从数据中心的角度推进基于深度学习的声学泄漏检测方法,从数据中心的角度将其应用于水分系统.
Yipeng Wu1, Xingke Ma1, Guancheng Guo1
1School of Environment, Tsinghua University, 100084, Beijing, China.
Water research
|June 28, 2024
概括
数据增强显著改善了在水分系统 (WDSs) 中声泄漏检测的深度学习. 像IAAFT和掩盖这样的技术通过增加数据多样性和专注于全球特征来提高准确性.
科学领域:
- 工程 工程师 工程师 工程师
- 计算机科学 计算机科学
- 环境科学 环境科学
背景情况:
- 水分系统 (WDS) 面临着严重的泄漏挑战.
- 对声学泄漏检测的深度学习是有希望的,但依赖于数据.
- 现有的研究往往忽视了数据增强的作用.
研究的目的:
- 调查数据增强对基于深度学习的声学泄漏检测的影响.
- 评估五种基于随机转换的增强技术.
- 提高人工智能驱动的WDS泄漏检测的性能和可靠性.
主要方法:
- 应用了震动,缩放,扭曲,代振幅调整的里埃变换 (IAAFT),并对来自现实世界WDS的声学信号进行掩盖.
- 使用卷积神经网络 (CNN) 分类器来分析增强声信号谱图.
- 在数据分割之前实施数据增强,以避免数据泄露.
主要成果:
- 数据增强对于防止数据泄露和过度乐观的结果至关重要.
- 通过增加数据量和多样性,IAAFT提高了识别准确度7%以上.
- 通过鼓励CNN学习全球光谱图特征,掩盖提高了性能.
- 连续应用IAAFT和掩盖进一步提高了泄漏检测性能.
- 数据增强提高了复杂模型转移学习的有效性.
结论:
- 数据增强对于推进人工智能驱动的声学泄漏检测技术至关重要.
- IAAFT和掩盖是提高泄漏检测准确性的有效技术.
- 以数据为中心的方法,包括增强,是WDS中成熟应用的关键.
相关概念视频
Multiple Pipe Systems
738
Multipipe systems consist of complex configurations of interconnected pipes designed to transport fluids efficiently across intricate networks. They are essential in engineering applications requiring precise control over flow distribution, pressure, and head loss. They are categorized into series, parallel, loop, and network configurations, each distinguished by unique flow characteristics and applications.
Series Configuration
In a series configuration, fluid flows sequentially from one pipe...
Series Configuration
In a series configuration, fluid flows sequentially from one pipe...
738
Design Example: Designing a Residential Plumbing System
671
The design of residential plumbing systems requires carefully evaluating water demand, flow rates, and pressure dynamics to ensure both efficiency and reliability. The nature of water flow within pipes is defined by its Reynolds number, which classifies flow as either laminar (smooth) or turbulent.
671
Rapidly Varying Flow
57
Rapidly varying flow (RVF) in open channels is characterized by abrupt changes in flow depth over a short distance, with the rate of depth change relative to distance often approaching unity. These flows are inherently complex due to their transient and multi-dimensional nature, making exact analysis difficult. However, approximate solutions using simplified models provide valuable insights into their behavior.Key Features of Rapidly Varying FlowRVF is commonly observed in scenarios involving...
57


