Underwater vehicle noise forecasting method based on linearization FE-SEA
By performing linearization in the FE-SEA method, the accuracy and efficiency problems of the prior art when dealing with nonlinear dynamic responses are solved, and more efficient and accurate noise forecasting of underwater navigation bodies is achieved.
Patent Information
- Application Number
- CN202510037170.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-09
- Publication Date
- 2025-05-27
AI Technical Summary
When handling nonlinear dynamic responses, the existing FE-SEA methods have limitations in accuracy and computational efficiency. Especially when facing problems such as large deformation and nonlinear contact, the prediction results may be biased or inaccurate.
The noise forecasting method based on linearized FE-SEA is adopted to solve the simulated outflow field through CFD turbulence, and linearized processing is performed, including linearization of nonlinear stiffness and damping, composite excitation processing, separation of random excitation and harmonic excitation, and precise setting of materials and loss factors.
It improves computing efficiency, reduces instability in the calculation process, and can accurately predict noise radiation within a wider frequency range, improving the accuracy and applicability of noise forecasting.
Smart Images
Figure CN120046528A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method for predicting the noise of an underwater vehicle, and particularly to a method for predicting the noise of an underwater vehicle based on linearized FE-SEA. Background Art
[0002] In the field of noise prediction of underwater vehicles, finite element analysis (FE) and statistical energy analysis (SEA) have become two important methods widely used. Finite element analysis is used to describe the vibration characteristics of underwater vehicles, especially the response under complex structures and dynamic loads, while statistical energy analysis is used to calculate the acoustic energy propagation and noise radiation caused by structural vibration. The FE-SEA hybrid method has been widely used in the noise prediction of underwater vehicles by combining the two, and has shown high accuracy and applicability, especially in the field of dealing with noise of different frequencies and radiation noise. However, in practical applications, there are some technical challenges, especially in dealing with non-linear dynamic responses.
[0003] During the operation of an underwater vehicle, it is usually subjected to various non-linear loads, such as turbulent effects, mechanical vibrations, etc. These non-linear factors have a significant impact on the vibration response of the structure and the generated noise. The existing FE-SEA methods usually default that the calculated system is a linear system, or are only applicable within a certain linear range. Therefore, when facing problems such as large deformations and non-linear contacts, the prediction results often deviate or are inaccurate. So under non-linear excitation, the existing methods have limitations in terms of accuracy and computational efficiency. Summary of the Invention
[0004] Object of the Invention: The object of the present invention is to propose a method for predicting the noise of an underwater vehicle based on linearized FE-SEA, which improves the computational efficiency and reduces the instability during the calculation process.
[0005] Technical Solution: The present invention includes the following steps:
[0006] External Flow Field Calculation: By solving CFD turbulence, the flow state around the underwater vehicle is simulated, and the calculation of the external flow field provides dynamic input for noise prediction;
[0007] Noise Prediction: Model establishment and pre-processing; Linearization processing: Linearization of non-linear stiffness and damping, and processing of composite excitation, including: Random excitation: Calculating the spectrum of random pressure fluctuations in the external flow field using a turbulence model; Harmonic excitation: Representing the harmonic excitation caused by the vibration of mechanical equipment as a sine function load, and mapping the non-linear excitation to a linear sub-model; Other condition settings and noise prediction.
[0008] The external flow field calculation specifically includes: model pre-processing, steady flow field calculation, unsteady flow field calculation, and wavenumber decomposition.
[0009] The pre - processing of the model includes: importing the geometric model for mesh generation and selecting the mesh type and size, setting the turbulent flow calculation model, boundary conditions, and initial conditions.
[0010] The steady - state flow field calculation is as follows: using the CFD method to simulate the pressure distribution and fluid velocity field on the surface of the vehicle.
[0011] The unsteady - state flow field calculation is as follows: analyzing transient turbulence and turbulence characteristics, calculating the fluid pressure varying with time, and converting it into acoustic excitation input through frequency - domain or time - domain transformation.
