【作者】 陈欢；

【导师】 杨德森；

【作者基本信息】 哈尔滨工程大学 ， 水声工程， 2011， 博士

【摘要】 潜艇的声隐身性能是衡量潜艇的安全性和作战性能的重要指标。辐射噪声是各种无源声呐和声学武器探测潜艇的主要目标信息,是破坏其声隐身性能的主要因素,自噪声则直接影响本艇声呐站的工作性能,因此降低潜艇水下噪声具有重要意义。为了有针对性地开展噪声控制,同时为潜艇声学设计中的噪声指标提供实际依据,开展噪声源识别分离技术研究,查找系统中的主要噪声源及确定各类噪声源的贡献具有重要意义。随着减振降噪技术的发展,以往的噪声源定位识别方法和分析手段已无法有效获得噪声源信息,因此研究适用于潜艇大尺寸、复杂系统,同时便于工程实施的噪声源近场定位识别方法,成为潜艇噪声控制和减振降噪技术的关键。本文首先阐述了近场声聚焦技术在噪声源定位识别方面的优势,在回顾近场声聚焦技术理论发展及其工程应用研究的基础上,以线列阵近场声聚焦技术在潜艇噪声源定位识别上的应用为研究背景,对测量面为平面和柱面的噪声源近场高分辨定位识别方法、及能抑制基阵灵敏度误差和相位误差的稳健噪声源近场定位识别方法等方面进行了系统、全面的理论和试验研究。本文首先以基阵近场理论为基础,从理论上阐述了近场声聚焦的基本原理,并给出均匀线列阵聚焦空间分辨率的计算公式,分析测量距离和基阵孔径等参数对聚焦空间分辨率的影响,并针对近场聚焦的空间混叠问题提出了定位模糊判决方法,可以用于指导基阵的阵形设计。本文针对常规近场聚焦波束形成和MVDR近场聚焦波束形成等定位识别方法的不足,重点研究了基于幅度相位联合补偿的MVDR、MUSIC近场聚焦波束形成方法。该方法可以实现噪声源相对强度估计,且在一定条件下可以有效降低信号频段和测量距离对算法性能的限制,提高基阵在低频段的聚焦空间分辨率,在高频段具有更强抑制空间混叠能力,具有更低的旁瓣级,及更强的抗干扰能力,更小的声源强度估计误差,更高的定位精度。针对极值搜索类算法不适用于相干源近场定位识别问题,提出了基于最大似然估计相干源近场高分辨定位识别方法,并利用遗传算法寻求最大似然估计全局最优解。该方法具有较高定位精度,且仅利用小孔径基阵就可实现相干源近场高分辨定位,同时不受阵列几何结构的限制。针对近场声聚焦技术难以实现端射方向噪声源定位识别问题,提出了基于虚拟旋转变换技术的端射方向噪声源近场定位识别方法,分析虚拟阵列孔径、测量距离、信噪比、观测区间等参数对算法性能的影响,并给出该方法的具体应用条件。本文针对潜艇辐射噪声的宽带连续谱成份,研究了基于子带分解和聚焦变换技术的宽带噪声源近场高分辨定位识别方法。该方法将子带分解和聚焦变换技术,与幅度相位联合补偿的极值搜索类算法有效结合,从而可实现宽带非相干源和相干源的相对强度估计和声源空间位置估计,并对上述方法的空间谱特性、定位性能和聚焦空间分辨率进行了对比分析。本文研究了适用于潜艇模型的测量面为柱面的噪声源近场高分辨定位识别方法,并对算法的性能进行详细分析。为解决线列阵近场定位中的左右舷模糊问题,提出矢量声压组合基阵的柱面近场高分辨定位识别方法,该方法通过在声压基阵中配置单只矢量水听器,将矢量水听器与柱面高分辨近场定位识别方法的优势成功结合,在实现水下噪声源近场高分辨定位识别的同时,利用矢量水听器单边指向性,去除近场定位中的左右舷模糊,同时对算法的舷侧模糊压制能力、声源相对强度估计误差等性能进行对比分析。本文研究了阵列流形失配下的稳健噪声源近场定位识别方法,该方法可以准确获得所需要的对角加载值,抑制阵列流形失配对基于幅度补偿的MVDR近场聚焦波束形成性能的影响,提高定位识别方法的稳健性,对算法的性能进行了详细的对比分析。提出了基于单辅助矢量水听器的基阵误差校正方法,利用精确校正的单只矢量水听器作为辅助基阵,可以在远场或近场对声源空间位置和基阵误差进行无模糊联合估计,只需要参数的一维搜索,运算量小不存在参数联合估计的局部收敛问题,分析了信噪比、声源空间位置等参数对校正结果的影响。本文开展了消声水池试验和实艇海试测量试验,验证了本文算法的有效性,理论分析结果和试验数据处理结果基本一致,为噪声源近场高分辨定位识别方法下一步的实际工程应用提供了试验基础。更多还原

