Modularized inverse design method for metamaterial acoustic barrier based on targeted noise attenuation

Abstract

The application discloses a metamaterial sound barrier modular reverse design method based on targeted noise attenuation, comprising the following steps: carrying out noise measurement on a target noise source; collecting, classifying and exporting the measured noise; filtering and Fourier transforming the collected noise to export an explicit target image and a target function, and importing the target function into a conditional function; controlling a wave finite element program through a particle algorithm to iteratively optimize and solve a band gap range; at the beginning of iteration, N-dimensional position variable initial values and N-speed variable initial values are randomly generated, wherein each position variable corresponds to a current structure evolution result and an evolution direction, the wave finite element method is combined to evaluate the wave dispersion characteristics of each candidate structure, the optimal position of each individual is discriminated through the conditional function, and the optimal position of the population in the optimal individual is discriminated through the target function; and the application has significant advantages in design efficiency, equivalent volume, attenuation frequency band and attenuation amount.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of vibration and noise reduction, and particularly relates to a metamaterial sound barrier modular reverse design method based on targeted noise attenuation. BACKGROUND

[0002] With the rapid development of high-speed rail, subway and other transportation fields, the residents along the railway are increasingly disturbed by wheel-rail noise, which has gradually become one of the most important factors affecting the health index of residents in terms of physiology, psychology and other aspects [1] . For the prevention and control of noise pollution, the relevant government departments have introduced relevant noise standards and control measures. Taking a typical noise-prone section of the residential area along the railway as an example, the noise control means mainly include: (1) taking measures at the sound source, such as using elastic wheels, sound-absorbing wheels, heavy welded long rails, etc.; (2) taking measures on the sound propagation path, such as setting up sound barriers to block the propagation of noise; (3) taking protective measures at the receiving end, such as setting up sound-absorbing layers, sound-insulating windows, etc. By comparing the technical requirements of the three methods, it can be known that noise reduction at the sound source is difficult to implement and accurately control, noise reduction at the receiving end is often too passive, and the cost of both is high [2] . Therefore, cutting off the propagation path is the most practical, that is, setting up sound barriers at noise-sensitive places along the railway to effectively block the propagation of noise. However, the current railway sound barriers usually adopt the plug-in or integral type sound barriers with sound-absorbing medium [3] , which has a certain noise reduction effect, but does not solve the root problem of noise generation, that is, the essence of noise is the coupling effect of elastic waves and environmental medium [4] . Conventional or improved damping and sound-absorbing materials cannot avoid the propagation nature of elastic waves or sound waves, so they cannot overcome the limitations of narrow frequency range and low noise attenuation [5] . Therefore, the elastic wave band gap sound isolation characteristics of acoustic metamaterials with unconventional material properties have injected new vitality into the research and application of wideband-high-attenuation sound barriers.

[0003] In the passive method, the Helmholtz resonator (HR) is usually a resonant unit composed of a sealed cavity and a narrow neck, forming an equivalent spring-mass system, which can achieve noise attenuation. For example, the document [6] proposed an array of several Helmholtz resonators with subwavelength size, and research found that the structure has a negative equivalent bulk modulus near the resonance frequency, and experimental verification was carried out. However, this type of resonator is only effective at its single resonance peak, with a very narrow frequency band. Based on this, many scholars have nested open rings or spiraled the neck [7] , adjusted the combined size of the Helmholtz resonator and the plate [8]But still not meet through a general, efficient, convenient intelligent design method, can realize the optimization design of Helmholtz resonator in a certain application scene, so for any noise scene to design acoustic metamaterials, realize the target of targeted noise reduction, become more urgent demand; at the same time, the development of intelligent optimization algorithm and machine learning [9-10] Make the properties of various structures can be targeted optimization design.

