---

Snap‐through instability in mechanical metamaterials

Deleted Journal, Год журнала: 2025, Номер unknown

Опубликована: Янв. 9, 2025

Abstract Snap‐through instability, a rapid transition between equilibrium states, has emerged as crucial mechanism for designing mechanical metamaterials with novel functionalities, including fast motion, energy modulation, and bistable deformation. Metamaterials snap‐through known snapping metamaterials, have enabled diverse applications, such robotics, sensing, absorption, shape reconfiguration, intelligence. Given the importance of these advancements, comprehensive review this field is highly desired. This paper provides an overview recent research on focusing their design strategies applications. Here, we summarized in several respects, beam‐based structures, shell‐based origami/kirigami designs, according to basic elements, alongside brief discussion unique deformation mechanisms. Furthermore, potential applications are presented terms energy, To conclude, perspectives challenges opportunities emerging highlighted, offering insights into future development metamaterials.

Язык: Английский

---

Optimizing Metamaterial Inverse Design with 3D Conditional Diffusion Model and Data Augmentation

Advanced Materials Technologies, Год журнала: 2025, Номер unknown

Опубликована: Апрель 3, 2025

Abstract The inverse design of metamaterials is critical for advancing their practical applications. Although deep learning has transformed this process, challenges remain, particularly with insufficient data and less realistic, diverse generation 3D represented as voxels. To address these limitations, a augmentation technique developed based on topological perturbation introduced conditional diffusion model (3D‐CDM) to optimize metamaterial generation. This original dataset, comprising 200 voxel representations lattices triply periodic minimal surfaces, labeled effective physical properties computed using homogenization methods. dataset expanded 5000 entries the proposed technique. Training 3D‐CDM augmented significantly improved quality accuracy generated designs. successfully produces realistic targeted properties, including volume fraction, Young's modulus, thermal conductivity, outperforming existing voxel‐based generative models in terms fidelity diversity. can be further optimized extended broader range material microstructures.

Язык: Английский

---

Embedding Physics Information into Neural Networks to Enhance the Accuracy of Star-Shaped Elastic Metamaterial Design

Physics Letters A, Год журнала: 2025, Номер unknown, С. 130213 - 130213

Опубликована: Янв. 1, 2025

Язык: Английский

---

Design Strategy of Curved-Beam Based Metamaterial with Unprecedented Lateral Deformation Mode Transition

Опубликована: Янв. 1, 2025

Язык: Английский

---

Predicting Stress–Strain Curve with Confidence: Balance Between Data Minimization and Uncertainty Quantification by a Dual Bayesian Model

Polymers, Год журнала: 2025, Номер 17(4), С. 550 - 550

Опубликована: Фев. 19, 2025

Driven by polymer processing-property data, machine learning (ML) presents an efficient paradigm in predicting the stress-strain curve. However, it is generally challenged (i) deficiency of training (ii) one-to-many issue relationship (i.e., aleatoric uncertainty), and (iii) unawareness model uncertainty epistemic uncertainty). Here, leveraging a Bayesian neural network (BNN) recently proposed dual-architected for curve prediction, we introduce dual that enables accurate prediction while distinguishing between at each processing condition. The trained using Taguchi array dataset minimizes data size maximizing representativeness 27 samples 4D parameter space, significantly reducing requirements. By incorporating hidden layers output-distribution layers, quantifies both uncertainty, aligning with experimental fluctuations, provides 95% confidence interval predictions Overall, this study establishes uncertainty-aware framework property reliable, modest small size, thus balancing minimization quantification.

Язык: Английский

---

Impact protection mechanism and failure prediction of modular hierarchical honeycomb system with self-locking effect

International Journal of Impact Engineering, Год журнала: 2025, Номер unknown, С. 105274 - 105274

Опубликована: Март 1, 2025

Язык: Английский

---

Inverse design of multistable kirigami metamaterial via geometry-enabled shape programming and transforming

Physical Review Applied, Год журнала: 2025, Номер 23(3)

Опубликована: Март 25, 2025

---