基于近场声阵列的旋转机械噪声源识别

摘要

In order to control the noise of the rotating machinery effectively, the signal processing techniques are usually used to identify the noise sources of the whole machine or component parts for fault diagnosis and structural optimization. In this paper, a model of uniform linear near-field acoustic pressure array is constructed to obtain the sound field data. Then, the method of Deconvolution Approach for the Mapping of Acoustic Sources (DAMAS) is employed to extract the information of the sound field from the results of traditional beam-forming analysis to realize the visual reconstruction of the sound source surface. Finally, the rotor vibration test rig based on the near-filed acoustic pressure array measuring is built to obtain the experimental vibration signals under different conditions. The bearings and connection between the turntable and the shaft are found to be the positions of noise sources. Therefore, the feasibility of the proposed method is verified.%为了有效控制旋转机械噪声，利用信号处理技术对整机或部件进行噪声源识别是十分必要的，噪声源准确识别可以为故障诊断和结构优化提供依据。首先论述建立均匀线性近场声阵列模型以获得空间声场数据的方法。其次，在传统波束形成结果基础上，利用反卷积法从中提取所需声场信息以实现对声源面可视化重构。接着，在所搭建转子噪声试验台上，利用近场声阵列提取各种工况下噪声信号，并识别出轴承以及盘轴连接处为转子主要噪声源，验证了基于声源成像反卷积法均匀线性近场声阵列在旋转机械噪声源识别方面的可行性。