文档介绍：山东科技大学硕士学位论文摘要
摘要
机械设备噪声中蕴含着丰富的机器运行状态信息,因此对机械设备进行噪声测量是对机械设备进行噪声控制和故障诊断的有效手段。利用传声器阵列对机械设备的声辐射分布进行声成像测量,可以直观地找到机械设备的主要噪声源位置,从而为噪声控制提供可靠依据。因此,基于传声器阵列的声源定位技术具有重要的研究意义和广阔的应用前景。在声源定位技术中,声源定位算法、阵列优化设计和实时声源定位系统设计是三个最为关键的部分。
本文基于这些问题进行了以下几个方面的研究:
(1) 基于机械噪声特性,给出了阵列模型和信号模型,研究声源定位理论和算法, 包括常规波束形成(CBF)算法、MVDR 算法和 MUSIC 算法。对不同算法进行了仿真分析,比较不同算法的优缺点以及适用性。
(2) 利用粒子群智能优化算法对平面和空间传声器阵列阵型进行优化排布,设计了
一种优化的 64 阵元平面螺旋阵列和一种 64 阵元星形立体传声器阵列,将优化的阵列阵型与其他几种常见的阵列进行了性能仿真比较,仿真结果显示出优化设计的阵型的优越性。
(3) 基于优化设计的阵列设计了阵列声成像系统,实现了声场的可视化和声像图的实时显示,动态地观测被测物体的噪声辐射状况。在系统硬件方面,对电源模块、传声器模块、ADC 采集模块、视频采集模块和处理显示模块进行了设计。在系统软件方面, 采用 Labview 开发了系统软件,可以实现信号采集,波形显示,实时声像图显示等功能。
(4) 在半消声室内对阵列声成像系统进行了性能测试,对阵列的一些性能参数进行了研究,利用阵列声成像系统对各种机械设备进行了噪声源测量,验证了该系统在声源定位应用中的有效性。
关键词:传声器阵列,粒子群智能优化算法,阵列阵型优化,阵列声成像系统,声
源定位实验
ABSTRACT
Mechanical noise contains a lot of running status information of the machine, so the measurement of the mechanical noise is an effective means of noise control and fault diagnosis. microphone array was used to identify the sound sources of the tested machines. Sound source location could be determined and provide reliable clues for noise control and fault diagnosis. So this technology has great research significance and broad application prospects., Beamforming algorithm, array geometric optimization and design of real-time sound source localization system were three key issues of sound source location.
This dissertation studied the three issues mentioned above and the main contents were summerised as follows:
Firstly, based on the mechanical noise, array model and signal model were introduced. Delay and sum beamforming algorithm, MVDR algorithm and MUSIC algorithm were discussed in sound source localization. For some important issues, such as resolution, SNR and sound sources number, simulations were carried out. Advantages and defects of these algorithms were studied and their effectiveness in sound source localization was also learned.
Secondly, Particle swarm optimization algorithm is employed t