A method and system for optimizing the design of an integrated lattice structure for both wave absorption and wave carrying.

Abstract

This invention belongs to the interdisciplinary field of integrated structural and functional design and advanced manufacturing. It relates to an optimization design method and system for an integrated lattice structure for both microwave absorption and load bearing. The invention first completes the screening of multiple topological lattice structures and designs a three-layer gradient honeycomb configuration. It then uses a DeepSDF implicit neural network with a size-aware mechanism to encode the structure, combined with PCA dimensionality reduction to extract core features. A multi-output GPR surrogate model is constructed to establish the mapping relationship between geometric features and microwave absorption and load bearing performance. The optimal structural parameters are obtained through multi-objective electromagnetic and mechanical optimization using the NSGA-II algorithm. Finally, integrated fabrication is achieved through FDM fused deposition modeling combined with a secondary impregnation process. This invention forms a complete technical closed loop of screening-design-encoding-optimization-manufacturing. The structure possesses advantages such as wideband microwave absorption, high load bearing, and lightweight design, and can be widely applied in stealth equipment fields such as aerospace, drones, and weaponry.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of integrated structural and functional design and advanced manufacturing. Specifically, it relates to an optimization design method and system for an integrated lattice structure for both microwave absorption and load bearing. In particular, it relates to a collaborative technology based on multi-topology screening, DeepSDF implicit encoding, GPR proxy model optimization, and FDM integrated manufacturing. It is applicable to equipment fields such as aerospace, UAVs, and weaponry that have stringent requirements for lightweight structures, broadband microwave absorption, and high load bearing performance. It can effectively solve the technical problems of collaborative optimization of microwave absorption and load bearing performance and the disconnect between design and manufacturing. Background Technology

[0002] With the rapid development of aerospace technology, the survivability and penetration capability of stealth equipment have become core competitive indicators. Integrated lattice structures combining radar absorption and load-bearing capacity have become the preferred form of core structures for stealth equipment due to their lightweight, high specific strength, wideband radar absorption, and ease of integrated manufacturing. These structures must simultaneously meet three core requirements: wideband radar absorption, high load-bearing capacity, and lightweight design. The synergistic optimization of these two requirements and the integrated design and manufacturing have become a technological bottleneck in the industry.

[0003] Existing integrated wave-absorbing and load-bearing structural optimization technologies have several shortcomings: First, most schemes only optimize a single topology without comparing and screening multiple topologies, resulting in poor overall structural adaptability. Second, traditional explicit modeling methods are difficult to accurately represent complex gradient lattice structures, and surrogate models have low prediction accuracy, making it impossible to achieve precise synergistic optimization of wave-absorbing and load-bearing performance. Third, deep learning modeling, intelligent optimization algorithms, and 3D printing manufacturing technologies are independent of each other and have not formed a complete technical chain, leading to a disconnect between optimized design schemes and actual manufacturing processes, making it difficult to implement simulation optimization results.

[0004] For example, patent document CN121479924A discloses a method for optimizing the design of minimal surface structures. This method uses a Kriging surrogate model combined with a sequential quadratic programming algorithm, but it only optimizes a single topology structure, lacks a multi-topology screening step, and is not related to actual manufacturing processes, thus lacking engineering practicality. Meanwhile, although techniques such as DeepSDF implicit coding and PCA dimensionality reduction have been applied to structural modeling, they have not yet formed a closed-loop system with topology screening, surrogate models, multi-objective optimization, and FDM manufacturing. Furthermore, the FDM fused deposition modeling process is not deeply matched with the structural optimization parameters, further exacerbating the disconnect between design and manufacturing.

[0005] To address the shortcomings of existing technologies, there is an urgent need in this field for a complete technical solution that relies on a technical path of multi-topology screening → gradient configuration design → DeepSDF encoding → PCA dimensionality reduction → GPR surrogate model → NSGA-II optimization → FDM manufacturing to achieve synergistic unity of topology adaptation, precise optimization, and efficient manufacturing, thus filling the gaps in existing technologies. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of existing technologies, such as poor topological adaptability, low accuracy in characterizing complex structures, low efficiency in multi-objective optimization, and disconnect between design and manufacturing. It provides a method and system for optimizing the design of integrated lattice structures that combine wave absorption and load bearing, achieving synergistic optimization of the wave absorption performance, load bearing performance, and lightweight performance of the lattice structure. At the same time, it connects the entire process of design, optimization, and manufacturing, improving the engineering feasibility of the solution.

