Intelligent design and manufacturing method of mosaic type electromagnetic frequency selection laminated composite material

Abstract

This invention provides an intelligent design and manufacturing method for mosaic-type electromagnetic frequency-selective laminated composite materials, relating to the field of electromagnetic frequency-selective composite material design and manufacturing technology. By establishing a mapping relationship database, training a deep neural network model, and invoking optimization algorithms for automatic optimization design, combined with pixelated manufacturing processes, it efficiently covers the global design domain, reduces iteration cycles, lowers trial-and-error costs, and solves the problem of low manufacturing reliability.

Description

Technical Field

[0001] This application relates to the field of electromagnetic frequency selective composite material design and manufacturing technology, and in particular to a method for intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials. Background Technology

[0002] With the rapid development of aerospace, 5G communication, and terahertz technology, the demand for precise electromagnetic environment control and multifunctional material integration continues to increase. Electromagnetic frequency-selective composite materials, possessing both frequency selectivity and structural load-bearing capabilities, have become a key research direction, with significant application value in areas such as aircraft stealth and communication equipment anti-interference. Within the existing technological framework, to avoid delamination failure of laminated composite materials caused by metal-based mosaic-type frequency-selective surfaces, researchers generally adopt resin matrix schemes loaded with functionalized fillers to achieve structural-functional integration. However, for complex electromagnetic frequency selection requirements, traditional design methods mainly rely on experimental measurements, theoretical derivations, and finite element simulation techniques, exploring electromagnetic performance by manually adjusting the material combinations and in-plane patterns of mosaic units. This approach faces multiple bottlenecks: mosaic structures involve diverse combinations of electromagnetically absorbing and electromagnetically transparent materials, with a parameter space dimension far exceeding that of conventional laminated structures, including numerous variables such as material type, filler particle size, load ratio, unit size, and spatial arrangement. This makes it difficult for traditional experimental and simulation methods to efficiently cover the global design domain. The design process relies heavily on individual engineers' experience and lacks a systematic intelligent optimization mechanism, resulting in lengthy iteration cycles and high trial-and-error costs, severely hindering the rapid development and practical deployment of materials. Simultaneously, in the manufacturing stage, significant differences in the thermodynamic properties and interfacial characteristics of different functional unit materials easily lead to stress concentration and insufficient bonding strength during layup, causing reliability issues such as interlayer delamination, interfacial debonding, and electromagnetic performance degradation. Therefore, overcoming the design efficiency bottleneck under large parameter spaces, establishing an automated mapping mechanism from electromagnetic performance requirements to structural patterns, and simultaneously solving the high-reliability manufacturing challenges of multi-material systems have become core challenges urgently needing to be addressed in the field of electromagnetic frequency selective composite materials.

[0003] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0004] To reduce the difficulty and cost of searching large parameter spaces for mosaic-type electromagnetic frequency selective composite materials in terms of material types and in-plane design composition, and to achieve highly reliable manufacturing, this application provides a smart design and manufacturing method for mosaic-type electromagnetic frequency selective laminated composite materials.

[0005] Firstly, the intelligent design and manufacturing method for mosaic-type electromagnetic frequency selective laminated composite materials provided in this application adopts the following technical solution: A method for intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials, comprising: The electromagnetic parameters of the basic material are obtained through experimental measurement, and the electromagnetic parameters include at least the permittivity and permeability. Based on the electromagnetic parameters, a database of mapping relationships between the geometric patterns and electromagnetic performance parameters of in-plane mosaic structures is established through batch electromagnetic simulation. Using the mapping database as training samples, a deep neural network model is trained using the database to obtain a target model that can predict the electromagnetic properties of an input mosaic structure pattern. Based on application requirements, the wave transmission performance index and wave absorption performance index of the laminated composite material in the target frequency band are set, and the objective function is optimized based on this definition. Using the target model as a fast performance evaluation tool, the progressive multi-objective Bayesian optimization algorithm is invoked, and the mosaic structure pattern is automatically optimized based on the optimization objective function to obtain the optimal structure pattern that meets the performance requirements. Based on the optimal structural pattern, a mosaic-type electromagnetic frequency selective laminated composite material is manufactured by spatially pixelating and compositing the electromagnetic absorbing base material and the electromagnetic transparent base material through a multilayer resin film melt impregnation process.

[0006] Optionally, the electromagnetic absorbing material comprises at least one absorbent filler selected from iron carbonyl, iron-cobalt alloy, or iron oxide, with a particle size of 5 μm–20 μm and a filler loading percentage of 1%–80% by mass. The electromagnetically transparent material contains at least one transparent filler selected from silicon dioxide, aluminum oxide, or boron nitride, with a particle size of 3μm–20μm; The base material is prepared into a laminated composite material by blending a resin matrix with fillers and then performing coating, hot rolling, and molding processes.

[0007] Optionally, the batch electromagnetic simulation includes: The in-plane mosaic structure is encoded as a binary matrix, where the first value represents an electromagnetic absorption unit and the second value represents an electromagnetic transparency unit. The electromagnetic simulation software is driven by automated scripts to perform parametric modeling and simulation of the encoded mosaic structure. The output electromagnetic performance data includes reflectivity R, transmittance T, and absorptivity A, and is expressed by the formula... The calculation yielded, where and These are the scattering parameters output from the simulation, in dB.

[0008] Optionally, the deep neural network model adopts a stacked residual network architecture and introduces a Dropout layer during training to prevent overfitting; Using the mosaic structure code as input, the output is the interpolation points of the electromagnetic transmittance curve and the absorptivity curve within the frequency band. The model training uses mean squared error as the loss function, and the network hyperparameters are optimized using a validation set.

