基于相位共轭方法的多目标噪声源在复杂声场中分离和识别研究

The identification of multi - target sound source in complex sound field was a hot issue in the field of noise control in recent years. The higher resolution of sound source identification would be achieved by introducing the acoustic surface plasmon resonance body for analysis the evanescent wave to make the phase conjugation method could be applied to the more realistic far field measurement to break the diffraction limit. Combined the advantages between the phase conjugation methods and blind source separation technique for separation the sound field would obtain the accurate results for identification and location the underwater multiple sound sources. For the complex condition of multi-target noise sources condition, using a closed enveloping sound source instead of combination of the actual multiple sound sources, the source identification could be implemented based on the phase conjugation method for the overall enveloping sound source, and then the blind source separation technique could improve the target sound source reconstruction to accurately reconstruct the sound pressure amplitude of each noise source. The project provide a new method for the multiple sound source identification, the sound field reconstruction and the acoustic active control which combined the phase conjugation method and the blind separation technology to achieve the purpose of reducing the noise level. The experimental research on the multi-target sound source on the basis of the above theory would be developed and the target random measurement array product would be designed and optimized for identification and location of the sound source applied for the sound source reconstruction ultimately. This project was aims to find the practical and effective noise source identification algorithm in complex sound field, and to provide a new method for noise control.

多目标噪声源在复杂声场中的识别问题是近年来噪声控制领域的热点问题。本项目拟将表面等离子共振理论引入声学领域，建立声学表面等离子共振体对倏逝波的声学参数进行量化分析，在远场测量条件下突破声波衍射极限，得到更高的识别分辨率以满足工程应用。在此研究基础上，结合相位共轭方法和盲源分离技术进行多目标噪声源的声场分离，具体研究内容包括：采用一个封闭的包络面包络所有的声源，基于相位共轭方法对该包络面的辐射声场进行识别以确定整体的噪声源位置，然后按照盲源分离技术的求解方法对已经确定位置的包络面噪声源进行逐个分离，准确识别各个噪声源声压幅值。将相位共轭方法与盲分离技术结合，可为多目标噪声源识别、声场重构乃至声学主动控制提供一种新的方法。此外，本项目还将对多目标噪声源识别进行实验研究，采用目标随机阵列进行复杂声场环境下的声源识别。本项目旨在寻找实用有效的噪声源识别算法，为减振降噪提供新的手段和技术支持。

针对水下噪声源识别和定位的军事需求，以降低舰船的辐射噪声为考核验证对象，开展了基于相位共轭方法实现噪声源定位和辐射特性分析研究，发展一种新的将实验和计算相结合的声场分析技术，建立基于相位共轭方法的噪声源识别与定位方法和基于相位共轭方法的实验和计算相结合的辐射特性分析方法并进行多目标噪声源在复杂声场中的识别问题。针对不同类型的平面相位共轭阵聚焦效果不同的问题，通过理论分析和数值仿真讨论了圆周阵、十字阵和矩形等目标阵列阵聚焦自由场点声源的特点，并进行实验验证。结果表明：测量振速采用偶极子阵发射的圆周阵聚焦效果最好，扩大圆周阵和十字阵的阵列尺寸能够提高其聚焦的空间分辨率。针对多声源声场成像与分离问题，在对不同强度的两个点声源声场和平板声源声场识别的研究基础上，选取目标场点并分别计算了各个声源的贡献量系数以实现声源分离技术。结果表明：通过对目标声源进行声学贡献量分析，可以判断出各个声源对目标场点的贡献权重，通过对目标场点贡献量为正的声源或平板区域进行控制就可以减小对目标场点的辐射噪声值。以提高声源识别分辨率为目标，分别开展基于迭代相位共轭方法的声源识别研究和基于声学超透镜的远场声源定位研究：相位共轭方法进行声源聚焦时，声场聚焦后会继续发散，导致声源识别分辨率不高。通过聚焦点的声场叠加，对聚焦后发散的声场进行时间反转计算与原相位共轭声场在聚焦点处叠加，通过数值模拟和实验验证了迭代相位共轭方法的有效性。声学超透镜能够对声波的倏逝波成分进行无损传播甚至放大传播，实现声源的完美成像，完成声源重构。