Research On Sound Field Reconstruction Based On Statistically Optimal Nearfield Acoustical Holography

Noise source identification and sound field visualization are the important steps in noise analysis and diagnosis.As a high-resolution acoustic imaging technique,nearfield acoustic holography(NAH)has an important application in this field.By measuring in the near field of the sound sources,the evanescent wave component can be obtained,thus the result of reconstruction is not restricted by Rayleigh criterion and reconstruction resolution is greatly improved.As an important branch of NAH,statistically optimized nearfield acoustical holography(SONAH)has its unique advantages in the reconstruction of large-size sound sources.In this thesis,the development of NAH technology and its research status quo are briefly introduced,and some shortcomings are pointed out.This thesis mainly focuses on some problems existing in the reconstruction of the sound field by using the SONAH method.Improvement measures are put forward to overcome these shortcomings.The conventional statistically optimized cylindrical nearfield acoustical holography(SOCNAH)and statistically optimized spherical nearfield acoustical holography(SOSNAH)methods are usually required to be conformal with the shape of sound sources.The scheme based on planar measurement has been developed to overcome this limitation.Numerical simulations are used to demonstrate the validity and applicability of reconstruction by using this measurement scheme.In the reconstruction of the sound fields generated by multiple spatial sound sources,the conventional SONAH methods need a large number of higher-order elementary wave functions to represent the sound field,and the results of reconstruction are generally unsatisfactory.To address this problem,an improved SONAH method based on planar measurement is proposed.In conventional methods,a single wave function type is often used to reconstruct the sound fields,while in the improved SONAH method set of elementary wave functions(such as elementary planar wave functions,elementary cylindrical wave functions and elementary spherical wave functions)which were selected depending on the features of sound sources to calculate the sound fields transition matrix,is used instead.The validity and applicability of this improved SONAH method are verified by numerical simulations,and are compared with the conventional method in detail.In general,the results show that the proposed method achieves more accurate results than the conventional SONAH methods,and the relative errors have a small disturbance with the frequency of sound sources changing,which suggests that the improved method has a strong robustness to frequency domain.Moreover,with the increase of frequency,the relative errors decrease gradually.The selection of the sets of elementary wave functions has a great influence on the accuracy of a reconstruction.It is found that the improved method has an accurate reconstruction when the sets of elementary wave functions are conformal with the sound source shapes and equal to the number of sound sources.Finally,the impact of reconstruction distance on reconstruction results was particularly analyzed.For non-free sound field reconstruction,a sound field separation technique based on SONAH with double-layer measurement is introduced in detail.The sound field on both sides of the holographic measurement surface can be reconstructed by using this sound field separation technique.Through these studies,a relatively complete SONAH theory is preliminarily established.