最优小波基的改进阈值函数在管道泄漏检测中的研究

Application of Improved Multivariable Threshold Function in Pipeline Leak Detection Based on Optimal Wavelet Basis

以管道泄漏信号为研究对象,根据小波阈值降噪的基本原理,对泄漏信号进行短时傅里叶变换来判定泄漏信号特征频率主要集中的频段,并利用该频段泄漏波形对比常用小波基函数波形,选择最优小波基函数,随后确定出常用分解层数。针对传统阈值函数存在抑制噪声污染与保留信号细节之间的矛盾,对小波阈值降噪中的传统阈值函数进行改进,提出一种有效的多变量阈值函数,并与传统小波阈值降噪算法进行对比,改进后的小波阈值降噪在保障信号信噪比的同时能有效地滤除噪声。

In this paper, the leakage signal of pipeline was taken as the object of study. According to the bas- ic principle of wavelet threshold de-noising, the short-time Fourier transform of the leakage signal was used to determine frequency band of the leakage signal＇ s characteristic frequency and through using the frequency band＇ s leakage waveform, the wavelet basis function was selected, including the selection of optimal wavelet basis function and the determination of the number of commonly-used decomposition levels. Considering the traditional threshold function＇ s bewilderment in suppressing the noise pollution and reserving the details of sig-nals, the traditional threshold function in wavelet threshold noise reduction was improved to propose an effective multivariable threshold function and to compare it with the traditional wavelet threshold noise reduction algorithm. The results show that, the improved wavelet threshold noise reduction can effectively filter the noise out while keeping the signal to noise ratio.