[0009] Optionally, the objective function includes: The electromagnetic frequency selection range refers to the values ​​of a and b in the formula, expressed in GHz, and falls within the frequency range of a GHz to b GHz. and These represent the number of interpolation points that satisfy the conditions. and It can be set to a fixed value between 0 and 1; This is a function to determine whether the transmittance or absorptivity at a certain frequency satisfies electromagnetic absorption or terahertz communication requirements.

[0010] Optionally, the progressive multi-objective Bayesian optimization algorithm includes: Iteratively call the prediction model to evaluate the electromagnetic properties of the mosaic structure; The potential optimal structure is selected based on the acquisition function and then verified through simulation. New data is fed back into the prediction model to incrementally update the training set, gradually improving model accuracy and optimization efficiency; The optimization process synchronously adjusts the frequency band threshold in the objective function. and To balance the wave transmission and absorption properties.

[0011] Optional, pixel-level manufacturing includes: The electromagnetic absorbing resin film and the electromagnetic transparent resin film are cut and spliced ​​according to the optimized mosaic pattern. The adhesive film is laid layer by layer on the fiber cloth and prepreg is formed by hot rolling process; Multilayer prepregs are cured using a molding machine at 50℃–300℃ and 0.1MPa–6MPa to form a periodic mosaic-structured laminated composite material.

[0012] Secondly, this application provides a mosaic-type electromagnetic frequency selective laminated composite material intelligent design and manufacturing system, comprising: The data acquisition module is used to acquire the electromagnetic parameters of the base material through experimental measurement, and the electromagnetic parameters include at least the dielectric constant and magnetic permeability. A database construction module is used to establish a database of mapping relationships between the geometric patterns and electromagnetic performance parameters of in-plane mosaic structures through batch electromagnetic simulation based on the electromagnetic parameters. The model training module is used to use the mapping relationship database as training samples to train a deep neural network model and obtain a target model that can predict the electromagnetic properties of an input mosaic structure pattern. The function module is used to set the wave transmission performance index and wave absorption performance index of the laminated composite material in the target frequency band according to the application requirements, and to optimize the objective function based on this definition. The pattern output module is used to use the target model as a fast performance evaluation tool, call the progressive multi-objective Bayesian optimization algorithm, and use the optimization objective function as the criterion to automatically optimize the design of the mosaic structure pattern to obtain the optimal structure pattern that meets the performance requirements. The manufacturing output module is used to manufacture a mosaic-type electromagnetic frequency selective laminated composite material by spatially pixelating and compositing electromagnetic absorbing base materials and electromagnetic transparent base materials through a multilayer resin film melt impregnation process according to the optimal structural pattern.

[0013] Thirdly, this application provides a computer device, the device comprising: a memory and a processor, wherein the processor, when executing computer instructions stored in the memory, performs the method described above.

[0014] Fourthly, this application provides a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the method described above.

[0015] In summary, this application obtains the electromagnetic parameters of the basic material through experimental measurement and establishes a database of mapping relationships between the geometric patterns and electromagnetic performance parameters of the in-plane mosaic structure through batch electromagnetic simulation; it trains a deep neural network model using the database to obtain a target model; according to application requirements, it sets the wave transmission performance index and wave absorption performance index of the laminated composite material in the target frequency band and defines the optimization objective function; using the target model as a rapid performance evaluation tool, it automatically optimizes the mosaic structure pattern to obtain the optimal structure pattern that meets the performance requirements; and based on the optimal structure pattern, it manufactures the mosaic-type electromagnetic frequency-selective laminated composite material. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiments of this application; Figure 2 This is a flowchart illustrating the first embodiment of the intelligent design and manufacturing method for mosaic-type electromagnetic frequency selective laminated composite materials of this application; Figure 3 This is a comparative analysis diagram of the prediction results of the prediction model for the T-curve and A-curve of the first random mosaic laminate composite material in the first embodiment of this application and the simulation data. Figure 4 This is a comparative analysis diagram of the prediction results of the prediction model for the T-curve and A-curve of the second random mosaic laminate composite material in the first embodiment of this application and the simulation data. Figure 5 These are the electromagnetic absorption and electromagnetic transmission performance curves and the corresponding periodic structural units of the mosaic-type laminated composite material in the first embodiment of this application. Figure 6 The second embodiment of this application shows the electromagnetic performance simulation curves of the mosaic-type laminated composite material and the corresponding periodic structural unit for achieving wider frequency electromagnetic projection. Figure 7 This is a schematic diagram of the internal structure of the pixel-level laminated composite material based on the multilayer resin film melt impregnation process in the second embodiment of this application; Figure 8 This is a structural block diagram of the first embodiment of the mosaic-type electromagnetic frequency selective laminated composite intelligent design and manufacturing system of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application.

[0018] Reference Figure 1 , Figure 1 This is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiments of this application.

[0019] like Figure 1As shown, the computer device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen or an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0020] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0021] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and a mosaic-type electromagnetic frequency selective laminated composite intelligent design and manufacturing program.

[0022] exist Figure 1 In the computer device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in this application can be set in the computer device. The computer device calls the mosaic-type electromagnetic frequency selective laminated composite intelligent design and manufacturing program stored in the memory 1005 through the processor 1001, and executes the mosaic-type electromagnetic frequency selective laminated composite intelligent design and manufacturing method provided in the embodiment of this application.

[0023] Existing designs for mosaic-type electromagnetic frequency-selective laminated composite materials primarily rely on experiments, theoretical analysis, and finite element simulations. Faced with a vast parameter space, efficient search for the optimal design solution is difficult, and the design process heavily depends on expert experience, resulting in long iteration cycles, low efficiency, and high R&D costs. Furthermore, existing manufacturing processes struggle to guarantee the bonding strength and compatibility of the layup interfaces when dealing with the performance differences between different material units, easily leading to interlayer delamination and performance degradation, posing challenges to fabrication reliability.

