Bandgap phononic crystal design method, device, electronic device and storage medium

Abstract

The present application discloses a bandgap phononic crystal design method, device, electronic device, and storage medium, belonging to the field of phononic crystal design technology. The method includes: obtaining a bandgap phononic crystal and generating a bandgap vector of the bandgap phononic crystal; constructing a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determining the structure of the expected bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector; constructing an image relationship model between the structure of the bandgap phononic crystal and hidden variables, and determining the expected hidden variables based on the image relationship model and the structure of the expected bandgap phononic crystal, generating image features of the expected phononic crystal based on the expected hidden variables; and generating a target phononic crystal based on the structure of the expected bandgap phononic crystal and the image features of the expected phononic crystal. The present invention constructs a phononic crystal design model based on the mapping relationship between the phononic crystal structure and the bandgap vector and the hidden variables, and can flexibly and accurately generate phononic crystals with bandgaps within the target frequency range.

Description

Technical Field

[0001] The present invention relates to the technical field of phononic crystal design, and in particular to a bandgap phononic crystal design method, device, equipment and storage medium. Background Art

[0002] Phononic crystals are a new type of artificial periodic structure that can control the propagation of elastic waves within a structure. Research has shown that when elastic waves propagate through phononic crystal structures with an infinite period, they are unable to propagate within certain frequency bands, known as the band gap, due to wave scattering or localized resonance. Even for phononic crystal structures with a finite period, which have been fabricated, elastic wave propagation is significantly suppressed within the band gap, manifesting as high transmission losses within this frequency band. Therefore, phononic crystal structures with excellent band gap performance hold great promise for applications in vibration and noise reduction.

[0003] In existing technologies, many researchers have devoted themselves to optimizing the structure, materials, and other parameters of phononic crystals to regulate and control elastic waves within different frequency bands. This traditional optimization often requires multiple iterations to achieve the desired structure, inevitably consuming significant computing resources and presenting significant drawbacks. In recent years, with the advancement of computer performance, artificial neural networks have demonstrated promising results in the design and optimization of structures. However, at present, artificial neural networks are primarily used for topological optimization of existing structures. They are unable to flexibly customize the bandgap distribution within the target frequency band and accurately generate the desired bandgap phononic crystal structure.

[0004] Therefore, there is an urgent need for a method for designing bandgap phononic crystals that can flexibly set the bandgap distribution within the target frequency band as needed and accurately generate the target bandgap phononic crystal structure. Summary of the Invention

[0005] In view of this, it is necessary to provide a bandgap phononic crystal design method, device, electronic device and storage medium to solve the problem in the prior art that it is impossible to set a target bandgap phononic crystal structure with a bandgap distribution within the target frequency band as required.

[0006] In order to solve the above technical problems, on the one hand, the present invention provides a method for designing a bandgap phononic crystal, comprising:

[0007] Obtaining a band gap phononic crystal and generating a band gap vector of the band gap phononic crystal;

[0008] Constructing a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determining the expected structure of the bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector;

[0009] Constructing an image relationship model between the structure of the bandgap phononic crystal and latent variables, determining expected latent variables based on the image relationship model and the structure of the expected bandgap phononic crystal, and generating image features of the expected phononic crystal based on the expected latent variables;

[0010] A target phononic crystal is generated according to the structure of the expected band gap phononic crystal and the image features of the expected phononic crystal.

[0011] In a possible implementation, obtaining a bandgap phononic crystal and generating a bandgap vector of the bandgap phononic crystal include:

[0012] Generate multiple random phononic crystals;

[0013] Analyzing the structure of the random phononic crystal based on a convolutional neural network and constructing a band gap phononic crystal screening model;

[0014] Based on the band gap phononic crystal screening model, the band gap phononic crystal and the band gap vector of the band gap phononic crystal are output according to the structure of the random phononic crystal.

