【作者】 梅继丹；

【导师】 惠俊英；

【作者基本信息】 哈尔滨工程大学 ， 水声工程， 2010， 博士

【摘要】 声图是一种特殊的被动声成像方式，可以给出噪声源的空间分布。最早应用于空气声学，在对轨道列车噪声分布测量、大型发动机内部噪声测量等方面已取得了良好的效果，并成功的指导了这些大型机器设备的减振降噪。水声中的舰艇目标降噪问题同样备受关注，受空气中的声图测量法启发，论文研究了水声中的声图测量技术，用于分析舰艇等目标的噪声源空间分布情况。声图测量采用的核心算法是聚焦波束形成技术，其原理与常规波束形成类似。不同的是由于声图测量的测量范围在近场，采用的是近场球面波补偿，而非平面波补偿，补偿量与空间位置有关。利用空间扫描的方式，来根据不同位置的波束形成输出功率强度来判定噪声源的空间分布情况。由于水声物理环境复杂，使得高精度的大型半波间距阵布放代价很大，声图测量一般都采用稀疏阵的测量形式。文中采用了分频段综合的方法和虚拟阵元法来抗阵列稀疏给测量带来的影响。水声信道是相干多途信道，文中结合了虚拟时间反转镜算法和虚阵算法用来抗多途信道的影响。以上措施取得了良好的效果。文中分析了声图测量法的分辨力与其阵长、目标距离、处理信号频率有关，相同距离和阵长条件下方法对低频声源的分辨能力较差，为此研究了近场MVDR聚焦波束形成算法和STMV聚焦波束形成算法来提高声图测量对低频声源的分辨能力。声图测量过程中，若测量区内存在多个声源，则它们之间会形成相干干扰，强声源对弱声源的干扰尤为严重，为此研究了基于自适应干扰抵消的近场波束零陷算法和基于Bartlett权的近场波束零陷算法，抗各声源间的相干干扰，进一步提高多声源的声图测量质量。此外，在进行水平阵声图测量时，扫描平面深度是否准确同样会影响到声图测量的定位精度。对于水平阵来说，声图测量是在某一深度平面内进行的二维扫描，扫描深度要与声源实际深度一致，否则就会引人测量误差，这使得对声源深度的测量成为了影响水平阵声图测量精度的一个关键问题。因为水平阵在垂直方向是无指向性的，所以关于水平阵的声源深度测量问题文献中一直未有涉及。论文提出对于过阵航行的特殊航行过程，利用某一深度的声图测量的定位结果，根据目标与阵的几何关系是可以解得声源深度的。文中研究了几何法和圆交汇法两种用于水平阵的声源深度测量方法。前一种方法适用于阵水平情况，而后一种方法用于阵有一定角度倾斜的情况。在对声源被动定位的基础上，对声图测量在辐射噪声源分离中的应用进行了初步探讨。研究了各噪声源信号的分离提取的方法，并分别从时域和频域的角度分析了方法的分离性能。论文基于水平阵、垂直阵以及面阵对声图测量算法进行了仿真研究，并通过4个航次的水平阵海试，验证了声图测量中的各种算法。更多还原

【Abstract】 Underwater acoustic image is a special passive acoustic imaging technology,which can give the spatial distribution of noise sources Applied in aero acousticsat the very beginning，underwater acoustic image has good results in noise-sourcedistribution measurements of railway trains and internal noise measurements inthe large engines，and is successful to guide the vibration and noise reduction ofthese large engines The vibration and noise reduction of ship also receives muchconcern Enlighten by aero acoustic image measurement，the paper researches onunderwater acoustic image measurement to maalyze the spatial distribution ofnoise sources in shipsThe core algorithm of underwater acoustic image measurement is dynamicsfocus beam forming whose principle is similar to normal beam forming Thedifference between the two methods is that the underwater acoustic imagemeasurement is applied in near field in which spherical wave compensationwhose value is related to spatial position is taken；whereas the normal beamforming takes plane wave compensation By using spatial scan，the spatialdistribution of noise sources is judged by the beam forming power output indifferent pixel pointsBecause of the complex underwater environment，it is difficult to lay a largehigh-accuracy half-wavelength array．Thus it is normal to use sparse array inunderwater acoustic image measurement In the paper,~equency-divisionintegrated technology and virtual array technology are applied to overcome theinfluence of sparse array．The underwater acoustic channel is coherent multi-pathchannel The paper combines virtual time reversal mirror algorithm and virtualarray algorithm to overcome multi-path interference，which achieved good effectThe resolution of underwater acoustic image measurement is related to array length，the distance to the target and the band of signal The resolution is not SOgood to low frequency sources w-hen the same target distance and array length aJ’efixed Thus researches on MVDR focus beam forming and STMV focus beamforming are applied to improve the resolution of low frequency sound sourcesThere will be coherent interference for each other when there are multi soundsources The influence of the strong sources to weak ones is severe Therefore，two anti-interference technologies are researched One is null beam forming innear field based on adaptive interference cancellation，and the other is based onBartlett weight The methods can overcome coherent interference and enhance thequality of underwater acoustic image in multi sound sourcesMoreover,the scanning depth will afl’ect the accuracy of the underwateracoustic image w-hen the measurement array is horizontal For the horizontal arraysince the acoustic image measurement is a two-dimensional image in certainscanning depth，scanning depth and actual depth of sources must be the Sa2rle，otherwise there will be measurement error，which 1eads the fact that the hemeasurement of the sound source depth is a key issue which afl’ects the accuracyof measuring for horizontal array．Because a horizontal array is non-directional inthe vertical direction，the measurement of source depth of horizontal array isseldom discussed in literatures The paper proposes the viewpoint that for thespecial navigation across the array using the location result of underwater acousticimage measurement to certain depth the source dcepth can be achieved fromgeometrical relationship of the sources and array Two depth measuring methodsare discussed in the paper One is geometric method and the other is circleintersection method The former is suitable for the horizontal array and the latteris for the array with a certain angle oftiltBased on the passive location，a preliminary study is done on the radiationnoise sources separation by underwater acoustic image measurement The paperresearches on the extraction methods of noise source signal，and the performance of signal separation is examined in the time domain and~equency domain，respectivelyThe paper does simulation research on underwater acoustic imagemeasurement Ngorithms based on the horizontal array,ve~ical array and planararray,and the various underwater acoustic image measurement algorithms areverified through four horizontal array voyage tests更多还原