[0024] This application provides a method for the intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the intelligent design and manufacturing method for mosaic-type electromagnetic frequency-selective laminated composite materials according to this application. It aims to solve the aforementioned problems. This method, by combining data-driven intelligent design with a highly reliable manufacturing process, achieves rapid design and fabrication of mosaic-type electromagnetic frequency-selective laminated composite materials.

[0025] In the first embodiment, the intelligent design and manufacturing method of mosaic-type electromagnetic frequency selective laminated composite material includes the following steps: Step S10: Obtain the electromagnetic parameters of the base material through experimental measurement. The electromagnetic parameters include at least the dielectric constant and the magnetic permeability.

[0026] For ease of understanding, the following explains some key terms in this embodiment: Mosaic-type electromagnetic frequency-selective laminated composite material refers to a material composed of material units with different electromagnetic properties arranged in a specific geometric pattern in a plane and laminated together. This material can selectively transmit or absorb electromagnetic waves within a specific frequency range, while also possessing structural load-bearing capabilities.

[0027] Electromagnetic parameters are physical quantities that describe the interaction characteristics of a material with electromagnetic waves, mainly including dielectric constant and permeability. Dielectric constant characterizes a material's ability to store electrical energy under the influence of an external electric field, while permeability characterizes a material's ability to be magnetized under the influence of an external magnetic field. These parameters are fundamental to evaluating the electromagnetic response of materials.

[0028] Batch electromagnetic simulation refers to the process of using computer software to automatically calculate and analyze the electromagnetic fields of a large number of mosaic structures with different geometric patterns. This process aims to efficiently obtain performance data of different structures under the influence of specific electromagnetic waves.

[0029] A mapping database is a collection of data that stores the relationships between the geometric patterns of in-plane mosaic structures and their corresponding electromagnetic performance parameters. This database is generated through batch simulations and used for training subsequent machine learning models.

[0030] Deep neural network models are machine learning models containing multiple hidden layers, capable of learning complex nonlinear mappings from large amounts of data. In this method, the model is trained to predict the electromagnetic properties of a mosaic structure pattern, and the trained model becomes the target model.

[0031] The optimization objective function is a mathematical expression set according to application requirements to quantify the wave transmission and absorption performance indicators of the mosaic-type electromagnetic frequency-selective laminated composite material. This function serves as the evaluation criterion for the optimization algorithm, guiding the optimization process.

[0032] The progressive multi-objective Bayesian optimization algorithm is a global optimization algorithm based on Bayesian theory, suitable for handling high-dimensional, black-box, multi-objective optimization problems. This algorithm constructs a surrogate model and uses a sampling function to select the next evaluation point, efficiently finding an approximate optimal solution within a finite number of evaluations.

[0033] The multilayer resin film melt impregnation process refers to a manufacturing method in which electromagnetic absorbing base material and electromagnetic transparent base material are cut and laid in the form of resin film according to a preset pixelated pattern, and then melted and impregnated by heating and pressurizing to solidify and form a laminated composite material.

[0034] It should be noted that the base material includes electromagnetic absorbing materials and electromagnetic transparent materials; the electromagnetic absorbing material contains at least one absorber filler selected from carbonyl iron, iron-cobalt alloy, or iron(II,III) oxide, with a particle size of 5μm–20μm and a filler loading mass percentage of 1%–80%; the electromagnetic transparent material contains at least one transparent filler selected from silicon dioxide, aluminum oxide, or boron nitride, with a particle size of 3μm–20μm; the base material is prepared into a laminated composite material by blending the resin matrix and the filler, followed by coating, hot rolling, and molding processes.

[0035] In this specific implementation, the electromagnetic parameters of the basic electromagnetic absorbing and electromagnetic transparent materials are obtained, including: using fiber-reinforced resin matrix composites with extremely low dielectric loss as the base material, including but not limited to glass fiber (GF), quartz fiber (QF) reinforcing phases, epoxy (EP), polyimide (PI), and other resin systems; and achieving electromagnetic functionalization by loading electromagnetic absorber fillers or electromagnetic transparent fillers into the matrix, wherein the electromagnetic absorbers include carbonyl iron (CIP) and iron-cobalt alloy (Co). x Fe 1-x The fillers include one or a mixture of granular fillers such as iron(II,III) oxide (Fe3O4) with a particle size of 5μm-20μm. The electromagnetically transparent fillers include one or a mixture of granular fillers such as silicon dioxide (SiO2), aluminum oxide (Al2O3), and boron nitride (BN) with a particle size of 3μm-20μm. The loading percentage of the electromagnetically functionalized fillers is between 1% and 80%, which can be adjusted according to the electromagnetic parameter requirements.

[0036] The resin matrix and filler particles are mixed and stirred to obtain a uniform gel. Then, the curing agent and curing catalyst are mixed and stirred with the gel to remove air bubbles, resulting in a filler-loaded preformed resin. The viscosity is controlled between 100 mPa·s and 50,000 mPa·s to ensure good film-forming properties. The uniformly blended preformed resin is then used to prepare an electromagnetically functionalized resin film using a coating machine. The loaded filler is uniformly distributed, the thickness is controlled between 0.02 and 0.2 mm, and the areal density is 60 g / m³. 2 -300g / m 2 To ensure sufficient impregnation of the underlying fiber cloth by the resin film, it is then laid on top of the fiber cloth layer and hot-rolled to form a single-layer prepreg. The hot roller temperature is 60℃-150℃, and the rolling speed is 0.01m / s-0.5m / s. According to the electromagnetic performance testing requirements, the filler-loaded prepreg is cut to a suitable size and then stacked in multiple layers in a mold. Based on the curing regime of the resin matrix, a molding press is used to composite the single-material prepreg into multiple layers to prepare a laminated composite material with electromagnetic absorption or electromagnetic transparency functions, with a thickness of 2mm-5mm. Electromagnetic parameters, including dielectric constant and permeability, of the electromagnetically transparent and electromagnetically absorbing materials are obtained through electromagnetic performance testing.

