CN115762685B - Sound transmission layer for reducing acoustic transmission loss of dissimilar material interface and optimization method - Google Patents
Sound transmission layer for reducing acoustic transmission loss of dissimilar material interface and optimization method Download PDFInfo
- Publication number
- CN115762685B CN115762685B CN202211543461.4A CN202211543461A CN115762685B CN 115762685 B CN115762685 B CN 115762685B CN 202211543461 A CN202211543461 A CN 202211543461A CN 115762685 B CN115762685 B CN 115762685B
- Authority
- CN
- China
- Prior art keywords
- layer
- sound
- acoustic transmission
- transmission loss
- dielectric
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The invention discloses a sound transmission layer for reducing acoustic transmission loss of a dissimilar material interface and an optimization method, and belongs to the field of optimization algorithms. The invention applies the concept of digital materials to disperse the whole optimization space into small material units, thereby parameterizing the whole design space and facilitating the optimization of the whole design space. The method adopts machine learning to obtain the prediction function of the relation between the digital material configuration and the average transmission loss, so that the prediction function is used for replacing finite element analysis in a genetic algorithm, and the calculation time is greatly saved. The method can design a more novel configuration, and is particularly suitable for designing the sound transmission layer in a low-frequency region.
Description
Technical Field
The invention belongs to the field of applied acoustics, and particularly relates to a sound transmission layer for reducing acoustic transmission loss of a dissimilar material interface and an optimization method.
Background
Acoustic transmission loss (STL) is mainly caused by impedance mismatch of the acoustically conductive medium without dissipation of the medium. The sound is reflected at the interface of dissimilar materials with mismatched impedance, resulting in less transmitted acoustic energy than incident acoustic energy, and the energy loss caused by reflection is the acoustic transmission loss. Acoustic transmission loss is a critical issue in applications such as acoustic sensor, transducer, and sound absorbing material design. For example, transmission losses between the sensor and the material being measured can cause the measured acoustic pressure to be inaccurate; also, the energy available from the acoustic transducer will be significantly less than the input acoustic energy; the absorption coefficient of the material is reduced by the transmission loss of the sound absorption material (such as underwater anechoic tiles) due to the rising chemistry caused by reflection at the incident interface. An effective way to reduce acoustic transmission loss is to lay an acoustic cover layer, i.e. an acoustically transparent layer, between two media with mismatched impedances. The acoustically transparent layer is typically composed of a composite material or structure with graded resistance, such as a wedge structure, a multi-layer structure with graded resistance, and a composite material with graded particles. Such a graded-impedance composite material allows for a uniform transition in impedance between the original impedance mismatched materials, thereby reducing transmission losses and is therefore also referred to as an impedance matching layer.
Although these acoustically transparent layers have been studied intensively, the optimal design of the acoustically transparent layers is not very clear. In terms of theoretical analysis, existing theoretical methods, including transfer matrix methods and non-uniform transmission line theory, are applicable only to some simple composite materials (e.g., multi-layer plates with discrete impedance, media with exponential impedance distribution). In terms of numerical computation, the conventional numerical optimization computation method is generally limited to human experience, and the optimization can only optimize geometric parameters of a specific composite material model, and other possible composite material models cannot be explored, so that an optimal solution in a true sense (all-domain) cannot be obtained. In addition, the conventional numerical calculation method requires a large amount of calculation resources.
Disclosure of Invention
The problems in the prior art are: 1) The traditional optimization design method of the sound transmission layer for reducing the acoustic transmission loss can only aim at a specific composite material model, such as a wedge-shaped structure and a gradient inclusion structure, and can not explore all possible material composite modes, so that the whole optimization space can not be traversed, and therefore, the global optimal solution can not be obtained; 2) Conventional numerical optimization calculation methods (e.g., finite element, topology optimization) are time consuming. The invention aims to solve the technical problems in the prior art and provide a sound transmission layer for reducing the acoustic transmission loss of a dissimilar material interface and an optimization method.
The specific technical scheme adopted by the invention is as follows:
in a first aspect, the present invention provides an optimization method for reducing acoustic transmission loss of a dissimilar material interface, where the acoustic transmission layer is disposed between a first dielectric layer and a second dielectric layer that are not matched in impedance, the optimization method includes:
s1, taking a single periodic unit in an acoustic transmission layer as an optimization target, discretizing a longitudinal section of the periodic unit into a binary matrix, and distinguishing and representing a first dielectric material and a second dielectric material according to different binary element values of the matrix, wherein the first dielectric material is the same as or has similar impedance to the material of the first dielectric layer, and the second dielectric material is the same as or has similar impedance to the material of the second dielectric layer; generating a binary matrix through randomization, generating a random sound-transmitting layer number code material model in batches, wherein each number in the model corresponds to a material unit, and the specific material of each unit is determined by the element value in the corresponding binary matrix;
s2, obtaining an acoustic transmission loss STL curve and an acoustic transmission loss average value ATL in a target frequency band range by utilizing finite element simulation for each sound transmission layer number code material model generated in the S1;
s3, converting a binary matrix of each sound transmission layer number code material model into a binary vector v, taking the binary vector v as the input of a training sample, and taking the corresponding acoustic transmission loss average value ATL calculated in S2 as a label of the training sample to construct a training data set; performing machine learning training by using a training data set to obtain a prediction function for predicting an acoustic transmission loss average value ATL according to a binary vector v;
s4, taking training samples in a training data set as an initial population, taking the prediction function as an fitness function, and optimizing through a genetic algorithm to obtain an optimal binary vector v with the minimum acoustic transmission loss average value ATL;
s5, converting the optimal binary vector v into a sound transmission layer number code material model to obtain an optimal configuration, obtaining an acoustic transmission loss STL curve and an acoustic transmission loss average value ATL in a target frequency band range through finite element simulation, judging whether the optimal configuration meets expected optimization requirements according to finite element simulation results, if so, completing optimization, if not, expanding training samples in a training data set, then performing machine learning training again to generate a new prediction function, and performing genetic algorithm optimization again until the optimal configuration meets the optimization requirements;
s6, smoothing the boundary between the two dielectric materials aiming at the optimal configuration meeting the optimization requirement to form the final configuration of the sound-transmitting layer.
