摘要
波束形成声源识别技术因具有在中高频的分辨率高的特点而被广泛应用于工程实际中。基于平面波和球面波的假设,以扇形轮阵列为例,进行了延时求和算法的单声源、多声源的声源识别性能仿真,结果表明:自主开发的算法程序是正确的,能够准确地识别单个及多个声源。进行了互功率谱及除自谱的互功率谱算法的声源识别性能仿真对比,结果表明:两种算法均可以准确定位声源,且除自谱的互功率谱算法可以有效地抑制旁瓣,提高了声源识别性能。为进一步验证算法程序的准确性,进行了工程试验,得到单一频率的声源成像图,开发了便于查看任意频带下数据结果的用户界面,同时与基于PULSE获得的声源成像结果进行对比,验证了算法程序的准确性。
As the sound source identification technology by beamforming had high resolution in medium-high frequency,so it was widely used in engineering practice.Based on the plane wave assumption and the spherical wave assumption and taken the combo array for example,made the performance simulation of simple sound source,multiple sound sources identification by using delay-and-sum algorithm,the results show that the independent designed algorithm program is correct and can identify the single sound source and multiple sound sources.And then,compared the performance simulation between the cross-spectral and the cross-spec-tral with exclusion of autospectra,the results show that two algorithms can identify sound source accurately while the latter algo-rithm can suppress sidelobe effectively and improve the performance of sound source identification.In order to verify the correct-ness of algorithm program,engineering test was made,then the imaging results of single frequence were acquired,meanwhile an interface for the convenience of viewing data was developed.The correctness was verified by the comparison the imaging results by the cross-spectral beamforming method and the ones acquired by PULSE.
出处
《现代制造工程》
CSCD
北大核心
2014年第11期47-52,共6页
Modern Manufacturing Engineering
关键词
波束形成
噪声源识别
性能仿真
应用
beamforming
noise source identification
performance simulation
application