[0037] By clarifying the composition and preparation process of the basic materials through the above technical solutions, it is possible to ensure that the electromagnetic absorbing and electromagnetic transparent materials used in intelligent design and manufacturing possess stable and controllable electromagnetic properties. Specifically, by selecting specific types of absorber fillers and transparent fillers, and precisely controlling their particle size and loading, the dielectric constant and magnetic permeability of the materials can be effectively regulated, thereby providing accurate input parameters for subsequent electromagnetic simulation and optimization design. Furthermore, the use of standardized preparation processes such as resin matrix and filler blending, coating, hot rolling, and molding ensures the uniformity and density of the basic materials, avoiding the impact of inherent material uncertainties on the accuracy of electromagnetic performance prediction of mosaic structure patterns and the final manufacturing quality of composite materials. This makes the entire intelligent design and manufacturing process more reliable, improves design efficiency and the predictability of material properties, and ultimately achieves precise control and high-performance manufacturing of mosaic-type electromagnetic frequency-selective laminated composite materials.

[0038] Step S20: Based on the electromagnetic parameters, establish a database of mapping relationships between the geometric patterns and electromagnetic performance parameters of the in-plane mosaic structure through batch electromagnetic simulation.

[0039] In specific implementation, the batch electromagnetic simulation includes: encoding the in-plane mosaic structure into a binary matrix, where the first value represents an electromagnetic absorption unit and the second value represents an electromagnetic transparency unit; using automated scripts to drive electromagnetic simulation software to perform parametric modeling and simulation of the encoded mosaic structure; outputting electromagnetic performance data including reflectivity R, transmittance T, and absorptivity A, and expressing this data using formulas. The calculation yielded, where and These are the scattering parameters output from the simulation, in dB.

[0040] In practical implementation, the aforementioned electromagnetically transparent and electromagnetically absorbing materials are equivalent to a basic cuboid design unit with in-plane length and width of 5mm-10mm and thickness of 2mm-10mm. This unit is numerically encoded. Within a specified 10×10 square in-plane design space, random in-plane mosaic structure codes are generated through hypercube sampling. Simultaneously, special in-plane mosaic structure codes such as in-plane toroidals and centrally symmetric structures are supplemented. The 10×10 square coding matrix is ​​transformed into a 100×1 one-dimensional matrix, and duplicate codes are removed. A batch simulation script, integrated with CST Studio finite element software, is designed to establish the aforementioned in-plane mosaic codes and corresponding electromagnetic performance database. The batch simulation script includes the operation and related settings of the CST Studio electromagnetic simulation software (setting background materials, boundary conditions, and the number of modes; importing and modeling electromagnetic parameters of the material model) as well as the output and storage of simulation results. The database includes in-plane mosaic structure codes and frequency-varying interpolation points representing electromagnetic performance.

[0041] By employing the aforementioned technical solution, the in-plane mosaic structure is encoded into a binary matrix, achieving a digital and standardized representation of complex geometric patterns and greatly simplifying the complexity of structural modeling. Based on this, automated scripts are used to drive electromagnetic simulation software for parametric modeling and simulation, effectively avoiding errors and inefficiencies caused by manual operation and ensuring the accuracy and consistency of the batch simulation process. Simultaneously, by accurately converting the scattering parameters S11 and S21 from the simulation output into reflectivity R, transmittance T, and absorptivity A, standardized and physically meaningful electromagnetic performance data are provided. This systematic batch electromagnetic simulation method not only significantly improves the efficiency and quality of establishing a database of mapping relationships between geometric patterns and electromagnetic performance parameters but also provides high-quality training samples for the accurate training of subsequent deep neural network models, thus laying a solid foundation for the intelligent design of mosaic-type electromagnetic frequency-selective laminated composite materials.

[0042] Step S30: Use the mapping database as training samples to train a deep neural network model and obtain a target model that can predict the electromagnetic properties of the input mosaic structure pattern.

[0043] In its implementation, the deep neural network model employs a stacked residual network architecture and introduces a Dropout layer during training to prevent overfitting. The mosaic structure encoding is used as input, and the output consists of interpolation points of the electromagnetic transmittance and absorptivity curves within the frequency band. The model training uses mean squared error as the loss function, and the network hyperparameters are optimized using a validation set. The introduction of a Dropout layer during dimensionality transformation reduces model complexity and computational cost while increasing robustness and generalization ability. Minimizing the mean squared error is used as the loss function to improve the model's prediction accuracy. The prediction model is tested on a randomly partitioned test set from the database to obtain the mean squared error.

[0044] In specific implementation, refer to Figure 3 and Figure 4 , Figure 3 and Figure 4 This is a visualization diagram showing the accuracy of the deep neural network prediction model provided in Embodiment 1 of the intelligent design and manufacturing method for mosaic-type electromagnetic frequency selective laminated composite materials of this application. Figure 3 and Figure 4 The comparison and analysis of the predicted T-curves and A-curves of the prediction model with simulation data for two random mosaic laminated composite materials are presented. In this embodiment, the first and second random mosaic laminated composite materials are used for differentiation. The solid lines in the figure represent the simulation curves, and the dashed lines represent the predicted curves. It can be seen that for a certain unknown mosaic structure, the prediction results can fit the simulation results well, demonstrating high prediction accuracy. This indicates that the prediction model is effective and reliable in evaluating the electromagnetic absorption and transmission performance of materials, and can provide an accurate prediction tool for subsequent optimization design.