Preferably, in the first aspect, the second medium layer is water, and the second medium material is hydrogel.
As a preferred aspect of the first aspect, in S5, the expanding training samples in the training data set includes: an empirical configuration determined from an artificial summary experience is added to the original randomly generated training dataset.
As a preference of the first aspect, the empirical configurations include a random gradient inclusion structure, a wedge gradient structure and a random wedge structure, each empirical configuration is an acoustic-transmitting layer number code material model, and also has a binary matrix formed by discretizing a longitudinal section of the periodic unit;
in the sound transmission layer number code material model of the random gradient inclusion structure, the material type of each code is randomly generated, but the probability of the second medium material in the code is gradually decreased line by line along the direction from the side of the second medium layer to the side of the first medium layer, and the probability of the first medium material in the code is gradually increased line by line gradient;
in the sound-transmitting layer number material model of the wedge-shaped gradient structure, the whole first dielectric material is in a continuous sheet and has relatively regular wedge-shaped outer outline, and the wedge-shaped width gradually increases along the direction from the side of the second dielectric layer to the side of the first dielectric layer.
In the sound transmission layer number code material model of the random wedge-shaped structure, the whole first dielectric material is in wedge-shaped distribution with non-connected pieces and irregular outer contours, the second dielectric material randomly appears in the wedge-shaped interior, and the number of the numbers of the first dielectric material appearing in a single line number is gradually increased line by line along the direction from the side of the second dielectric layer to the side of the first dielectric layer.
As a preference of the first aspect, in the random gradient inclusion structure, the probability growth curve of gradient increment is in a form of linearity, polynomial, logarithm or exponential.
Preferably, in the above first aspect, in the wedge-shaped gradient structure, two side contour lines of the first dielectric material distributed in a wedge shape are straight lines, polynomial curves, logarithmic curves or exponential curves.
As a preferable mode of the first aspect, the machine learning model used in the machine learning training is a back propagation artificial neural network, a support vector machine, a convolutional neural network, or a linear regression.
As a preferred feature of the first aspect, the machine learning model employs a BP neural network.
Preferably, in the first aspect, the initial population is a portion of training samples in the training dataset, where an average value of acoustic transmission losses is below a threshold value.
Preferably, the target frequency band is in the range of 0 to 10kHz.
In a second aspect, the present invention provides an acoustic layer optimized by the acoustic layer optimizing method according to any one of the first aspect, where the acoustic layer is a bilayer structure between a first dielectric layer and a second dielectric layer with mismatched impedance, and one side of the acoustic layer contacting the second dielectric layer is an impedance jump layer, and one side of the acoustic layer contacting the first dielectric layer is a sandwich layer, where the impedance jump layer is entirely made of a first dielectric material; the second dielectric material is used as a main material of the sandwich layer, a first dielectric material distribution area is arranged in the main material, and the first dielectric material distribution area in the sandwich layer is completely surrounded and wrapped by the second dielectric material.
Compared with the prior art, the invention has the following beneficial effects:
1) While the traditional optimization method can only optimize a small amount of geometric parameters aiming at a fixed configuration, the method of the invention applies the concept of digital materials to disperse the whole optimization space into small material units, thereby parameterizing the whole design space and facilitating the optimization of the whole design space.
2) Compared with the traditional finite element calculation method combined with the genetic algorithm optimization method to traverse all possible digital material configurations, the method provided by the invention firstly utilizes machine learning to obtain the prediction function of the relation between the digital material configuration and the average transmission loss, so that the prediction function is used for replacing finite element analysis in the genetic algorithm, and the calculation time is greatly saved.
3) By means of the method according to the invention, it is possible to design more novel configurations, in particular sound-transmitting layer designs between low frequency regions, for example between water and other media, in the range of (0-10 kHz) complex configurations with impedance jumps are obtained, which reduce the average sound transmission losses by 31% compared to conventional impedance-matching configurations.
Drawings
FIG. 1 is a schematic diagram of a method for generating a model of an acoustic-transparent layer digital material;
fig. 2 is a schematic diagram of a finite element analysis method of acoustic transfer loss and an average acoustic transfer loss calculation method;
FIG. 3 is a schematic diagram of an artificial neural network;
FIG. 4 is a schematic diagram of a genetic algorithm;
FIG. 5 shows four different types of digital material configuration generation effects, the former being a fully random binary configuration, the latter three being an empirical configuration based on human experience, randomly generated using features that reduce STL values;
FIG. 6 is a flowchart of a method for optimizing the design of a sound-transmitting layer in an embodiment of the present invention;
fig. 7 shows an acoustic layer configuration obtained by optimizing design in the embodiment of the present invention.
Detailed Description
The invention is further illustrated and described below with reference to the drawings and detailed description. The technical features of the embodiments of the invention can be combined correspondingly on the premise of no mutual conflict.