[0012] The wave - number decomposition specifically is: converting the unsteady pressure field into a frequency - domain distribution and analyzing the contribution of different wave - number components in the flow field to the noise.
[0013] The wave - number decomposition is to perform a two - dimensional double Fourier transform on the surface pressure field in two - dimensional space, and distinguish acoustic components and turbulent components according to the different propagation speeds and propagation direction characteristics of different waves.
[0014] The model establishment and pre - processing specifically are: performing mesh generation on the main structure of the vehicle, simulating the structural modal, vibration, and deformation behaviors, and performing statistical modeling on the secondary components and high - frequency parts of the vehicle, mainly focusing on the energy transfer and dissipation processes.
[0015] The setting of other conditions includes: host excitation, material, and loss factor settings. The main host excitation sources mainly include the vibrations of various mechanical power systems. By accurately modeling the frequencies and amplitudes of these excitation sources in the finite - element model, reliable input data can be provided for noise prediction. Material properties (such as elastic modulus, density, and damping, etc.) have an important impact on the vibration and noise radiation of the underwater vehicle. In this invention, by accurately inputting these parameters and combining linearization processing, it is ensured that the response characteristics of the material under different working conditions are consistent with the actual situation, thereby improving the accuracy of vibration response. At the same time, the loss factor (such as internal damping and vibration energy loss) is set through experimental data or empirical models and effectively simulates the energy dissipation in noise prediction, further improving the accuracy of noise prediction. By reasonably setting these parameters, this invention can provide support for the noise control and optimal design of the underwater vehicle.
[0016] The noise prediction includes: based on structural - acoustic theory, calculating the vibration - acoustic radiation in the mid - low frequency band, including structural modal frequencies, amplitude distributions, and acoustic powers; statistical energy analysis: based on the energy - flow theory, calculating the noise energy distribution in the high - frequency band and predicting the energy transfer between different subsystems; wave - number decomposition and spectrum analysis: combining the flow - field pressure and structural vibration, calculating the sound - power spectra and acoustic - field distributions in each frequency band; hybrid calculation and analysis: taking the high - frequency band noise output of each cabin section of the underwater vehicle as the prediction and providing noise control and optimization suggestions for the vehicle design.
[0017] Beneficial effects: The present invention has the following advantages:
[0018] (1) High calculation accuracy and high calculation efficiency: By linearizing the nonlinear system, while ensuring the calculation accuracy, the calculation efficiency is greatly improved. By linearizing at the nonlinear operating point, the complex numerical solution process in traditional nonlinear simulation is avoided, significantly reducing the calculation time and resource consumption. By accurately linearizing the nonlinear dynamic response, the present invention can more accurately capture the characteristics of noise sources caused by nonlinear effects, improving the accuracy of noise prediction. Compared with the possible errors in traditional methods, the linearized FE-SEA method proposed by the present invention can accurately predict noise radiation in a wider frequency range, thus providing more reliable data support for the noise control and acoustic optimization design of underwater vehicles;
[0019] (2) Wide applicability and high stability: Nonlinear dynamic response can lead to unstable calculation results of traditional methods, especially when the vibration amplitude of the system is large or the load is strong. The linearization process enables the system analysis and solution to remain stable even in the face of large-amplitude vibrations or complex nonlinear behaviors, effectively avoiding divergence or non-convergence phenomena during the solution process. This advantage enables the present invention to be widely applicable to the noise prediction of underwater vehicles under various working conditions, especially in complex environments in practical engineering;
[0020] (3) Simplify the nonlinear solution: Solving traditional nonlinear problems usually requires a large number of iterative calculations and complex algorithm support, while the linearized model transforms these problems into standard linear problems, simplifying the solution process. This not only greatly shortens the design and optimization cycle but also reduces the engineering implementation difficulty and cost brought by nonlinear processing;
[0021] (4) High engineering application value: By applying the linearized FE-SEA method to the noise prediction of underwater vehicles, comprehensive optimization of noise sources and vibration noise can be achieved. The implementation of this method makes the noise control of underwater vehicles more efficient and accurate, with extremely high engineering application value, capable of meeting the requirements of high-precision noise prediction and acoustic optimization design, and is widely applicable to the noise control and design of various underwater vehicles such as ships and submarines. Description of the drawings
[0022] Figure 1 is the flow chart of the present invention;
[0023] Figure 2 is the numerical calculation model diagram of this embodiment;
[0024] Figure 3 is the curve graph of the calculated value of the propulsion cabin noise. Detailed implementation manners
[0025] The present invention will be further described below with reference to the accompanying drawings.
