Underwater structure low-frequency radiation sound power forecasting method based on blind source separation

Abstract

The invention discloses an underwater structure low-frequency radiation sound power forecasting method based on blind source separation, and the method comprises the following steps: (1), setting a housing sound source in a shallow sea environment, and obtaining discrete sound pressure; (2) obtaining an estimation signal of the source signal through a blind source separation algorithm; (3) windowing the estimated signal space data of the source signal in the z direction to obtain the sound pressure propagating outwards after windowing; (4) the sound pressure propagating outwards is converted into a wave number domain through discrete space Fourier transform, and a holographic surface sound pressure cylindrical wave spectrum is obtained; (5) filtering evanescent waves in a wavenumber domain by using a wavenumber domain filter to obtain propagation waves on a holographic surface; (6) transmitting the filtered sound pressure to a far field in a spherical shape through a near-field acoustical holography technology by means of stable-phase estimation; (7) deducing the sound intensity by using the far-field sound pressure so as to obtain the radiation sound power; according to the invention, the measurement workload is reduced; measurement is carried out on a holographic surface close to a sound source, the influence of boundary reflection is small, and a calculation result is accurate.

Description

Technical Field

[0001] The present invention relates to the field of underwater acoustic technology, and particularly to a method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation. Background Art

[0002] Radiation sound power is a parameter objectively measuring the sound radiation characteristics of a structure. To ensure the accuracy, consistency, and comparability of the sound power, the test should be carried out in a free-field environment. Given the vast continental shelf and rich shallow sea resources in China, where the shallow sea area with a water depth of 0 - 15 meters accounts for 2.6% of the total offshore area, in the sound radiation prediction of an underwater vehicle, the shallow sea environment is often used to simulate the free-field environment. However, the influence of external interference sources in the low-frequency shallow sea environment is usually unpredictable. Therefore, how to accurately predict the low-frequency radiation sound power of the shell structure in the shallow sea environment has practical engineering significance.

[0003] In recent years, near-field acoustic holography (NAH) has become one of the key technologies for predicting the vibration and sound radiation of underwater structures, and the radiation sound power of a sound source can be predicted through the acoustic quantities near the sound source surface. NAH requires that all sound sources must be on the same side of the holographic plane, and it is difficult to meet its requirements for the sound field environment in actual engineering applications. In response to such problems, the sound field separation technology has emerged. However, for an underwater cylindrical sound source, the sound field separation is generally carried out using the complex sound pressure on both holographic planes or the complex sound pressure and vibration velocity on a single holographic plane, which greatly increases the measurement workload. Summary of the Invention

[0004] Object of the Invention: The object of the present invention is to provide a method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation, which can separate the interference part and predict the radiation sound power only by measuring the sound pressure on a single cylindrical holographic plane, reducing the workload, having a fast calculation speed, and more accurate results.

[0005] Technical Solution: A method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation according to the present invention includes the following steps:

[0006] (1) Set a shell sound source in the shallow sea environment, and obtain discrete sound pressure through the shell sound source;

[0007] (2) Obtain the separated signal, i.e., the estimated signal of the source signal, through a blind source separation algorithm;

[0008] (3) Window the spatial data of the estimated signal of the source signal in the Z direction to obtain the windowed sound pressure propagating outward;

[0009] (4) Convert the windowed outward-propagating sound pressure to the wavenumber domain through discrete spatial Fourier transform to obtain the holographic surface sound pressure cylindrical wave spectrum;

[0010] (5) Filter the evanescent wave in the wavenumber domain using a wavenumber domain filter to obtain the propagating wave on the holographic surface;

[0011] (6) After transforming the cylindrical coordinate system to the spherical coordinate system using the transformation relationship, transfer the filtered sound pressure to the far field in a spherical shape through stationary phase estimation by near-field acoustic holography technology;

[0012] (7) Derive the sound intensity from the far-field sound pressure, thereby obtaining the radiated sound power.

[0013] Further, the specific steps of step (1) are as follows: Set a cylindrical holographic surface with a radius of r S at a distance d H from the sound source. There are interference sources outside the holographic surface; measure the complex sound pressure at discrete points on the holographic surface to obtain the sound pressure p(r H , m△θ, n△z) of (M + 1)×(N + 1) discrete points on the cylindrical holographic surface with a radius of r H ; where d S is required to satisfy: kd S <1, k = 2πf / c, f is the frequency, and c is the speed of sound; the axial interval on the cylindrical holographic surface is △z = 2L / (M + 1), and the circumferential interval angle is △θ = 2π / (N + 1); where M and N are positive integers, M is 1 less than the number of field points on the circumference and is an even number, and N is 1 less than the number of field points in the z direction; m is an integer from 0 to M, and n is an integer from 0 to N. △z is the field point interval in the z direction, and △θ is the field point interval of the circumference; Shallow sea environment: The sea surface is set as an absolutely soft boundary, and the seabed is set as an absolutely hard boundary.

