CN104091085A - Cavitation noise feature estimation method based on propeller wake flow pressure fluctuation computing - Google Patents

Abstract

The invention discloses a cavitation noise feature estimation method based on propeller wake pressure fluctuation calculation, and belongs to the field of underwater acoustic target feature extraction. The cavitation noise characteristic estimation method based on the calculation of propeller wake pressure fluctuations includes the following steps: (1) generate a spare grid and import it into the calculation program to generate a calculation example file, (2) set the cavitation model and turbulence model, ( 3) Numerical calculation parameter setting, (4) Numerical calculation, (5) Numerical method reliability verification and grid determination, (6) Cavitation wake pressure fluctuation unsteady numerical calculation, (7) Pressure fluctuation signal power spectrum transformation and low-frequency line spectrum amplitude extraction, (8) line spectrum feature estimation and analysis. The invention introduces relevant research achievements in the fields of modern fluid mechanics, cavitation dynamics and signal processing into the noise characteristic analysis of underwater targets, reflecting the interdisciplinary nature of multiple disciplines and multiple fields.

Description

Estimation Method of Cavitation Noise Characteristics Based on Propeller Wake Pressure Fluctuation Calculation

technical field

The invention relates to the field of underwater acoustic target feature extraction, in particular to a cavitation noise feature estimation method based on calculation of propeller wake pressure fluctuations. the

Background technique

Propeller noise is one of the three major noise sources of ships. It contains information on the type of propeller and structural features of the target. These features are highly tolerant and separable, and are the main features and important basis for identifying underwater targets. Once cavitation occurs, cavitation noise becomes the main noise of the propeller. Due to the interference of these target source noises by the ocean environment noise and the distortion generated in the complex underwater acoustic channel propagation, the characteristics of the noise signal received by the passive sonar are not obvious, and the signal-to-noise ratio is reduced. Therefore, it is more and more difficult to recognize underwater targets by extracting noise features with traditional signal processing methods. Further excavating the essential characteristics of propeller noise is an urgent problem to be solved in underwater target recognition. the

For the research on feature extraction of measured propeller noise using signal processing technology, foreign scholars have already started a long time ago. As early as 1971, Whalen had proposed the maximum likelihood modulation receiver. With the development of this technology, time-frequency processing methods such as high-order spectrum, AR spectrum, dual spectrum and wavelet analysis, as well as nonlinear processing methods such as fractal, chaos, limit cycle and modal decomposition, are all used in propeller noise feature extraction. Try widely. In recent years, Li Qihu and other scholars have used theoretical analysis and numerical simulation to study the single-frequency signal component detection method and detection system performance under strong interference background noise. Nanjing University Bao Fei et al. combined empirical mode decomposition and singular value decomposition to extract the cavitation noise modulation component of the propeller from the strong interference background noise. The modern signal processing method extracts the characteristics of the measured noise signal under the background noise, and has achieved good results. However, due to the lack of mechanism characteristics in the measured signal, this method has low adaptability to strong interference background noise. the

Therefore, some scholars have carried out model-based noise feature analysis and method research. Tao Duchun treats the noise modulation envelope as an impulsive random process with the same shape, equal repetition period, random amplitude, and group structure. And from the power spectral density and autocorrelation function of the ship radiation noise modulation envelope, rich rhythmic information related to various physical properties of the ship is extracted. Jiang Guojian and Lin Jianheng et al. used the theoretical model of exponentially decaying random pulse sequences to analyze the cavitation noise of ship propellers, and obtained the propeller cavitation noise spectrum. In recent years, Shi Guangzhi and other scholars have established a cavitation noise signal model for the identification of the number of propeller blades. Model the noise envelope of the twin-propeller ship, study the structural problem of the harmonic family characteristics of the modulation spectrum of the twin-propeller target, and further use the model feature extraction technology to study the noise feature fine analysis method based on model matching. In addition to blade frequency characteristics, these models do not consider parameters such as propeller geometry and operating conditions, and it is difficult to reflect the mechanism characteristics of cavitation noise. the

Propeller cavitation is the direct sound source of cavitation noise, and propeller cavitation wake is an important transmission path of cavitation noise. Since the wake is affected by the periodic rotation beat of the propeller, it has the characteristics of periodic pulsation. These features reflect characteristic information such as propeller operating conditions and geometry. At the same time, the propeller rotation beat has an obvious amplitude modulation effect on the cavitation noise radiated by it, and the line spectrum features of its power spectrum also reflect the propeller rhythm information including characteristic information such as propeller operating conditions and geometry. Therefore, due to the beat effect of the propeller blades, the propeller cavitation wake and cavitation noise have characteristic correlations, and their characteristics reflect the propeller operating condition parameters and geometric shape parameters. Since the acoustic mechanism research of the propeller cavitation noise is not perfect at present, the present invention starts from the origin of the cavitation noise, that is, the cavitation wake, to describe a prediction method of its noise characteristics. the

Many scholars at home and abroad have conducted research on propeller cavitation wake. Francesc et al. from the Italian ship model pool laboratory used RANS, LES and BEM methods to simulate the wake flow field of E779A propeller under cavitation and non-cavitation conditions respectively. Based on the mixed two-phase flow model, Rickard and Goran of Sweden used the implicit LES method and the Kunz cavitation model to simulate the dynamic behavior of E779A propeller cavitation in a non-uniform flow field. The simulation was more successful. Scholars such as Bin Ji from Tsinghua University used the Rayleigh–Plessete equation and the k-ω Shear Stress Transport (SST) turbulence model to numerically simulate the cavitation wake of uniform and non-uniform inflow on a high-skewed propeller. The blade cavitation and tip vortex cavitation are well predicted, and the pressure fluctuation characteristics of the wake field induced by cavitation are consistent with the propeller shaft frequency and blade frequency characteristics. Yang Qiongfang from Naval Engineering University evaluated and analyzed the cavitation model and turbulence model in propeller cavitation simulation, and selected the improved Sauer cavitation model and the revised SST k-ω turbulence model to predict the propeller cavitation bucket map more accurately. For the non-uniform inflow with seven-blade large-skew propeller, the thrust and moment collapse performance caused by cavitation and the influence on the incipient cavitation of the blade tip vortex are analyzed, and the pulsation characteristics of cavitation thrust and moment, propeller The change of blade cavitation area and cavitation shape with the circumferential position is given, and the interval division of whether blade cavitation occurs in the propeller in wake is given. At present, the research on cavitation wake at home and abroad mainly focuses on the numerical prediction of some kind of propeller module cavitation. It is even less to use wake flow to study cavitation noise characteristics. the

