【摘要】 天然气管道泄漏检测采集的音波序列中含有大量噪声,在进行泄漏特征提取之前必须进行降噪处理。通过频率分布和混沌性分析,引入泄漏音波序列局部投影方法。该方法寻找相空间重构过程中由噪声形成的子空间,通过抑制该空间中的噪声投影来达到降噪的目的。仿真信号测量和现场试验表明:局部投影方法适用于泄漏音波序列降噪,可以解决降噪过程中频率混叠和混沌性带来的问题;与其他方法相比,局部投影方法在时域统计指标、频域和功率谱分析、自相关函数分析以及提高信噪比等方面明显得到改善,大大提高了泄漏特征提取的准确率。更多还原

【Abstract】 The acoustic series collected from natural gas pipeline leakage detection includes much noise,and thus,the noise must be reduced before extracting leakage characteristic index. The noise reduction method of local projection for leakage acoustic series was introduced by analyzing frequency distribution and chaos characteristics. The subspace formed by noise was given in the phase space reconstruction process in this method,and through inhibiting noise projection in that space,the noise was reduced. Simulation signal test and field experiment results indicate that the local projection method is suitable for leakage acoustic series noise reduction and can solve the problems brought by frequency alias and chaos in noise reduction processing. The results show that compared with other methods,the local projection method can get obvious improvement in time domain statistical indexes,in analyzing frequency domain,power spectrum and autocorrelation function,as well as in raising signal-to-noise ratio,which can largely increase the accuracy when extracting leakage characteristic index.更多还原