[0014] Further, the specific steps of step (2) are as follows: First, express the discrete sound pressure p(r H , m△θ, n△z) received on the holographic surface as p(r H , m△θ, n△z) = Ap1(r H , m△θ, n△z); where p1(r H , m△θ, n△z) is the outward sound pressure emitted by the sound source, that is, the source signal;

[0015] Then, obtain the estimated signal of the source signal through the blind source separation algorithm. The formula is as follows:

[0016] p I (r H , m△θ, n△z) = Bp(r H , m△θ, n△z) (1).

[0017] Further, in the step (3), the window function is:

[0018]

[0019] where z c is the cut-off distance, set to 80% of the scanning half-length; α is the steepness coefficient, usually taking values from 0.01 to 0.2; then the sound pressure propagating outwards after windowing is:

[0020] p II (r H , m△θ, n△z) = p I (r H , m△θ, n△z)∏(z / 2z c ) (3).

[0021] Further, in the step (4), the formula is as follows:

[0022]

[0023] where M is 1 less than the number of field points on the circumference and is an even number, N is 1 less than the number of field points in the z direction, m is an integer from 0 to M, n is an integer from 0 to N, △z is the field point interval in the z direction, △θ is the field point interval on the circumference, the discrete circumferential component k m and the discrete axial component k zn are defined as:

[0024]

[0025]

[0026] Further, in the step (5), the wavenumber domain filter is:

[0027]

[0028] where k t is the wavenumber vector, defined as α is the steepness coefficient, set to 0.01 - 0.2; k c is the cut-off wavenumber of the filter, determined by the wavenumber in the surrounding medium, then the holographic surface propagating wave sound pressure cylindrical wave spectrum formula is as follows:

[0029]

[0030] Further, in the step (6), the cylindrical coordinate system is transformed into the spherical coordinate system using the transformation relationship as follows:

[0031] The transformation formula for converting the cylindrical coordinate system to the spherical coordinate system is: Among them, the cylindrical coordinate system and the spherical coordinate system share a θ. When transmitted to the far field in spherical form, θ and remain unchanged, and R becomes r F ;

[0032] Under the stationary phase estimation formula:

[0033]

[0034] Among them, r F is set to 10 times the wavelength of the lowest frequency; when the frequency is very low, only the radiation of the mode with n = 0 can reach the far field.

[0035] Furthermore, the formula of the step (7) is as follows:

[0036]

[0037]

[0038] SWL = 10log 10 W(ω) / W0(ω) (10)

[0039] Among them, c is the sound speed in the water medium, ρ0 is the density of the water medium, is the sound intensity on the far-field spherical prediction surface with a radius of r F W0(ω) is the reference value of the radiated sound power, usually set to 0.67 * 10 -18 w, W(ω) is the radiated sound power, with the unit of W, and SWL is the reference level of the radiated sound power, with the unit of dB.

[0040] Beneficial effects: Compared with the prior art, the present invention has the following remarkable advantages: the sound field can be separated on a single holographic measurement surface, and only the complex sound pressure on the holographic measurement surface is required; the measurement workload is reduced; the measurement is carried out on a holographic surface very close to the sound source, and the influence of boundary reflection is very small, and the calculation result is accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is a schematic diagram of the shallow sea sound field environment of the present invention;

[0042] Figure 2 is a schematic diagram of the discretization of the holographic surface of the present invention;

[0043] Figure 3 is the basic framework of the blind source separation algorithm of the present invention;

[0044] Figure 4 is the wavenumber domain filter of the present invention;

[0045] Figure 5 is a schematic diagram of the coordinate system conversion of the present invention;

[0046] Figure 6 This is the simulation model of the present invention;

[0047] Figure 7 This is the cloud map of the sound pressure level of the sound field of 210 Hz of the present invention;

[0048] Figure 8 This is the cloud map of the sound pressure level of the sound field of 330 Hz of the present invention;

[0049] Figure 9 This is the cloud map of the sound pressure level of the sound field of 940 Hz of the present invention;

[0050] Figure 10 This is the cloud map of the sound pressure level of the sound field of 130 Hz of the present invention;

[0051] Figure 11 This is the cloud map of the sound pressure level of the sound field of 210 Hz of the interference source of the present invention;

[0052] Figure 12 This is the cloud map of the sound pressure level of the sound field of 330 Hz of the interference source of the present invention;

