【摘要】 为了从麦克风阵列测量中得到更清晰的声成像图,在阵列波束成形数据处理上运用反卷积技术抑制旁瓣越来越普及。传统的反卷积技术假设声源图是由阵列点扩散函数建立起来的,但在航空领域内实际被测声源波束图通常与合成得到的点扩散函数不一样,使得该技术在应用中受到较大限制。本文介绍另一种基于空间声源相干的波束成形反卷积技术,用迭代方法逐步地将声源图中与峰值声源空间相干的部分去掉,从而将旁瓣从实际被测的波束图中移除。基于该技术原理,在波束成形技术基础上发展了用于阵列数据处理的CLEAN-SC算法,应用于风洞声学测量中获得了开/闭口风洞中的起落架和翼型的声源分布特性。实验对比分析表明该阵列数据处理优化技术在空间声源定位和抑制旁瓣能力上比传统波束成形技术都有显著提高。更多还原

【Abstract】 Deconvolution technique has been widely used to improve spatial resolution of noise source plots from microphone array measurements.However,most deconvolution methods which are based on the assumption that source plots are built up by point spread functions fall short in real applications since the actual beam patterns of noise sources are often not identical to the synthetically obtained PSF′s.To overcome this problem,an improved deconvolution technique based on Spatial Source Coherence,briefly called"CLEAN-SC"algorithm,has been developed and applied to identify and localize noise sources in both closed-section and open wind tunnel aeroacoustic measurements.The CLEAN-SC algorithm is firstly verified by using apoint source of a loud speaker located on the centre-line of the microphone array,and then applied to measure the noise sources of a 1/4scale landing gear model in a closed-section wind tunnel and a NACA23018 airfoil model in an open wind tunnel.It is shown that with the CLEAN-SC algorithm the array resolution can be improved by about 27% and the sidelobe level can be significantly reduced when compared with both the conventional beamforming and CLEAN-PSF technique.The noise sources of the landing gear structures and airfoil can also be identified distinctly.更多还原