摘要
管道运输是五大运输方式之一,环境影响和人为破坏等因素会造成管道的泄漏现象,管道发生泄漏会对人民财产安全和环境造成严重危害;目前国内的管道泄漏定位研究多以单点泄漏为主,而真实的管道运行工况中往往会发生多点泄漏现象。提出一种基于时差定位法的管道多点泄漏定位法;以两点泄漏为例,在管道上下游两端分别放置两个传感器,为降低管道运行过程中的噪声干扰,利用互谱结合变分模态分解算法将采集信号分离成多个模态并提出一种基于信息熵和滤波后信号相关系数等多指标结合的模态选取法;将去噪后的信号通过快速独立成分分析(fast independent component anaysis,Fast ICA)算法分离出独立的信号源,再基于相关系数匹配出对应泄漏源产生的声发射信号,最后通过时差定位得出泄漏源位置。在此研究基础上基于实验室管道和设备开展多点泄漏定位试验并取得了可靠的试验结果,为工业实际的多点泄漏定位提供了理论思路和解决方案。
Pipeline transportation is one of the five modes of transportation.Corrosion,environmental impacts,man-made damages,and other factors lead to pipeline leakage phenomenon in the process of pipeline transportation.Pipeline leakage causes property damage and is also harmful to environment.At present,many studies in China are focused on single-point leakage.However,in real pipeline operating conditions,multi-points leakages are often occurring.This paper presents a method of multi-point leakage locations,which is based on the time difference location method.Taking two-point leakages as an example,two sensors were placed at both ends of the pipeline respectively.The acquired signals were separated into multiple modes by using the cross-spectrum combined with a variational mode decomposition algorithm to reduce the noise interference during the pipeline operation.A mode selection method combining the correlation coefficient after cross-spectrum filtering and information entropy and other multiple indicators was proposed.An independent signal source was separated from the denoised signal through fast independent component anaysis(Fast ICA),and then acoustic emission signals caused by the corresponding leakage source were matched based on the correlation coefficient.Finally,the location of the leakage source was obtained through the time difference location.The multi-point leakage location experiments were carried out in the laboratory by using pipelines and required equipment.Results of this study support the theoretical ideas and provide potential solutions for multi-point leakage inindustrial practice.
作者
韩文翔
章兰珠
HAN Wenxiang;ZHANG Lanzhu(East China University of Science and Technology,Shanghai 200237,China)
出处
《振动与冲击》
EI
CSCD
北大核心
2022年第24期210-217,共8页
Journal of Vibration and Shock
基金
国家自然科学基金(51975213)。
关键词
多点泄漏
互谱
变分模态分解
多指标
multi-point leakage
cross-spectrum
variational mode decomposition
multi-index