[0053] Figure 13 This is the cloud map of the sound pressure level of the sound field of 940 Hz of the interference source of the present invention;

[0054] Figure 14 This is the blind source separation effect of 100 Hz of the present invention;

[0055] Figure 15 This is the blind source separation effect of 150 Hz of the present invention;

[0056] Figure 16 This is the blind source separation effect of 610 Hz of the present invention;

[0057] Figure 17 This is the comparison of the calculation results under the interference environment of the present invention. Specific embodiments

[0058] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0059] As Figure 1 shown, an underwater structure low-frequency radiation sound power prediction method based on blind source separation provided by an embodiment of the present invention includes the following steps:

[0060] (1) Set a shell sound source in a shallow sea environment, and obtain discrete sound pressure through the shell sound source; specifically as follows: at a distance d S from the sound source, set a cylindrical holographic surface with a radius of r H , and there is an interference source outside the holographic surface; measure the complex sound pressure at discrete points on the holographic surface, and obtain the discrete sound pressure p(r H of (M + 1) × (N + 1) points on the cylindrical holographic surface with a radius of rH , mΔθ, nΔz); As Figure 2 shown, where d S is required to be: kd S where k = 2πf / c, f is the frequency, c is the speed of sound; the axial interval on the cylindrical holographic surface is Δz = 2L / (M + 1), and the circumferential interval angle is Δθ = 2π / (N + 1); where M and N are positive integers, M is 1 less than the number of field points on the circumference and is an even number, N is 1 less than the number of field points in the z direction; m is an integer from 0 to M, n is an integer from 0 to N, Δz is the field point interval in the z direction, and Δθ is the field point interval of the circumference; the shallow sea environment is: the sea surface is set as an absolutely soft boundary, and the seabed is set as an absolutely hard boundary.

[0061] (2) Obtain the separated signal, that is, the estimated signal of the source signal, through the blind source separation algorithm; As Figure 3 shown, specifically as follows: First, represent the discrete sound pressure p(r H , mΔθ, nΔz) received on the holographic surface as p(r H , mΔθ, nΔz) = Ap1(r H , mΔθ, nΔz); where p1(r H , mΔθ, nΔz) is the outward sound pressure emitted by the sound source, that is, the source signal.

[0062] Then, obtain the estimated signal of the source signal through the blind source separation algorithm, and the formula is as follows:

[0063] p I (r H , mΔθ, nΔz) = Bp(r H , mΔθ, nΔz) (1)

[0064] (3) Window the spatial data of the estimated signal of the source signal in the Z direction to obtain the windowed outward-propagating sound pressure; The window function is:

[0065]

[0066] where z c is the cut-off distance, set to 80% of the scanning half-length; α is the steepness coefficient, usually taking values from 0.01 to 0.2; then the windowed outward-propagating sound pressure is:

[0067] p II (r H , mΔθ, nΔz) = p I (r H , mΔθ, nΔz) ∏(z / 2z c ) (3)

[0068] (4) The windowed outward-propagating sound pressure is transformed into the wavenumber domain by discrete spatial Fourier transform to obtain the holographic surface sound pressure cylindrical wave spectrum. The formula is as follows:

[0069]

[0070] where M is 1 less than the number of field points on the circumference and is even, N is 1 less than the number of field points in the z direction, m is an integer from 0 to M, n is an integer from 0 to N, △z is the field point interval in the z direction, △θ is the field point interval on the circumference, the discrete circumferential component k m and the discrete axial component k zn are defined as:

[0071]

[0072]

[0073] (5) The evanescent wave is filtered in the wavenumber domain using a wavenumber domain filter to obtain the propagating wave on the holographic surface. As Figure 4 shown, the wavenumber domain filter is:

[0074]

[0075] where k t is the wavenumber vector, defined as α is the steepness coefficient, set to 0.01 - 0.2; k c is the cut-off wavenumber of the filter, determined by the wavenumber in the surrounding medium. Then, the formula for the holographic surface propagating wave sound pressure cylindrical wave spectrum is as follows:

[0076] P n I (r H ,k zn ,ω) = P n (r H ,k zn ,ω)Π(n,k z ) (8)

[0077] (6) After transforming the cylindrical coordinate system to the spherical coordinate system using the transformation relationship, the filtered sound pressure is transferred to the far field in a spherical shape through stationary phase estimation by near-field acoustic holography technology. As Figure 5 shown, the specific transformation of the cylindrical coordinate system to the spherical coordinate system using the transformation relationship is as follows:

[0078] The transformation formula for converting the cylindrical coordinate system to the spherical coordinate system is: where the cylindrical coordinate system and the spherical coordinate system share a θ. When transferring to the far field in a spherical form, θ and remain unchanged, and R becomes r F ;

[0079] Under the steady-phase estimation formula:

[0080]

[0081] where r F is set to 10 times the wavelength of the lowest frequency; when the frequency is very low, only the radiation of the mode with n = 0 can reach the far field.

