信息缺失条件下管道泄漏信号识别研究

Recognition of pipeline leakage signals with incomplete information

在管道泄漏诊断过程中,负压波信息缺失会导致误判或漏判,采用改进奇异值降噪技术对负压波信号进行预处理,以提高定性评判泄漏的准确性。介绍了奇异值降噪技术的基本原理及其改进方法,并利用信号互相关技术对泄漏点位置进行了定量计算,实现了对液体输送管线的泄漏故障诊断。对室内液体管道在泄漏诊断过程中负压波特征明显和模糊情况下的奇异值降噪结果进行了对比。结果表明,相关分析方法对信息缺失条件下负压波信号拐点识别的准确率较高。该方法能突出强噪声背景下的负压波信号,并且使泄漏点的定位精度提高到1.52%。该研究结果也可用于液体管线小泄漏的诊断分析。

In the pipeline leakage diagnosis, incomplete negative pressure wave will generate false alarm or overlooked leakage alarm. The negative pressure wave signals preprocessed by the improved singularity value decomposition (SVD) noise-reducing method can qualitatively enhance the accuracy of diagnostic results. The basic principles and the improvement method of SVD technique were presented, and the leak point positions were determined quantitatively using signal correlation technology. The leakage faults in liquid delivery pipelines were diagnosed. The experimental leakage data of negative pressure wave with various fuzzy extents were analyzed with these methods. A high precision of leak positioning is up to 1.52 %, and a good way for enhancing the negative pressure wave under high-noise condition was achieved. The results prove that the signal correlation analysis is favorable for diagnosing the pipeline leakage with incomplete information.