[0004] A kind of acoustic metamaterial barrier design method for transformer noise reduction developed by Institute of Electrical Engineering, Chinese Academy of Sciences provides an acoustic metamaterial barrier design method for transformer noise reduction, first, the parameters of acoustic metamaterial barrier (1) are equivalent to a single uniform material plate, then the transformer noise reduction model is established, the multi-physical field coupling calculation is carried out, the average sound pressure of the noise reduction target area is taken as the target quantity, the mass density and elastic modulus of the equivalent acoustic metamaterial barrier are inversed, and finally the parameters of acoustic metamaterial unit, the size of acoustic metamaterial unit and the mass density and elastic modulus of the material are optimized.

[0005] This method only simplifies the optimization design of mass and density to single, and does not highlight the advantages of acoustic material; the optimization design technology for resonator is only effective on its single resonance peak, and the frequency band is very narrow. The commonly used method is to replace the traditional model by nesting open ring or neck spiralization, adjusting the joint size of Helmholtz resonator and plate and other measures to improve the sound attenuation characteristics in the limited space. But still not meet through a general, efficient, convenient intelligent design method, can realize the optimization design of Helmholtz resonator in a certain application scene, so for any noise scene to design acoustic metamaterials, realize the target of targeted noise reduction, become more urgent demand.

[0006] [1] Lv J, Jiang Y. Mechanism Analysis of the Impact on Chinese Urban Rail Transit Construction on the Surrounding Real Estate Prices. Applied Mechanics & Materials, 2012, 178-181: 1866-1869.

[0007] [2]Yan L.Study on the Impact of High-speed Rail on Regional Economic Development[J].Modern Industrial Economy and Informationization,2018.

[0008] [3]HE W,HE K W,ZOU C,et al.Experimental noise and vibration characteristics of elevated urban rail transit considering the effect of track structures and noise barriers[J].Environmental Science and Pollution Research,2021,28:45903-45919.

[0009] [4]Wang P,Yi Q,Zhao C,Xing M.Elastic wave propagation characteristics of periodic track structure in high-speed railway.Journal of Vibration and Control,2019,25(3):517-528.

[0010] [5]LI Xiaozhen,ZHAO Qiu-chen,ZHANG Xun,YANG Dewang.Testing and analysis of noise reduction effect of semi-enclosed sound barrier for high-speed railroad[J].Journal of Southwest Jiaotong University,2018,53(04):661-669+755.

[0011] [6]Fang N,Xi D,Xu J,et al.Ultrasonic metamaterials with negative modulus[J].Natural Materials.2006,5:452-456.

[0012] [7] S.-H. Park, J. Sound Vib. 2013, 332, 4895.

[0013] [8] Gebrekidan S B, Kim H J, Song S J. Investigation of Helmholtz resonator-based composite acoustic metamaterial [J]. Applied Physics, 2019, 125 (1): 65-72.

[0014] [9] Garland A P, White B C, Jensen S C, et al. Pragmatic generative optimization of novel structural lattice meta-materials with machine learning [J]. Materials & Design, 2021, 203: 109632.

[0015]

[10] Wang L, Liu H T. Parameter optimization of bidirectional re-entrant auxetic honeycomb meta-material based on genetic algorithm [J]. Composite Structures, 2021, 267: 113915.

[0016] Structural Dynamics, 2012, 41(5): 987-1000.

[0017]

[11] Zhang C, Liu GQ, Zhao XZ, Li CL, Lu ZM, Wang TZ. A design method of acoustic metamaterial barrier for transformer noise reduction [P]. Beijing: CN109117578A, 2019-01-01. SUMMARY

[0018] To solve the problems existing in the prior art, the purpose of the present application is to provide a metamaterial acoustic barrier modular reverse design method based on targeted noise attenuation. The present application extracts the objective function and constraint conditions for any engineering noise background, and then combines the fusion algorithm based on the generalized particle swarm algorithm and the wave finite element method to reversely design the required acoustic metamaterial cell, and verifies it combined with acoustic test, so as to realize the targeted noise attenuation.