[0045] The above technical solution utilizes a stacked residual network architecture to effectively construct a deep neural network, enabling it to possess powerful feature extraction capabilities and more accurately capture the nonlinear mapping relationship between mosaic patterns and complex electromagnetic properties. Simultaneously, introducing a Dropout layer during training effectively suppresses overfitting of the model to training data, significantly improving the model's generalization prediction ability for unknown mosaic patterns and ensuring the reliability of prediction results. Using the mosaic structure encoding as model input and the interpolation points of the electromagnetic transmittance and absorptivity curves within the frequency band as output allows the model to directly learn and predict key performance indicators, improving the relevance and efficiency of prediction. Furthermore, using mean squared error as the loss function accurately measures the deviation between predicted and true values, guiding the model's optimization direction. Optimizing network hyperparameters using a validation set further ensures the model's stability and prediction accuracy in practical applications, providing an efficient and accurate performance evaluation tool for subsequent automatic optimization design, significantly improving the efficiency and success rate of the entire intelligent design and manufacturing method.

[0046] Step S40: Based on application requirements, set the wave transmission performance index and wave absorption performance index of the laminated composite material in the target frequency band, and optimize the objective function based on this definition.

[0047] It should be noted that electromagnetic transparency and electromagnetic absorption bands are set according to the electromagnetic frequency selection requirements, and the simulation database is analyzed to extract objective functions that characterize the performance of electromagnetic frequency selection. Suitable objective functions are then selected for subsequent optimization.

[0048] In specific implementation, the objective function includes: The electromagnetic frequency selection range refers to the values ​​of a and b in the formula, expressed in GHz, and falls within the frequency range of a GHz to b GHz. and These represent the number of interpolation points that satisfy the conditions. and It can be set to a fixed value between 0 and 1; This is a function to determine whether the transmittance or absorptivity at a certain frequency satisfies electromagnetic absorption or terahertz communication requirements.

[0049] It should be noted that this embodiment provides a specific and quantifiable optimization objective function, effectively solving the problem of how to accurately define and evaluate frequency selectivity performance in the intelligent design process. This objective function introduces... and It can simultaneously measure the wave transmission and absorption performance of a material in the target frequency band, and allows for setting a, b, ... and Parameters such as these allow for flexible definition of the electromagnetic frequency selection range and performance threshold. This quantification method enables the progressive multi-objective Bayesian optimization algorithm to automatically optimize the design of mosaic structure patterns using clear and explicit criteria. Specifically, the optimization algorithm can... and The numerical value is used to determine the gap between the current structural pattern and the ideal frequency selectivity performance, and the design parameters are adjusted accordingly to efficiently converge to the optimal structural pattern that meets the transmission and absorption performance indicators. This not only improves design efficiency, but also ensures that the final mosaic-type electromagnetic frequency selective laminated composite material can accurately achieve the expected frequency selection function, such as achieving high transmittance in a specific frequency band for terahertz communication, and high absorption in other frequency bands for electromagnetic shielding or absorption.

[0050] Step S50: Using the target model as a fast performance evaluation tool, the progressive multi-objective Bayesian optimization algorithm is invoked, and the mosaic structure pattern is automatically optimized based on the optimization objective function to obtain the optimal structure pattern that meets the performance requirements.

[0051] In its implementation, the progressive multi-objective Bayesian optimization algorithm includes: iteratively calling the prediction model to evaluate the electromagnetic performance of the mosaic structure; selecting the potential optimal structure based on the acquisition function for simulation verification; feeding new data back to the prediction model to incrementally update the training set, gradually improving model accuracy and optimization efficiency; and synchronously adjusting the frequency band threshold in the objective function during the optimization process. and To balance the wave transmission and absorption properties.

[0052] In practical implementation, an iteratively trainable prediction model, dynamically expandable electromagnetic targets, and a multi-objective Bayesian iterative optimization process are integrated into a data-driven machine learning framework. The optimization process uses each round of Bayesian optimization as a loop, continuously iterating and training the prediction model, constantly expanding the electromagnetic frequency selection indicators, and ultimately achieving a synergistic optimization design of electromagnetic transmission and electromagnetic absorption.

[0053] In practical implementation, the process of the progressive multi-objective Bayesian optimization algorithm includes: Mosaic structure encoding based on hypercubic sampling and centrally symmetric random sampling was used for batch CST Studio finite element simulations, generating a simulation database. Based on the transmittance and absorptivity calculation formulas, this database was integrated into a collection containing different mosaic structure encodings and their corresponding electromagnetic performance A and T curves. Next, two data-driven deep neural network prediction models were trained using this data. Each model consists of an input layer, multiple hidden layers, and an output layer, capable of predicting the A and T curves of a given mosaic structure. Then, the values ​​of a and b in the electromagnetic frequency selection requirements were set, and N was calculated based on the magnitude of the X value in the formula. T and N A The relevant metrics are then used to iteratively search for the optimal ply configuration that satisfies these objectives using a multi-objective Bayesian optimization algorithm. In each iteration, the algorithm selects the most promising mosaic structure for evaluation based on the output of the prediction model and the uncertainty estimate, and feeds the new evaluation results back to the prediction model, while updating the model based on whether the judgment is correct or not. and The magnitude of the value. Through continuous iteration, the electromagnetic transmittance or electromagnetic absorption rate within the target frequency band of electromagnetic frequency selection is gradually approached, and finally, the best mosaic structure is selected, and the target mosaic structure code that meets the preset frequency selection requirements is determined.

[0054] like Figure 5 As shown, the electromagnetic frequency is selected as N. A The values ​​of a and b for the function are set to 8GHz and 18GHz respectively, while N... T The values ​​of a and b in the function are set to 2-4 GHz, meaning that the absorption of electromagnetic waves in the 8-18 GHz frequency range and the transmission of electromagnetic waves in the 2-4 GHz frequency range are achieved within the 2-18 GHz frequency range. The electromagnetic absorption and transmission performance curves, and the corresponding periodic structural unit of the mosaic-type laminated composite material, are shown below. Figure 5 As shown, its mosaic feature is that within a 10×10 periodic unit, the dark electromagnetic absorbing material is distributed in the outer ring position, achieving an absorption effect of more than 70% for electromagnetic waves in the 8-18GHz frequency range. At the same time, it achieves a transmission rate greater than the absorption rate for electromagnetic waves in the 2-4GHz frequency range, especially in the 2-3GHz frequency range, the transmission rate exceeds 50%.

