摘要
在输油管线泄漏诊断中,压力信号常常夹杂大量的噪声信号,进行工况识别时需对原始信号进行降噪处理。应用小波Mallat算法,对压力信号分解和重构。在各个子带上分析信号的时域和频域特征,结合压力信号和噪声的时频分布规律,提出了对压力信号低频近似部分和分解到一定尺度的高频细节信号重构的降噪方法。对比小波阈值降噪结果,验证了小波重构具有良好的降噪效果,并简化降噪步骤、降低计算量,适合工程计算。
In oil pipeline leaks detecting, much noise interference always exists in pressure signals. It is necessary to identify the working status of pipeline that original signals must be de- noised. Pressure signals are decomposed and reconstructed by Mallat algorithm of wavelet. Timedomain characteristics and frequency-domain characteristics are analyzed on every sub-band. Based on time-frequency distribution of pressure signals and noise, a new denoising method is presented. Original signals are denoised by wavelet reconstructed with lowfrequency approximate part and high-frequency detail part by decomposed in some scale. The denoising effect is verified by the method of wavelet threshold. Wavelet reconstruction can be denoised efficiently, simplify denoising steps and decrease the computational complexity greatly. This algorithm is suitable for engineering application.
出处
《石油矿场机械》
2009年第11期72-76,共5页
Oil Field Equipment
关键词
小波变换
子带重构
压力波
降噪
wavelet transform
sub-band reconstruction
pressure wave
denoising