[0019] In order to achieve the above object, the technical scheme adopted by the present application is: a metamaterial sound barrier modular reverse design method based on targeted noise attenuation, comprising the following steps:

[0020] Step 1, noise measurement is performed on the target noise source;

[0021] Step 2, the measured noise is collected, classified and exported by the acquisition instrument;

[0022] Step 3, after filtering and Fourier transform of the collected noise, an explicit target image and a target function are exported, and the target function is imported into the conditional function in step 4;

[0023] Step 4, the wave finite element program is controlled by the particle algorithm, and the band gap range is solved by iterative optimization; N-dimensional position variable initial values and N-speed variable initial values are randomly generated at the beginning of iteration, wherein each position variable corresponds to a current structure evolution result and evolution direction, the wave dispersion characteristics of each candidate structure are evaluated by combining the wave finite element method, and the individual optimal position gbest_i is discriminated through the conditional function, and the population optimal position Gbest in the individual optimal is discriminated through the target function, that is, the optimal cell of the acoustic metamaterial.

[0024] As a further improvement of the present application, the following steps are further included:

[0025] Step 5, the obtained optimal cell is 3D printed, and the actual acoustic performance of the obtained optimal cell is verified through experiments combined with the restriction conditions of the application scene.

[0026] As a further improvement of the present application, in step 1, the acoustic sensor or sound pressure microphone is used to measure the noise of the application scene requiring noise reduction, and the target noise source includes mechanical noise, factory noise, building secondary noise, rail transit noise and vehicle noise.

[0027] As a further improvement of the present application, in step 4, the wave finite element program is controlled by the particle algorithm, and the band gap range is solved by iterative optimization, which is specifically as follows:

[0028] The stiffness and mass matrix under the finite element modeling are extracted, and the Matlab wave analysis program is developed, the acoustic performance of the acoustic metamaterial structure is calculated, and finally the performance information of the structure is controlled evolution by the generalized particle swarm algorithm program.

[0029] As a further improvement of the present application, the generalized particle swarm algorithm program is specifically as follows:

[0030] Suppose there are n particles forming a community, where the i-th particle is represented as a D-dimensional vector. In each iteration, the particles update the structure by tracking two extrema: the first is the optimal solution found by the particle itself, called the individual extremum; the second is the optimal solution found by the entire population so far, called the global extremum.

[0031] The beneficial effects of this invention are:

[0032] 1. Compared with the secondary development of the genetic algorithm (GA) for finite element analysis, the generalized particle swarm fusion wave finite element algorithm (G-PSO-WF) of this invention has more convenient logic and faster convergence stability, making it more suitable for engineering design applications of acoustic metamaterials. It provides a powerful tool for the reverse design and targeted optimization process of acoustic metamaterials.

[0033] 2. This invention yields a basic configuration of a spatially misaligned Helmholtz resonant acoustic metamaterial (SIHRAM) to improve sound attenuation in confined spaces. A self-programmed G-PSO program is applied to the targeted optimization design of acoustic metamaterial sound barriers, incorporating noise source information as the objective function for particle swarm optimization to obtain the metamaterial's basic cell with the optimal absolute bandgap frequency band. Compared to semi-analytical methods (SAM), traditional finite element methods (FM), and wave finite element methods (WFM), the Generalized Particle Swarm Optimization-Wave finite element method (G-PSO-WF) significantly reduces computational time and cost. Compared to neural network inverse design methods (NNDM), G-PSO-WF exhibits higher stability, is easier to operate, and demonstrates significant effectiveness. Attached Figure Description

[0034] Figure 1 This is a flowchart of the AM-MRD intelligent target design and testing process in an embodiment of the present invention;

[0035] Figure 2 The diagram shows the acoustic radiation and acoustic power flow streamlines of the three-layer B_SIHRAM structure under the characteristic wave mode in this embodiment of the invention.