In addition, the Chinese patent application number ZL201310538724.7 also discloses a feature extraction method based on the numerical prediction of propeller cavitation noise in non-uniform inflow. And define the boundary conditions; Next, in the CFD software, set up the calculation model, and perform steady-state iterative calculation of the running water performance parameters and inlet velocity to verify the accuracy of the model; then, in the CFD software, use the steady-state calculation as the unsteady-state The calculated initial value is calculated by unsteady iterative calculation, and the cavitation cycle shape of the propeller blade and the change of the cavitation area of the recording blade are displayed through post-processing; finally, the cavitation radiation noise of the propeller is calculated from the cavitation area of the propeller blade according to the single cavitation radiation noise theory , for feature extraction. The method used in this application document converts the cavitation area into a spherical volume, and obtains the radius of the spherical volume, and then brings the radius change into the spherical single-cavity noise radiation model to predict cavitation noise and its characteristics. Since propeller cavitation is very different from spherical single cavitation, the accuracy of this conversion method needs to be further tested. The method of the present invention uses the characteristic correlation between the propeller cavitation wake pressure fluctuation and the cavitation noise to estimate the noise feature. Specifically, the cavitation wake pressure fluctuation information contains propeller operating conditions and geometric shape parameters, and cavitation noise also has this attribute. Therefore, they have the same original relationship, that is, the rotation of the propeller causes the cavitation wake and generates noise, and the cavitation noise is also modulated by the rotating blade. The root cause of cavitation wake and noise is the rotation of the propeller in the fluid. the

Contents of the invention

The main characteristics of propeller cavitation noise are:

1. Propeller cavitation is the source of cavitation noise, and propeller cavitation noise always appears with the appearance of propeller cavitation;

2. Propeller cavitation wake is not only the sound source of cavitation noise, but also an important carrier of cavitation noise transmission;

3. The sound intensity of cavitation noise is closely related to the volume change of the cavitation bubble, especially the volume change at the moment of collapse is the largest, and the radiation noise is also the strongest;

4. The volume change and position distribution of the propeller cavitation have periodic characteristics with the rotation of the blade, so that the cavitation noise also has periodic characteristics, which will be reflected in its noise spectrum distribution;

5. The cavitation wake and cavitation noise will be modulated by the rotation beat of the propeller blade at the same time;

6. The above characteristics make propeller cavitation and its noise characteristics have a close and essential correlation;

7. Chip cavitation noise is generally distributed in the low frequency band, and its spectrum presents a line spectrum feature, while the noise emitted by tip vortex cavitation is generally distributed in the middle and high frequency bands, and its spectrum presents a continuous feature. the

The principle of the present invention is exactly based on the main characteristics of the above-mentioned propeller cavitation, based on the viscous multiphase flow theory, using modern computational fluid dynamics methods to construct the N-S equation for the underwater propeller wake field, and combining the turbulence model and the cavitation model to the equations Carry out numerical solutions to obtain relevant information such as the volume fraction of the vapor phase around the blade surface of the underwater propeller and the pressure fluctuation in the wake field; and then use signal processing methods such as power spectrum to extract and analyze the low-frequency characteristics of the numerically calculated flow field information data ; Finally, the characteristic correlation between the pressure fluctuation of the flow field and the noise is used to estimate and judge the noise characteristics of the underwater target propeller. Although pressure pulsation is a mechanical parameter in the wake field and noise sound pressure is an acoustic parameter, their physical concepts are different, but their characteristics in some aspects, such as the distribution of low-frequency line spectrum amplitude, have some similarities. These common points essentially reflect the characteristics of propeller geometry and working condition parameters, which are called feature correlations here. the

The present invention adopts following technical scheme:

The cavitation noise feature estimation method based on propeller wake pressure fluctuation calculation includes the following steps:

(1) After the backup grid is generated and imported into the calculation program, the calculation example file is generated:

Use professional modeling software to create a three-dimensional geometric model of the propeller and import it into the mesh generation software, and establish three alternative meshes in the mesh division software. The calculation domains of these three alternative meshes are the same, and the distance from the velocity inflow boundary to the center of the propeller is 1D, D is the diameter of the propeller, the distance from the downstream pressure outlet boundary is 5D, and the distance from the center of the propeller to the periphery of the side is 2.5D, the number of grid cells of these three grids is according to The multiples are gradually increased, and the grid cell size of the adjacent boundary in the grid is reasonably transitioned at the boundary point, so that the difference in grid size is small, and finally the skew of all volume grid cells in the grid is limited within 0.9, so that Ensure the stability of subsequent numerical calculations;

(2) Setting of cavitation model and turbulence model:

The full cavitation model and the renormalized group turbulence model are used, and their important parameters are corrected. The correction of the phase change rate parameter in the cavitation model and the correction of the turbulent viscosity coefficient in the turbulent flow model are written in C language, and then called by a macro (DEFINE_TURBULENT_VISCOSITY, etc.) embedded in the calculation program;

(3) Numerical calculation parameter setting:

Set the relevant parameters of working conditions, boundary conditions and numerical algorithms;

(4) Numerical calculation:

Since the cavitation model is added to the RANS equation, the stability of the calculation is reduced, and singular phenomena are prone to occur. Therefore, in order to make the numerical calculation run smoothly, a step-by-step calculation process is adopted. Specifically, in the propeller operating condition parameters, the ambient pressure and inflow velocity can be directly set to the operating condition values, while the propeller speed is increased by stages. , until it increases to the value of the predetermined working condition; first calculate the flow field distribution of the no-cavitation model, and then open the cavitation model after the calculation is stable; first perform the first-order precision discrete format calculation on parameters such as pressure, density, momentum and vapor fraction, After the calculation is stable, the discrete precision is increased to the second order or QUCIK, etc. Since the calculation of multiphase flow model, cavitation model and sliding grid consumes a lot of computer resources, parallel computing technology is used to shorten the calculation time. the

(5) Numerical method reliability verification and grid determination:

