Study On Noise Source Identification And Sound Field Visualization Based On Beamforming Techniques

Beamforming is an advanced technique for accurate identifying noise sources and visualizing spatial acoustic field. The technique is particularly suitable for analyzing sound sources location from medium to long measurement distances and large high-frequency sources. What's more, beamforming supports three different calculation algorithms, which are stationary, quasi-stationary and non-stationary sound sources. With a fairly low number of microphones, beamforming provide the finest possible resolution. Beamforming can estimate the distribution on source surface, find the sound source location and get the field's dominating feature through make use of spatial acoustic field signals radiating from the target object. The main contents in this dissertation are summarized as follows:In chapter one, the history and current status of beamforming has been reviewed, and the significance has also been discussed. At the same time, this article points out that beamforming has a prospecting future in the field of locating noise sources. Finally the key problems existed in beamforming and main research contents have been determined in the dissertation.In chapter two, the Delay-and-Sum beamforming has been implemented for plane and spherical wave incident. An innovative and optimised 3D array geometry is designed on basis of traditional planar array. The novel 3D array is easy to realize and has a stable and reliable performance. Without increasing the number of microphones, the designed 3D array can suppress ghost images and well in the case of plane wave incident. Meanwhile, it also improves the resolution with near-field sound sources.In chapter three, in order to overcome shortcomings of low resolution in Delay-and- Sum beamforming, the MVDR and MUSIC algorithms of spatial spectrum estimatiom are introduced in area of noise-source location. As a result, the resolution accuracy for spatial source signals is improved greatly. Besides, for the problem of unreal source causing by spatial aliasing, an effective method has been proposed according to unreal sources'unstable property, thus the real direction of far-field sources can be identified accurately.In chapter four, a novel adaptive weighting scheme is proposed and a new adaptive near-field beamforming technique is implemented simultaneously. Through choosing feasible weighting, the proposed technique can enhance signal radiating from the target source and decrease the influence of disturbing signals. With these measures taken, this article solves the problems of low resolution accuracy in conventional near-field beamforming and realizes accurately prediction of the sound field.In chapter five, the pressure-pressure-based, pressure-velosity-based and sound intencity-based spatial sound field sepration technique has been proposed, combined with beamforming which can be used in excluding the impact of background disturbing noise, realizing separation and accurate prediction of the target source, widening the applicable scope of beamforming technique. Numerical simulation and experiments for two actual loudspeaker sources verify the validity and correctness of the proposed methods.In chapter six, main researches in this dissertation have been summarized, and the topics need further study have been pointed out as well.