[0026] Embodiment 1
[0027] As Figure 1 shown, the linearized FE-SEA underwater vehicle noise prediction method of this embodiment includes the following steps:
[0028] S1. External flow field calculation: The external flow field is an important noise source for underwater vehicles, and its calculation involves the hydrodynamic process of the interaction between water flow and the vehicle. Through CFD turbulence solving, the flow state around the underwater vehicle is accurately simulated. These flow disturbances generate noise during propagation, and the calculation of the external flow field provides accurate dynamic input for further noise prediction. Specifically, it includes the following steps:
[0029] S11. Model preprocessing
[0030] Perform preprocessing for flow field calculation in STAR-CCM+, including: importing the geometric model for mesh generation and selecting appropriate mesh types and sizes, setting the turbulence calculation model, boundary conditions (velocity inlet, pressure outlet, wall conditions) and initial conditions (initial flow velocity), and selecting an appropriate solver for computing resource configuration. And perform mesh independence verification to ensure the accuracy of the simulation model.
[0031] S12. Steady flow field calculation
[0032] Adopt the computational fluid dynamics (CFD) method to simulate the pressure distribution on the vehicle surface and the fluid velocity field, which is suitable for the analysis of flow characteristics under stable operating conditions. This part of the calculation provides basic data for eddy current noise and surface pressure excitation.
[0033] S13. Unsteady flow field calculation
[0034] Analyze transient turbulence and turbulence characteristics, and calculate the fluid pressure varying with time. The unsteady flow field data is converted into acoustic excitation input through frequency domain or time domain conversion, which is suitable for the calculation of mid-high frequency flow-induced noise.
[0035] S14. Wavenumber decomposition
[0036] Convert the unsteady pressure field into a frequency domain distribution, and analyze the contribution of different wavenumber components in the flow field to noise. Wavenumber decomposition is to perform a two-dimensional double Fourier transform on the surface pressure field in two-dimensional space, and distinguish acoustic components and turbulence components according to the different propagation speeds and propagation direction characteristics of different waves.
[0037] S2. Noise prediction
[0038] S21. Model Establishment and Preprocessing
[0039] Model establishment: Mesh generation is performed on the main structures of the vehicle (such as the hull and compartments) to simulate the structural modes, vibrations, and deformation behaviors, mainly for calculating the acoustic characteristics in the medium and low frequency bands. SEA model construction: Statistical modeling is carried out on the secondary components and high-frequency parts of the vehicle, mainly focusing on the energy transfer and dissipation processes.
[0040] S22. Linearization processing, specifically including:
[0041] S221. Linearization of Nonlinear Stiffness and Damping
[0042] When a nonlinear dynamic system responds, it exhibits complex characteristics such as large displacements, contact nonlinearities, and material nonlinearities. The linearization process transforms the nonlinear behavior of the system into an approximate linear model through approximation methods. For example, the nonlinear stiffness is linearized using Taylor expansion near certain operating points to obtain a linear stiffness matrix.
[0043] S222. Treatment of Composite Excitations
[0044] An unmanned underwater vehicle (UUV) is often subjected to multiple actions of random turbulent noise and harmonic excitations during operation. The following separation treatment scheme is adopted:
[0045] The linearized finite element - statistical energy hybrid analysis (FE - SEA) method proposed in this embodiment is applicable to nonlinear node combination systems excited by random and harmonic loads. In each system, each plate component is modeled. Linearization processing is performed according to the load type. In the case of random loading, statistical linearization (SL) is adopted, while in the case of harmonic loading, the harmonic balance method (HBM) is adopted.