Compare the numerical calculation results of the hydrodynamic parameters and cavitation of the propeller under typical working conditions with the relevant experimental results to verify the grid independence and the reliability of the numerical method used; Select the grid to set according to the method of steps 2 and 3 and perform numerical calculations, and compare the hydrodynamic parameters and cavitation in the calculation results, when these results tend to be stable with the increase in the number of grids and are consistent with the experimental results , select the grid that satisfies the condition with the least number of grid units as the selected grid for the following numerical calculation; otherwise, increase the number of grids appropriately, and repeat step 1 to start again;

(6) Calculation of unsteady value of cavitation wake pressure fluctuation:

Using the grid selected in step 5, the unsteady value calculation of the wake field of the propeller is carried out under the required working conditions. In the calculation program, the pressure fluctuation at a specific position (point A) in the wake field is detected and Save its detection data, and at the same time make the data dimensionless;

(7) Transformation of the power spectrum of the pressure pulsation signal and extraction of the amplitude of the low-frequency line spectrum:

The fast Fourier transform method in signal processing is used to transform the power spectrum of physical quantities such as pressure pulsation and noise signal data in the flow field, and extract the amplitude of the low-frequency line spectrum. The low-frequency line spectrum includes axial frequency, double axial frequency, and triple axial frequency. frequency and blade frequency; and then using the similarity between the wake field pressure fluctuation characteristics and the cavitation noise modulated by the blade, the characteristic correspondence relationship from the low-frequency line spectrum amplitude of the pressure fluctuation to the low-frequency line spectrum amplitude of the noise is established;

(8) Estimation and analysis of line spectrum features:

Corresponding the amplitude of the low-frequency line spectrum in step 7 to the amplitude of the low-frequency line spectrum of the noise signal power spectrum, as an estimate of the distribution characteristics of the low-frequency line spectrum amplitude of the cavitation noise signal, specifically, using the power spectrum of the pressure pulsation signal The amplitudes of low frequency components such as shaft frequency, double shaft frequency, triple shaft frequency and leaf frequency are used to estimate the amplitudes of low frequency components such as shaft frequency, double shaft frequency, triple shaft frequency and leaf frequency of the noise signal respectively. the

Furthermore, the unsteady calculation of cavitation wake pressure fluctuation in step 6 includes the following steps:

(6-1) Import the mesh generation calculation example file selected in step 5;

(6-2) Setting of cavitation model and turbulence model;

(6-3) Numerical calculation parameter setting;

(6-4) Numerical calculation;

(6-5) Pressure pulsation signal extraction: detect and save the pressure pulsation at a specific position (point A) in the wake field and save the detected data. the

The cavitation model and turbulence model setting in step (6-2), the numerical calculation parameter setting in step (6-3) and the numerical calculation in (6-4) are respectively the same as the cavitation model and turbulence model setting in step 2 , The numerical calculation parameter setting of step 3 is the same as the numerical calculation of step 4. the

Furthermore, the alternative grid in step 1 adopts the regional mixed grid division method: the flow field area around the propeller is divided by the unstructured grid method, and the grid gradually decreases from the hub to the blade tip, and the blade The surface grid at the tip is triangular, the side length of the grid unit is about 0.001D, and the unit at the blade hub is about 0.02D; since cavitation is mainly distributed in the blade surface and tip vortex area, the grid quality requirements in this area are relatively high. high. In order to better adapt to the wall function, a boundary layer grid is established on the blade surface; the calculation domain with a regular shape around the propeller is divided by a structural grid; based on the above method, three calculation domain grids with different numbers of grid cells are simultaneously generated as an alternative grid; the grid in the tip vortex area is encrypted, and the surface of the blade adopts the boundary layer grid to improve the prediction accuracy of the tip vortex cavitation. The grid unit size in the tip vortex area is about 0.001D, and the boundary layer grid has 4 layers , the height ratio of the two adjacent layers is 1.1, and the height of the grid cells in the first layer is about 0.001D, so that the dimensionless parameter 20<y ⁺ <300. In addition, the difference in the number of grid units of the three candidate grids is mainly reflected in the area of the propeller wake within its disk, especially the area near the propeller. Because the mesh quality in this area is crucial to the accuracy of propeller cavitation performance and wake pressure fluctuation calculation.

Furthermore, the setting of the full cavitation model and its parameters in step 2 are as follows:

When p<p _v , the steam generation rate is:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; e</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Ce</span>}\frac{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; 3</span>}}\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</span>}}{{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; )</span>$

When p>p _v , the vapor phase changes into liquid phase, and the steam solidification rate R _c is also obtained:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Cc</span>}\frac{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; 3</span>}}\frac{\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}{{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}}}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}$

Among them, f _v =α _v ρ _v /ρ _m is the vapor phase mass fraction, vaporization coefficient Ce=0.02 and condensation coefficient Cc=0.01 are empirical parameters.

According to dimensional analysis, k is used instead of In the FLUENT software environment, the user-defined function UDF can be used to correct the parameter phase change rate _Re in the cavitation model. The corrected model is written in C language and transferred to the calculation program.

Furthermore, the RNG k-ε turbulence model and its parameters in step 2 are modified as follows: the renormalized group turbulence model is the RNG k-ε turbulence model, and the renormalized group turbulence model RNG k-ε is the The transient N-S equation is a model derived by the mathematical method of the Renormalization Group (RNG for short). It systematically removes these small-scale motions from the governing equations by embodying these small-scale effects in both the large-scale motion term and the modified viscosity term. Its k equation and ε equation are respectively:

$\frac{<span\; class="notranslate">\; \&PartialD;</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; ku</span>}}_{\mathrm{<span\; class="notranslate">\; mj</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; [</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; )</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; k</span>}}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; G</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}$

$\frac{<span\; class="notranslate">\; \&PartialD;</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; mj</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; [</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; )</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}\frac{\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}{\mathrm{<span\; class="notranslate">\; k</span>}}\mathrm{<span\; class="notranslate">\; G</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\frac{{\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; k</span>}}$

In the formula, the turbulent kinetic energy dissipation rate (Turbulent Dissipation Rate) The reciprocal of the effective turbulent Prandtl number a _k = a _ε = 1.39 for the turbulent kinetic energy k and dissipation rate ε; the model parameters C _1ε = 1.47, C _2ε = 1.68; the viscosity coefficient is μ = μ _t + μ _m , μ _m is the viscosity coefficient of the mixed flow; the modified turbulent viscosity coefficient μ _t = [ρ _v +α _l ¹⁰ (ρ _l -ρ _v )]C _μ k ² /ε, C _μ = 0.085 is more suitable for unsteady two Phase flow simulation for better simulation of propeller cavitation.