[0015] In one possible implementation, generating a plurality of random phononic crystals includes:

[0016] Determine the two-dimensional single-sided pixels and construct a basic unit containing a random number of bases and scatterers;

[0017] The basic unit is symmetrically and folded to generate a two-dimensional phononic crystal;

[0018] The two-dimensional phononic crystals whose scatterers satisfy connectivity are screened out, and random phononic crystals are generated.

[0019] In one possible implementation, a convolutional neural network is used to analyze the structure of random phononic crystals and construct a bandgap phononic crystal screening model, including:

[0020] Analyzing the structure of the random phononic crystal based on finite element analysis, performing frequency testing on the random phononic crystal, and determining the band gap distribution of the random phononic crystal;

[0021] generating a band gap label of the random phononic crystal according to the band gap distribution;

[0022] A convolutional neural network is trained based on the random phononic crystals, band gap labels, and band gap distribution to generate a band gap phononic crystal screening model.

[0023] In one possible implementation, a mapping relationship model between the structure of a bandgap phononic crystal and a bandgap vector is constructed, and the structure of an expected bandgap phononic crystal is determined based on the mapping relationship model and the expected bandgap vector, including:

[0024] Build an initial model based on a fully connected neural network;

[0025] Using the band gap distribution as conditional information, and based on the trained initial model of the structure and band gap vector of the band gap phononic crystal, a mapping relationship model between the structure and band gap vector of the band gap phononic crystal is generated;

[0026] Based on the mapping relationship model, the structure of the expected band gap phononic crystal is generated with the expected band gap vector as input.

[0027] In one possible implementation, a graphical relationship model between the structure of a bandgap phononic crystal and hidden variables is constructed, including:

[0028] Based on the encoder and decoder, build the initial autoencoder model;

[0029] Extracting a first image feature of the band gap phononic crystal structure of the training sample according to the encoder;

[0030] Mapping the first image feature into a latent variable;

[0031] Based on the decoder, reconstruct the image feature according to the latent variable and the band gap phononic crystal structure of the training sample to generate a second image feature;

[0032] According to the comparison result of the first image feature and the second image feature, the initial autoencoder model is updated and iterated to generate an image relationship model between the structure of the bandgap phononic crystal and the latent variables.

[0033] In one possible implementation, determining expected latent variables based on the image relationship model and the structure of the expected band gap phononic crystal, and generating image features of the phononic crystal according to the expected latent variables include:

[0034] Based on the image relationship model, determining the corresponding hidden variables according to the structure of the expected band gap phononic crystal;

[0035] According to the structural mapping relationship between the hidden variables and the band gap phononic crystal, the image features of the expected phononic crystal are generated.

[0036] In a second aspect, the present invention further provides an electronic device, comprising a memory and a processor, including:

[0037] A training data generation module is used to obtain a bandgap phononic crystal and generate a bandgap vector of the bandgap phononic crystal;

[0038] A phononic crystal structure generation module is used to construct a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determine the structure of the expected bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector;

[0039] An image feature generation module is used to construct an image relationship model between the structure of the band gap phononic crystal and latent variables, determine the expected latent variables based on the image relationship model and the structure of the expected band gap phononic crystal, and generate image features of the expected phononic crystal based on the expected latent variables;

[0040] The target generation module is used to generate a target phononic crystal according to the structure of the expected band gap phononic crystal and the image characteristics of the expected phononic crystal.

[0041] In a third aspect, the present invention also provides an electronic device comprising a memory and a processor, wherein the memory is used to store programs and data; the processor is coupled to the memory and is used to execute the program stored in the memory to implement the band gap phononic crystal design method as described above.

[0042] In a fourth aspect, the present invention further provides a computer storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement the band gap phononic crystal design as described above.

[0043] The beneficial effects of the present invention are as follows: first, the present invention trains a model based on the bandgap phononic crystal and the bandgap vector of the bandgap phononic crystal, and the model consists of two parts: a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and an image relationship model between the structure of the bandgap phononic crystal and the hidden variables; then, the mapping relationship between the structure of the bandgap phononic crystal and the bandgap vector is analyzed, a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector is constructed, and the structure of the expected bandgap phononic crystal is generated; the mapping relationship between the structure of the bandgap phononic crystal and the hidden variables is analyzed, an image relationship model between the structure of the bandgap phononic crystal and the hidden variables is constructed, and the image features of the expected phononic crystal are generated; finally, according to the structure of the expected bandgap phononic crystal and the image features of the expected phononic crystal, a target phononic crystal is generated. The present application constructs a phononic crystal design model through the mapping relationship between the structure of the phononic crystal and the bandgap vector and the mapping relationship between the phononic crystal structure and the hidden variables, takes the expected bandgap vector as input, and finally outputs the target bandgap phononic crystal. The band gap distribution within the target frequency band can be flexibly set according to expected needs, its band gap vector can be determined, and the target band gap phononic crystal structure can be accurately generated based on the band gap vector. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0045] Figure 1A schematic flow chart of an embodiment of the phononic crystal design method provided by the present invention;

[0046] Figure 2 For the present invention Figure 1 Flow chart of the embodiment of the partial content of step S101;

[0047] Figure 3 For the present invention Figure 1 Flow chart of the embodiment of part of step S101;

[0048] Figure 4 For the present invention Figure 3 A flow chart of an embodiment of step S301;

[0049] Figure 5 For the present invention Figure 1 Flow chart of the embodiment of part of step S102;

[0050] Figure 6 For the present invention Figure 1 Flow chart of the embodiment of part of step S102;

[0051] Figure 7 A schematic structural diagram of an embodiment of the phononic crystal design device provided by the present invention;

[0052] Figure 8 A schematic structural diagram of an embodiment of an electronic device provided by the present invention;

[0053] Figure 9 A schematic diagram of the convolutional neural network model structure provided by the present invention;

[0054] Figure 10 This is a schematic diagram of the band gap label and band gap distribution of the phononic crystal provided by the present invention. DETAILED DESCRIPTION

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0056] In the description of the embodiments of the present invention, unless otherwise specified, "plurality" means two or more. "And / or" describes the association relationship between associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone.

[0057] The terms "first," "second," and so on, used in the embodiments of the present invention are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature designated as "first" or "second" may explicitly or implicitly include at least one such feature.

[0058] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0059] Before describing the embodiments, the following definitions are given for the relevant terms:

[0060] A latent variable is an auxiliary variable used to indicate the probability that a sample belongs to a specific distribution. It is a discrete random variable used to indicate that a sample belongs to a specific distribution or category.

[0061] A bandgap phononic crystal is an artificial structure or material that can influence the propagation characteristics of mechanical waves by designing its periodic structure or geometric properties. The important characteristic of a bandgap phononic crystal is the ability to design a special bandgap, which can isolate vibrations within a specific frequency range.

[0062] The present invention provides a bandgap phononic crystal design method, device, electronic device and storage medium, which are described below respectively.

[0063] Figure 1 A schematic flow chart of an embodiment of the bandgap phononic crystal design method provided by the present invention is shown in FIG. Figure 1 As shown, the bandgap phononic crystal design method includes:

[0064] S101, obtaining a bandgap phononic crystal and generating a bandgap vector of the bandgap phononic crystal;

[0065] S102, constructing a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determining the expected structure of the bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector;

[0066] S103, constructing an image relationship model between the structure of the band gap phononic crystal and the latent variables, determining the expected latent variables based on the image relationship model and the structure of the expected band gap phononic crystal, and generating image features of the expected phononic crystal based on the expected latent variables;

[0067] S104 , generating a target phononic crystal according to the structure of the expected band gap phononic crystal and the image features of the expected phononic crystal.

[0068] It should be noted that the model for generating the target phononic crystal consists of two parts: a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and an image relationship model between the structure of the bandgap phononic crystal and the latent variables. The mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector is constructed based on a fully connected neural network, and the image relationship model between the structure of the bandgap phononic crystal and the latent variables is constructed by an autoencoder, including an encoder and a decoder.

[0069] It should be further explained that the band gap vector is a quantitative description of the band gap characteristics of the phononic crystal, including key parameters such as the band gap width and position. These parameters comprehensively reflect the vibration suppression ability of the phononic crystal within a specific frequency range.

[0070] Compared with the prior art, this embodiment trains a model based on bandgap phononic crystals and the bandgap vectors of the bandgap phononic crystals. The model consists of two parts: a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and an image relationship model between the structure of the bandgap phononic crystal and the hidden variables. Then, the mapping relationship between the structure of the bandgap phononic crystal and the bandgap vector is analyzed, and a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector is constructed to generate the structure of the expected bandgap phononic crystal. The mapping relationship between the structure of the bandgap phononic crystal and the hidden variables is analyzed, and an image relationship model between the structure of the bandgap phononic crystal and the hidden variables is constructed to generate the image features of the expected phononic crystal. Finally, according to the structure of the expected bandgap phononic crystal and the image features of the expected phononic crystal, the target phononic crystal is generated. The phononic crystal design model constructed by the present application through the mapping relationship between the structure of the phononic crystal and the bandgap vector and the mapping relationship between the phononic crystal structure and the hidden variables can flexibly set the bandgap distribution vector within the target frequency band as the model input as needed, and accurately generate the target bandgap phononic crystal structure.

[0071] This embodiment constructs a phononic crystal design model by mapping the structure of the bandgap phononic crystal to the bandgap vector and the implicit variables of the bandgap phononic crystal structure. The model takes the expected bandgap vector as input and ultimately outputs the target bandgap phononic crystal. The bandgap distribution within the target frequency band can be flexibly set according to expected needs, and its bandgap vector can be determined. The target bandgap phononic crystal structure can then be accurately generated based on the bandgap vector.

[0072] In some embodiments of the present invention, Figure 2 As shown, Figure 2 The present invention provides Figure 1 The flowchart of an embodiment of step S101 includes:

[0073] S201, generating multiple random phononic crystals;

[0074] S202. Analyze the structure of random phononic crystals based on convolutional neural networks and construct a bandgap phononic crystal screening model.

[0075] S203 , based on the band gap phononic crystal screening model, outputting the band gap phononic crystal and the band gap vector of the band gap phononic crystal according to the random phononic crystal.

[0076] It should be noted that the convolutional neural network consists of an input layer, a convolution layer, a pooling layer, a fully connected layer, and an output layer. The specific structure is shown in the following figure. Figure 9 , input the generated random phononic crystal into the input layer, and output the band gap phononic crystal and the band gap vector of the band gap phononic crystal.

[0077] In some embodiments of the present invention, Figure 3 As shown, Figure 3 The present invention provides Figure 2 The flowchart of an embodiment of step S201 includes:

[0078] S201, determining two-dimensional single-sided pixels and constructing a basic unit including a random number of bases and scatterers;

[0079] S202, symmetric and folding the basic unit to generate a two-dimensional phononic crystal;

[0080] S203, screening out two-dimensional phononic crystals whose scatterers meet connectivity requirements, and generating random phononic crystals.

[0081] Specifically, first, determine that the single-sided pixel of the two-dimensional phononic crystal is 18; construct a 9×9 all-zero matrix a as the basic unit in MATLAB; then, change the 0s in random n positions in the basic unit to 1s, where 0 represents the matrix rubber material and 1 represents the scatterer aluminum material; perform symmetry and fold operations on the basic unit to generate a complete two-dimensional phononic crystal structure; then judge the connectivity of the aluminum material, and if it meets the requirements, output the two-dimensional phononic crystal structure, otherwise regenerate a new two-dimensional phononic crystal structure; finally, stretch the output two-dimensional phononic crystal structure along the normal direction to obtain a three-dimensional phononic crystal structure, which is a random phononic crystal.

[0082] It should be noted that 50% of the generated random phononic crystals are positive data and 50% are negative data. The positive data are gapped phononic crystals and the negative data are gapless phononic crystals. 50% of the positive data are screened out, and 80% of these positive data are used as training data sets, and 20% of the positive data are used as test data sets to train and test the model.

[0083] In some embodiments of the present invention, Figure 4 As shown, Figure 4 The present invention provides Figure 3 The flowchart of step S301 in an embodiment includes:

[0084] S401, analyzing the structure of the random phononic crystal based on finite element analysis, performing a frequency test on the random phononic crystal, and determining the band gap distribution of the random phononic crystal;

[0085] S402, generating a band gap label of the random phononic crystal according to the band gap distribution;

[0086] S403: Training a convolutional neural network based on the random phononic crystals, the band gap labels, and the band gap distribution to generate a band gap phononic crystal screening model.

[0087] It should be noted that, first, the COMSOL Multiphysics finite element software is called through Matlab to define two labels for the phononic crystal structure: the first label and the second label. The first label represents whether the structure has a band gap. If the phononic crystal structure has a band gap, the label is set to 1, otherwise it is set to 0. The second label represents the structure's band gap distribution. A 1×2000 vector is defined as the band gap distribution, corresponding to frequencies of 1 to 2000 Hz. The phononic crystal is tested for frequency. If there is a band gap at the corresponding frequency, the frequency position is set to 1, otherwise it is set to 0. Based on the band gap distribution, a band gap distribution frequency waveform is plotted. Combining the band gap phononic crystal structure and the band gap distribution, the band gap vector can be obtained.

[0088] Further, if Figure 10 As shown, Figure 10 The present invention provides a schematic diagram of the band gap label and band gap distribution of the phononic crystal, wherein the band gap phononic crystal is subjected to a frequency test, and the band gap of the band gap phononic crystal at a frequency of 1243 to 1901 Hz is obtained by finite element statistics. According to the band gap distribution values ​​​​in the frequency range of 1243 to 1901 Hz, the band gap distribution values ​​​​in the frequency range of 1243 to 1901 Hz are all set to 1, and the band gap distribution values ​​​​in the remaining frequency ranges are set to 0.

[0089] Specifically, this embodiment uses 2000 sets of phononic crystals and band gap vectors as input training samples, sets the initial learning rate to 0.008, reduces the learning rate to half of the original after every 1800 training times, and shuffles the samples into the neural network for training, where 80% of the samples are used as training sets and 20% of the samples are used as test sets. At the same time, to avoid overfitting of the neural network, a Dropout layer is added after each convolutional layer, and 20% of the neurons are randomly selected and not activated.

[0090] This embodiment trains a convolutional neural network using random phononic crystals, bandgap labels, and bandgap distributions to construct a bandgap phononic crystal screening model. The bandgap phononic crystal screening model can quickly and accurately analyze randomly generated phononic crystal structures and output bandgap phononic crystals and their bandgap vectors. Although the above specific implementation steps can also screen out bandgap phononic crystals and their bandgap vectors through finite element analysis, they require a large amount of computing resources and have poor computational efficiency. Therefore, a convolutional neural network is used to construct a model that can quickly and accurately screen randomly generated phononic crystals and output bandgap phononic crystals and their bandgap vectors.

[0091] In some embodiments of the present invention, Figure 5 As shown, Figure 5 The present invention provides Figure 1 The flowchart of step S102 in an embodiment includes:

[0092] S501. Build an initial model based on a fully connected neural network;

[0093] S502, using the band gap distribution as conditional information, generating a mapping relationship model between the structure of the band gap phononic crystal and the band gap vector according to the trained initial model of the structure of the band gap phononic crystal and the band gap vector;

[0094] S503 : Based on the mapping relationship model, the structure of the expected band gap phononic crystal is generated with the expected band gap vector as input.

[0095] It should be noted that the mapping between the structure and bandgap vector of a bandgap phononic crystal is highly nonlinear and multivariate. First, the structural parameters and bandgap vector of the bandgap phononic crystal were normalized to ensure consistency in the input of the fully connected neural network. A fully connected neural network model was trained using 8,000 training samples. The model consisted of seven layers: an input layer with 2,000 neurons, three hidden layers with 256 neurons each, and an output layer with 128 neurons. A nonlinear activation function, tanh, was used after each hidden and output layer to enhance the model's expressiveness. The mean squared error (MSE) function was chosen as the loss function to measure the difference between the predicted bandgap vector and the true vector, serving as the optimization objective. Nadam was used as the optimizer. During training, the fully connected network was trained using mini-batch gradient descent, with 1,000 samples trained at a time. Backpropagation was used to update the network parameters. After training, the structural parameters of the test phononic crystal were input into the fully connected neural network. The predicted results were compared with the true bandgap vector to optimize the network model parameters.

[0096] This embodiment trains a fully connected neural network, analyzes the mapping relationship between the structure of the bandgap phononic crystal and the bandgap vector, and constructs a mapping relationship model. Through the mapping relationship model, the structural characteristics of the bandgap phononic crystal can be obtained by inputting the bandgap vector. By adjusting the input parameters of the bandgap vector, the expected bandgap phononic crystal structure can be obtained, thereby achieving a fast, accurate and flexible way to obtain the expected bandgap phononic crystal structure through the network model.

[0097] In some embodiments of the present invention, Figure 6 As shown, Figure 6 The present invention provides Figure 1 The flowchart of the embodiment of part of step S103 is as follows: constructing an image relationship model between the structure of the bandgap phononic crystal and the hidden variables, including:

[0098] S601. Construct an initial autoencoder model based on the encoder and decoder;

[0099] S602, extracting a first image feature of a bandgap phononic crystal structure of a training sample according to the encoder;

[0100] S603, mapping the first image feature into a latent variable;

[0101] S604: Based on the decoder, reconstruct the image feature according to the latent variable and the band gap phononic crystal structure of the training sample to generate a second image feature;

[0102] S605 . Based on the comparison result of the first image feature and the second image feature, the initial autoencoder model is updated and iterated to generate an image relationship model between the structure of the bandgap phononic crystal and the latent variables.

[0103] Specifically, the encoder first extracts the image features of the training sample phononic crystals through multiple convolutional layers. The extracted image features can reflect the physical structure and band gap characteristics of the phononic crystals, reflecting the local characteristics and spatial hierarchy of the phononic crystal structure image. The image is mapped to its internal representation, that is, mapped into latent variables. The relationship between the latent variables and the training sample band gap phononic crystal structure is analyzed, and the latent variables that have a significant impact on the quality of the reconstructed image are identified. Then, the extracted latent variables are used as input to the decoder. The image features are reconstructed based on the latent variables and the training sample band gap phononic crystal structure. The reconstructed image features are compared with the original image features of the phononic crystal. The model is iteratively updated based on the comparison results, and finally an image relationship model between the structure of the band gap phononic crystal and the latent variables is generated. Using the phononic crystal structure as the model input, the image features of the phononic crystal structure can be generated.

[0104] This embodiment extracts the image features of the phononic crystal by training the autoencoder, maps them into latent variables, and reconstructs the image based on the relationship between the latent variables and the phononic crystal structure, ensuring that the generated target phononic crystal has the same image features as the expected phononic crystal, making the band gap phononic crystal structure generated by the phononic crystal design model more accurate.

[0105] In some embodiments of the present invention, generating a target phononic crystal according to an expected band gap vector based on the phononic crystal design model includes:

[0106] Based on the image relationship model, determining the corresponding hidden variables according to the structure of the expected band gap phononic crystal;

[0107] According to the structural mapping relationship between the hidden variables and the band gap phononic crystal, the image features of the expected phononic crystal are generated.

[0108] Specifically, based on the trained image relationship model, the structure of the expected band gap phononic crystal is used as the input of the model, and the image features of the band gap phononic crystal, that is, the pixel features of the band gap phononic crystal, can be output.

[0109] This embodiment uses a trained autoencoder to generate image features of the bandgap phononic crystal based on the structure of the bandgap phononic crystal, ensuring that the generated target phononic crystal has the same image features as the expected phononic crystal, making the bandgap phononic crystal structure generated by the phononic crystal design model more accurate.

[0110] In order to better implement the phononic crystal design method in the embodiment of the present invention, based on the phononic crystal design method, correspondingly, Figure 7 As shown, an embodiment of the present invention further provides a phononic crystal design device, and the phononic crystal design device 700 includes:

[0111] The training data generation module 701 is used to obtain a bandgap phononic crystal and generate a bandgap vector of the bandgap phononic crystal;

[0112] The phononic crystal structure generation module 702 is used to construct a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determine the expected structure of the bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector;

[0113] An image feature generation module 703 is configured to construct an image relationship model between the structure of the bandgap phononic crystal and latent variables, determine the expected latent variables based on the image relationship model and the structure of the expected bandgap phononic crystal, and generate image features of the expected phononic crystal based on the expected latent variables;

[0114] The target generation module 704 is configured to generate a target phononic crystal according to the structure of the expected band gap phononic crystal and the image features of the expected phononic crystal.

[0115] The phononic crystal design device 700 provided in the above embodiment can implement the technical solution described in the above embodiment of the phononic crystal design method. The specific implementation principles of the above modules or units can be found in the corresponding contents in the above embodiment of the phononic crystal design method, which will not be repeated here.

[0116] In the embodiments of the present invention, the device for designing phononic crystals may be a standalone server, or a server network or server cluster composed of servers. For example, the device for designing phononic crystals described in the embodiments of the present invention includes, but is not limited to, a computer, a network host, a single network server, a set of multiple network servers, or a cloud server composed of multiple servers. A cloud server is composed of a large number of computers or network servers based on cloud computing.

[0117] The present invention also provides a phononic crystal design device, such as Figure 8 As shown, Figure 8 This is a block diagram of an embodiment of a phononic crystal design device provided by the present invention. The phononic crystal design device 800 can be a computing device such as a mobile terminal, desktop computer, notebook, PDA, or server. The phononic crystal design device 800 includes a processor 801 and a memory 802 , wherein the memory 802 stores a phononic crystal design program 803 .

[0118] In some embodiments, memory 802 may be an internal storage unit of a computer device, such as a hard drive or memory of the computer device. In other embodiments, memory 802 may also be an external storage device of the computer device, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. equipped on the computer device. Furthermore, memory 802 may include both an internal storage unit of the computer device and an external storage device. Memory 802 is used to store application software installed on the computer device and various types of data, such as program code installed on the computer device. Memory 802 may also be used to temporarily store data that has been output or is about to be output. In one embodiment, the phononic crystal design program 803 may be executed by the processor 801, thereby implementing the phononic crystal design method, apparatus, device, and storage medium of various embodiments of the present invention.

[0119] In some embodiments, the processor 801 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program codes or process data stored in the memory 802 , such as executing a phononic crystal design program.

[0120] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0121] The above is a detailed introduction to the phononic crystal design method, device and image processing equipment provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A bandgap phononic crystal design method, characterized in that: include: Acquire a bandgap phononic crystal and generate a bandgap vector of the bandgap phononic crystal, wherein the bandgap vector includes a bandgap width and a position; Constructing a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determining the expected structure of the bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector; Based on the encoder and decoder, build the initial autoencoder model; Extracting a first image feature of the band gap phononic crystal structure of the training sample according to the encoder; Mapping the first image feature into a latent variable; Based on the decoder, reconstruct the image feature according to the latent variable and the band gap phononic crystal structure of the training sample to generate a second image feature; According to the comparison result of the first image feature and the second image feature, the initial autoencoder model is updated and iterated to generate an image relationship model between the structure of the bandgap phononic crystal and the latent variable; Based on the image relationship model, determining the corresponding hidden variables according to the structure of the expected band gap phononic crystal; generating an expected image feature of the phononic crystal according to a structural mapping relationship between the hidden variables and the bandgap phononic crystal; A target phononic crystal is generated according to the structure of the expected band gap phononic crystal and the image features of the expected phononic crystal.

2. The bandgap phononic crystal design method according to claim 1, characterized in that: Obtain a bandgap phononic crystal and generate the bandgap vector of the bandgap phononic crystal, including: Generate multiple random phononic crystals; Analyzing the structure of the random phononic crystal based on a convolutional neural network and constructing a band gap phononic crystal screening model; Based on the band gap phononic crystal screening model, a band gap phononic crystal and a band gap vector of the band gap phononic crystal are output according to the structure of the random phononic crystal.

3. The bandgap phononic crystal design method according to claim 2, characterized in that: Generate multiple random phononic crystals, including: Determine the two-dimensional single-sided pixels and construct a basic unit containing a random number of bases and scatterers; The basic unit is symmetrically and folded to generate a two-dimensional phononic crystal; The two-dimensional phononic crystals whose scatterers satisfy connectivity are screened out, and random phononic crystals are generated.

4. The bandgap phononic crystal design method according to claim 3, characterized in that: Based on convolutional neural networks, the structure of random phononic crystals is analyzed and a bandgap phononic crystal screening model is constructed, including: Analyzing the structure of the random phononic crystal based on finite element analysis, performing frequency testing on the random phononic crystal, and determining the band gap distribution of the random phononic crystal; generating a band gap label of the random phononic crystal according to the band gap distribution; A convolutional neural network is trained based on the random phononic crystals, band gap labels, and band gap distribution to generate a band gap phononic crystal screening model.

5. The bandgap phononic crystal design method according to claim 4, characterized in that: Constructing a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determining the expected structure of the bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector, including: Build an initial model based on a fully connected neural network; Using the band gap distribution as conditional information, and based on the trained initial model of the structure and band gap vector of the band gap phononic crystal, a mapping relationship model between the structure and band gap vector of the band gap phononic crystal is generated; Based on the mapping relationship model, the structure of the expected band gap phononic crystal is generated with the expected band gap vector as input.

6. A bandgap phononic crystal design device, characterized in that: include: A training data generation module is used to obtain a bandgap phononic crystal and generate a bandgap vector of the bandgap phononic crystal, wherein the bandgap vector includes a bandgap width and a position; A phononic crystal structure generation module is used to construct a mapping relationship model between the structure of the bandgap phononic crystal and the bandgap vector, and determine the structure of the expected bandgap phononic crystal based on the mapping relationship model and the expected bandgap vector; An image feature generation module is used to construct an initial autoencoder model based on an encoder and a decoder; and extract a first image feature of the bandgap phononic crystal structure of the training sample according to the encoder; Mapping the first image feature into a latent variable; Based on the decoder, an image feature is reconstructed according to the latent variable and the structure of the bandgap phononic crystal of the training sample to generate a second image feature; based on the comparison result of the first image feature and the second image feature, an initial autoencoder model is updated and iterated to generate an image relationship model between the structure of the bandgap phononic crystal and the latent variable; and an expected latent variable is determined based on the image relationship model and the structure of the expected bandgap phononic crystal, and an image feature of the expected phononic crystal is generated according to the expected latent variable; The target generation module is used to generate a target phononic crystal according to the structure of the expected band gap phononic crystal and the image characteristics of the expected phononic crystal.

7. A bandgap phononic crystal design device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the phononic crystal design method according to any one of claims 1 to 5 are implemented.

8. A storage medium, characterized in that: The storage medium stores computer program instructions, and when the computer program instructions are executed by a computer, the computer is caused to execute the band gap phononic crystal design method according to any one of claims 1 to 5.

Images