基于独立分量分析的管道异常振动事件定位方法  

A locating method for petroleum pipeline abnormal vibration events based on independent component analysis

作　　者：张景川[1] 曾周末[1] 曹庆松[1] 封皓[1] 靳世久[1] 

机构地区：[1]天津大学精密测试技术与仪器国家重点实验室,天津300072

出　　处：《振动与冲击》2011年第4期81-85,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金重点资助项目(60534050);教育部博士点基金资助项目(200800560020)

摘　　要：针对传统互相关时延估计法定位管道故障点的局限性,提出了基于独立分量分析(ICA)的互相关时延估计方法用于管道异常振动事件定位。利用ICA滤波后的管道沿线振动信号进行互相关运算,抑制了相关高斯噪声引起的时延估计误差,为管道安全监测系统提供了一种新的高精度时延估计方法。仿真结果和现场实验数据表明,该方法可以快速精确地定位管道周围异常事件的发生位置,改善了相关运算的时延估计性能,相对于传统互相关时延估计法,可进一步降低时延估计的平均误差和均方差,具有更高的定位精度和定位一致性。Aiming at the limitation of conventional cross-correlation algorithms for time-delay estimation,a cross-correlation algorithm for time-delay estimation based on independent component analysis（ICA） for locating petroleum pipeline abnormal vibration events was proposed,it was used in the distributed optical fiber petroleum pipeline safety detection system for the safety of pipeline.During the pipeline detection,the environment was very complicated,the correlated Gaussian noises impacted the accuracy of time-delay estimation.ICA had good effect on removing the Gaussian interference and could protect available signals effectively.Using the de-noised signals with ICA,the cross-correlation analysis for time-delay estimation was conducted with the proposed method,the correlated Gaussian noises were suppressed effectively.Finally,compared with the results of the conventional cross-correlation algorithms,the simulation results and the signals obtained from tests proved that the proposed method can decrease the average error and mean square deviation of the time-delay estimation and improve locating accuracy and consistency.

关 键 词：油气管道 分布式光纤传感器 独立分量分析 互相关 时延估计 

分 类 号：TN247[电子电信—物理电子学] TN911.6