Basis-to-basis operator learning using function encoders DOI

Tyler Ingebrand,

Adam J. Thorpe, Somdatta Goswami

и другие.

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2024, Номер 435, С. 117646 - 117646

Опубликована: Дек. 16, 2024

Язык: Английский

Separable physics-informed DeepONet: Breaking the curse of dimensionality in physics-informed machine learning DOI
Luis Mandl, Somdatta Goswami, Lena Lambers

и другие.

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2024, Номер 434, С. 117586 - 117586

Опубликована: Ноя. 28, 2024

Язык: Английский

Процитировано

3
Causality enforcing parametric heat transfer solvers for evolving geometries in advanced manufacturing DOI
Akshay J. Thomas, Ilias Bilionis, Eduardo Barocio

и другие.

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2025, Номер 437, С. 117764 - 117764

Опубликована: Янв. 22, 2025

Язык: Английский

Процитировано

0
Differentiable Deep Learning Surrogate Models Applied to the Optimization of the IFMIF-DONES Facility DOI Creative Commons

Giovanni E. Romero,

Guillermo Rodríguez-Llorente,

L. Pacheco Rodriguez

и другие.

Particles, Год журнала: 2025, Номер 8(1), С. 21 - 21

Опубликована: Фев. 25, 2025

One of the primary challenges for future nuclear fusion power plants is understanding how neutron irradiation affects reactor materials. To tackle this issue, IFMIF-DONES project aims to build a facility capable generating source in order irradiate different material samples. This will be achieved by colliding deuteron beam with lithium jet. In work, within DONES-FLUX project, deep learning surrogate models are applied design and optimization linear accelerator. Specifically, neural operators employed predict envelopes along longitudinal axis accelerator effects at end, after collision. approach has resulted that able approximating complex simulations high accuracy (less than 17% percentage error worst case) significantly reduced inference time (ranging from 2 6 orders magnitude) while being differentiable. The substantial speed-up factors enable application online reinforcement algorithms, differentiable nature allows seamless integration programming techniques, facilitating solving inverse problems find optimal parameters given objective. Overall, these results demonstrate synergy between programming, offering promising collaboration among physicists computer scientists further improve other facilities. research lay foundations projects, where efforts performed.

Язык: Английский

Процитировано

0
Rapid estimation of residual stress in composite laminates using a deep operator network DOI
Seung-Woo Lee, T. C. Smit,

Kyusoon Jung

и другие.

Composites Part B Engineering, Год журнала: 2025, Номер unknown, С. 112409 - 112409

Опубликована: Март 1, 2025

Язык: Английский

Процитировано

0
Populating cellular metamaterials on the extrema of attainable elasticity through neuroevolution DOI
Min Yan, Ruicheng Wang, Ke Liu

и другие.

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2025, Номер 440, С. 117950 - 117950

Опубликована: Март 30, 2025

Язык: Английский

Процитировано

0
An attention-based architecture for predicting thermal stress on turbine blade film hole surface using partial temperature data DOI
Junjie Huang,

Jianqin Zhu,

Zeyuan Cheng

и другие.

Journal of Thermal Stresses, Год журнала: 2024, Номер 47(10), С. 1386 - 1409

Опубликована: Сен. 23, 2024

Язык: Английский

Процитировано

0
Basis-to-basis operator learning using function encoders DOI

Tyler Ingebrand,

Adam J. Thorpe, Somdatta Goswami

и другие.

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2024, Номер 435, С. 117646 - 117646

Опубликована: Дек. 16, 2024

Язык: Английский

Процитировано

0