Noise source identification is essential for the reduction of engine noises. The medium and high frequency noise sources can be significantly identified by the Beamforming method, but the low frequency noise source identification results are not satisfactory. The statistically optimized near-field acoustical holography (SONAH) method is suitable for low frequency noise source identification;however, the error is large for high frequency noise source identification. In order to identify the noise source of a diesel engine accurately and provide a clear direction for further low noise improvement, the most prominent inlet side noise source of the diesel engine was identified, combining Beamforming and the SONAH sound source identification methods. In the experiment, the distance between the measurement array and the engine surface was 1 m with the Beamforming method, while the SONAH measurement distance was 0.25 m. The sound power spectrum and the sound intensity contour maps were analyzed. The sound intensity contour maps produced by the Beamforming method showed that the acoustical center of 920-1 450 Hz frequency band appeared in the fuel injector position, indicating that the injector was the main noise source in this frequency band. And in 1 650-2 200 Hz frequency band, the acoustical center mainly focused on the top of the intake manifold, and the sound intensity contour lines attenuated slowly upward. Therefore, it can be concluded that part of the noise comes from the intake manifold, and that the others come from the cylinder head cover. The sound intensity contour maps produced by the SONAH method showed that the first acoustical center of the 760-776 Hz frequency band appeared in the lower right of the white casing, and that the second one appeared in the oil pump governor, while for the 920-936 Hz frequency, it mainly appeared in the oil pump governor. Analysis showed that the white casing was used to cut off the oil pump shaft noise so that the noise leaked from the lower right corner gaps due to bad sealing performance. So the oil pump shaft was the true noise source in the 760-776 Hz frequency band, and the oil pump governor was the main noise source in the 920-936 Hz frequency band. Further acoustical contribution analysis results showed that the sound power contribution of the intake manifold to the inlet side noise source was 15.38%, the sound power contribution of the fuel injector and the oil pump governor were 5.47%and 5.11%, respectively, and the cylinder head cover and the oil pump drive shaft corresponded to 4.85%and 4.26%, respectively. In conclusion, a noise source within a wide band can be identified with high precision by combining the advantages of Beamforming and SONAH, and the test is simple and easy to use.%为了准确识别某柴油机的噪声源，为进一步的低噪声改进指明方向，综合采用Beamforming（波束形成）和SONAH（统计最优近场声全息）2种阵列的声源识别方法对该柴油机噪声辐射最突出的进气侧噪声源进行识别。结果表明：噪声贡献量较大的1650～2200 Hz频率范围内对应的噪声源为进气总管、汽缸盖罩，920～1450 Hz对应喷油器，而760～776 Hz和920～936 Hz分别对应油泵传动轴和油泵调速器。进一步的声贡献量分析结果显示：进气总管对该发动机进气侧辐射噪声的声功率贡献量达15.38%，喷油器、油泵调速器声功率贡献度量分别为5.47%和5.11%；汽缸盖罩和油泵传动轴声功率贡献量分别为4.85%和4.26%。综上，结合Beamforming和SONAH在不同频段内具有高分辨率的优点，可以在宽频带内进行声源识别，且试验实现简单，操作方便。
展开▼