[0082] (7) Derive the sound intensity from the far-field sound pressure, and thus obtain the radiated sound power. The formula is as follows:

[0083]

[0084]

[0085] SWL = 10log 10 W(ω) / W0(ω) (10)

[0086] where c is the sound speed in the water medium, ρ0 is the density of the water medium, is the sound intensity on the far-field spherical prediction surface with a radius of r F , W0(ω) is the reference value of the radiated sound power, usually set to 0.67*10 -18 w, W(ω) is the radiated sound power, with the unit of W, and SWL is the reference level of the radiated sound power, with the unit of dB. Specific embodiments:

[0088] Establish a simulation model as shown in Figure 6 in the ACTRAN finite element software. The sea surface is set as an absolutely soft boundary, and the seabed is set as an absolutely hard boundary. The sea depth is 10m, the radius of the cylindrical shell is 0.4m, the axial length is 1m, the shell thickness is 0.006m, the coordinate origin is located at the center of the cylindrical shell, and a harmonic force with a magnitude of 1N is applied radially on the shell at the Cartesian coordinate (0m, 0.4m, 0m). The center of the holographic surface coincides with the center of the cylindrical shell, with a radius of 0.5m, an axial length of 3.2m, an axial interval of 0.2m, and a circumferential interval of 24°. Since there are uncontrollable interference sound sources in the actual test environment, for the convenience of analysis, single-pole sources with a magnitude of 1Pa are added at distances of 0.6m, 1.6m, and 2.6m from the sound source, that is, at coordinates (0, 1, 0), (0, 2, 0), and (0, 3, 0), respectively, to analyze the anti-interference ability of this method in a complex sound field environment. The frequency range is 100Hz - 1000Hz, the step size is 10Hz, and the settings of each parameter are shown in Table 1.

[0089] Table 1 Model parameter settings

[0090]

[0091] The sound pressure level cloud maps corresponding to 210 Hz, 330 Hz, and 940 Hz frequencies under free-field conditions and boundary-influenced conditions are respectively drawn. As Figures 7 - 9 shown, since the measurement of the present invention is carried out on a holographic surface very close to the sound source, the boundary influence can be ignored.

[0092] The influence of interference sources on the sound field: The sound pressure level cloud maps corresponding to 130 Hz, 210 Hz, 340 Hz, and 940 Hz frequencies of interference sources at different positions are as Figures 10 - 13 shown.

[0093] After the sound field separation is carried out by the present invention, the blind source separation results of each measurement point at 100 Hz, 150 Hz, and 610 Hz are as Figures 14 - 16 shown. Before calculating the discrete spatial Fourier transform, the spatial data must be windowed in the direction, so only the measurement points from 45 to 210 are concerned.

[0094] The comparison of the calculated results of the radiated sound power is as Figure 17 shown. At low frequencies, the interference sources and the boundary have a greater influence on the sound field. This is because the low-frequency wavelength is longer, the sound field is uneven, and more interference is generated with the interference waveform. As the frequency increases, the influence of the interference source on the sound field becomes smaller, and the prediction effect is better. This method is effective when the distance between the interference source and the sound source is greater than.

[0095] An embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is loaded into the processor, it implements a method for predicting the low-frequency radiated sound power of an underwater structure based on blind source separation as described in any one of the above.

[0096] An embodiment of the present invention further provides a storage medium. The storage medium stores a computer program, and when the computer program is executed by a processor, it implements a method for predicting the low-frequency radiated sound power of an underwater structure based on blind source separation as described in any one of the above.

Claims

1. A method for predicting the low-frequency radiated sound power of an underwater structure based on blind source separation, characterized in that, It includes the following steps: (1) Set a shell sound source in the shallow sea environment and obtain discrete sound pressures through the shell sound source; (2) Obtain the estimated signal of the separated signal, i.e., the source signal, through the blind source separation algorithm; (3) Window the spatial data of the estimated signal of the source signal in the z direction to obtain the windowed outwards-propagating sound pressure; (4) Convert the windowed outwards-propagating sound pressure to the wavenumber domain through discrete spatial Fourier transform to obtain the holographic surface sound pressure cylindrical wave spectrum; (5) Use a wavenumber domain filter to filter evanescent waves in the wavenumber domain to obtain the propagating waves on the holographic surface; (6) After using the transformation relationship to transform the cylindrical coordinate system to the spherical coordinate system, transfer the filtered sound pressure to the far field in a spherical shape through stationary phase estimation by near-field acoustic holography; (7) Derive the sound intensity from the far-field sound pressure, thereby obtaining the radiated sound power.