In order that the above objects, features and advantages of the invention will be readily understood, a more particular description of the invention will be rendered by reference to the appended drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. The present invention may be embodied in many other forms than described herein and similarly modified by those skilled in the art without departing from the spirit of the invention, whereby the invention is not limited to the specific embodiments disclosed below. The technical features of the embodiments of the invention can be combined correspondingly on the premise of no mutual conflict.
In the description of the present invention, it will be understood that when an element is referred to as being "connected" to another element, it can be directly connected to the other element or be indirectly connected with intervening elements present. In contrast, when an element is referred to as being "directly connected" to another element, there are no intervening elements present.
In the description of the present invention, it should be understood that the terms "first" and "second" are used solely for the purpose of distinguishing between the descriptions and not necessarily for the purpose of indicating or implying a relative importance or implicitly indicating the number of features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature.
The invention provides an optimization method of an acoustic transmission layer for reducing acoustic transmission loss of a dissimilar material interface. Because of the difference of the impedance between the first medium layer A and the second medium layer B, the sound is reflected at the interface of dissimilar materials with unmatched impedance, so that the transmitted acoustic energy is smaller than the incident acoustic energy, and the energy loss caused by reflection is the acoustic transmission loss. The invention is provided with the sound transmission layer, so that the impedance has transition between materials with unmatched original impedance, thereby reducing acoustic transmission loss and transmission loss.
It should be noted that, since the sound-transmitting layer is generally planar and continuous, the structure therein presents a periodic variation, the entire sound-transmitting layer need not be optimized in the present invention, but only the periodically-varying units thereof need to be extracted for optimization. Since the configuration (which may also be referred to as structure) change of the sound-transmitting layer is mainly manifested in the longitudinal section, the configuration can be optimized in the present invention for the longitudinal section of the individual periodic units in the sound-transmitting layer.
In a preferred embodiment of the present invention, the method for optimizing the sound-transmitting layer for reducing the acoustic transmission loss of the dissimilar material interface includes the following steps:
s1, taking a single periodic unit in an acoustic transmission layer as an optimization target, discretizing a longitudinal section of the periodic unit into a binary matrix, and distinguishing and representing a first dielectric material and a second dielectric material according to different binary element values of the matrix, wherein the first dielectric material is the same as or has similar impedance to the material of the first dielectric layer, and the second dielectric material is the same as or has similar impedance to the material of the second dielectric layer; a binary matrix is generated through randomization, a random sound-transmitting layer number code material model is generated in batches, each code in the model corresponds to a material unit, and the specific material of each unit is determined by the element value in the corresponding binary matrix.
It should be noted that, the binary matrix is constructed based on a physical model of the material acoustic transmission problem, as shown in fig. 1, the acoustic transmission layer is located between the dielectric material a and the dielectric material B, and a periodic unit longitudinal section of the acoustic transmission layer may be discretized into a 0-1 binary matrix. The value of each matrix element in the binary matrix is a binary value, namely 0 or 1, and the invention can use 1 for representing the first dielectric material, 0 for representing the second dielectric material, 0 for representing the first dielectric material, and 1 for representing the second dielectric material, which is not limited. Based on the binary matrix, elements in the matrix can be mapped according to binary element values to form a unit filled with a specified medium material in the sound transmission layer, and the unit is called a digital code, so that a sound transmission layer digital material model for numerical simulation is formed. If the thickness of the sound transmission layer is H, the binary matrix is an n multiplied by m matrix, and the height of one number in the sound transmission layer number material model is H/m.
It should be noted that the first dielectric material is preferably the same as the material of the first dielectric layer, and the same second dielectric material is preferably the material of the second dielectric layer, but if the material resistances of the two materials are not greatly different, two different materials may be used. For example, if the second dielectric layer is water, the second dielectric material may be replaced with a hydrogel because the water is in a liquid state and cannot be used to construct the layer, because the impedance difference between the hydrogel and water is small. The first dielectric material also needs to be determined according to the actual simulated first dielectric layer material, and rubber, polyurethane (PU) and the like can be adopted.
S2, obtaining an acoustic transmission loss (STL) curve and an acoustic transmission loss average value (ATL) in a target frequency band range by using finite element simulation for each sound transmission layer number code material model generated in the S1.
It is to be noted that calculating an acoustic transfer loss (STL) curve in a target frequency band range by finite element simulation belongs to the prior art. As shown in fig. 2, a finite element analysis method of acoustic transmission loss and a flow of an average acoustic transmission loss calculation method are shown, wherein the construction of a finite element model is a key step. When the physical model based on the sound transmission problem is used for constructing the finite element model, a hard sound field boundary and a low reflection boundary are required to be respectively arranged up and down, then symmetry boundaries are arranged on two sides, and a perfect matching layer, a second medium material layer B, a pressure acoustic and solid coupling boundary, a sound transmission layer and a first medium material layer A are sequentially arranged from top to bottom within the four boundaries. The sound-transmitting layer can be constructed by introducing the sound-transmitting layer number material model. In the concerned frequency band range, namely in the target frequency band range, the incidence process of sound waves can be simulated through finite element simulation to obtain STL curves formed by STL values at different channels, and the STL values corresponding to all channels in the target frequency band range are averaged to obtain the ATL value. For example, in fig. 2, since the target band range of interest is a low frequency region of 0 to 10kHz, this region is divided into 10 bands, and STL values of the 10 bands are averaged to obtain corresponding ATL values.