Further, the numerical calculation parameter setting of the step 3 includes the relevant parameter setting of working condition, boundary condition and numerical algorithm;

The working conditions mainly set the propeller rotation speed, ambient pressure and inflow velocity values, and determine the dimensionless parameters of the propeller, that is, the advance coefficient (J) and the cavitation number (σ _n ); for the boundary condition setting, the velocity inlet boundary adopts the inflow Velocity value, the far-field boundary condition is set by the inflow velocity, and the outlet pressure of the downstream pressure outlet interface is set as static pressure; parameter settings in the numerical algorithm: the convection item in the control equation is discretized by the second-order upwind method, and the diffusion item is by the second-order central difference The format is discrete, and the speed-pressure coupling adopts the SIMPLE algorithm suitable for unstructured grids, and uses point-by-point Gauss-Seidel iterations to solve discrete equations; uses algebraic multi-grid to accelerate calculation convergence, and uses sliding grid calculation technology for unsteady calculations to improve calculation efficiency. accuracy. The second-order precision discrete scheme is adopted. In order to ensure the stability of the second-order calculation, the under-relaxation factor is appropriately reduced. Set as: 0.25, 0.6, 0.2, 0.7, 0.7, 0.9, mass conservation continuity (continuity) residual convergence standard is third order, other physical quantity residual convergence standard in the equation is fourth order.

Furthermore, the specific position A point in the step 6 is located at the propeller wake radial direction r=0.5R and the axial direction x=2R, according to the dimension conversion principle, the formula Dimensionless, where ΔP is the total pressure pulsation value of the numerical calculation results, ρ is the density of the mixed fluid, n is the propeller speed, and D is the propeller diameter; in the unsteady calculation, the TIME STEP time step is set as T=0.0125T _P , T _P is the rotation period of the propeller, and the time length involved in the data is 30T _P .

The method of the present invention can be implemented in common CFD fluid calculation software (CFX, FLUENT, etc.), and grid division can be implemented by software such as GAMBIT. Firstly, import the preliminary designed propeller digital model into the grid division software, and perform grid division according to the method in the present invention. After the grid model is imported into the computing platform, a numerical example file is formed, and numerical parameters are set in the example file, numerical calculation is performed according to the design working conditions, and the pressure pulsation signal is output to a text file. Write the power spectral density transformation program in MATLAB software to realize the signal transformation from time domain to frequency domain, and finally extract the low-frequency line spectrum amplitude parameters. In addition, the invention also utilizes batch files to perform parallel numerical calculations on the operating system platform. the

Beneficial effect:

(1) The present invention introduces relevant research achievements in the fields of modern fluid mechanics, cavitation dynamics and signal processing into the noise characteristic analysis of underwater targets, reflecting the interdisciplinary nature of multiple disciplines and multiple fields. the

(2) At present, due to the influence of strong interference background noise and complex underwater acoustic channels, it is difficult to meet the requirements of underwater target recognition technology only by extracting target features from detection noise. Therefore, the present invention has important application value and application prospect in the field of underwater acoustic target recognition . the

(3) Since the wake field is affected by the rotation beat of the propeller blades and the radiation noise is also affected by the blade frequency modulation, the characteristics of the pressure fluctuation in the wake field and the slow-varying component of the sound pressure signal are closely related, and the power spectral densities of the two are low-frequency Features such as line spectrum amplitude distribution have correlation, and the present invention utilizes the correlation between flow field characteristic parameters and fluid noise characteristic parameters to perform characteristic estimation on propeller cavitation noise. the

(4) The method of the present invention corrects the relevant parameters of the cavitation model and the turbulence model, establishes a boundary layer on the surface of the blade, and finely processes the grid in the tip vortex cavitation region. Through the comparison with the experiment (see Figure 5), these improvement measures have significantly improved the prediction accuracy of tip vortex cavitation, and better solved a difficult problem in the numerical prediction of cavitation. At the same time, it provides a strong guarantee for the numerical calculation of cavitation wake pressure fluctuation in the next step. the

(5) Use the determined grid to calculate the unsteady value of the cavitation wake of the propellers E779A and E779B, and extract the wake pressure fluctuation signal, and then perform power spectrum transformation on the pressure fluctuation to obtain the distribution characteristics of the low-frequency line spectrum amplitude, Comparing this distribution with the low-frequency line spectrum distribution of the measured noise (see Figure 6 and Figure 7), the characteristic correlation between the two is verified. Finally, this characteristic correlation can be used to estimate the noise characteristics under other working conditions through numerical calculation of pressure fluctuations, thus providing important directional value for underwater target classification and recognition technology based on propeller noise characteristics. the

Description of drawings

Figure 1(a) is the geometric model of the E779A propeller;

Figure 1(b) is the geometric model of the E779B propeller;

Fig. 2 is the schematic diagram of the hybrid mesh of the calculation domain of the full channel propeller of the present invention;

Fig. 3 (a) is the structural representation of boundary layer grid of the present invention;

Figure 3(b) is an enlarged view at A in Figure 3(a);

Fig. 4 (a) is method flowchart of the present invention;

Fig. 4 (b) is the calculation flowchart of the unsteady value of cavitation wake pressure fluctuation of the present invention;

Figure 5 shows the numerical prediction results and experimental results of E779A propeller cavitation;

Figure 6 shows the cavitation numerical prediction results and experimental results of the counterclockwise rotating E779B propeller at different positions and times under the condition of non-uniform inflow;

Figure 7 is the normalized measurement noise power spectrum and pressure pulsation numerical calculation signal power spectrum of the E779A propeller;

Figure 8(a) is the normalized power spectral density low-frequency line spectrum amplitude of the E779B propeller cavitation wake pressure pulsation and measurement noise signal with n=15rps;

Figure 8(b) is the normalized power spectral density low-frequency line spectrum amplitude of the E779B propeller cavitation wake pressure pulsation and measurement noise signal at n=20rps;