【Abstract】 The acoustic stealthy performance of the submarine is a important index arm for judging its safety and operational performances.Radiated noise is the main target information of various passive sonar and detecting submarines with acoustic weapons,and is the most important factor for damaging stealthy performance.And self-noise is directly affecting its own sonar station working performance.So reducing submarine underwater noise has an important significance.For pertinently carrying out to control noises, as well as providing the actual basis for noise indicators of the acoustic design for submarines,it has a significant meaning to carry out studying the technology of noise source identification separation, find the main noise sources in the system and determine the contribution of various noise sources.With the development of the damp vibration and noise reduction technology, previous methods of the noise source location identification and analysis tools have been unable to effectively receive informations of noise sources,so researching a facilitate engineering implementation method for locating and identifying near-field noise source to apply to the large-size submarine, complex systems,which is the key of the submarine noise control and damp vibration and noise reduction technology.This paper describes the advantage of the near-field acoustic focused technology in the noise source location identification.In the basic of reviewing the theory development of the near field acoustic focusing technology and its engineering application, based on the linear array near-field acoustic focus technology applying on locating and identifying submarine noise sources, the paper progresses a systemic comprehensive theory research and experimental research on high-resolution near-field noise source location identification method whose measurement surface is flat and cylindrical,and the stability near-field noise source location identification method which can inhibit the array sensitivity error and phase error.Firstly based on near-field basic matrix theory, the paper describes the basic principles of the near-field sound focused technology, and gives the calculation formula of uniform linear array focused spatial resolution, analyse the effect of measuring distance, arrays aperture and other parameters on the focused spatial resolution.For the near-field focused spacel aliasing,advance a position fuzzy sentence method to guide designing the base matrix’s form.In this paper, focusing on the lack of the conventional beamforming,MVDR near-field focused beamforming and other location identification methods, significantly researching MVDR, MUSIC near-field focused beamforming methods in the basic of the combination compensation of the amplitude and the phase.This method can implement the relative intensity estimate for noise sources, and in certain conditions it can reduce the signal frequency and restrictions of the measuring distance on the algorithm performance, increase the focused spatial resolution when the basic array works at the low frequency, make it have better capacity to inhibit space aliasing at high frequencies, with lower side lobe level, stronger anti-interference ability, smaller estimation error of sound source intensity, higher positioning accuracy.For the problem that the class algorithm of extremum seeking does not apply to coherent sources near-field location and identification, give a coherent sources near-field high-resolution location identification method based on maximum likelihood estimation,and find a global optimal solution of maximum likelihood estimation with the heredity algorithm.The method has high positioning accuracy, and only with a small aperture arrays it can achieve coherent sources near-field high-resolution location, while not subject to restrictions on the geometry structure of the array.Because near-field sound focused technology is difficult to locate and identify the noise source at the end-fire direction, the paper raise a end-fire direction near-field noise source location identification method based on the virtual rotation transformation technology to analyse that the aperture of the virtual array,the measurement distance, SNR, the observation intervaland and other parameters affect the algorithm performance, and give its specific application conditions.For components of the broadband continuum spectrum of a submarine radiated noise, the paper studies a wideband noise source high-resolution near-field location identification method based on sub-band decomposition and focused transformation technology.The method effectively combines sub-band decomposition and transformation technology with the extremum seeking algorithm by using the amplitude and phase compensation, which enables broadband non-coherent sources and coherent sources to achieve the relative intensity estimation and the spatial location estimation of sound sources,also compare and analyse the the characteristics of thespacial spectrum, the location performance and focused spatial resolution by this method.In this paper, study a noise source near-field high-resolution location identification method for the submarine model’s measurement surface being cylindrical, and particularly analysis the performance of this algorithm.To distinguish the larboard and starboard at the near field line array positioning, propose a cylindrical near-field high-resolution location identification method with combination basic arrays, which successfully combines the advantages of the hydrophone and the cylindrical high-resolution near-field location identification method by configuring a single vector hydrophone in the sound pressure basic array, while achieving the underwater noise source near-field high resolution location identification,use the unilateral directive property of vector hydrophone to remove the confusion of the larboard and starboard in near-field location, at the same time compare and analyse the performances of the hull fuzzy suppression ability and the relative intensity estimation errors of the sound source.This paper research the stability noise source near-field location identification method at the array manifold mismatch. It can accurately obtain the required load values of across corners, inhibit that the array manifold mismatch affects performances of MVDR near-field focused beamforming rate based on the amplitude compensation, improve the solidity property of location identification methods, compare and analyse the performance of the algorithm in detail.Proposed a arrays error correction methods based on a single auxiliary vector hydrophone, use a single vector hydrophone by correcting precisely as an auxiliary basic array,which can progress a no confusion combination estimation in the far field or near-field to the spatial location of sound source arrays and the basic error, only need one-dimensional parameter search, computational load is not existed the local convergence issue at the parameters associated estimation, analyse the effect of the SNR, spatial location of the sound source and other parameters on the calibration results.This paper carries out anechoic pool and the measurement trial for the submarine on the sea, verify the effectiveness of the algorithm, theoretical analysis results and the trial results by treated are nearly the same, which provides the trial basis for further more practical engineering application of the noise source high-resolution near-field location identification method.更多还原