Research On Improving Accuracy Of Near-field Acoustic Holography Reconstruction

Near-field acoustic holography is a very effective noise source identification,positioning and sound field visualization technology developed in recent years.The technology has a wide range of applications in many fields such as automobiles,aviation,machinery,submarines,ships,weapons,railways,and electronics.Application background.However,there are still problems in the application of the technology in practical engineering,which hinders the promotion and application of near-field acoustic holography technology.The improvement of reconstruction accuracy is a problem to be studied.Simulation analysis shows that the improved algorithm of this paper has a greater improvement than the original algorithm.Firstly,in the space Fourier transform method,in the process of near-field acoustic holography reconstruction,since the inverse transfer operator grows exponentially in the high-wave-number region,the reconstruction accuracy must be improved through the low-pass filter.The selection of the cutoff wave-number in the low-pass filter is the key.By assuming that the sound pressure on the reconstruction surface is not submerged by noise,the selection formulae for the SNR,the distance between the holographic surface and the reconstruction surface,and the sound wave frequency are obtained.The improved algorithm of this paper has a larger improvement than the original algorithm when the signal-to-noise ratio is small,the distance between the holographic surface and the reconstruction surface is relatively large,and the sound wave frequency is small.Secondly,in the statistical optimization method,the selection of wave number vectors is generally regular.This paper studies the method of selecting irregular wave number vectors.First,the wave number vectors are divided at equal intervals,and then the wave number vectors are performed in an exponential manner.After unequally spaced partitions,the propagating wave components within the equally spaced radiating circles are combined with the evanescent wave components other than the radiating circles divided at unequally spaced intervals.Finally,the regularized parameters are obtained through the L-curve method.Sound field reconstruction.Simulation analysis shows that the selection of a more regular wave number vector based on the irregular wave number vector has a higher reconstruction accuracy.Thirdly,in the wave superposition method,when the equivalent source is disposed onthe sound source surface,the transfer matrix will be diagonally dominant and the symmetry satisfaction degree will be as low as possible.At this time,the sound source intensity obtained by the traditional regularization method is used.Unable to correctly match the real sound field,resulting in reduced reconstruction accuracy.Through in-depth study found that you can use compressive sensing methods to solve the sound source intensity.First,the problem of solving the sound source intensity is converted into a combinatorial optimization problem to solve the sound source intensity,and then The combinatorial optimization problem is transformed into a convex optimization problem that solves the source intensity,and the sound source intensity is solved by the convex optimization toolbox to reconstruct the sound field.The simulation analysis shows that the improved algorithm in this paper has a greater improvement in the reconstruction accuracy than the regularization method for solving the sound field information.In summary,the improved algorithm in this paper has a greater improvement in reconstruction accuracy than the original algorithm,so that the improved algorithm can be used to solve the problems in engineering practice.