Figure 8(c) shows the normalized power spectral density low-frequency line spectrum amplitude of the E779B propeller cavitation wake pressure pulsation and measurement noise signal with n=25rps. the

Detailed ways

The present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments. the

Example

As shown in Figure 4(a), the cavitation noise feature estimation method based on propeller wake pressure fluctuation calculation includes the following steps:

(1) After the spare grid is generated and imported into the calculation program, the calculation example file is generated:

Use professional modeling software to create a 3D geometric model of the propeller and import it into the mesh generation software, as shown in Figure 1(a) and Figure 1(b), for the E778A and E779B propeller models, three alternative meshes are established in the meshing software The calculation domains of these three alternative grids are the same, the speed inflow boundary is 1D from the center of the propeller, D is the diameter of the propeller, the distance from the downstream pressure outlet boundary is 5D, and the distance from the center of the propeller to the side periphery is 2.5D. The number of grid cells of the grid is according to The multiples are gradually increased, and the grid cell size of the adjacent boundary in the grid is reasonably transitioned at the boundary point, so that the difference in grid size is small, and finally the skew of all volume grid cells in the grid is limited within 0.9, so that Ensure the stability of subsequent numerical calculations;

The grid adopts the regional mixed grid division method (as shown in Figure 2): the flow field area around the propeller is divided by the unstructured grid method, and the grid gradually decreases from the hub to the blade tip, and the surface grid at the blade tip is Triangular, the side length of the grid unit is about 0.001D, and the unit at the blade hub is about 0.02D; since cavitation is mainly distributed in the blade surface and tip vortex area, the grid quality requirements in this area are relatively high. In order to better adapt to the wall function, a boundary layer grid is established on the surface of the blade (as shown in Figure 3); the structural grid is used to divide the calculation domain of the regular shape around the propeller; The calculation domain grid is used as an alternative grid; the grid in the tip vortex area is encrypted, and the surface of the blade uses a boundary layer grid to improve the prediction accuracy of the tip vortex cavitation. The grid cell size in the tip vortex area is about 0.001D , the boundary layer grid has 4 layers in total, the height ratio of the two adjacent layers is 1.1, and the height of the grid cells in the first layer is about 0.001D, so that the dimensionless parameter 20<y ⁺ <300. In addition, the difference in the number of grid units of the three candidate grids is mainly reflected in the area of the propeller wake within its disk, especially the area near the propeller. Because the mesh quality in this area is crucial to the accuracy of propeller cavitation performance and wake pressure fluctuation calculation.

(2) Setting of cavitation model and turbulence model:

The full cavitation model and the renormalized group turbulence model are used, and their important parameters are corrected. The correction of the phase change rate parameter in the cavitation model and the correction of the turbulent viscosity coefficient in the turbulent flow model are written in C language, and then called by a macro (DEFINE_TURBULENT_VISCOSITY, etc.) embedded in the calculation program;

The full cavitation model setting and its parameters are corrected as follows:

When p<p _v , the steam generation rate is:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; e</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Ce</span>}\frac{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; 3</span>}}\frac{{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; p</span>}}{{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}<span\; class="notranslate">\; )</span>$

When p>p _v , the vapor phase changes into liquid phase, and the steam solidification rate R _c is also obtained:

${\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Cc</span>}\frac{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&gamma;</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}\sqrt{\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; 3</span>}}\frac{\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}{{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; l</span>}}}}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}$

Among them, f _v =α _v ρ _v /ρ _m is the mass fraction of the vapor phase, the vaporization coefficient Ce=0.02 and the condensation coefficient Cc=0.01 are empirical parameters;

According to dimensional analysis, k is used instead of In the FLUENT software environment, the user-defined function UDF can be used to correct the parameter phase change rate _Re in the cavitation model. The corrected model is written in C language and transferred to the calculation program.

The renormalization group turbulence model is the RNG k-ε turbulence model, and the renormalization group turbulence model RNG k-ε is a model derived from the mathematical method of the renormalization group (Renormalization Group, referred to as RNG) for the transient N-S equation . It systematically removes these small-scale motions from the governing equations by embodying these small-scale effects in both the large-scale motion term and the modified viscosity term. Its k equation and ε equation are respectively:

$\frac{<span\; class="notranslate">\; \&PartialD;</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; ku</span>}}_{\mathrm{<span\; class="notranslate">\; mj</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; [</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; )</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; k</span>}}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; G</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}$

$\frac{<span\; class="notranslate">\; \&PartialD;</span>}{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; mj</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; [</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}_{\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}\mathrm{<span\; class="notranslate">\; \&mu;</span>}<span\; class="notranslate">\; )</span>\frac{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}{{<span\; class="notranslate">\; \&PartialD;</span>\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}\frac{\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}{\mathrm{<span\; class="notranslate">\; k</span>}}\mathrm{<span\; class="notranslate">\; G</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}}\frac{{\mathrm{<span\; class="notranslate">\; \&epsiv;</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; k</span>}}$

In the formula, the turbulent kinetic energy dissipation rate (Turbulent Dissipation Rate) The reciprocal of the effective turbulent Prandtl number a _k = a _ε = 1.39 for the turbulent kinetic energy k and dissipation rate ε; the model parameters C _1ε = 1.47, C _2ε = 1.68; the viscosity coefficient is μ = μ _t + μ _m , μ _m is the viscosity coefficient of the mixed flow; the modified turbulent viscosity coefficient μ _t = [ρ _v +α _l ¹⁰ (ρ _l -ρ _v )]C _μ k ² /ε, C _μ = 0.085 is more suitable for unsteady two Phase flow simulation for better simulation of propeller cavitation.

(3) Numerical calculation parameter setting:

Set the working conditions, boundary conditions and relevant parameters of the numerical algorithm; the working conditions mainly set the propeller rotation speed, ambient pressure and inflow velocity values, and determine the dimensionless parameters of the propeller, namely the advance speed coefficient (J) and cavitation number (σ _n ); for the boundary condition setting, the velocity inlet boundary adopts the inflow velocity value, the far field boundary condition adopts the inflow velocity setting, and the outlet pressure of the downstream pressure outlet interface is set as the static pressure; the parameter setting in the numerical algorithm: the control equation The convection item is discretized by the second-order upwind scheme, the diffusion item is discretized by the second-order central difference scheme, the velocity-pressure coupling adopts the SIMPLE algorithm suitable for unstructured grids, and the discrete equations are solved by point-by-point Gauss-Seidel iteration; the algebraic multigrid is used to accelerate Calculation convergence, using sliding grid computing technology for unsteady calculations to improve the accuracy of calculations. The second-order precision discrete scheme is adopted. In order to ensure the stability of the second-order calculation, the under-relaxation factor is appropriately reduced. Set as: 0.25, 0.6, 0.2, 0.7, 0.7, 0.9, mass conservation continuity (continuity) residual convergence standard is third order, other physical quantity residual convergence standard in the equation is fourth order.