[0055] Through the above technical solution, this embodiment can effectively overcome the potential accuracy limitations of the initial prediction model. Iterative calls to the prediction model are used for rapid evaluation, and a high-precision simulation verification is performed using a data acquisition function to intelligently select the potentially optimal structure. Newly acquired simulation data is promptly fed back to the prediction model, achieving incremental updates to the training set. This continuously improves the model's prediction accuracy and optimization efficiency, preventing the optimization process from getting stuck in local optima. Simultaneously, the frequency band threshold in the objective function is adjusted concurrently during the optimization process. and This allows for dynamic and precise control of the balance between wave transmission and absorption properties, ensuring that the final mosaic-type electromagnetic frequency selective laminated composite material can more accurately meet complex and ever-changing application requirements, and significantly improving the practicality and adaptability of the design scheme.

[0056] Step S60: According to the optimal structural pattern, the electromagnetic absorbing base material and the electromagnetic transparent base material are arranged and composited in space in a pixelated manner through a multilayer resin film melt impregnation process to produce a mosaic-type electromagnetic frequency selective laminated composite material.

[0057] It should be noted that pixel-level manufacturing includes: cutting and splicing electromagnetic absorbing resin film and electromagnetic transparent resin film according to the optimized mosaic pattern; laying the film layer by layer on the fiber cloth and forming a prepreg through hot rolling process; and using a molding press to cure the multi-layer prepreg under conditions of 50℃–300℃ and 0.1MPa–6MPa to form a periodic mosaic structure laminated composite material.

[0058] In practical implementation, the pixel-level manufacturing of the mosaic-type electromagnetic frequency-selective laminated composite material includes: according to the above-optimized mosaic-type in-plane pattern, cutting a uniform resin film prepared by blending electromagnetic absorber filler with a low-dielectric resin matrix into a specified pixelated pattern, with the smallest unit size of the pattern being 5mm×5mm, serving as the electromagnetic absorption portion in the mosaic-type in-plane pattern; cutting the uniform resin film prepared by blending electromagnetic transparent filler with a low-dielectric resin film into the remaining electromagnetic transparent portion. A designed periodic unit size of 50mm×50mm×5mm is periodically arranged in the in-plane direction to prepare a laminated composite material with a size of 200mm×200mm×5mm.

[0059] Specifically, the resin matrix slurry and functionalized particles are placed in a mixing tank according to a preset ratio and stirred at a temperature of 40℃-100℃ for 5-120 minutes to obtain a resin system containing particles. This resin system is then coated onto the surface of release paper using a coating machine, and covered with a plastic film to obtain a resin film loaded with electromagnetic functionalized particles. The coating speed is 1-10 mm / s. Based on an optimized mosaic pattern, the electromagnetic absorption resin film and electromagnetic transparency resin film are cut and spliced ​​into a complete pattern, then laid on a fiber cloth and processed using a hot rolling method. After cooling, a mosaic-type particle layer-fiber layer resin-based prepreg is obtained. Multiple layers of the above mosaic-type particle layer-fiber layer resin-based prepreg are laid in a mold and then hot-pressed to form an alternating layer of fiber reinforcement and mosaic pattern particles, resulting in a laminated composite material with in-plane electromagnetic absorption and electromagnetic transparency pixelated arrangement.

[0060] The preferred method for preparing the mosaic-patterned particle-loaded fiber-reinforced resin matrix composite material is as follows: the hot rolling temperature is 40℃-200℃, and the rolling speed is 0.01m / s-0.5m / s. Preferably, the hot pressing method includes vacuum bag film hot pressing, molding machine hot pressing, or vacuum tank hot pressing; the hot pressing temperature is 50℃-300℃, the hot pressing time is 5min-24h, and the hot pressing pressure is 0.1MPa-6MPa. Preferably, when the resin-based prepreg exceeds 3 layers, a vacuum bag film is used to wrap every 3-10 layers of resin-based prepreg, and vacuum extraction and pre-compacting are performed to remove air between the prepreg layers.

[0061] The aforementioned pixel-level manufacturing method enables the precise conversion of optimized mosaic patterns obtained from intelligent design into physical entities, effectively solving the technical challenge of transforming digital designs into high-precision physical structures. Specifically, electromagnetic absorbing resin films and electromagnetic transparent resin films are cut and spliced ​​according to the optimized mosaic pattern, ensuring pixel-level precision in the mosaic structure in space. Layers of the films are laid on the fiber cloth and prepreg is formed through a hot rolling process, providing necessary mechanical support and structural stability for the composite material and laying the foundation for subsequent curing. Finally, the multi-layer prepreg is cured under precisely controlled temperature and pressure using a molding press, ensuring full cross-linking of the resin and material density, thus forming a laminated composite material with a periodic mosaic structure. This refined manufacturing process ensures that the mosaic-type electromagnetic frequency-selective laminated composite material can accurately reproduce the wave transmission and absorption performance indicators predicted in the design stage, thereby achieving the expected electromagnetic frequency selection function.

[0062] This embodiment achieves automatic optimization design by establishing a mapping relationship database, training a deep neural network model, and calling optimization algorithms. Combined with pixelated manufacturing processes, it efficiently covers the global design domain, reduces iteration cycles, lowers trial and error costs, and solves the problem of low manufacturing reliability.

[0063] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description and will not be repeated hereafter. The difference is that, based on the above, in order to achieve wider frequency band electromagnetic transmission, N... T The b-value of the function expands towards higher frequencies.