Each of the above-mentioned acoustic layer number material models can be used as a sample for machine learning by obtaining a corresponding ATL value through finite element simulation. Therefore, when the models of the acoustic transmission layer number code materials are generated in batches, the number of the models needs to ensure that the machine learning has enough samples for training.
S3, converting a binary matrix of each sound transmission layer number code material model into a binary vector v, taking the binary vector v as the input of a training sample, and taking the corresponding acoustic transmission loss average value ATL calculated in S2 as a label of the training sample to construct a training data set; machine learning training is performed using the training data set to obtain a prediction function for predicting an acoustic transfer loss average ATL from the binary vector v.
It should be noted that, the conversion of the binary matrix into the binary vector is mainly for facilitating the input of the machine learning model. The machine learning model which can be adopted by the invention is not limited, and a back propagation artificial neural network, a support vector machine, a convolution neural network, linear regression and the like can be used as the machine learning model in the invention. The training of machine learning models belongs to the prior art and will not be described further.
As shown in fig. 3, a schematic diagram of a machine learning model in the form of a BP neural network used in one embodiment to construct a prediction function is shown, where the input layer dimension is a binary vector v dimension, i.e. 1×400, the hidden layer has three layers, the dimensions are 10-5-5, respectively, and the output layer dimension is 1, i.e. an ATL value. Thus, the model obtained after the machine learning model is trained can be regarded as a prediction function f for predicting the acoustic transmission loss average ATL from the binary vector v, that is, atl=f (v).
And S4, taking training samples in the training data set as an initial population, taking the prediction function ATL=f (v) as an fitness function, and optimizing through a genetic algorithm to obtain the optimal binary vector v with the minimum acoustic transmission loss average value ATL.
The genetic algorithm (Genetic Algorithm, GA for short) in the invention is a self-adaptive random search heuristic algorithm and is widely applied to the fields of complex function system optimization, machine learning, system identification, fault diagnosis, classification system and the like. As shown in fig. 4, the GA algorithm uses computer simulation operation to convert the solving process of the problem into processes like crossing, mutation, etc. of chromosome genes in biological evolution, wherein each iteration needs to use fitness function to distinguish the quality of individuals in the population, and the individuals are screened based on the fitness value, so as to ensure that the individuals with good fitness value have a chance to generate more children in the next generation. Therefore, when solving a specific problem using GA, the selection of the adaptive value function has a large influence on the convergence of the algorithm and the convergence speed. In the present invention, the prediction function atl=f (v) is recorded as an fitness function in the genetic algorithm optimization process. The objective of the final optimization is to minimize the fitness function, i.e. to ensure that the acoustic transfer loss average ATL is minimal. After the optimal binary vector v is obtained through optimization, the binary vector v can be converted into a binary matrix form again, and then an optimal sound-transmitting layer number code material model is formed.
In order to accelerate the convergence of the genetic algorithm, when an initial population is selected, training samples in the training data set can be screened, partial samples with excessively high ATL values are removed, and partial samples with smaller ATL values are reserved.
S5, converting the optimal binary vector v into a sound transmission layer number code material model to obtain an optimal configuration, obtaining an acoustic transmission loss STL curve and an acoustic transmission loss average value ATL in a target frequency band range through finite element simulation, judging whether the optimal configuration meets expected optimization requirements according to finite element simulation results, if so, completing optimization, if not, expanding training samples in a training data set, then performing machine learning training again to generate a new prediction function, and performing genetic algorithm optimization again until the optimal configuration meets the optimization requirements.
It should be noted that the finite element simulation in this step is the same as that in S2, and will not be described here again. The objective is to determine whether the optimal configuration meets the expected optimization requirement, where the expected optimization requirement needs to be determined according to the actual situation, and the STL curve that the sound-transmitting layer is expected to finally meet and the ATL value that needs to be met may be preset, and the specific value is not limited herein.
However, if the optimal configuration is found after finite element simulation to not meet the expected optimization requirement, the number of the samples of the acoustic transmission layer number code material model generated in batch initially may be insufficient, and the optimizing process falls into a local optimal solution, but does not find a global optimal solution. Therefore, in this case, it is necessary to expand training samples in the training data set, then perform machine learning again to generate a new prediction function f, perform genetic algorithm optimization of S4 and S5 again, and then determine whether the expected result is reached again. The process is cycled until the optimal configuration meets the desired optimization requirements.
It should be noted that when the training samples in the training data set are expanded, the method in S1 may be continuously adopted to generate the model of the acoustic layer number code material in a random generation manner and obtain the tag through finite element simulation, but because the binary matrices corresponding to the models of the acoustic layer number code material are generated randomly, the model may not necessarily conform to the configuration form required for reducing the ATL.
Therefore, as a preferred mode of the embodiment of the invention, when expanding training samples in the training data set, the method is adopted to generate partial experience configurations in advance according to artificially summarized experiences, and then perform finite element simulation on the experience configurations to obtain ATL values, and then add the ATL values into the training data set generated at original random in a sample mode. The form of the empirical configuration added here is not limited and may be selected according to related studies or actual experiments in the prior art. In this embodiment, three empirical configurations are provided, including a random gradient inclusion structure, a wedge gradient structure, and a random wedge structure, each of which is a model of an acoustically transparent layer number material, and also has a binary matrix formed by discretizing a longitudinal section of the periodic unit, where the dimension of the binary matrix is the same as the binary matrix generated randomly. Specific forms of these three empirical configurations are described in detail below.