(4) Numerical calculation:

Since the cavitation model is added to the RANS equation, the stability of the calculation is reduced, and singular phenomena are prone to occur. Therefore, in order to make the numerical calculation run smoothly, a step-by-step calculation process is adopted. Specifically, in the propeller operating condition parameters, the ambient pressure and inflow velocity can be directly set to the operating condition values, while the propeller speed is increased by stages. , until it increases to the value of the predetermined working condition; first calculate the flow field distribution of the no-cavitation model, and then open the cavitation model after the calculation is stable; first perform the first-order precision discrete format calculation on parameters such as pressure, density, momentum and vapor fraction, After the calculation is stable, the discrete precision is increased to the second order or QUCIK, etc. Since the calculation of multiphase flow model, cavitation model and sliding grid consumes a lot of computer resources, parallel computing technology is used to shorten the calculation time. the

(5) Numerical method reliability verification and grid determination:

Compare the numerical calculation results of the hydrodynamic parameters and cavitation of the propeller under typical working conditions with the relevant experimental results to verify the grid independence and the reliability of the numerical method used; Select the grid to set according to the method of steps 2 and 3 and perform numerical calculations, and compare the hydrodynamic parameters and cavitation in the calculation results, when these results tend to be stable with the increase in the number of grids and are consistent with the experimental results , select the grid that satisfies the condition with the least number of grid units as the selected grid for the following numerical calculation; otherwise, increase the number of grids appropriately, and repeat step 1 to start again. Figure 5 shows the numerical prediction results and Experimental results, Figure 6 is the cavitation numerical prediction results and experimental results of the E779B propeller rotating counterclockwise under the non-uniform inflow condition at different positions, the results of Figure 5 and Figure 6 show that the method used in the present invention predicts propeller cavitation The effect is better;

(6) Calculation of unsteady value of cavitation wake pressure fluctuation:

Using the grid selected in step 5, the unsteady value calculation of the wake field of the propeller is carried out under the required working conditions. In the calculation program, the pressure fluctuation at a specific position (point A) in the wake field is detected and Save its detection data, and at the same time make the data dimensionless;

Point A at the specific position is located at the propeller wake radial direction r=0.5R and axial direction x=2R, according to the dimension conversion principle, the formula Dimensionless, where ΔP is the total pressure pulsation value of the numerical calculation result, ρ is the density of the mixed fluid, n is the propeller speed, and D is the propeller diameter; in the unsteady calculation, the TIME STEP time step is set as T=0.0125T _P , T _P is the rotation period of the propeller, and the time length involved in the data is 30T _P .

(7) Transformation of the power spectrum of the pressure pulsation signal and extraction of the amplitude of the low-frequency line spectrum:

The fast Fourier transform method in signal processing is used to transform the power spectrum of physical quantities such as pressure pulsation and noise signal data in the flow field, and the amplitude of the low-frequency line spectrum (shaft frequency, double shaft frequency, triple shaft frequency and leaf frequency) is carried out. Extraction; and then use the similarity between the wake field pressure fluctuation characteristics and the cavitation noise modulated by the blade to establish the characteristic correspondence relationship from the low-frequency line spectrum amplitude of the pressure fluctuation to the low-frequency line spectrum amplitude of the noise;

(8) Estimation and analysis of line spectrum features:

The low-frequency line spectrum amplitude in step 7 is one-to-one corresponding to the low-frequency line spectrum amplitude of the noise signal power spectrum, as an estimation of the distribution characteristics of the low-frequency line spectrum amplitude of the cavitation noise signal. (a) and (b) in Fig. 7 are the normalized power spectral densities of cavitation wake pressure fluctuation and cavitation noise under the uniform inflow condition of E779A propeller when J=0.88, n=25rps. Through the comparison of Figure 7(a) and (b), it is found that the normalized power spectral density of the pressure pulsation signal and the power spectrum of the measurement noise pressure signal have some common features: (1) The signal in the low frequency band of 10-100Hz is dominated by line spectrum , and is consistent with the propeller blade number and rotational speed parameter values. Among them, the peak of the leaf frequency line spectrum (100Hz) is the highest, the peak value of the axis frequency line spectrum (25Hz) is next, and then the 75Hz and 50Hz line spectrum. Except for the lower signal strength at 75Hz, the other features are basically the same as those in Fig. 7(a). (2) In the middle frequency range of 100-1000Hz, there are also rich line spectrum features, the line spectrum is dominated by the double frequency of the axis frequency and the leaf frequency, and the continuous spectrum line also begins to decline, which is similar to that in Figure 7(a) The spectral shape changes are basically the same. However, the line spectrum amplitude in this frequency region in the numerical results is smaller than the experimental value in Fig. 7(a), and the frequency resolution is lower. (3) The amplitude range of the power spectrum is between 10 ⁰ and 10 ^-9 , which is basically consistent with the experimental data in Fig. 7(a). The common features mentioned above indicate that the E779A paddle model has similar characteristics between pressure pulsation and noise under the condition of uniform inflow, that is, they have characteristic correlation.

