基于VMD和窄带包络相关的船舶辐射噪声改进DEMON分析

Improved DEMON analysis of ship radiated noise based on VMD and narrow-band envelope correlation

船舶辐射噪声的包络谱中蕴含着轴频和桨叶数等船舶固有特征信息,对船舶目标识别具有重要意义。为了提高船舶辐射噪声包络谱解调性能,提出了基于变分模态分解(Variational Mode Decomposition,VMD)和窄带包络相关的改进DEMON分析方法。首先利用VMD算法代替传统带通滤波器,将船舶辐射噪声信号分解为若干个子带;然后对各子带进行希尔伯特(Hilbert)检波并计算平均窄带包络相关系数,用于衡量信号的包络调制在频域上的非均匀性;最后提取各子带信号包络谱并按照平均窄带包络相关系数进行加权融合,从而得出宽带噪声信号的包络谱。利用该方法对实测不同类型和不同航速船舶辐射噪声信号进行了处理,结果均表明所提方法能有效提高包络谱解调效果,较传统方法更为有效。

The envelope spectrum of ship radiated noise contains characteristic information such as shaft frequency and blade number,which is of great significance for ship target recognition.In order to improve the demodulation performance of envelope spectrum of ship radiated noise,an improved DEMON analysis method based on variational mode decomposition(VMD)and narrow-band envelope correlation is proposed in this paper.Firstly,VMD is used to replace the traditional band-pass filter to decompose the ship radiated noise signal into several sub-bands.And then,by Hilbert detection of each sub-band,the average narrow-band envelope correlation coefficient is calculated to measure the non-uniformity of modulation in frequency-domain.Finally,the envelope spectrum of each sub-band signal is extracted and weighted according to the average narrow-band envelope correlation coefficient to obtain the envelope spectrum of broadband noise signal.The measured noise signals radiated by different types and speeds of ships are processed.The results show that the proposed method can effectively improve the demodulation effect of envelope spectrum,and is more effective than the traditional method.