摘要
针对在管道泄漏检测中,泄漏信号和噪声都与管道状态和环境密切相关,且变分模态分解(VMD)方法不能较好地分解高频部分的信号,存在一定的噪声干扰问题,提出了一种基于互信息的VMD自适应噪声消除算法。首先,通过VMD算法将泄漏信号分解成具有不同特征时间尺度的固有模态函数;然后根据互信息准则,通过依次计算相邻分量之间的互信息值来区分高频和低频信号;最后将经过小波去噪后的高频信号和低频信号一起进行信号重构,得到特征增强信号。试验结果表明,对于信号去噪方面,相对于VMD方法,该方法更具有优越性。
In the light of the problem that leakage signal and noise are closely related to state and environment of pipeline in pipeline leakage detection,and most conventional noise elimination methods have to depend on signal intensity or empirical parameters of noise,so a variational mode decomposition（ VMD）adaptive de-noising algorithm based on mutual information is proposed. Firstly,the signal is decomposed into intrinsic mode function with different characteristic time scales through VMD. Then,according to the mutual information criterion,the high frequency signals and low frequency signals are distinguished by calculating the mutual information values between adjacent components in proper sequence. Finally,the high frequency signals and low frequency signals after wavelet de-noising are reconstructed,and the feature-enhanced signal can be got. Based on the experimental results,this method is more advantageous than VMD method.
出处
《压力容器》
2017年第8期75-80,共6页
Pressure Vessel Technology
基金
黑龙江省教育厅科技计划项目(12541063)
黑龙江省自然基金项目(E2016013
E201332)
关键词
泄漏检测
VMD算法
互信息
信号重构
leakage detection
variational mode decomposition（VMD） algorithm
mutual Information
signal reconstruction
