| The noise level of machinery can reflect the reliability and stability of equipment.Mechanical products with low vibration and low noise have strong competitiveness in the market.To reduce the noise level,accurate localization and identification of the source is necessary.In canonical localization method,the space resolution is proportional to the frequency of noise and this will lead to aliasing of close resources in low frequency.With the development of complexity and integration in the field of mechanical design,proposing a high-resolution localization method is important to analyze the propagation of noise.This research will promote the reduction of noise in facilities.This paper mainly researches the location and identification method of noise source in device with microphone array.The main points are as follows:(1)Researching source localization method based on microphone array and analyzing the influence of formation and localization algorithms on the accuracy and resolution with simulation.(2)Using virtual array method to increase the equivalent physical aperture of microphone array without changing formation,which is based on the relationship between array element distribution and noise wavelength.Combining virtual array algorithm and MUSIC algorithm to propose a VA-MUSIC(Virtual Array MUSIC)method.(3)Developing a 64-channel microphone array noise source localization imaging system based on VA-MUSIC method and collecting data in a semi-anechoic laboratory.Comparing MUSIC and VA-MUSIC method and proving that VA-MUSIC method has higher localization resolution and better spatial resolution to low-frequency noise sources. |