Fig. 8(a), (b) and (c) are the normalized power spectral density low-frequency line spectrum amplitude comparison of the cavitation wake pressure fluctuation and the measurement noise signal under the non-uniform inflow of the E779B propeller. Figure 7 shows that the amplitude of the low-frequency line spectrum of pressure pulsation and noise is at 15, 20 and 25rps, and the variation trends of the shaft frequency, double shaft frequency and triple shaft frequency amplitude of the two are basically the same. When the blade frequency is at 15rps, the difference between the two is the largest, at 20rps, the difference between the two decreases, and at 25rps, the difference is basically the same. This shows that when the propeller rotates at 25rps, cavitation obviously occurs, and cavitation noise becomes the main noise. At this time, the correlation between the two blade frequency amplitude characteristics is strengthened. However, when the speed is low (15rps), no cavitation occurs, and the ambient noise becomes the main sound source at this time, so the correlation between the two leaf frequency characteristics is weakened. This shows that the accuracy of this method is significantly improved when cavitation occurs and cavitation noise becomes the main noise of the propeller. the

As shown in Figure 4(b), the unsteady calculation of cavitation wake pressure fluctuations in step 6 includes the following steps:

(6-1) Import the mesh generation calculation example file selected in step 5;

(6-2) Setting of cavitation model and turbulence model;

(6-3) Numerical calculation parameter setting;

(6-4) Numerical calculation;

(6-5) Pressure pulsation signal extraction: detect and save the pressure pulsation at a specific position (point A) in the wake field and save the detected data. the

The cavitation model and turbulence model setting in step (6-2), the numerical calculation parameter setting in step (6-3) and the numerical calculation in (6-4) are respectively the same as the cavitation model and turbulence model setting in step 2 , The numerical calculation parameter setting of step 3 is the same as the numerical calculation of step 4. the

The method of the present invention can be implemented in common CFD fluid calculation software (CFX, FLUENT, etc.), and grid division can be implemented by software such as GAMBIT. Firstly, import the preliminary designed propeller digital model into the grid division software, and perform grid division according to the method in the present invention. After the grid model is imported into the computing platform, a numerical example file is formed, and numerical parameters are set in the example file, numerical calculation is performed according to the design working conditions, and the pressure pulsation signal is output to a text file. Write the power spectral density transformation program in MATLAB software to realize the signal transformation from time domain to frequency domain, and finally extract the low-frequency line spectrum amplitude parameters. In addition, the invention also utilizes batch files to perform parallel numerical calculations on the operating system platform. the

Claims (9)

1. the cavitation noise feature of calculating based on screw current pressure fluctuation is estimated access method, it is characterized in that: comprise the following steps:

(1) standby grid generates and imports and generates example file after calculation procedure:

After utilizing professional software to make screw propeller 3-D geometric model, import Grid Generation Software, at grid, divide in software and set up three kinds of alternative grids, the computational fields of these three kinds of alternative grids is identical, the speed frontier distance propeller center that becomes a mandarin is 1D, D is airscrew diameter, downstream pressure outlet frontier distance is 5D, and propeller center is 2.5D to side periphery distance, the grid cell quantity of these three grids according to multiple increases gradually, to the grid cell size of adjacent boundary in grid, in the reasonable transition of frontier point, makes the skew of all volume mesh unit in grid all be limited in 0.9;

(2) cavitation model and turbulence model are set:

Adopt full cavitation model and Renormalization Group turbulence model, and its important parameter is revised;

(3) numerical evaluation setting parameter:

Correlation parameter to working condition, boundary condition and numerical algorithm is set;

(4) numerical evaluation:

Adopt step by step computation process step by step, in screw propeller duty parameter, environmental pressure and inflow velocity can directly be set to operating mode value, and revolution speed of propeller adopts classification to increase, until be increased to predetermined operating mode value; First calculate non-cavitating model Flow Field Distribution, by the time after calculation stability, open again cavitation model; First the parameters such as pressure, density, momentum and vapour phase mark are carried out to single order precision discrete scheme and calculate, after calculation stability, more discrete precision is brought up to second order or QUCIK etc., and adopt parallel computing to calculate;

(5) numerical method reliability demonstration and grid are determined:

To under typical condition, the numerical result of the hydrodynamic parameter of screw propeller oar and cavitation and related experiment result be compared, to verify the reliability of grid independence and the numerical method that adopted; And in logarithm value result of calculation, hydrodynamic parameter and cavitation compare, when result tends towards stability along with the increase of number of grid and be consistent with experimental result, the minimum grid of the selected middle grid cell quantity that satisfies condition is as the selected grid of numerical evaluation below; Otherwise suitably increase number of grid, repeating step 1 restarts;

(6) the non-permanent numerical evaluation of cavitation wake flow pressure fluctuation:

Adopt the selected grid in step 5, the tail flow field of screw propeller is carried out under required working condition to non-permanent numerical evaluation, in calculation procedure, to a certain ad-hoc location in tail flow field (A point), pressure fluctuation detects and preserves it and detects data, data is carried out to nondimensionalization simultaneously;

(7) conversion of pressure fluctuation signal power spectrum and low frequency spectrum lines magnitude extraction:

Adopt signal process in fast fourier transform method stream field the physical quantity such as pressure fluctuation and noise signal data carry out power spectrum conversion, and low frequency spectrum lines amplitude is extracted; Recycling tail flow field pressure fluctuation characteristic and cavitation noise be by the similarity of blade modulation signature, the feature corresponding relation of foundation from the low frequency spectrum lines amplitude of pressure fluctuation to the low frequency spectrum lines amplitude of noise;

(8) line spectrum feature is estimated and is analyzed:

Step 7 medium and low frequency line spectrum amplitude is corresponded to one by one to the low frequency spectrum lines amplitude of noise signal power spectrum, as the estimation to cavitation noise signal low frequency spectrum lines characteristics of amplitude distribution.

2. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as claimed in claim 1, is characterized in that: the non-permanent calculating of cavitation wake flow pressure fluctuation in described step 6 comprises the following steps:

(6-1) the selected grid importing in step 5 generates example file;

(6-2) cavitation model and turbulence model are set;

(6-3) numerical evaluation setting parameter;

(6-4) numerical evaluation;

(6-5) pressure fluctuation signal is extracted: to a certain ad-hoc location in tail flow field (A point), pressure fluctuation detects and preserves it and detects data.

3. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as claimed in claim 1, is characterized in that: the alternative grid in described step 1 adopts subregion mixed mesh method:

Screw propeller around flow field regions adopts non-structured grid method to divide, and grid is reduced to blade tip gradually by propeller hub, and blade tip place surface grids is triangle, and grid cell length of side size is about unit, 0.001D ， Jiang Grains place and is about 0.02D; At blade surface, set up boundary layer grid; Adopt structured grid to divide the computational fields of the peripheral regular shape of screw propeller; Tip whirlpool area grid is encrypted, blade surface adopts boundary layer grid to improve the forecast precision to tip vortex cavitation simultaneously, tip whirlpool area grid unit size is about 0.001D, boundary layer grid has 4 layers, its adjacent two layers aspect ratio is 1.1, ground floor grid cell height is about 0.001D, makes dimensionless group 20<y ⁺<300.

4. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as claimed in claim 1, is characterized in that: the full cavitation model in described step 2 sets and parameter is modified to:

Work as p<p _vtime, steam generation rate is:

$R_e=\mathrm{Ce}\frac{k}{\mathrm{\&gamma;}}{\mathrm{\&rho;}}_l{\mathrm{\&rho;}}_v\sqrt{\frac{2}{3}\frac{p_v-p}{{\mathrm{\&rho;}}_l}}(1-f_v)$

Work as p>p _vtime, vapour phase becomes liquid phase, obtains equally steam solidification rate R _c:

$R_c=\mathrm{Cc}\frac{k}{\mathrm{\&gamma;}}{\mathrm{\&rho;}}_l{\mathrm{\&rho;}}_v\sqrt{\frac{2}{3}\frac{p-p_v}{{\mathrm{\&rho;}}_l}}{f}_v$

Wherein, f _v=α _vρ _v/ ρ _mfor vapour phase massfraction, vaporization coefficient Ce=0.02 and condensation coefficient Cc=0.01 are empirical parameter.

5. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as claimed in claim 1, it is characterized in that: RNG k-ε turbulence model and parameter thereof in described step 2 are modified to: Renormalization Group turbulence model is RNG k-ε turbulence model, k equation and the ε equation of RNG k-ε turbulence model are respectively:

$\frac{\&PartialD;}{\&PartialD;t}\left({\mathrm{\&rho;}}_{m}k\right)+\frac{\&PartialD;}{{\&PartialD;x}_j}\left({\mathrm{\&rho;}}_m{\mathrm{ku}}_{\mathrm{mj}}\right)=\frac{\&PartialD;}{{\&PartialD;x}_j}\left[\left({\mathrm{\&alpha;}}_k\mathrm{\&mu;}\right)\frac{\&PartialD;k}{{\&PartialD;x}_j}\right]+G-{\mathrm{\&rho;}}_m\mathrm{\&epsiv;}$

$\frac{\&PartialD;}{\&PartialD;t}\left({\mathrm{\&rho;}}_m\mathrm{\&epsiv;}\right)+\frac{\&PartialD;}{{\&PartialD;x}_j}\left({\mathrm{\&rho;}}_m\mathrm{\&epsiv;}{u}_{\mathrm{mj}}\right)=\frac{\&PartialD;}{{\&PartialD;x}_j}\left[\left({\mathrm{\&alpha;}}_{\mathrm{\&epsiv;}}\mathrm{\&mu;}\right)\frac{\&PartialD;\mathrm{\&epsiv;}}{{\&PartialD;x}_j}\right]-R+C_{1\mathrm{\&epsiv;}}\frac{\mathrm{\&epsiv;}}{k}G-C_{2\mathrm{\&epsiv;}}{\mathrm{\&rho;}}_m\frac{{\mathrm{\&epsiv;}}^2}{k}$

In formula, Turbulent Kinetic dissipative shock wave (Turbulent Dissipation Rate) the a reciprocal of effective turbulent prandtl number of Turbulent Kinetic k and dissipative shock wave ε _k=a _ε=1.39; Model parameter C _{1 ε}=1.47, C _{2 ε}=1.68; Viscosity coefficient is μ=μ _t+ μ _m, μ _mfor mixed flow coefficient of viscosity; Revise turbulent viscosity coefficient μ _t=[ρ _v+ α _l ¹⁰(ρ _l-ρ _v)] C _μk ²/ ε, C _μ=0.085.

6. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as claimed in claim 1, is characterized in that: the numerical evaluation setting parameter of described step 3, comprises that the correlation parameter of working condition, boundary condition and numerical algorithm is set;

Working condition is mainly set screw propeller rotational speed, and environmental pressure and inflow velocity value, determine screw propeller dimensionless group, i.e. advance coefficient (J) and cavitation number (σ _n); For boundary condition, set, speed inlet boundary adopts inflow velocity value, and far field boundary condition adopts inflow velocity to set, and the top hole pressure at downstream pressure outlet interface is set to static pressure; Parameter setting in numerical algorithm: it is discrete that in Na Wei-Stokes (N-S) equation, convective term adopts Second-order Up-wind form, diffusion term adopts Using Second-Order Central Difference form discrete, velocity pressure coupling adopts the SIMPLE algorithm that is applicable to non-structured grid, uses pointwise Gauss-Seidel iterative discrete equation; Utilize the convergence of algebraic multigrid speed-up computation, for non-permanent calculating, adopt Sliding mesh computing technique, adopt second order accuracy discrete scheme, for the stability that guarantees that second order calculates, sub-relaxation factor is suitably reduced, the isoparametric sub-relaxation factor of pressure, momentum, vapour phase mark, Turbulent Kinetic, turbulence dissipation rate and turbulent flow stickiness is set as respectively: 0.25,0.6,0.2,0.7,0.7,0.9, mass conservation continuity (continuity) residual error convergence is three rank, and in equation, other physical quantity residual error convergence is quadravalence.

7. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as claimed in claim 1, it is characterized in that: the ad-hoc location A point of described step 6 is positioned at screw current radially r=0.5R and axial x=2R place, according to dimension conversion principle, adopt formula carry out nondimensionalization, the total pressure pulsation value that wherein Δ P is numerical result, ρ is fluid-mixing density, n is revolution speed of propeller, D position airscrew diameter; In non-permanent calculating, TIME STEP time step is set as T=0.0125T _p, T _pfor screw propeller swing circle, it is 30T that data relate to time span _p.

8. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as claimed in claim 1, is characterized in that: the low frequency spectrum lines in the power spectrum conversion in described step 7 comprises axle frequently, and frequently, three times of axles frequently and leaf frequency for two times of axles.

9. the cavitation noise feature method of estimation of calculating based on screw current pressure fluctuation as described in claim 1 or 4 or 5, it is characterized in that: the correction of turbulent viscosity coefficient in the correction of transformation ratio parameter in cavitation model and turbulence model is adopted to C language compilation, and recycling macro call (DEFINE_TURBULENT_VISCOSITY etc.) form embeds calculation procedure.