[0036] Figure 3 This is a diagram showing the actual setup of the test in an embodiment of the present invention;

[0037] Figure 4 This is a comparison chart of the test results in the embodiments of the present invention. Detailed Implementation

[0038] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0039] Example

[0040] like Figure 1As shown, a metamaterial sound barrier modular reverse design method (AM-MRD) based on targeted noise attenuation, AM-MRD has five core steps, as follows:

[0041] Step one, for the noise of the target noise source, that is, for the application scenarios such as mechanical noise, factory noise, building secondary noise, rail transit noise, vehicle noise, etc. which need to be noise reduced, use acoustic sensors or sound pressure microphones for testing;

[0042] Step two, collect, classify and export the measured noise by the acquisition instrument;

[0043] Step three, filter and Fourier transform programming post-processing by computer processing software, and export the explicit target image and target function. The target function in step three is in the form of code, which is imported into the conditional function in step four, that is, the AMMRD input end is completed;

[0044] Step four, the core algorithm layer G-PSO-WF of AM-MRD is a fusion algorithm based on generalized particle swarm optimization and wave finite element method: the particle swarm optimization is a search algorithm based on group cooperation developed by simulating the foraging behavior of bird flock, and its basic idea is to find the optimal solution through the cooperation and information sharing between individuals in the group. The above joint process is divided into three parts: wave analysis program, finite element analysis and generalized particle swarm optimization control optimization, which is used for the structure evolution and targeted optimization of the acoustic performance of metamaterials. The targeted optimization part extracts the stiffness and mass matrix under the traditional finite element modeling in COMSOL Multiphysics, and combines the wave analysis program in Matlab for secondary development and calculation, which can quickly calculate the acoustic performance of acoustic metamaterial structure. Finally, with the help of COMSOL Multiphysics with MATLAB window, the performance information of the structure is controlled and evolved by the generalized particle swarm optimization algorithm program, so as to achieve the purpose of efficient and fast optimization. The reverse design process of G-PSO-WF randomly generates N-dimensional position variable initial value and N-speed variable initial value at the beginning of iteration, where each position variable corresponds to a current structure evolution result and evolution direction. The initialization of the algorithm is a group of random particles (random solution), and then the optimal solution is found through iteration. There are n particles in a colony, where the i-th particle is represented as a D-dimensional vector. In each iteration, the particle updates the structure by tracking two extreme values: the first one is the optimal solution found by the particle itself, which is called individual extreme value; the second one is the optimal solution found by the whole population at present, which is called global extreme value. In addition, the wave finite element program can be controlled by the particle algorithm to continuously iterate and optimize the solution of the band gap range. The wave dispersion characteristics of each candidate structure are evaluated by combining the wave finite element method, and the individual optimal position gbest_i is determined by the conditional function through the cycle, and the population optimal position Gbest in the individual optimal position is determined by the objective function, which is the best scheme for the design of acoustic metamaterial cell;

[0045] Step five, the output end of AMTD, first 3D print the optimized cell in step four, then select the acoustic sensor room test of acoustic impedance tube or the sound attenuation test in the anechoic chamber according to the restriction conditions of the application scene, so as to verify the actual acoustic performance of the metamaterial cell designed by AMTD, and provide the basis or premise for the subsequent engineering application of the metamaterial.

[0046] The embodiment utilizes an intelligent acoustic metamaterial modular reverse design method (AM-MRD), which can reverse design an acoustic metamaterial cell that meets the target noise attenuation by only inputting the objective function of any noise source and the constraint conditions of the application scenario. The AM-MRD input is the actual test noise index, the core algorithm layer G-PSO-WF is a fusion algorithm based on the generalized particle swarm optimization algorithm and the wave finite element method, which is used for structure evolution and targeted optimization of the acoustic performance of the metamaterial, and the output is the optimized metamaterial cell and its sound barrier structure obtained by reverse design. Taking the actual noise background of rail transit environment as an example, a spatially disoriented Helmholtz resonant acoustic metamaterial cell B_SIHRAM is obtained according to the above process, and the structure evolution process and its acoustic performance effectively show the targeted optimization effect of the G-PSO-WF algorithm. The experimental and simulation results prove the excellent noise reduction performance of B_SIHRAM, and the 438-2087Hz effective bandgap characteristic of B_SIHRAM can perfectly block the 500-1500Hz main frequency band of the measured wheel-rail noise, and the measured sound attenuation at the bandgap valley is 57.8dB. In addition, the comprehensive comparison shows that the acoustic metamaterial designed based on AM-MRD has significant advantages in design efficiency, equivalent volume, attenuation frequency band and attenuation amount.

