A neural network for the prediction of damage to reinforced cylindrical shells subjected to non-contact underwater explosions DOI Open Access
Kai Zhou, Shichao Ding, Y.J. Xie

и другие.

Journal of Physics Conference Series, Год журнала: 2024, Номер 2891(6), С. 062007 - 062007

Опубликована: Дек. 1, 2024

Abstract Explosion tests and numerical simulations are of great significance for the study submarine other underwater target damage characteristics. However, cost real ship test is high, implementation difficult, time simulation calculation which presents some difficulties in attempting to quickly assess structures. Currently, combination machine learning has become a more effective means addressing aforementioned issues. In this paper, acoustic-structural arithmetic employed realise non-contact explosion reinforced cylindrical shell section. The maximum deflection reinforcement bar extracted as output parameter from results. distance, explosive equivalent, thickness taken input parameter. prediction analysis carried out based on back-propagation neural network algorithm learning. data generated by processed analysed error analysis. processing revealed that exhibited superior effect, establishing an accurate efficient model characteristics exploding shells.

Язык: Английский

Evolution of the microstructure of MWCNT-modified SBS asphalt under salt-freezing coupling DOI
Yu Zhang, Yingjun Jiang, Chenfan Bai

и другие.

Construction and Building Materials, Год журнала: 2025, Номер 481, С. 141410 - 141410

Опубликована: Май 6, 2025

Язык: Английский

Процитировано

0

High-precision weld width detection in laser transmission welding via crow and wolf optimized neural networks DOI
Ning Jiang,

Rundong Qian,

Haiyu Qiao

и другие.

Optics & Laser Technology, Год журнала: 2025, Номер 190, С. 113211 - 113211

Опубликована: Май 18, 2025

Язык: Английский

Процитировано

0

Integrated diagnosis optimization design of the electronic equipment based on spatial mapping DOI Creative Commons
X. J. Gu, Xianjun Shi

Science Progress, Год журнала: 2024, Номер 107(4)

Опубликована: Окт. 1, 2024

The complexity of test and fault information within electronic devices makes their integrated diagnosis a challenging problem when designing equipment reliability. Current is analyzed for optimization resource optimization. However, this neglects the connection between them. This paper proposes design strategy based on spatial mapping principle to quantitatively describe constraint relationship model established by constructing logical space, optimal configuration are sought grey wolf algorithm. Seven high-dimensional benchmark functions an used verify efficiency algorithm proposed in paper. compared with other four terms algorithm’s speed accuracy. results indicate that after has critical detection, isolation, false alarm rates 100%, 99.99%, 98.99%, 0.2993%, respectively. After optimization, number tests reduced 88.9%, cost saved 89%. Compared algorithms, achieves best results, reduces 42%–55%, decreases 77.63%–83.91%. not only considers resources but also dramatically while improving efficiency.

Язык: Английский

Процитировано

0

A neural network for the prediction of damage to reinforced cylindrical shells subjected to non-contact underwater explosions DOI Open Access
Kai Zhou, Shichao Ding, Y.J. Xie

и другие.

Journal of Physics Conference Series, Год журнала: 2024, Номер 2891(6), С. 062007 - 062007

Опубликована: Дек. 1, 2024

Abstract Explosion tests and numerical simulations are of great significance for the study submarine other underwater target damage characteristics. However, cost real ship test is high, implementation difficult, time simulation calculation which presents some difficulties in attempting to quickly assess structures. Currently, combination machine learning has become a more effective means addressing aforementioned issues. In this paper, acoustic-structural arithmetic employed realise non-contact explosion reinforced cylindrical shell section. The maximum deflection reinforcement bar extracted as output parameter from results. distance, explosive equivalent, thickness taken input parameter. prediction analysis carried out based on back-propagation neural network algorithm learning. data generated by processed analysed error analysis. processing revealed that exhibited superior effect, establishing an accurate efficient model characteristics exploding shells.

Язык: Английский

Процитировано

0