相关概念视频
Fiber Reinforced Concrete
340
Fiber-reinforced concrete significantly enhances the structural and nonstructural properties of traditional concrete by incorporating fibers like steel, glass, and polymers. These fibers, varying from natural ones such as sisal and cellulose to manufactured ones like polypropylene and Kevlar, are mixed into hydraulic cement with aggregates. Steel fibers, often preferred for their robustness, contribute to improved ductility, toughness, and post-cracking performance. The concrete is classified...
340
Design Example: Managing Concrete Workability
300
This example deals with managing the workability of concrete for a raft foundation project under hot weather conditions. Workability is crucial for ensuring the concrete is easy to place, compact, and finish. In this scenario, a slump test — a common method to measure the workability of fresh concrete — initially indicated low workability. This was attributed to the rapid water loss from the concrete mix, exacerbated by the high temperatures causing the course aggregates to heat up.
300
您也可能阅读
相关文章
通过共同作者、期刊和引用图与本文相关的文章。
排序
Same author
ForestSAR-YOLO: real-time detection of small human targets in UAV imagery.
Scientific reports·2026
Same author
Assessing the onset of spring water-level rise in snowmelt-dominated rivers of northeastern Russia using machine learning.
Scientific reports·2026
Same author
Multi-objective optimization of a regional biogas supply chain using organic waste.
Scientific reports·2026
使用机器学习的不连续和连续纤维复合工艺的数据驱动优化:一篇评论
Ivan Malashin1, Dmitry Martysyuk1, Vadim Tynchenko1
1Artificial Intelligence Technology Scientific and Education Center, Bauman Moscow State Technical University, 105005 Moscow, Russia.
Polymers
|September 27, 2025
概括
机器学习通过自适应过程控制和缺陷检测来增强纤维复合材料制造. 这项调查探讨了数据驱动的方法,并提出了一种混合人工智能模型,以改善实时质量保证.
科学领域:
- 材料科学 材料科学 材料科学
- 制造业 工程 制造工程
- 人工智能的人工智能
背景情况:
- 复合材料制造面临着材料异质性和工艺变异性的挑战.
- 传统方法在实时质量保证和缺陷检测方面扎.
- 机器学习 (ML) 为自适应过程控制提供数据驱动的解决方案.
研究的目的:
- 调查ML在纤维复合材料制造中的应用.
- 审查过程控制和质量保证的数据驱动方法.
- 为天然纤维复合材料提出一种新的混合人工智能模型架构.
主要方法:
- 预测建模,传感器融合和自适应控制技术的审查.
- 分析了六个案例研究,包括机器人涂布,粘附预测和成型优化.
- 开发一种混合人工智能模型,集成基于物理的GNN,3D Spectral-UNet和交叉注意力控制器.
主要成果:
- 在合成数据上展示了拟议的混合AI模型的端到端可操作性.
- 验证了ML在缺陷检测和实时质量保证方面的有效性.
- 突出了基于物理学的神经网络和数字双胞胎的潜力.
结论:
- ML为推进复合材料制造提供了一个全面的路线图.
- 未来的工作应该解决小型数据制度和工业可扩展性的挑战.
- 拟议的混合人工智能模型显示了对闭环参数调整和改善制造结果的承诺.