[0064] The second embodiment provides a smart design and manufacturing method for mosaic-type laminated composite materials for wider frequency electromagnetic communication: the communication frequency band achieves more than 50% electromagnetic transmission effect in the 2-4GHz frequency range, while maintaining more than 70% absorption effect for electromagnetic waves in the 12-18GHz frequency range.

[0065] Step 1: Prepare basic electromagnetic transparent laminated composite materials and electromagnetic absorbing laminated composite materials through a multilayer resin film melt impregnation process, and obtain the basic electromagnetic parameters of the two design units through electromagnetic performance testing experiments, including the real and imaginary parts of the dielectric constant and the real and imaginary parts of the magnetic permeability.

[0066] Step 2: Use a self-coded script to perform batch processing of CST Studio finite element simulation, establish approximately 5000 sets of data, and perform calculations and integration according to the formula in the first embodiment.

[0067] Step 3: Randomly divide the integrated database into three parts: 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; the validation set is used to find a better architecture; and the test set is used to evaluate performance, using the MSE loss function to evaluate the mean squared error between simulated and predicted values. The prediction model is trained using the integrated database from Step 2. A stacked residual network architecture is employed for dimensionality transformation in deep neural networks, solving the gradient vanishing and exploding problems in deep networks and improving the network's depth and expressive power. A Dropout layer is introduced during the dimensionality transformation process, reducing model complexity and computational cost while increasing the model's robustness and generalization ability.

[0068] Step 3 uses a trained prediction model to approximate the electromagnetic transmission and absorption properties of different mosaic structures. Compared to CST Studio finite element simulation, this method improves speed by approximately 10% while maintaining high prediction accuracy. 6 times (0.02ms).

[0069] Step 4: As in the first embodiment, select electromagnetic frequency N. A The values ​​of a and b for the function are set to 8GHz and 18GHz respectively, while N... T The values ​​of a and b for the function are set to 2-4 GHz respectively.

[0070] Step 5: A data-driven machine learning framework combining an incremental multi-objective Bayesian optimization algorithm was used for the optimization design of extended electromagnetic transmission performance. This framework integrates: 1) Iteratively train the prediction model; 2) Continuously update and expand electromagnetic transmission targets; 3) Multi-objective Bayesian iterative optimization process; 4) Determine the optimal design from the machine learning model; 5) The iterative process of the first 10 candidates. In the second embodiment, N... T The value of X in the function expands from 0.5 to 1, eventually balancing electromagnetic transmission in the 2-4 GHz range and electromagnetic absorption in the 12-18 GHz range.

[0071] In step 5, after every 10 multi-objective Bayesian optimizations, the optimization results are simulated using CST Studio and added to the database to retrain the prediction model, continuously improving the accuracy and generalization ability of the prediction model.

[0072] It should be noted that, Figure 6 The simulation curves of the electromagnetic performance of the mosaic-type laminated composite material and the corresponding periodic structural unit in the second embodiment of this application, which achieve wider-band electromagnetic projection, demonstrate that electromagnetic transmission efficiency exceeding 50% is achieved in the 2-4 GHz frequency range, with even higher efficiency exceeding 70% in the 2-3 GHz frequency range. Simultaneously, it maintains electromagnetic absorption exceeding 70% for electromagnetic waves in the 12-18 GHz frequency range. This indicates that the intelligent design method of this invention strives to balance the inherent contradiction between electromagnetic transmission and electromagnetic communication, achieving optimized design for both.

[0073] Step 6: Similar to the manufacturing method in the first embodiment, manufacture the optimized mosaic-type electromagnetic frequency-selective laminated composite material from the second embodiment. See also... Figure 7 This paper presents a schematic diagram of the internal structure of pixel-level laminated composite materials based on a multilayer resin film melt impregnation process.

[0074] Furthermore, embodiments of this application also propose a computer-readable storage medium storing a program for the intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials. When the program for the intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials is executed by a processor, it implements the steps of the method for the intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials as described above.

[0075] Reference Figure 8 , Figure 8 This is a structural block diagram of the first embodiment of the mosaic-type electromagnetic frequency selective laminated composite intelligent design and manufacturing system of this application.

[0076] like Figure 8 As shown in the embodiments of this application, the mosaic-type electromagnetic frequency selective laminated composite intelligent design and manufacturing system includes: Data acquisition module 10 is used to acquire the electromagnetic parameters of the base material through experimental measurement, wherein the electromagnetic parameters include at least the dielectric constant and the magnetic permeability; Database construction module 20 is used to establish a mapping relationship database between the geometric pattern of the in-plane mosaic structure and the electromagnetic performance parameters based on the electromagnetic parameters through batch electromagnetic simulation; The model training module 30 is used to use the mapping relationship database as training samples to train a deep neural network model and obtain a target model that can predict the electromagnetic properties of the input mosaic structure pattern. Function module 40 is used to set the wave transmission performance index and wave absorption performance index of the laminated composite material in the target frequency band according to application requirements, and optimize the objective function based on this definition. The pattern output module 50 is used to use the target model as a fast performance evaluation tool, call the progressive multi-objective Bayesian optimization algorithm, and use the optimization objective function as the criterion to automatically optimize the design of the mosaic structure pattern to obtain the optimal structure pattern that meets the performance requirements. The manufacturing output module 60 is used to manufacture a mosaic-type electromagnetic frequency selective laminated composite material by spatially pixelating and compositing electromagnetic absorbing base material and electromagnetic transparent base material through a multilayer resin film melt impregnation process according to the optimal structural pattern.

[0077] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solution of this application. In specific applications, those skilled in the art can make settings as needed, and this application does not impose any restrictions on this.

[0078] This embodiment achieves automatic optimization design by establishing a mapping relationship database, training a deep neural network model, and calling optimization algorithms. Combined with pixelated manufacturing processes, it efficiently covers the global design domain, reduces iteration cycles, lowers trial and error costs, and solves the problem of low manufacturing reliability.

