采用SVD-NMF的管道泄漏信号去噪算法

发布时间：2018-05-14 11:59

本文选题：管道泄漏定位 + 非负矩阵分解(NMF)　；参考：《传感技术学报》2017年01期

【摘要】：在管道泄漏检测中,压力信号中的噪声干扰会降低传统互相关法的定位精度。传统的去噪算法对环境的适应性差,去噪效果不理想。为此,提出了一种奇异值分解SVD(Singular Value Decomposition)与非负矩阵分解NMF(Nonnegative Matrix Factorization)相结合的管道泄漏信号去噪算法。该方法首先通过奇异值分解确定非负矩阵分解的阶数并对其初始化;然后,采用改进的非负矩阵分解算法对原信号进行迭代分解,获得去噪信号;最后,对去噪信号进行处理后通过互相关计算时延,并结合泄漏信号的传播速度实现泄漏定位。大量实验结果表明,SVD-NMF算法能够显著降低迭代次数,提高去噪速度;同时在泄漏检测中,能够达到去除噪声干扰,提高定位精度的目的。
[Abstract]:In pipeline leakage detection, the noise interference in the pressure signal will reduce the location accuracy of the traditional cross correlation method. The traditional de-noising algorithm has poor adaptability to the environment, and the effect of de-noising is not ideal. In this paper, a new method of pipeline leak signal denoising based on singular value decomposition (SVD(Singular Value decomposition) and nonnegative matrix factorization (NMF(Nonnegative Matrix factorization) is proposed. In this method, the order of nonnegative matrix factorization is determined by singular value decomposition and initialized. Then, the original signal is iteratively decomposed by the improved non-negative matrix decomposition algorithm, and the denoising signal is obtained. After processing the denoising signal, the time delay is calculated by cross-correlation, and the leakage location is realized by combining the propagation speed of the leakage signal. A large number of experimental results show that SVD-NMF algorithm can significantly reduce the number of iterations and improve the speed of de-noising, at the same time, in the leakage detection, the noise interference can be removed and the positioning accuracy can be improved.
【作者单位】：天津大学精密测试技术与仪器国家重点实验室;
【分类号】：U178;TN911.7

【相似文献】