[0046] Random excitation: Calculate the random pressure pulsation spectrum of the external flow field using a turbulence model.
[0047] Harmonic excitation: Represent the harmonic excitation caused by the vibration of mechanical equipment as a sine function load.
[0048] S223. Mapping Nonlinear Excitations to Linear Sub - models
[0049] For complex nonlinear behaviors (such as the nonlinearity of springs and contacts), this method processes by decomposing the nonlinear part of the system into a linear part and a nonlinear perturbation part. By linearizing the influence of the nonlinear part, the overall system is approximately a linear system during the calculation process.
[0050] Linearized FE-SEA Hybrid Method: The finite element method (FE) is usually used to solve the dynamic response of structures, while the statistical energy method is used to describe the propagation and distribution of acoustic energy. Through linearization, the nonlinear behaviors of the structural part (such as large deformations, material nonlinearities, etc.) can be effectively simplified into a linear model, and combined with the SEA method for overall system analysis. Specifically, after the vibration response of the structure is linearized, the transfer and calculation of acoustic energy are then carried out through SEA. Through these linearization steps, the analysis of nonlinear dynamic systems is effectively simplified, enabling the handling of complex problems under random and harmonic loads in a more efficient computational framework.
[0051] S23. Other Condition Settings: Host Excitation, Material, Loss Factor Settings
[0052] The main host excitation sources include the vibrations of various mechanical power systems. By accurately modeling the frequencies and amplitudes of these excitation sources in the finite element model, reliable input data can be provided for noise prediction. Material properties (such as elastic modulus, density, and damping, etc.) have an important impact on the vibration and noise radiation of underwater vehicles. In this invention, by accurately inputting these parameters and combining with linearization, it is ensured that the response characteristics of the material under different working conditions are consistent with the actual situation, thereby improving the accuracy of the vibration response. At the same time, the loss factor (such as internal damping and vibration energy dissipation) is set through experimental data or empirical models, and the dissipation of energy is effectively simulated in noise prediction, further improving the accuracy of noise prediction. By reasonably setting these parameters, this invention can provide support for the noise control and optimal design of underwater vehicles.
[0053] S24. Noise Prediction
[0054] Finite Element Method Calculation (FE): Based on structural acoustics theory, calculate the vibration and sound radiation in the mid- and low-frequency bands, including structural modal frequencies, amplitude distributions, and sound power. Statistical Energy Analysis (SEA): Based on the energy flow theory, calculate the noise energy distribution in the high-frequency band and predict the energy transfer between different subsystems. Wavenumber Decomposition and Spectrum Analysis: Combine the flow field pressure and structural vibration to calculate the sound power spectrum and sound field distribution in each frequency band. Hybrid Calculation and Analysis. Output the high- and mid-frequency band noise of each cabin section of the underwater vehicle as a prediction, providing noise control and optimization suggestions for the design of the vehicle.
[0055] By linearizing the non - linear factors of the underwater vehicle system, the present invention can effectively simplify the solution process of non - linear dynamic problems while having sufficient computational accuracy, especially when facing large - amplitude vibrations or complex non - linear effects. The linearization process transforms the analysis of the non - linear system into a linear analysis, which not only improves the computational efficiency but also reduces the instability during the calculation process. Especially in noise prediction, the linearized model can accurately describe the noise sources and acoustic characteristics caused by non - linear effects, thus providing a more reliable basis for the noise control and design optimization of underwater vehicles.
