管道泄漏信号自适应增强技术

Research on Adaptive Enhancement of Pipeline Leak Signal

基于动态压力传感器的管道泄漏检测,由于工况的多变使得泄漏信号经常被背景信号淹没,固定分解尺度的小波去噪得不到理想的效果。在分析泄漏信号特征的基础上,提出了一种简单、实用的泄漏信号自适应增强方法。它首先对动态压力信号进行正负区间划分,分别计算各个正负区间信号的累加和,并计算正、负累加和信号的3σ,对于累加和大于3σ的信号区间,保留幅值小于1/2峰值部分信号,并对剩余部分进行3次样条插值,从而自适应地实现泄漏信号增强。信噪比计算、功率谱分析和定位结果表明,该方法可以最大程度地保留泄漏信号的固有变化趋势,并提高泄漏信号的信噪比,降低定位误差。

Dynamic pressure sensor based pipeline leak detection encounters lower SNR problems because the signals were submerged by background noise caused by inconstant working condition.Even de-noising with fixed-scale wavelet analysis can not solve the problem.With analysing the characteristic of leak signal,a simple,practical and adaptive signal enhancement is proposed.Dynamic pressure signal is divided into positive and negative intervals,the sums of every intervals and the 3σ of either positive or negative signal sums are calculated.If the signal sum of one interval is bigger than the 3σ,the signals below one half of the peak in the interval are reserved,and the others are interpolated with cubic spline,with which the leak signal is enhanced adaptively.Calculations of SNR,analysis of power spectrum and leak location show that the SNR of leak signals are improved,the location errors are reduced and intrinsic trends are reserved.