Abstract
The acoustic field reconstruction using the statistical optimal nearfield acoustic holography can avoid the winding error, but the finite aperture error cannot be avoided. In order to reduce the finite aperture error, the data extrapolation technique is used before the separation of the sound field. Numerical simulation shows that the method can effectively reduce the reconstruction error.
1. Introduction
The statistical optimal nearfield acoustical holography algorithm can effectively avoid the winding error in the twodimensional Fourier transform, and the precision of the reconstructed sound field is high [15]. At the same time, the acoustic field separation algorithm based on the statistical optimal nearfield acoustic holography has been paid more and more attention [6, 7]. In order to effectively improve the effect of sound field reconstruction, this paper presents a statistical optimal nearfield acoustoacoustic field separation technique based on wavenumber domain extrapolation, and the effectiveness of the method is verified by numerical simulation.
2. Theoretical deduction
In the case of noise source interference, the holographic extrapolation technique is usually the first sound field separation, and then data extrapolation, this paper proposes the first data extrapolation, and then the sound field separation, and finally the use of statistical optimal nearfield acoustic holography Principle of sound source reconstruction.
2.1. Holographic surface data extrapolation
In the actual measurement, taking into account the cost of measurement and budget issues, holographic aperture is generally small, if the direct use of measured data for sound field separation and reconstruction of the sound source surface will bring a larger error, it is necessary to extrapolate the holographic data. With the holographic data extrapolation, we can extend the holographic aperture to the sound source surface more than 4 times, effectively reduce the error. The data extrapolation diagram is shown in Fig. 1.
The data extrapolation process is essentially a data smoothing process. The key to the success of data extrapolation is the design of the filter. The highwavenumber components generated during the data extrapolation are filtered out and fitted to the enlarged aperture.
The specific steps are as follows:
Step 1: In the real space domain, the holographic surface data is zeroed according to the rule of Eq. (1):
Fig. 1The data extrapolation diagram
Step 2: The twodimensional Fourier transform is performed and the results of the twodimensional Fourier transform are lowpass filtered:
where the ${\lambda}^{\alpha}$ is a low pass filter. An improved regularization algorithm is used to determine the parameters of the lowpass filter:
Step 3: The inverse Fourier transform is performed:
Step 4: Holographic surface actual measurement data replacement:
Step 5: The first to fourth steps are repeated until the number of iterations is reached.
2.2. Sound field separation technology
For a steady state singlefrequency sound field, the sound pressures on the holographic surfaces 1 and 2 are ${p}_{1}\left(r\right)$ and ${p}_{2}\left(r\text{'}\right)$, respectively. Holographic sound pressure is generated by the sound source and noise source. Sound source in the holographic surface 1 produces a sound pressure of ${p}_{11}\left(r\right)$, the sound source in holographic surface 2 on the sound pressure is ${p}_{12}\left(r\right)$, according to the sound pressure can be superimposed principle, we can obtain:
Similarly, we can obtain:
According to the principle of the statistical optimal nearfield acoustooptic holography, the weight coefficients of the plane waves and the sound field are independent of each other, the acoustic pressure ${p}_{11}\left(r\right)$ and ${p}_{12}\left(r\right)$ generated by the holographic surface can be solved. At the same time, the resulting sound pressure is ${p}_{21}\left(r\text{'}\right)$ and ${p}_{22}\left(r\text{'}\right)$, and the sound field is separated.
2.3. Statistical optimal nearfield acoustic holography
The princeple of the SONAH: in the free sound field, the wave field consists of plane waves, plane waves can be countless units of plane waves in a linear combination of said. which is:
According to the principle of wave superposition:
Eq. (8) establishes the relationship between the holographic sound pressure and sound source surface sound pressure. Eq. (7) is written in the form of a matrix:
According to the matrix theory combined with the regularization of the relevant knowledge can be solved:
where $\theta $ is a regularization parameter, which is determined according to the method of Document 3.
Fig. 2The diagram of sound field separation
3. Numerical simulation
The simulation uses the point source to simulate, and the spatial position relation between the point source, the noise source and the holographic surface is shown in Fig. 2. Sound source, noise sources are located on both sides of the holographic surface. The specific parameters are as follows: the source and the noise source distance 0.5 m, respectively, the radius of the sound source and interface sound source are 0.01 m, and their frequency are 1000 Hz, surface vibration velocity $v=$ 0.5 m/s, holographic and reconstruction surface size of 1 m by 1 m. And the distance between the reconstruction surface and the sound source surface is 0.2 m and 0.1 m. The two holographic planes use 32×32 array elements. Fig. 36 shows the reconstructed sound field using the method proposed in this paper. Fig. 710 shows the theoretical sound pressure of the reconstruction surface.
Reconstruction error is defined as:
The error versus frequency curve is shown in Fig. 11.
Fig. 3Reconstructed surface acoustic pressure (a)
Fig. 4Reconstructed surface acoustic pressure (b)
Fig. 5The clouds of the reconstructed Surface Sound Pressure (a)
Fig. 6The clouds of the reconstructed Surface Sound Pressure (b)
Fig. 7Reconstruction surface theory sound pressure (a)
Fig. 8Reconstruction surface theory sound pressure (b)
From the error curve, it can be seen that the error of reconstructed sound field is increasing with the vibration frequency of the sound source increasing. When the vibration frequency reaches 3000 Hz, the error has reached 50 %, which indicates the reconstruction failure. However, reconstruction at 3000 Hz is effective. The error varies with the vibration frequency, because the higher the vibration frequency is, the higher the wave component value is, and the more effective wave number filtered by the lowpass filter, the lower the passthrough point of the lowpass filter can reduce the error. The adaptive lowpass filter algorithm is the focus of the next step.
Fig. 9Reconstruction surface theory sound pressure cloud (a)
Fig. 10Reconstruction surface theory sound pressure cloud (a)
Fig. 11Error versus frequency curve
4. Conclusions
First, the holographic data is extrapolated, then the sound field on both sides of the holographic surface is separated. Finally, the sound field is reconstructed by using the principle of statistical optimal acoustic holography. The theoretical derivation shows the correctness of the method and the numerical simulation verifies the effectiveness of the method.
References

Du Xianghua, Zhu Haichao, Mao Rongfu, et al. Patch nearfield acoustic holography based on support vector regression. Acta Acoustic, Vol. 37, Issue 3, 2012, p. 286293.

Hald J. Patch nearfield acoustical holography using a new statistically optimal method. Proceedings of Internoise, 2003, p. 22032210.

Li Weibing, Chen Jian, Yu Fei, et al. Study on statistically optimal cylindrical nearfield acoustical holography. Chinese Journal of Mechanical Engineering, Vol. 41, Issue 4, 2005, p. 123127.

Finn J., Virginie J. Statistically optimized near field acoustic holography using an array of pressurevelocity probes. Journal of Acoustical Society of America, Vol. 121, Issue 3, 2006, p. 15501558.

Finn J., Chen Xinyi A comparison of statistically optimized near field acoustic holography using single layer pressurevelocity measurements and using double layer pressure measurements. Journal of Acoustical Society of America, Vol. 123, Issue 4, 2008, p. 18421845.

Hald J. Basic theory and properties of statistically optimized nearfield acoustical holography. Journal of Acoustical Society of America, Vol. 125, Issue 4, 2009, p. 21052120.

Jia W. Q., Chen J., Yang C., et al. Study of a sound field separation technique based on a single holographic surface and wave superposition method. Journal of Mechanical Engineering Science Part C, Vol. 223, 2009, p. 18271835.