[0056] Embodiment 2
[0057] For Figure 2 the numerical calculation model of the underwater vehicle, for components with a relatively large average number of modes, such as the outer shell wall, SEA modeling is used, while for components with a relatively small average number of modes, such as the connection of plates, FE modeling and analysis are used to form a complete calculation model. The external flow field calculated by STRA - CCM is loaded into the numerical model, and main engine excitation, other parameter settings, etc. are added to predict the vibration and noise in the medium - high frequency band within the scope of engineering applications.
[0058] 1. The simulation analysis of the vehicle noise in the medium - high frequency band is based on the finite element - statistical energy analysis method to establish the entire FE - SEA simulation model for acoustic calculation. The acoustic load is the convective pressure and acoustic pressure on the model obtained from the external flow field calculation.
[0059] 2. Import the finite element model of the vehicle. For the large external structure part, the SEA subsystem is used for modeling; the ribs, etc. are separately decomposed and co - node connection processing is carried out at the same time to ensure the complete connection between different structures, and this part of the structure is modeled using the FE subsystem. Due to the transmission of the noise path, in order to fully consider the influence of medium - frequency structure vibration on sound propagation, the same FE subsystem modeling is also used. As Figure 3 shown are the noise values of each frequency band calculated for the propulsion cabin.
Claims
1. A method for predicting underwater vehicle noise based on linearized FE-SEA, characterized in that: The following steps are involved: External flow field calculation: Through CFD turbulence solution, the flow state around the underwater vehicle is simulated. The calculation of the external flow field provides dynamic input for noise prediction; Noise prediction: model building and pre-processing; linearization processing: linearization of nonlinear stiffness and damping, processing of composite excitation, including: random excitation: using turbulence model to calculate the random pressure pulsation spectrum of the external flow field; Harmonic excitation: Represent the harmonic excitation caused by mechanical equipment vibration as a sinusoidal function load, map the nonlinear excitation to the linear sub-model; other condition settings and noise prediction.
2. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 1, characterized in that: The external flow field calculation specifically includes: model pre-processing, steady flow field calculation, unsteady flow field calculation and wave number decomposition.
3. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 2, characterized in that: The model pre-processing includes: importing the geometric model to perform mesh division and selecting the mesh type and size, setting the turbulence calculation model, boundary conditions and initial conditions.
4. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 3 is characterized in that: The steady flow field calculation is as follows: using the CFD method to simulate the pressure distribution and fluid velocity field on the surface of the navigation body.
5. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 4, characterized in that: The unsteady flow field calculation is as follows: analyzing transient turbulence and turbulence characteristics, calculating the fluid pressure that varies with time, and converting it into acoustic excitation input through frequency domain or time domain.
6. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 5, characterized in that: The wave number decomposition is specifically: converting the unsteady pressure field into a frequency domain distribution, and analyzing the contribution of different wave number components in the flow field to the noise.
7. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 2, characterized in that: The wave number decomposition is to perform a two-dimensional double Fourier transform on the surface pressure field in the two-dimensional space, and distinguish the acoustic component and the turbulent component according to the different propagation speeds and propagation direction characteristics of different waves.
8. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 1, characterized in that: The model establishment and pre-processing are specifically as follows: meshing the main structure of the vehicle, simulating the structural modes, vibration and deformation behaviors, and statistically modeling the secondary components and high-frequency parts of the vehicle, focusing mainly on the energy transfer and dissipation processes.
9. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 1, characterized in that: The other condition settings include: host excitation, material, and loss factor settings.
10. The method for predicting underwater vehicle noise based on linearized FE-SEA according to claim 1, characterized in that: The noise prediction includes: based on the structural acoustics theory, calculating the vibration sound radiation in the medium and low frequency bands, including the structural modal frequency, amplitude distribution and sound power; statistical energy analysis: based on the energy flow theory, calculating the noise energy distribution in the high frequency band and predicting the energy transfer between different subsystems; wave number decomposition and spectrum analysis: combining the flow field pressure and structural vibration to calculate the sound power spectrum and sound field distribution in each frequency band; hybrid calculation and analysis: using the noise output in the medium and high frequency bands of each compartment of the underwater vehicle as a prediction.