As shown in the first graph of fig. 5, the binary matrix of the completely randomly generated model of the acoustic-transparent layer number code material is used as a contrast, so that the type of the medium material in each code is also disordered.
As shown in the second graph in fig. 5, in the model of the acoustic-transmitting layer number code material of the random gradient inclusion structure, the material type of each code is randomly generated, that is, each code is randomly set to be the first dielectric material or the second dielectric material, but macroscopically, the probability of the second dielectric material appearing in the code decreases line by line along the direction from the side of the second dielectric layer to the side of the first dielectric layer, and the probability of the first dielectric material appearing in the code increases gradient line by line. From the macroscopic view, along the direction from the side of the second dielectric layer to the side of the first dielectric layer, the probability of the second dielectric material appearing in the digits decreases row by row, the number of digits of the first dielectric material appearing in each row of digits is continuously increased, but the positions of the digits are not constant. In the random gradient inclusion structure, the probability of the first dielectric material in the digital code needs to be gradually increased row by row, so that a corresponding probability growth curve is required to be set to control the probability of the first dielectric material in each row in actual implementation, and the selectable probability growth curve is in a linear, polynomial, logarithmic or exponential form, so that the random gradient inclusion structure is not limited, and different models can be generated as abundantly as possible so as to facilitate global optimization.
As shown in the third graph in fig. 5, in the model of the acoustic layer number code material with the wedge-shaped gradient structure, the whole first dielectric material is in a continuous piece, the outer contour of the first dielectric material is distributed in a relatively regular wedge shape, and the wedge-shaped width is gradually increased line by line along the direction from the side of the second dielectric layer to the side of the first dielectric layer. This configuration is relative to the first two configurations in which the dielectric material distribution is no longer disordered but rather assumes a relatively regular state. In the wedge-shaped gradient structure, the contour lines on two sides of the first dielectric material distributed in a wedge shape can be a straight line, a polynomial curve, a logarithmic curve or an exponential curve which are specified according to experience, and the method is not limited to the above, and different models can be generated as abundantly as possible so as to facilitate global optimization.
As shown in the fourth graph in fig. 5, in the model of the acoustic transmission layer number code material with the random wedge-shaped structure, the whole first dielectric material is in a wedge-shaped distribution with non-connected pieces and irregular outer contours, the second dielectric material randomly appears in the wedge, and the number of the first dielectric material appearing in a single row of numbers is gradually increased along the direction from the side of the second dielectric layer to the side of the first dielectric layer. The difference between this random wedge structure and the previous wedge gradient structure is that the outer contour is no longer empirically specified but is randomly generated, but remains overall wedge-shaped, but the second dielectric material may also occur inside the wedge-shaped region due to the random generation, the wedge-shaped distribution of the first dielectric material no longer being entirely continuous.
S6, smoothing the boundary between the two dielectric materials aiming at the optimal configuration meeting the optimization requirement to form the final configuration of the sound-transmitting layer.
It should be noted that the purpose of performing the boundary smoothing in this step is to consider the technical requirements of the actual processing of the material, since a material boundary with too discrete boundaries is not practical in the actual processing. After the boundary between the two dielectric materials is smoothed, the boundary is ensured to be relatively smooth, and the processing of the sound transmission layer is facilitated.
In addition, after the final configuration of the sound-transmitting layer is obtained, the final configuration can be simplified to a certain extent before the sound-transmitting layer is actually processed, partial structural features which do not meet the processing technology requirement can be removed, or the mechanism research can be further carried out on the features in the final configuration obtained by optimization, so that the structural features belong to key features and the features belong to non-key features, and further the actual processing configuration can be further adjusted.
The method for optimizing the sound transmission layer for reducing the acoustic transmission loss of the dissimilar material interface shown in the above steps S1 to S6 is applied to a specific example, so as to show the specific implementation process and technical effects thereof.
Examples
In this embodiment, the process of the method for optimizing the sound-transmitting layer for reducing the acoustic transmission loss of the interface of the dissimilar materials is shown in fig. 6, and the specific steps are as follows:
the first step: for two media A and B with unmatched impedance and a given sound-transmitting layer thickness H, a random sound-transmitting layer number code material model is generated in batches based on the concept of digital materials, and the sound-transmitting layer is a composite material formed by a first medium material A and a second medium material B. In this embodiment, the media on both sides of the sound-transmitting layer are water and rubber, respectively, so that the media materials constituting the sound-transmitting layer are respectively selected from acrylamide hydrogel and Polyurethane (PU). The generation method comprises the following steps: firstly, randomly generating a series of n multiplied by M order binary matrixes M, wherein 1 and 0 in the matrixes represent a dielectric material A and a dielectric material B respectively; then, according to the corresponding relation of 1, 0 and A, B in each binary matrix M, a digital material model of the sound-transmitting layer is generated, wherein each digital represents a material unit, and the height of each material unit is H/M.
And a second step of: for each of the configurations of the acoustically transparent layer digital materials generated in batch, acoustic transmission loss (STL) in the frequency band of interest is calculated by simulation using finite element software, and an average value ATL of the acoustic transmission loss in the frequency band is calculated. In this embodiment, the concerned band is a low frequency band ranging from 0kHz to 10kHz, and the band is divided into 10 bands, and the STL values of the 10 bands on the STL curve are averaged to obtain the corresponding ATL value.
And a third step of: the binary matrix of each sound-transmitting layer number material model is converted into a binary vector v, and together with the calculated ATL values, a training sample is constructed, and all the training samples are used as a training data set to train the BP neural network, so that a trained model is obtained and used as a prediction function between the binary vector v and the ATL values, namely, atl=f (v).