[0047] To further illustrate the effect of the embodiment, in terms of noise reduction effect, the sound radiation field and the sound wave power flow stream under the key wave mode of Mode 1 to Mode 6 are extracted respectively, and the results are shown in Figure 2 It can be found that in the passband frequency range, such as Figure 3 the sound wave power flow at the frequencies of 320Hz and 2090Hz in (a) and (f) of Figure 3 can freely pass through the metamaterial barrier, and the sound pressure level difference between the input and output ends is small; while in the bandgap frequency range of 500Hz, 1000Hz, 1500Hz and 2000Hz shown in (b)-(e) of Figure 3 , the sound radiation field intensity of the output end after the sound wave passes through the B_SIHRAM barrier is much smaller than that of the input end, and due to the bandgap impedance effect of the metamaterial, the sound power flow forms "sound mode resonance", "sound wave reflection" or "sound wave spiral wrapping", so it is difficult to pass through the super barrier, and appears "Stop" state; and with the increase of the number of B_SIHRAM layers passed by the sound wave, the sound radiation field intensity of the output end continuously attenuates.

[0048] In terms of design efficiency, the computer system of the AMTD co-simulation environment is Win10 64-bit; the processor is Intel(R) Core(TM) i7-126KFU CPU@1.70GHz 2.40GHz; the RAM is 128G; the solid-state drive is 512G; and the graphics card is GPU 3090Ti. In the AMTD co-simulation process, given that each information update involves 50 particle swarms and each particle undergoes 50 iterations in 12 dimensions, a large amount of finite element parameter adjustment and modeling calculation is required. The actual test time cost is shown in Table 1. Under the same complex working conditions, the AMTD co-simulation method can save at least 230 hours of time, and the design efficiency is improved by 12.5 times, demonstrating the high efficiency of this targeted design method.

[0049] Table 1 Comparison of time between the co-simulation method and the finite element method under the same working conditions.

[0050]

[0051] test:

[0052] The specific setup of the test is as follows: Figure 3 As shown, Figure 3 In the diagram, (a) the sample, (b) the overall setup for on-site testing, (c) the 3D-printed metamaterial cell and its details, (d) the input end of the impedance tube, (e) a schematic diagram of the sample location, (f) the output end of the impedance tube, (g) the porous sound-absorbing material at the port, (h) the high-precision INV acoustic sensor microphone, and (i) the INV3060S signal acquisition instrument. Based on the above approach, relevant metamaterial specimens and a customized thick-walled acoustic impedance tube were printed using 3D printing technology. Liquid sound-absorbing material (LASD) was added to the inside of the impedance tube beforehand to minimize the impact of sound transmission within the tube. Since the acoustic performance of this structure depends on the variation of the two-dimensional facade structure, the 3D structure printing height was selected according to actual needs and limitations. The built-in sound source, computer, acquisition instrument, data cable, four acoustic sensor microphones, and porous sound-absorbing material at both ends of the impedance tube were also arranged. Before the sound insulation test begins, the thickness of the test sample and environmental parameters are set in the software, the speaker is calibrated, the sound pressure sensor microphone is calibrated, and the sound pressure values ​​at the four sound pressure sensors in the sound source tube and the receiving tube are obtained through the test. The transmission loss of the test sample can then be calculated using equations (1) to (4).