[0079] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this application. In practical applications, those skilled in the art can select some or all of it to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0080] In addition, for technical details not described in detail in this embodiment, please refer to the method for intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials provided in any embodiment of this application, which will not be repeated here.

[0081] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0082] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0083] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application. The above are only preferred embodiments of this application and do not limit the patent scope of this application. All equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for intelligent design and manufacturing of mosaic-type electromagnetic frequency selective laminated composite materials, characterized in that, include: The electromagnetic parameters of the basic material are obtained through experimental measurement, and the electromagnetic parameters include at least the permittivity and permeability. Based on the electromagnetic parameters, a database of mapping relationships between the geometric patterns and electromagnetic performance parameters of in-plane mosaic structures is established through batch electromagnetic simulation. Using the mapping database as training samples, a deep neural network model is trained using the database to obtain a target model that can predict the electromagnetic properties of an input mosaic structure pattern. Based on application requirements, the wave transmission performance index and wave absorption performance index of the laminated composite material in the target frequency band are set, and the objective function is optimized based on this definition. Using the target model as a fast performance evaluation tool, the progressive multi-objective Bayesian optimization algorithm is invoked, and the mosaic structure pattern is automatically optimized based on the optimization objective function to obtain the optimal structure pattern that meets the performance requirements. Based on the optimal structural pattern, a mosaic-type electromagnetic frequency selective laminated composite material is manufactured by spatially pixelating and compositing the electromagnetic absorbing base material and the electromagnetic transparent base material through a multilayer resin film melt impregnation process.

2. The method according to claim 1, characterized in that, The electromagnetic absorbing material comprises at least one absorbent filler selected from carbonyl iron, iron-cobalt alloy or iron oxide, with a particle size of 5μm–20μm and a filler loading percentage of 1%–80% by mass. The electromagnetically transparent material contains at least one transparent filler selected from silicon dioxide, aluminum oxide, or boron nitride, with a particle size of 3μm–20μm; The base material is prepared into a laminated composite material by blending a resin matrix with fillers and then performing coating, hot rolling, and molding processes.

3. The method according to claim 1, characterized in that, The batch electromagnetic simulation includes: The in-plane mosaic structure is encoded as a binary matrix, where the first value represents an electromagnetic absorption unit and the second value represents an electromagnetic transparency unit. The electromagnetic simulation software is driven by automated scripts to perform parametric modeling and simulation of the encoded mosaic structure. The output electromagnetic performance data includes reflectivity R, transmittance T, and absorptivity A, and is expressed by the formula... The calculation yielded, where and These are the scattering parameters output from the simulation, in dB.

4. The method according to claim 1, characterized in that, The deep neural network model adopts a stacked residual network architecture and introduces a Dropout layer during training to prevent overfitting. Using the mosaic structure code as input, the output is the interpolation points of the electromagnetic transmittance curve and the absorptivity curve within the frequency band. The model training uses mean squared error as the loss function, and the network hyperparameters are optimized using a validation set.

5. The method according to claim 1, characterized in that, The objective function includes: The electromagnetic frequency selection range refers to the values ​​of a and b in the formula, expressed in GHz, and falls within the frequency range of a GHz to b GHz. and These represent the number of interpolation points that satisfy the conditions. and It can be set to a fixed value between 0 and 1; This is a function to determine whether the transmittance or absorptivity at a certain frequency satisfies electromagnetic absorption or terahertz communication requirements.

6. The method according to claim 1, characterized in that, The progressive multi-objective Bayesian optimization algorithm includes: Iteratively call the prediction model to evaluate the electromagnetic properties of the mosaic structure; The potential optimal structure is selected based on the acquisition function and then verified through simulation. New data is fed back into the prediction model to incrementally update the training set, gradually improving model accuracy and optimization efficiency; The optimization process synchronously adjusts the frequency band threshold in the objective function. and To balance the wave transmission and absorption properties.

7. The method according to claim 1, characterized in that, Pixel-level manufacturing includes: The electromagnetic absorbing resin film and the electromagnetic transparent resin film are cut and spliced ​​according to the optimized mosaic pattern. The adhesive film is laid layer by layer on the fiber cloth and prepreg is formed by hot rolling process; Multilayer prepregs are cured using a molding machine at 50℃–300℃ and 0.1MPa–6MPa to form a periodic mosaic-structured laminated composite material.

8. A mosaic-type electromagnetic frequency selective laminated composite intelligent design and manufacturing system, characterized in that, include: The data acquisition module is used to acquire the electromagnetic parameters of the base material through experimental measurement, and the electromagnetic parameters include at least the dielectric constant and magnetic permeability. A database construction module is used to establish a database of mapping relationships between the geometric patterns and electromagnetic performance parameters of in-plane mosaic structures through batch electromagnetic simulation based on the electromagnetic parameters. The model training module is used to use the mapping relationship database as training samples to train a deep neural network model and obtain a target model that can predict the electromagnetic properties of an input mosaic structure pattern. The function module is used to set the wave transmission performance index and wave absorption performance index of the laminated composite material in the target frequency band according to the application requirements, and to optimize the objective function based on this definition. The pattern output module is used to use the target model as a fast performance evaluation tool, call the progressive multi-objective Bayesian optimization algorithm, and use the optimization objective function as the criterion to automatically optimize the design of the mosaic structure pattern to obtain the optimal structure pattern that meets the performance requirements. The manufacturing output module is used to manufacture a mosaic-type electromagnetic frequency selective laminated composite material by spatially pixelating and compositing electromagnetic absorbing base materials and electromagnetic transparent base materials through a multilayer resin film melt impregnation process according to the optimal structural pattern.

9. A computer device, characterized in that, The device includes a memory and a processor, wherein the processor, when executing computer instructions stored in the memory, performs the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Includes instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 7.

Images