Fourth step: and screening a part of binary vectors v with lower ATL values from the training data set, taking the corresponding training samples as an initial population, and carrying out genetic algorithm optimization by taking ATL=f (v) as an fitness function by setting proper crossing rate, mutation rate and convergence criterion to obtain the lowest ATL predicted value and the corresponding binary vector v. In this embodiment, the genetic algorithm is directly implemented by using an optimization kit in MATLAB, the solver selects ga-Genetic Algorithm, and the crossover probability and the mutation probability in the genetic operator are set to 0.8 and 0.01, respectively.
Fifth step: and converting the binary vector v of the lowest ATL predicted value obtained by the genetic algorithm into a digital material configuration so as to obtain an optimal configuration, calculating an acoustic transmission loss STL curve of the optimal configuration by using a finite element method, and calculating an average ATL. And judging whether the STL curve and the ATL value of the optimal configuration meet the optimization expectation, if so, ending the optimization, otherwise, expanding the training data set, and adding some configurations generated according to human experience. Compared with a completely random configuration, the ATL value range of the empirical configuration is wider, more features are included, and the machine learning is facilitated to obtain a more accurate prediction model. The finite element simulation of the second step and the optimization process of the third to fifth steps are then re-performed.
The empirical configuration adopted in the embodiment comprises the random gradient inclusion structure, the wedge-shaped gradient structure and the random wedge-shaped structure shown in the foregoing fig. 5, wherein the probability of occurrence of the material a or the material B in the random gradient inclusion structure changes along the thickness direction in a gradient manner, and the gradient can be in the forms of linearity, polynomial, logarithm, index and the like; the wedge-shaped gradient structure can be used for wedge-shaped structures with different gradients, including linear type, parabolic type, exponential type and the like, and parameters of curvature change can be randomly generated; in addition, the outer contours of the wedge blocks in the material with random wedge-shaped configuration are randomly generated, and the inside of the wedge blocks has probability to generate gaps in the thickness direction.
Seventh step: and after the expected optimal configuration is obtained, performing material interface edge smoothing on the optimal configuration to form a final configuration.
As shown in fig. 7, the left graph shows the final configuration of the sound-transmitting layer obtained by optimizing the above-mentioned sound-transmitting layer optimizing method in this embodiment, where the sound-transmitting layer is a double-layer structure between a first dielectric layer and a second dielectric layer with mismatched impedance, where the side contacting the second dielectric layer is an impedance jump layer, and the side contacting the first dielectric layer is a sandwich layer, where the impedance jump layer is entirely made up of a first dielectric material a; the second dielectric material B is used as a main material of the sandwich layer, a first dielectric material A distribution area is arranged in the main material, and the first dielectric material A distribution area in the sandwich layer is completely surrounded and wrapped by the second dielectric material B.
Therefore, the configuration obtained by optimizing the optimizing method is inconsistent with the configuration obtained by summarizing the artificial experience, and a layer of impedance jump layer is formed on the side contacted with the second dielectric layer, which violates the concept of ensuring impedance gradual change in the artificial experience. This novel complex configuration of impedance jumps has better performance than conventional impedance matching configurations, especially in the low frequency range (0-10 kHz), which further reduces the average acoustic transmission loss by 31% compared to conventional impedance graded matching configurations.
In addition, in this embodiment, the distribution area of the material B in the inclusion layer in the final configuration after the edge smoothing is approximately elliptical. Based on the final configuration, it may also be subjected to structural feature analysis, extraction of primary structural features, filtering of secondary structural features, then simplification of the primary structural features, and verification of the simplified optimal configuration by means of finite elements, thus obtaining the final configuration. For example, referring to the right-hand drawing of FIG. 7, the oval material B distribution area may be reduced to a circle, the impedance jump layer may be smoothed to form a smooth layer, and the parameters including the total thickness D, width W, and the thickness D of the impedance jump layer may be parameterized 2 Radius d of material B region, distance h between the bottom of material B region and the bottom of sound-transmitting layer, and W, D 2 Ratio w of D, h to D r 、t r 、d r 、h r These parameters can be further explored to determine the mechanism by which they affect the final performance and whether there is a possibility of further optimization of the configuration.
The above embodiment is only a preferred embodiment of the present invention, but it is not intended to limit the present invention. Various changes and modifications may be made by one of ordinary skill in the pertinent art without departing from the spirit and scope of the present invention. Therefore, all the technical schemes obtained by adopting the equivalent substitution or equivalent transformation are within the protection scope of the invention.
