摘要
针对管道压力泄漏信号去噪的问题,提出基于敏感因子奇异值分解(Singular Value Decomposition,SVD)的管道泄漏压力信号去噪的方法。该方法首先对原始信号构造Hankel矩阵再进行SVD分解,将分解后得到的分量信号利用敏感因子找出敏感分量,最后通过定位因子选择敏感分量所对应的奇异值进行信号重构,并用该方法对矿浆管道实验平台运行中采集到的压力信号进行降噪处理。实验结果表明,该方法有效地去除矿浆管道的压力信号中的噪声,作为信号的预处理为管道泄漏检测和定位提供良好的基础。此外,该方法与小波去噪方法进行对比,结果表明,该方法具有更好的去噪效果。
Considering the problems of great difficulties in resolving pipeline pressure leak signal de-noising,a new method based on singular value decomposition（SVD）is proposed.First,Hankel matrix in original signal is built.Then SVD decomposition of the selected signals is done.The component signals obtained by decomposition are used to find out the sensitive component by using the sensitive factor.At last,the corresponding singular values of the sensitive components are reconstructed by the located factor.The experimental results show that this method,used as a pretreatment of signals,effectively removes the noise of pressure factor of mineral pipeline,which provides a good foundation for pipeline leak detection and location.Furthermore,a comparison is made between wavelet de-nosing and singular value decomposition,the method used in this paper has better de-noising effects.
出处
《计算机与数字工程》
2017年第4期768-772,787,共6页
Computer & Digital Engineering
关键词
敏感奇异值分解
管道泄漏
去噪
sensitive factors SVD
pipeline leak
de-noising