[0053] Without considering the effect of time, the sound pressure values ​​at four test points under approximate plane wave conditions, measured by a sound pressure sensor, are p1, p2, p3, and p4, respectively, and their expressions are as follows:

[0054]

[0055] Solving the system of equations simultaneously yields four undetermined coefficients:

[0056]

[0057] Where A and B are the complex sound pressure amplitudes of the incident and reflected sound waves in the source tube, respectively; C and D are the complex sound pressure amplitudes of the incident and reflected sound waves in the receiving tube, respectively; and d is the thickness of the test sample. Here, the sound pressure transmission coefficient t is defined as... p The ratio of the transmitted wave amplitude C to the incident wave amplitude A can be expressed as:

[0058]

[0059] Clearly, based on the definition of transmission loss, the sound insulation of the test sample can be expressed as:

[0060] TL = -10log(t) p (4)

[0061] from Figure 4 As shown in (a) (a is a comparison of the indoor test and simulation results for this structure), the STL trend, peak frequency, and overall attenuation of the four test results are highly consistent with the simulation results. The attenuation effect is particularly significant in the 439.0-2087.6Hz bandgap range, with a trough reaching 57.8dB. However, it is not difficult to find that while the STL curve trends of the tests and simulations are consistent, the sound attenuation in certain frequency bands, especially at the peak, differs significantly. The reasons for the deviation are as follows:

[0062] (1) The sealing state and installation conditions of the 3D printed test sample in the experiment will have a significant impact on the test results, resulting in a reduction in the acoustic attenuation of the superstructure due to gaps or acoustic diffraction noise; while the IAB or ASCB boundary fields selected in the simulation have a certain degree of idealization assumptions.

[0063] (2) The precision and parameters of materials used in 3D printing (such as nonlinear damping) can lead to deviations in the peak effect.

[0064] However, in summary, such as Figure 4 As shown in (b) (b is a comprehensive comparison of B_SIHRAM with CHRA and MFRAM), the experimental results and simulation results of the supercell sample have good consistency overall.

[0065] The embodiments described above are merely illustrative of specific implementations of the present invention, and while the descriptions are detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A metamaterial acoustic barrier modular inverse design method based on targeted noise attenuation, characterized in that, Comprising the following steps: Step 1, noise measurement on target noise source; Step 2, collecting, classifying and exporting the measured noise by the acquisition instrument; Step 3, filtering and Fourier transforming the collected noise to export the explicit target image and target function, and importing the target function into the conditional function in step 4; Step 4, controlling the wave finite element program by the particle algorithm, and iteratively optimizing to solve the band gap range; N-dimensional position variable initial values and N speed variable initial values are randomly generated at the beginning of iteration, wherein each position variable corresponds to a current structure evolution result and evolution direction, the wave dispersion characteristics of each candidate structure are evaluated by combining the wave finite element method, and the individual optimal position gbest_i is discriminated through the conditional function, and the population optimal position Gbest in the individual optimal is discriminated through the target function, that is, the optimal cell of the acoustic metamaterial; In step 4, the wave finite element program is controlled by the particle algorithm, and the band gap range is iteratively optimized to solve the band gap range as follows: The stiffness and mass matrix under finite element modeling are extracted, and the wave analysis program of Matlab is combined for secondary development to calculate the acoustic performance of the acoustic metamaterial structure, and finally the performance information of the structure is controlled evolution by the generalized particle swarm optimization program. The generalized particle swarm optimization program is as follows: There are n particles in a colony, and the i-th particle is represented as a D-dimensional vector In each iteration, the particle updates the structure by tracking two extreme values: the first is the optimal solution found by the particle itself, called individual extreme value; The second is the optimal solution found by the whole population at present, called global extreme value.

2. The metamaterial acoustic barrier modular inverse design method based on targeted noise attenuation of claim 1, wherein, Further comprising the following steps: Step 5, 3D printing the obtained optimal cell, and combining the restriction conditions of the application scene to verify the actual acoustic performance of the obtained optimal cell through experiments.

3. The metamaterial acoustic barrier modular inverse design method based on targeted noise attenuation according to claim 1 or 2, characterized in that, In step 1, the noise of the application scene requiring noise reduction is measured by using an acoustic sensor or a sound pressure microphone, and the target noise source includes mechanical noise, factory noise, building secondary noise, rail transit noise and vehicle noise.

Images