|
| PCCP管道断丝声纹信号时频特征研究 |
| Study on time-frequency characteristics of wire break acoustic signal of PCCP pipeline |
| 投稿时间：2025-02-15 修订日期：2025-05-22 |
| DOI： |
| 中文关键词: PCCP 断丝 声纹特征 分形维数 边际谱 时频特征 |
| 英文关键词:PCCP wire break fractal dimension marginal spectrum time-frequency characteristic |
| 基金项目: |
|
| 摘要点击次数: 51 |
| 全文下载次数: 0 |
| 中文摘要: |
| PCCP管道作为长距离输水工程的核心承压结构，其服役安全性与预应力钢丝断裂数量密切相关。针对基于分布式光纤的断丝声纹监测需求，提出了一套系统化的声纹特征提取方法。采用分形维数分析法对异常声波信号进行精准识别与时段截取，有效缩减非线性时频分析数据量。采用ICEEMDAN将断丝声纹信号分解为多个模态分量，并基于模态分量序列的突变系数分析选取声纹信号瞬时分量及长时分量。通过对瞬时分量和长时分量进行边际谱分析，提取断丝声纹信号频域特征。最后，通过与5组非断丝干扰冲击信号进行对比，结果表明：断丝声纹信号异常声波信号持续时长为0.25s左右，而干扰声纹的异常声波信号持续时长往往低于0.15s；断丝声纹信号瞬时分量主频集中在1400Hz以上，而干扰声纹信号的瞬时分量主频在1300Hz以下；断丝声纹信号长时分量主频在120Hz左右，而其他非断丝干扰声纹长时分量主频在80Hz以下。所提出的PCCP管道异常声纹信号时频特征的提取方法通过特征参数的自适应提取机制，实现了全流程计算机自动化处理，为PCCP断丝监测提供了有力技术支持。 |
| 英文摘要: |
| As the core pressure-bearing structure in long-distance water conveyance projects, the service safety of PCCP is closely related to the number of fractured prestressed steel wires. To address the demand for acoustic monitoring of wire breaks using distributed optical fiber sensing, a systematic acoustic feature extraction method was proposed. Fractal dimension analysis was employed to accurately identify and intercept abnormal acoustic signal segments, effectively minimizing the data volume for nonlinear time-frequency analysis. The ICEEMDAN algorithm was then applied to decompose acoustic signals into multiple intrinsic mode functions. Based on mutation coefficient analysis of intrinsic mode sequences, both instantaneous and long-term components of the acoustic signals were selected. Marginal spectrum analysis was conducted on these components to extract frequency-domain features of wire-break acoustic signals. Comparative experiments with five types of non-fracture interference signals demonstrated three key findings: (1) The duration of abnormal acoustic waves caused by wire breaks is approximately 0.25s, significantly longer than that of interference signals (<0.15s); (2) The dominant frequency of instantaneous components in wire-break signals concentrates above 1400Hz, while that of interference signals remains below 1300Hz; (3) The long-term component of wire-break signals exhibit a dominant frequency near 120Hz, contrasting with interference signals' long-term components below 80Hz. The proposed time-frequency characteristics extraction method achieves fully automated computer processing through adaptive parameter extraction mechanisms, providing robust technical support for PCCP wire-break monitoring. |
| 查看/发表评论 下载PDF阅读器 |
| 关闭 |