Claims (10)
1. The method for optimizing the sound transmission layer for reducing the acoustic transmission loss of the dissimilar material interface is characterized by comprising the following steps of:
s1, discretizing a longitudinal section of a periodic unit into a binary matrix by taking a single periodic unit in an acoustic transmission layer as an optimization target, and respectively representing a first dielectric material and a second dielectric material by different binary element values of the matrix, wherein the first dielectric material and the material of the first dielectric layer have the same impedance, and the second dielectric material and the material of the second dielectric layer have the same impedance; generating a binary matrix through randomization, generating a random sound-transmitting layer number code material model in batches, wherein each number in the model corresponds to a material unit, and the specific material of each unit is determined by the element value in the corresponding binary matrix;
s2, obtaining an acoustic transmission loss curve and an acoustic transmission loss average value in a target frequency band range by utilizing finite element simulation for each sound transmission layer number code material model generated in the S1;
s3, converting a binary matrix of each sound transmission layer number code material model into a binary vector, taking the binary vector as input of a training sample, and taking the corresponding acoustic transmission loss average value calculated in the S2 as a label of the training sample to construct a training data set; performing machine learning training by using a training data set to obtain a prediction function for predicting an acoustic transmission loss average value according to the binary vector;
s4, taking training samples in the training data set as an initial population, taking the prediction function as an fitness function, and optimizing through a genetic algorithm to obtain an optimal binary vector with the minimum acoustic transmission loss average value;
s5, converting the optimal binary vector into a sound transmission layer number material model to obtain an optimal configuration, obtaining an acoustic transmission loss curve and an acoustic transmission loss average value in a target frequency band range through finite element simulation, judging whether the optimal configuration meets expected optimization requirements according to finite element simulation results, if so, completing optimization, if not, expanding training samples in a training data set, then performing machine learning training again to generate a new prediction function, and performing genetic algorithm optimization again until the optimal configuration meets the optimization requirements;
s6, smoothing the boundary between the two dielectric materials aiming at the optimal configuration meeting the optimization requirement to form the final configuration of the sound-transmitting layer.
2. The method for optimizing an acoustic transmission layer for reducing interface acoustic transmission loss of a dissimilar material according to claim 1, wherein the second dielectric layer is water, and the second dielectric material is hydrogel.
3. The method for optimizing a sound-transmitting layer for reducing acoustic transmission loss of a dissimilar material interface according to claim 1, wherein in S5, the method for expanding training samples in a training data set is as follows: an empirical configuration determined from an artificial summary experience is added to the original randomly generated training dataset.
4. The method for optimizing a sound-transmitting layer for reducing acoustic transmission loss of a dissimilar material interface according to claim 3, wherein the empirical configurations comprise a random gradient inclusion structure, a wedge-shaped gradient structure and a random wedge-shaped structure, each empirical configuration is a sound-transmitting layer code material model and also comprises a binary matrix formed by discretizing a longitudinal section of a periodic unit;
in the sound transmission layer number code material model of the random gradient inclusion structure, the material type of each code is randomly generated, but the probability of the second medium material in the code is gradually decreased line by line along the direction from the side of the second medium layer to the side of the first medium layer, and the probability of the first medium material in the code is gradually increased line by line gradient;
in the sound-transmitting layer number material model of the wedge-shaped gradient structure, the whole first dielectric material is in continuous sheet and has regular wedge-shaped distribution of outer outline, and the wedge-shaped width gradually increases along the direction from the side of the second dielectric layer to the side of the first dielectric layer;
in the sound transmission layer number code material model of the random wedge-shaped structure, the whole first dielectric material is in wedge-shaped distribution with non-connected pieces and irregular outer contours, the second dielectric material randomly appears in the wedge-shaped interior, and the number of the numbers of the first dielectric material appearing in a single line number is gradually increased line by line along the direction from the side of the second dielectric layer to the side of the first dielectric layer.
5. The method for optimizing an acoustically transparent layer for reducing interfacial acoustic transmission loss of a dissimilar material according to claim 4, wherein said probability growth curve of gradient increment in said random gradient inclusion structure is in the form of a linear, polynomial, logarithmic or exponential.
6. The method for optimizing a sound-transmitting layer for reducing acoustic transmission loss of a dissimilar material interface according to claim 4, wherein contour lines on two sides of the first dielectric material distributed in a wedge-shaped gradient structure are straight lines, polynomial curves, logarithmic curves or exponential curves.
7. The method for optimizing a sound transmission layer for reducing acoustic transmission loss of a dissimilar material interface according to claim 1, wherein a machine learning model adopted in the machine learning training is a counter-propagating artificial neural network, a support vector machine, a convolutional neural network or a linear regression.
8. The method for optimizing an acoustic transmission layer for reducing acoustic transmission loss at a dissimilar material interface of claim 7, wherein said initial population is a portion of training samples in a training dataset having an average value of acoustic transmission loss below a threshold value.
9. The method for optimizing an acoustic transmission layer for reducing acoustic transmission loss of a dissimilar material interface according to claim 1, wherein the target frequency band range is 0-10 khz.