2. The method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation according to claim 1, wherein The specific steps of step (1) are as follows: At a distance d from the sound source S a cylindrical holographic surface with a radius of r H is set up, and there are interference sources outside the holographic surface; the complex sound pressure at discrete points on the holographic surface is measured to obtain the sound pressure p(r H ) at (M + 1)×(N + 1) discrete points on the cylindrical holographic surface with a radius of r H , m△θ, n△z); where d S is required to satisfy: kd S <1, k = 2πf / c, f is the frequency, c is the sound speed; the axial interval on the cylindrical holographic surface is △z = 2L / (M + 1), and the circumferential interval angle is △θ = 2π / (N + 1); where M and N are positive integers, M is 1 less than the number of field points on the circumference and is an even number, N is 1 less than the number of field points in the z direction; m is an integer from 0 to M, n is an integer from 0 to N, △z is the field point interval in the z direction, and △θ is the field point interval of the circumference; the shallow sea environment is: the sea surface is set as an absolutely soft boundary, and the seabed is set as an absolutely hard boundary.

3. A method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation according to claim 1, characterized in that, The specific steps of step (2) are as follows: First, represent the discrete sound pressure p(r H , mΔθ, nΔz) received on the holographic plane as p(r H , mΔθ, nΔz) = Ap1(r H , mΔθ, nΔz); where p1(r H , m△θ, n△z) is the outward sound pressure emitted by the sound source, that is, the source signal; Then, the estimated signal of the source signal is obtained through the blind source separation algorithm, and the formula is as follows: p I (r H , mΔθ, nΔz) = Bp(r H , mΔθ, nΔz) (1).

4. A method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation according to claim 1, characterized in that, In step (3), the window function is: where z c is the cut-off distance, set to 80% of the scanning half-length; α is the steepness coefficient, usually taking values from 0.01 to 0.2; then the sound pressure propagating outwards after windowing is: p II (r H , mΔθ, nΔz) = p I (r H , mΔθ, nΔz) ∏(z / 2z c ) (3).

5. A method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation according to claim 1, characterized in that, In step (4), the formula is as follows: where M is 1 less than the number of field points on the circumference and is even, N is 1 less than the number of field points in the z direction, m is an integer from 0 to M, n is an integer from 0 to N, △z is the field point interval in the z direction, △θ is the field point interval of the circumference, the discrete circumferential component k m and the discrete axial component k zn are defined as:

6. The underwater structure low-frequency radiation sound power prediction method based on blind source separation according to claim 1, wherein, In step (5), the wavenumber domain filter is: where k t is the wave number vector, defined as α is the steepness coefficient, set to 0.01 - 0.2; k c is the cut-off wave number of the filter, determined by the wave number in the surrounding medium, and the following formula for the cylindrical wave spectrum of the acoustic pressure of the propagating wave on the holographic surface is obtained: P n I (r H ,k zn ,ω) = P n (r H ,k zn ,ω)Π(n,k z ) (8) 7. A method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation according to claim 1, characterized in that In step (6), the specific process of using the transformation relationship to transform the cylindrical coordinate system to the spherical coordinate system is as follows: The transformation formula for converting the cylindrical coordinate system to the spherical coordinate system is as follows: Among them, the cylindrical coordinate system and the spherical coordinate system share a θ. When it is transmitted to the far field in spherical form, θ and remain unchanged, and R becomes r F ; Under the stationary phase estimation formula: Among them, r F is set to 10 times the lowest frequency wavelength; when the frequency is very low, only the radiation in the mode of n = 0 can reach the far field.

8. A method for predicting the low-frequency radiation sound power of an underwater structure based on blind source separation according to claim 1, characterized in that In step (7), the formula is as follows: SWL = 10 log 10 W(ω) / W0(ω) (10) where c is the sound speed in the water medium, and ρ0 is the density of the water medium, is the sound intensity on the far-field spherical prediction surface with a radius of r F W0(ω) is the reference value of the radiated sound power, usually set to 0.67*10 -18 w, W(ω) is the radiated sound power in watts, and SWL is the reference level of the radiated sound power in dB.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the computer program is loaded into the processor, it implements a method for predicting the low-frequency radiated sound power of an underwater structure based on blind source separation according to any one of claims 1-8.

10. A storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a method for predicting the low-frequency radiated sound power of an underwater structure based on blind source separation according to any one of claims 1-8.

Images