Laminated structure material electromagnetic performance design method and system based on artificial intelligence

Abstract

The invention discloses a laminated structure material electromagnetic performance design method and system based on artificial intelligence, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining basic electromagnetic parameters of layering units through an experiment, coding the layering units with different electromagnetic absorption performances, and generating random codes of layering configuration according to time dependence; through electromagnetic simulation software, establishing a simulation database of random codes of laying layer configuration and corresponding electromagnetic absorption performance; using the simulation database to train an agent model for predicting the electromagnetic absorption performance under different laying layer configurations to obtain a prediction model; and performing prediction through the prediction model to obtain the electromagnetic absorption performance of the candidate laying layer configuration, judging whether the electromagnetic absorption performance of the candidate laying layer configuration is an optimal value or not, and if yes, determining that the candidate laying layer configuration is the optimal laying layer configuration. The problems that a large-parameter space composed of laying units and material types of an electromagnetic absorption material is large in searching difficulty, high in cost and the like are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence, and particularly relates to a method and system for electromagnetic performance design of laminated structural materials based on artificial intelligence. Background Art

[0002] Laminated composite materials are a type of composite material formed by stacking and bonding multiple layers of different materials or sheets (plies) of the same material with different orientations. Its core feature is to optimize the combination of material properties through hierarchical design, and it is widely used in fields such as aerospace, ships, and electronic stealth.

[0003] The combination of ply units and material types of laminated structural materials constitutes an extremely large parameter space, and the search efficiency of traditional trial-and-error methods is extremely low. Existing design methods rely on expert knowledge to select ply configurations, which is highly subjective and difficult to systematize. Dependent on experiments and electromagnetic simulations to verify group by group, the time-consuming for a single simulation is relatively long, and the cost of exploring a large range of parameters is too high. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. For this reason, the present invention proposes a method and system for electromagnetic performance design of laminated structural materials based on artificial intelligence, which can solve the problems such as the high difficulty and cost of searching the large parameter space composed of ply units and material types faced by electromagnetic absorption materials.

[0005] A method for electromagnetic performance design of laminated structural materials based on artificial intelligence according to an embodiment of the first aspect of the present invention includes: S100. Obtain the basic electromagnetic parameters of ply units through experiments, encode the ply units with different electromagnetic absorption performances, and generate a random code of ply configuration according to time dependence; S200. Establish a simulation database of the random code of ply configuration and the corresponding electromagnetic absorption performance through electromagnetic simulation software; S300. Use the simulation database to train a surrogate model for predicting the electromagnetic absorption performance under different ply configurations to obtain a prediction model; S400. Predict through the prediction model to obtain the electromagnetic absorption performance of the candidate ply configuration, and determine whether the electromagnetic absorption performance of the candidate ply configuration is the optimal value. If so, determine the candidate ply configuration as the optimal ply configuration.

[0006] According to some embodiments of the present invention, in S300, the step of using the simulation database to train a surrogate model for predicting the electromagnetic absorption performance under different ply configurations to obtain a prediction model includes: Input the encoded ply configuration, output the interpolation points of the reflection loss curve, and perform dimensional transformation in the deep neural network using a stacked residual network architecture to obtain a prediction model.

[0007] According to some embodiments of the present invention, a Dropout layer is introduced during the dimensional transformation in the deep neural network, and a part of the neuron outputs are randomly discarded with a preset probability in each training batch.

[0008] According to some embodiments of the present invention, it further includes evaluating the prediction model, dividing a part of the data in the simulation database into a test set, using the test set to evaluate the prediction model, and determining the prediction model according to the evaluation results.

[0009] According to some embodiments of the present invention, dividing a part of the data in the simulation database into a test set, using the test set to evaluate the prediction model, and determining the prediction model according to the evaluation results includes: The MSE loss function is used to evaluate the mean square error between the simulation value and the predicted value. The mean square error gradient descent represents obtaining a better prediction model. If it no longer decreases after a preset number of validations or reaches the upper limit of the preset number of times, the evaluation is stopped and the prediction model is obtained.

[0010] According to some embodiments of the present invention, in S400, obtaining the electromagnetic absorption performance of the candidate ply configuration, and determining whether the electromagnetic absorption performance of the candidate ply configuration is the optimal value. If so, determining the candidate ply configuration as the optimal ply configuration includes the following steps: S401. Set the frequency band target of electromagnetic absorption; S402. At the frequency band target, evaluate the effective absorption bandwidth of the candidate ply configuration by the prediction model; S403. Through Bayesian iteration, simulate a preset number of groups of candidate ply configurations obtained in each round of Bayesian iteration through the electromagnetic simulation software to obtain simulation results; S404. Backfill the simulation results into the simulation database and retrain the model until there is no ply configuration with a better effective absorption bandwidth in the new round of candidate ply configurations, which is the ply configuration optimal for the frequency band target.

[0011] According to some embodiments of the present invention, the electromagnetic absorption performance is analyzed through the simulation database, and the indexes characterizing the electromagnetic absorption performance are extracted from the data. The distribution of the indexes characterizing the electromagnetic absorption performance in the parameter space conforms to a Gaussian distribution or a near-Gaussian distribution.

[0012] According to some embodiments of the present invention, the indicators characterizing the electromagnetic absorption performance include those composed of interpolation points in a frequency range based on the reflection loss curve representing the electromagnetic absorption performance, and the formula for the corresponding indicators is: In the formula: N EAB is the number of interpolation points that meet the conditions, EAB is the effective absorption bandwidth, N freq is a judgment function for whether the reflection loss at a certain frequency meets -10 dB, freq1 is the starting frequency of the target range, and freq2 is the cut-off frequency of the target range.

[0013] According to some embodiments of the present invention, the formula for the electromagnetic absorption performance is: Wherein: EAC is the electromagnetic absorption ability, RL freq is the reflection loss value at a certain frequency.

[0014] There is also disclosed an electromagnetic performance design system for laminated structural materials based on artificial intelligence, including a memory and a processor. A determination program for the electromagnetic performance design method of laminated structural materials based on artificial intelligence is stored in the memory, and the processor runs the determination program for the electromagnetic performance design method of laminated structural materials based on artificial intelligence, so that the electromagnetic performance design system for laminated structural materials based on artificial intelligence executes the electromagnetic performance design method of laminated structural materials based on artificial intelligence.

[0015] The electromagnetic performance design method and system for laminated structural materials based on artificial intelligence according to the embodiments of the present invention have at least the following beneficial effects: By establishing a prediction model, the prediction model predicts the electromagnetic absorption performance of different ply configurations, and quickly obtains the ply configuration with the optimal electromagnetic absorption performance among different candidate ply configurations. It solves the problems of large search difficulty and high cost in the large parameter space composed of ply units and material types faced by electromagnetic absorption materials, and provides an efficient and fast design method for broadband electromagnetic absorption materials that does not rely on expert experience.

[0016] The additional aspects and advantages of the present invention will be partially given in the following description, partially become obvious from the following description, or be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The following further describes the present invention in conjunction with the drawings and embodiments, where: Figure 1 is a flowchart of the electromagnetic performance design method for laminated structural materials based on artificial intelligence according to an embodiment of the present invention; Figure 2 Flow chart of the progressive Bayesian optimization algorithm according to an embodiment of the present invention; Figure 3 Probability distribution diagram of EAB in 10,000 groups of simulation data according to an embodiment of the present invention; Figure 4 Corresponding relationship diagram between EAB and EAC in 10,000 groups of simulation data according to an embodiment of the present invention; Figure 5 Probability distribution diagram of EAC in 10,000 groups of simulation data according to an embodiment of the present invention; Figure 6 Linear fitting curve of the data predicted by the prediction model and the simulation data after Bayesian iteration according to an embodiment of the present invention, and the linear correlation degree is represented by the coefficient of determination R and the relative error; Figure 7 Bayesian iteration process according to an embodiment of the present invention, and the prediction situation of the first 100 candidates in one round of iteration; Figure 8 Schematic diagram of a laminated structural material composed of three materials and a total of 34 layers according to an embodiment of the present invention; Figure 9 is Figure 8 Reflection loss curve diagram of the prediction and simulation results of the represented laminated structural material in the full frequency band of 1 - 18 GHz; Figure 10 Reflection loss curve diagram of the prediction and simulation of the laminated structural material optimized and designed for the 1 - 4 GHz sub - frequency band; Figure 11 Reflection loss curve diagram of the prediction and simulation of the laminated structural material optimized and designed for the 4 - 8 GHz sub - frequency band; Figure 12 Reflection loss curve diagram of the prediction and simulation of the laminated structural material optimized and designed for the 8 - 18 GHz sub - frequency band; Figure 13 Structural block diagram of the electromagnetic performance design system of the laminated structural material according to an embodiment of the present invention. Detailed implementation manners

[0018] The embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation of the present invention.

[0019] In the description of the present invention, it should be understood that for the orientation description, such as the upper and lower directions, the orientation or positional relationship indicated is based on the orientation or positional relationship shown in the drawings. This is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention.

[0020] In the description of the present invention, "a plurality of" refers to more than two. If there is a description of "first" and "second", it is only for the purpose of distinguishing technical features, and should not be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence of the indicated technical features.

[0021] In the description of the present invention, unless otherwise clearly defined, terms such as "set", "installed", "connected", etc. should be understood in a broad sense. Those skilled in the art can reasonably determine the specific meanings of the above terms in the present invention in combination with the specific content of the technical solution.

[0022] Term Explanation: Laminate configuration: The layering sequence of the laminated structural material from top to bottom, which shows how to lay the layers; the layering unit is a layer of material and is one in the layering sequence; layering means stacking layer by layer.

[0023] Refer to Figure 1 As shown, the present invention discloses an electromagnetic performance design method for laminated structural materials based on artificial intelligence, including: S100. Obtain the basic electromagnetic parameters of the layering unit through experiments, encode the layering units with different electromagnetic absorption performances, and generate a random code of the laminate configuration according to time dependence; S200. Establish a simulation database of the random code of the laminate configuration and the corresponding electromagnetic absorption performance through electromagnetic simulation software; S300. Use the simulation database to train a surrogate model for predicting the electromagnetic absorption performance under different laminate configurations to obtain a prediction model; S400. Predict through the prediction model to obtain the electromagnetic absorption performance of the candidate laminate configuration, and determine whether the electromagnetic absorption performance of the candidate laminate configuration is the optimal value. If so, determine the candidate laminate configuration as the optimal laminate configuration.

[0024] In some embodiments of the invention, the basic electromagnetic parameters of the layering unit are obtained through experiments, including the real part and the imaginary part of the permittivity, and the real part and the imaginary part of the permeability.

[0025] Batch processing of combined CST Studio finite element simulations is achieved using an auto - encoding script. Through the above - mentioned processing, a simulation database of approximately 10,000 groups of data is established. The purpose is that electromagnetic simulation for each randomly generated ply configuration is a cumbersome and repetitive task. To reduce the time cost and uncertain errors brought by manual processing, a script is used to import each ply configuration into the electromagnetic simulation software, assign material properties, and simultaneously set a series of electromagnetic simulation boundaries, modes, frequency ranges, etc.

[0026] The integrated simulation database is randomly divided into 80% as the training set, 15% as the validation set, and 5% as the test set. The training set is used to train the prediction model, which represents the implicit mapping relationship between the ply configuration and the electromagnetic absorption performance. The validation set is used to validate the trained prediction model, find a better architecture, and improve the generalization ability of the model. The test set is used to evaluate the performance. The MSE loss function is used to evaluate the mean square error between the simulation values and the predicted values. The mean square error gradient descent represents obtaining a better prediction model. If it no longer decreases after 50 validations or reaches the upper limit of 300 times, the evaluation stops and the prediction model is obtained.

[0027] The trained prediction model can predict the electromagnetic absorption performance under a specified ply code, achieving low - cost and high - efficiency prediction and subsequent iterative optimization of unknown ply configurations. By inputting the codes corresponding to different candidate ply configurations and outputting the corresponding electromagnetic absorption performance, the ply configuration with the best electromagnetic absorption performance is selected from them, and the corresponding ply configuration is the structure of the laminated structural material.

[0028] The present invention solves the problems of high difficulty and high cost in searching the large - parameter space composed of ply units and material types faced by electromagnetic absorption materials, and provides an efficient and rapid design method for broadband electromagnetic absorption materials that does not rely on expert experience.

[0029] In some embodiments of the present invention, in S300, training the surrogate model for predicting the electromagnetic absorption performance under different ply configurations using the simulation database to obtain the prediction model includes: Inputting the encoded ply configuration, with the output being the interpolation points of the reflection loss curve, and using a stacked residual network architecture for dimensional transformation in the deep neural network to obtain the prediction model.

[0030] In this embodiment, the simulation database trains a surrogate model for predicting the electromagnetic absorption performance under different ply configurations. The input is the encoded ply configuration, and the output is the interpolation points of the reflection loss curve. A stacked residual network architecture is used for dimensional transformation in the deep neural network, which includes an input layer, multiple hidden layers, and an output layer. The number of neurons in each layer changes in a certain multiple gradient with the number of layers, solving the problems of gradient vanishing and gradient explosion in deep networks. In backpropagation, the gradient is passed layer by layer through the chain rule. The absolute value of the derivative of the activation function is usually less than 1, and the gradient will decay exponentially after being multiplied by multiple layers, resulting in difficulty in updating the parameters of the shallow network; or if the weight initialization is too large or the derivative of the activation function is greater than 1, the gradient may grow exponentially, leading to unstable parameter updates. The inter-layer transformation of traditional networks is linear, while the residual connection introduces non-linear superposition, diversifying the gradient propagation path, enhancing stability, and improving the depth and expressive ability of the network. During this process, by inputting different data, the internal architecture of the neural network is continuously trained. Each time a dimensional transformation is completed, a prediction model is obtained, and the mean square error between the simulation value and the prediction value is evaluated using the MSE loss function. The mean square error gradient descent represents obtaining a better prediction model. If it no longer decreases after a preset number of validations or reaches the upper limit of 300 times, the evaluation is stopped and the prediction model is obtained.

[0031] Further, a Dropout layer is introduced during the dimensional transformation in the deep neural network, and a part of the neuron outputs are randomly discarded with a preset probability in each training batch.

[0032] A Dropout layer is introduced during the dimensional transformation, and a part of the neuron outputs are randomly discarded with a 25% probability in each training batch, reducing the model complexity and computational cost, while increasing the robustness and generalization ability of the model.

[0033] In some embodiments of the present invention, it also includes evaluating the prediction model. A part of the data in the simulation database is divided into a test set, and the test set is used to evaluate the prediction model, and the prediction model is determined according to the evaluation result.

[0034] Further, dividing a part of the data in the simulation database into a test set, using the test set to evaluate the prediction model, and determining the prediction model according to the evaluation result includes: The mean square error between the simulation value and the prediction value is evaluated using the MSE loss function. The mean square error gradient descent represents obtaining a better prediction model. If it no longer decreases after a preset number of validations or reaches the upper limit of the preset number of times, the evaluation is stopped and the prediction model is obtained.

[0035] In this embodiment, 80% of the integrated simulation database is randomly divided as the training set, 15% as the validation set, and 5% as the test set. The training set is used to train the prediction model, which represents the implicit mapping relationship between the ply configuration and the electromagnetic absorption performance. The validation set is used to validate the trained prediction model, find a better architecture, and improve the generalization ability of the model. The test set is used to evaluate the performance. The MSE loss function is used to evaluate the mean square error between the simulation values and the predicted values. The mean square error gradient descent represents obtaining a better prediction model. If it no longer decreases after 50 validations or reaches the upper limit of 300 times, the evaluation stops and the prediction model is obtained.

[0036] In some embodiments of the present invention, the electromagnetic absorption performance is analyzed by analyzing the simulation database, and the indexes characterizing the electromagnetic absorption performance in the data are extracted. The distribution of the indexes characterizing the electromagnetic absorption performance in the parameter space conforms to a Gaussian distribution or a near-Gaussian distribution.

[0037] Reference Figures 3 to 5 shown in Figure 3 shows the analysis of the probability distribution of the effective absorption bandwidth (EAB) of 10,000 groups of data. Among them, the proportion in the range of 4 - 5 GHz is the largest, which is 32.95%, while the probability in the range of 7 - 8 GHz is only 0.52%. In the present invention, in order to provide guidance for the ply design with excellent electromagnetic absorption ability, our focus is on the distribution proportion of the widest EAB in the histogram and the corresponding ply configuration. However, from the data distribution of EAB, the efficiency of exploring this focus through a large number of simulation calculations is low, and the obtained EAB is still relatively narrow.

[0038] Generally, the electromagnetic absorption performance can be characterized by the electromagnetic absorption capacity (EAC), and its formula is: where: EAC is the electromagnetic absorption capacity, and the overall absorption capacity of the laminated structure material is shown in the form of the integral area. RL freq is the reflection loss value at a certain frequency, freq1 is the starting frequency of the target range, and freq2 is the cut-off frequency of the target range.

[0039] By studying the scatter distribution state of the electromagnetic absorption capacity (EAC) with respect to the EAB, as Figure 4 shown, it can be found that the EAB corresponding to the EAC value varies within a relatively large numerical range. For example, when the EAC value is 133 dB·GHz, the variable range of the EAB is 0.76 - 7.08 GHz, and as Figure 5The probability distribution of EAC shown is mostly concentrated in the range of 121 - 141 dB·GHz, accounting for up to 82.42%. This indicates that among 10,000 groups of data, up to 82.42% of the ply configurations have similar EAC. The highly concentrated data will mask the characteristics and patterns of the ply configurations, thus making it difficult to predict the mapping relationship between electromagnetic performance and ply configuration.

[0040] Conversely, EAB shows a nearly Gaussian distribution in the data structure within the range from 0 to 8 GHz, indicating that using the EAB value can better characterize the electromagnetic absorption performance related to the ply configuration, which is conducive to exploring the implicit mapping relationship between the two. Therefore, in the following, we select the EAB represented by the N EAB value to perform the reverse design of the ply based on Bayesian optimization.

[0041] In some embodiments of the present invention, the indicators characterizing the electromagnetic absorption performance include being composed of interpolation points in the frequency range based on the reflection loss curve representing the electromagnetic absorption performance. The formula for the corresponding indicator is: In the formula: N EAB is the number of interpolation points meeting the conditions, EAB is the effective absorption bandwidth (Effective absorption bandwidth), N freq is the judgment function for whether the reflection loss at a certain frequency meets -10 dB, freq1 is the starting frequency of the target range, and freq2 is the cut-off frequency of the target range. Through the above formula, the effective bandwidth of the frequency band within which the reflection loss of the material meets the electromagnetic stealth requirement is calculated.

[0042] Reference Figure 2 As shown, in some embodiments of the present invention, in S400, through the prediction of the prediction model, the electromagnetic absorption performance of the candidate ply configuration is obtained, and it is judged whether the electromagnetic absorption performance of the candidate ply configuration is the optimal value. If so, determining the candidate ply configuration as the optimal ply configuration includes the following steps: S401. Set the frequency band target of electromagnetic absorption; S402. Evaluate the effective absorption bandwidth of the candidate ply configuration by the prediction model within the frequency band target; S403. Through Bayesian iteration, simulate the preset number of groups of candidate ply configurations obtained in each round of Bayesian iteration through the electromagnetic simulation software to obtain the simulation results; S404. Backfill the simulation results into the simulation database and retrain the model until there are no ply configurations with better effective absorption bandwidths in the newly selected candidate ply configurations, then it is the optimal ply configuration for the frequency band target.

[0043] ReferenceFigure 8 , which is a schematic diagram of the ply configuration of the laminated structural material provided by the specific embodiment of the present invention. As can be seen from the figure, this structure is composed of three materials and shows a certain gradient distribution.

[0044] Reference Figure 9 , in this embodiment, the electromagnetic absorption frequency band target is 1 - 18 GHz, that is, freq1 = 1 GHz and freq2 = 18 GHz. Reference Figure 7 As shown, during the progressive Bayesian optimization iteration process, the electromagnetic simulation results of the first 100 candidate ply configurations in the new round are continuously added to retrain the prediction model until no better electromagnetic absorption performance appears in the candidate ply configurations of the new round.

[0045] Use the trained prediction model to approximate the electromagnetic absorption performance of different ply configurations. Compared with the CST Studio finite element simulation, the speed is increased by about 10 6 times (0.02 ms) while maintaining a high prediction accuracy. However, considering the extensive parameter space (3 34 ), the challenge lies in how to find the mapping relationship between the ply configuration and the electromagnetic performance and further accelerate the inverse design of the electromagnetic performance.

[0046] After every 100 times of Bayesian optimization, the optimization results are simulated by CST Studio and added to the database, and the prediction model is retrained to continuously improve the accuracy and generalization ability of the prediction model.

[0047] Change and set the electromagnetic target according to the iteration effect, and use the optimal solution under the previous electromagnetic target as the prior condition for the subsequent process after each change.

[0048] By building a machine learning framework with a joint progressive Bayesian optimization algorithm, the prediction and optimal design of the unknown ply configuration in the total design space of 3 Figure 8 shown, which is composed of three electromagnetic absorption materials, are realized. As 34 shown, after progressive optimization, the coefficient of determination of the prediction model is 0.964, and the mean absolute error is 0.27 GHz, which proves the reliability of the prediction model. Figure 6

[0049] Figure 9 As Figure 9 shown, the electromagnetic performance prediction diagram of the laminated structural material provided by this embodiment. As can be seen from the figure, with the electromagnetic absorption frequency band target of 1 - 18 GHz, it can be known that the broadband electromagnetic absorption laminated structural material after optimal design realizes an effective absorption bandwidth of 9.6 GHz, and the simulation results and the prediction results have a good matching degree.

[0050] Expand the electromagnetic target to other ranges and cycle the above process to finally obtain the ply configuration of the effective absorption bandwidth of the target in this frequency band. Refer to Figures 10 to 12 , taking 1 - 4 GHz, 4 - 8 GHz, and 8 - 18 GHz as the frequency band targets respectively, after the same steps S402 to S404 as above, the optimal ply configuration of the target in this frequency band is obtained.

[0051] As Figures 10 to 12 shown, they are respectively the electromagnetic performance prediction and simulation diagrams of the laminated structure materials optimized for the frequency band targets of 1 - 4 GHz, 4 - 8 GHz, and 8 - 18 GHz. The width of the frequency band below - 10 dB shown in the figure indicates that good electromagnetic absorption has been achieved for the sub - band.

[0052] Refer to Figure 13 shown, the present invention also discloses an electromagnetic performance design system for laminated structure materials based on artificial intelligence, including a memory and a processor. A determination program for the electromagnetic performance design method of laminated structure materials based on artificial intelligence is stored in the memory. The processor runs the determination program for the electromagnetic performance design method of laminated structure materials based on artificial intelligence, so that the electromagnetic performance design system for laminated structure materials based on artificial intelligence executes the electromagnetic performance design method of laminated structure materials based on artificial intelligence.

[0053] Since the electromagnetic performance design system for laminated structure materials based on artificial intelligence adopts all the technical solutions of the electromagnetic performance design method of laminated structure materials based on artificial intelligence in the above - mentioned embodiment, it at least has all the beneficial effects brought by the technical solutions of the above - mentioned embodiment, which will not be elaborated here.

[0054] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above - mentioned embodiments. Various changes can be made without departing from the gist of the present invention within the scope of knowledge possessed by those of ordinary skill in the art to which the present invention pertains.

Claims

1. An electromagnetic property design method for laminated structural materials based on artificial intelligence, characterized in that, Including: S100. Obtain the basic electromagnetic parameters of the laminate unit through experiments, encode the laminate units with different electromagnetic absorption performances, and generate a random code for the laminate configuration according to time dependence; S200. Establish a simulation database of the random code of the laminate configuration and the corresponding electromagnetic absorption performance through an electromagnetic simulation software; S300. Use the simulation database to train a surrogate model for predicting the electromagnetic absorption performance under different laminate configurations to obtain a prediction model; S400. Predict through the prediction model to obtain the electromagnetic absorption performance of the candidate laminate configuration, and determine whether the electromagnetic absorption performance of the candidate laminate configuration is the optimal value. If so, determine the candidate laminate configuration as the optimal laminate configuration.

2. The electromagnetic property design method of the laminated structure material based on artificial intelligence according to claim 1, characterized in that In S300, the step of using the simulation database to train a surrogate model for predicting the electromagnetic absorption performance under different laminate configurations to obtain a prediction model includes: Input the encoded laminate configuration, and the output is the interpolation points of the reflection loss curve. Use a stacked residual network architecture to perform dimensional transformation in the deep neural network to obtain a prediction model.

3. The electromagnetic property design method of the laminated structure material based on artificial intelligence according to claim 2, characterized in that Introduce a Dropout layer during the dimensional transformation process in the deep neural network, and randomly discard a part of the neuron outputs with a preset probability in each training batch.

4. The electromagnetic property design method of the laminated structure material based on artificial intelligence according to claim 1, characterized in that It also includes evaluating the prediction model, dividing a part of the data in the simulation database into a test set, using the test set to evaluate the prediction model, and determining the prediction model according to the evaluation result.

5. The electromagnetic property design method of the laminated structure material based on artificial intelligence according to claim 4, characterized in that, The step of dividing a part of the data in the simulation database into a test set, using the test set to evaluate the prediction model, and determining the prediction model according to the evaluation result includes: Use the MSE loss function to evaluate the mean square error between the simulation value and the prediction value. The mean square error gradient descent represents obtaining a better prediction model. When it no longer decreases after a preset number of validations or reaches the upper limit of the preset number of times, stop the evaluation and obtain the prediction model.

6. The electromagnetic property design method of the laminated structure material based on artificial intelligence according to claim 1, characterized in that, The electromagnetic absorption performance is analyzed through the simulation database, and the indexes characterizing the electromagnetic absorption performance are extracted from the data. The distribution of the indexes characterizing the electromagnetic absorption performance in the parameter space conforms to a Gaussian distribution or a near-Gaussian distribution.

7. The electromagnetic property design method of the laminated structural material based on artificial intelligence according to claim 6, characterized in that The indexes characterizing the electromagnetic absorption performance are composed of interpolation points in the frequency range based on the reflection loss curve representing the electromagnetic absorption performance. The formula for the corresponding index is: Where: N EAB is the number of interpolation points that meet the conditions, EAB is the effective absorption bandwidth, N freq is a judgment function for whether the reflection loss at a certain frequency meets -10 dB, freq1 is the starting frequency of the target range, and freq2 is the cutoff frequency of the target range.

8. The electromagnetic property design method of the laminated structural material based on artificial intelligence according to claim 7, characterized in that In S400, the step of obtaining the electromagnetic absorption performance of the candidate laminate configuration, determining whether the electromagnetic absorption performance of the candidate laminate configuration is the optimal value, and if so, determining the candidate laminate configuration as the optimal laminate configuration includes the following steps: S401. Set the frequency band target of electromagnetic absorption; S402. Evaluate the effective absorption bandwidth of the candidate laminate configuration by the prediction model at the frequency band target; S403. Through Bayesian iteration, simulate the preset number of groups of candidate laminate configurations obtained in each round of Bayesian iteration through the electromagnetic simulation software to obtain simulation results; S404. Backfill the simulation results into the simulation database and retrain the model until there is no laminate configuration with a better effective absorption bandwidth in the new round of candidate laminate configurations, which is the optimal laminate configuration for the frequency band target.

9. The electromagnetic property design method of the laminated structure material based on artificial intelligence according to claim 1, characterized in that The formula for the electromagnetic absorption performance is as follows: Where: EAC is the electromagnetic absorption capacity, RL freq is the reflection loss value at a certain frequency, freq1 is the starting frequency of the target range, and freq2 is the cutoff frequency of the target range.

10. An electromagnetic property design system for laminated structure materials based on artificial intelligence, characterized in that: It includes a memory and a processor. A determination program for the electromagnetic performance design method of the laminated structure material based on artificial intelligence is stored in the memory. The processor runs the determination program for the electromagnetic performance design method of the laminated structure material based on artificial intelligence, so that the electromagnetic performance design system of the laminated structure material based on artificial intelligence executes the electromagnetic performance design method of the laminated structure material based on artificial intelligence according to any one of claims 1 to 9.

Images