10. The sound-transmitting layer optimized by the sound-transmitting layer optimizing method according to any one of claims 1 to 9, wherein the sound-transmitting layer has a double-layer structure between a first dielectric layer and a second dielectric layer with unmatched impedance, wherein the side contacting the second dielectric layer is an impedance jump layer, and the side contacting the first dielectric layer is a sandwich layer, wherein the impedance jump layer is entirely made of a first dielectric material; the second dielectric material is used as a main material of the sandwich layer, a first dielectric material distribution area is arranged in the main material, and the first dielectric material distribution area in the sandwich layer is completely surrounded and wrapped by the second dielectric material.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211543461.4A CN115762685B (en) | 2022-12-02 | 2022-12-02 | Sound transmission layer for reducing acoustic transmission loss of dissimilar material interface and optimization method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211543461.4A CN115762685B (en) | 2022-12-02 | 2022-12-02 | Sound transmission layer for reducing acoustic transmission loss of dissimilar material interface and optimization method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115762685A CN115762685A (en) | 2023-03-07 |
| CN115762685B true CN115762685B (en) | 2023-07-28 |
Family
ID=85343065
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211543461.4A Active CN115762685B (en) | 2022-12-02 | 2022-12-02 | Sound transmission layer for reducing acoustic transmission loss of dissimilar material interface and optimization method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115762685B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103915090A (en) * | 2012-12-31 | 2014-07-09 | 中国科学院声学研究所 | Broadband noise reduction porous-material acoustic liner and equipment |
| CN105184018A (en) * | 2015-10-13 | 2015-12-23 | 同济汽车设计研究院有限公司 | Method for calculating mid-frequency transmission loss of subsystems and optimizing acoustic packages |
| CN109165764A (en) * | 2018-06-26 | 2019-01-08 | 昆明理工大学 | A kind of line loss calculation method of genetic algorithm optimization BP neural network |
| CN113361383A (en) * | 2021-06-01 | 2021-09-07 | 北京工业大学 | Porcelain insulator damage neural network identification method based on genetic algorithm characteristic parameter optimization |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2490150C1 (en) * | 2011-12-16 | 2013-08-20 | Федеральное государственное бюджетное образовательное учреждение высшего профессионального образования "Тольяттинский государственный университет" | Modified laminar acoustic structure of vehicle body upholstery |
| CN110097185B (en) * | 2019-03-29 | 2021-03-23 | 北京大学 | Optimization model method based on generation of countermeasure network and application |
| US20210158211A1 (en) * | 2019-11-22 | 2021-05-27 | Google Llc | Linear time algorithms for privacy preserving convex optimization |
| CN112040382B (en) * | 2020-08-10 | 2021-07-30 | 上海船舶电子设备研究所(中国船舶重工集团公司第七二六研究所) | High-bandwidth underwater acoustic transducer based on acoustic impedance gradient matching layer |
| CN114692813A (en) * | 2020-12-29 | 2022-07-01 | 清远市富盈电子有限公司 | Calculation method for impedance value of PCB (printed Circuit Board) |
| CN114390427A (en) * | 2021-12-29 | 2022-04-22 | 瑞声光电科技(常州)有限公司 | Sound field optimization method, device and equipment and readable storage medium |
| CN114444390A (en) * | 2022-01-25 | 2022-05-06 | 东南大学 | Design method of vacuum electronic device slow wave structure based on deep learning and machine learning |
-
2022
- 2022-12-02 CN CN202211543461.4A patent/CN115762685B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103915090A (en) * | 2012-12-31 | 2014-07-09 | 中国科学院声学研究所 | Broadband noise reduction porous-material acoustic liner and equipment |
| CN105184018A (en) * | 2015-10-13 | 2015-12-23 | 同济汽车设计研究院有限公司 | Method for calculating mid-frequency transmission loss of subsystems and optimizing acoustic packages |
| CN109165764A (en) * | 2018-06-26 | 2019-01-08 | 昆明理工大学 | A kind of line loss calculation method of genetic algorithm optimization BP neural network |
| CN113361383A (en) * | 2021-06-01 | 2021-09-07 | 北京工业大学 | Porcelain insulator damage neural network identification method based on genetic algorithm characteristic parameter optimization |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115762685A (en) | 2023-03-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110599766B (en) | Road traffic jam propagation prediction method based on SAE-LSTM-SAD | |
| CN106650179A (en) | Method of designing Acoustic Metamaterials based on CMA-ES optimization algorithm | |
| Dong et al. | Nelder–Mead optimization of elastic metamaterials via machine-learning-aided surrogate modeling | |
| CN112819523B (en) | Marketing prediction method combining inner/outer product feature interaction and Bayesian neural network | |
| CN113822499A (en) | Train spare part loss prediction method based on model fusion | |
| CN111415008B (en) | Ship flow prediction method based on VMD-FOA-GRNN | |
| CN115762685B (en) | Sound transmission layer for reducing acoustic transmission loss of dissimilar material interface and optimization method | |
| Weeratunge et al. | A machine learning accelerated inverse design of underwater acoustic polyurethane coatings | |
| CN107358311A (en) | A kind of Time Series Forecasting Methods | |
| JP2001287516A (en) | Method for designing tire, method for designing mold for vulcanization of tire, manufacturing method of mold for vulcanization of tire, manufacturing method of tire, optimization analysis apparatus for tire, and storage medium recording optimization analysis program of tire | |
| CN107622354B (en) | Emergency capacity evaluation method for emergency events based on interval binary semantics | |
| Pan et al. | Accelerated inverse design of customizable acoustic metaporous structures using a CNN-GA-based hybrid optimization framework | |
| Pan et al. | Bottom-up approaches for rapid on-demand design of modular metaporous structures with tailored absorption | |
| CN115482893A (en) | Electromagnetic metamaterial design method based on deep learning and structural variables | |
| CN115983366A (en) | Model pruning method and system for federal learning | |
| CN113360716B (en) | Logical processing method and system for gas pipe network structure | |
| CN113505929A (en) | Topological optimal structure prediction method based on embedded physical constraint deep learning technology | |
| US11604973B1 (en) | Replication of neural network layers | |
| Zhang et al. | Learning to inversely design acoustic metamaterials for enhanced performance | |
| CN114186470A (en) | Design optimization system and design method of broadband transparent wave absorber based on multi-objective genetic algorithm | |
| Tantawy et al. | Performance investigation and element optimization of 2D array transducer using Bat Algorithm | |
| CN115099118B (en) | NSGA III-based high-dimensional multi-objective joint parallel simulation optimization method | |
| CN113946974B (en) | Multi-objective optimization-based self-organizing type one layered fuzzy prediction system | |
| CN111080053B (en) | EHF-WAS-based power grid development strategy efficiency and benefit evaluation method | |
| CN116844645B (en) | Gene regulation network inference method based on multi-view layered hypergraph |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |