Study On The Identification Method Of Source And The System Development Based On LabVIEW

It is the most effective and convenient way to control noise from source.The premise of noise control is to identify source accurately.Near-field acoustic holography and beamforming are two common methods for source identification.Near field acoustical holography uses the relationship between the holographic surface and the source surface to reconstruct the sound field.This method is suitable for short distance measurement and medium to low frequency source.Beamforming can generate a maximum response value of the desired signal by delay,weighting and summation of the sound signal,so as to obtain the real source distribution.This method is suitable for long distance measurement and high frequency source.With the rise of artificial intelligence,the combination of machine learning algorithm and sound source recognition can not only improve the calculation efficiency,but also improve the positioning accuracy.Then,a method of source identification based on extreme learning machine is proposed.After learning the existing data samples,a model is obtained.Taking the test data as input,the source position is classified according to the judgment of the model.This method has the advantages of high calculation efficiency and high positioning accuracy.Firstly,the derivation process of five source identification methods is studied,and the feasibility of these methods is verified by numerical simulation.Based on the theory of plane near-field acoustic holography,cross spectrum imaging and deconvolution approach for the mapping of acoustic sources,a set of source identification system is developed by using LabVIEW as software platform,microphone,NI acquisition card and computer as hardware.The system includes oscillographic module,calibration module,acquisition module and algorithm processing module.The oscilloscope module has time-domain display,frequencydomain display,octave band display and power spectrum display.The calibration module combines channel calibration and channel setting,which can achieve the calibration and setting of 70 channels.The acquisition module can realize the simultaneous acquisition of 70 channels,as well as the function of triggering and pre triggering.The algorithm processing module includes four kinds of sound source identification methods.And the system has a good scalability,and can add more source identification methods later.Finally,four integrated source identification methods and limit learning machine source identification method are verified.