[0007] The core technological innovations of the integrated lattice structure optimization design method and system for wave absorption and wave carrying are as follows:

[0008] A multi-topology joint screening system was established, and the basic topology was selected from a variety of typical lattice structures by combining electromagnetic and mechanical multi-physics field simulation and hierarchical analysis method to ensure the comprehensive performance of the structure from the source.

[0009] Gradient lattice encoding is completed using a DeepSDF implicit neural network with a size-aware mechanism, and PCA is used to reduce and simplify features, thus solving the problems of difficulty in representing complex structures at different scales and feature redundancy.

[0010] Construct a multi-output GPR proxy model to accurately establish the nonlinear mapping relationship between structural features and wave absorption and load-bearing performance, thereby improving prediction accuracy and optimization efficiency;

[0011] Embedding the NSGA-II algorithm achieves multi-objective collaborative optimization of wideband absorption, high load capacity, and lightweight design, obtaining Pareto optimal solution set while taking into account multiple performance requirements;

[0012] By linking FDM molding and secondary impregnation processes, a seamless transformation from optimized design parameters to physical prototypes is achieved, forming a closed-loop technology process.

[0013] To achieve the objectives of this invention, the following technical solutions are adopted.

[0014] S1. Multi-topology filtering:

[0015] Using microwave absorption performance, load-bearing capacity, and lightweight performance as evaluation criteria, and combining multiphysics simulation and hierarchical analysis, the topology with the best overall performance was selected from various typical lattice topologies. This step selected more than ten typical topologies, including honeycomb structures, rhombic lattices, tetrahedral lattices, octahedral lattices, Kagome lattices, pyramid lattices, cubic lattices, cross lattices, star lattices, Gyroid structures, and Kelvin structures. Evaluation weights were set as follows: microwave absorption performance 0.4, load-bearing capacity 0.4, and lightweight performance 0.2. Electromagnetic simulations were conducted in the 2–18 GHz frequency band using CST electromagnetic simulation software, and mechanical compression simulations were performed using ANSYS. After comprehensive scoring, the honeycomb structure was determined as the optimal basic topology.

[0016] S2, Gradient Configuration Design:

[0017] Based on the selected honeycomb topology, a multi-layer gradient lattice configuration with gradually varying wall thickness along the thickness direction is designed. Considering both performance and engineering feasibility, a three-layer gradient honeycomb sandwich configuration is preferred. The design variables, constraints, and multi-objective optimization function are clearly defined.

[0018] Design variables: lattice height, aperture, wall thickness, and impregnation thickness of microwave absorbing paste for each layer;

[0019] Constraints: Total height of the absorbing core layer ; cell wall thickness Single cell cycle The lattice-equivalent Young's modulus is not lower than the lower limit of the design domain.

[0020] Multi-objective optimization function: The optimization objectives are to maximize the effective absorption bandwidth of the 2-18 GHz wideband, optimize the reflection loss of the key frequency band of 6-12 GHz, minimize the structural mass, and maximize the compressive load strength.

[0021] Latin hypercube sampling was used to generate sample combinations, and a "structural parameter-performance index" dataset was constructed through simulation for subsequent model training. Gradient design significantly improved the effective absorption bandwidth of the cellular structure compared to a uniform configuration.

[0022] S3, DeepSDF implicit neural network encoding:

[0023] A DeepSDF implicit neural network model was constructed, introducing a size-aware mechanism to discretize and encode a three-layer gradient honeycomb structure, outputting a 128-dimensional latent vector. The network consists of an input layer, four fully connected layers, and an output layer. The size-aware mechanism normalizes the structural size parameters and co-encodes them with the three-dimensional coordinate data of the structure, achieving a unified representation of lattice structures at different scales. The model training uses mean squared error (MSE) as the loss function, and the convergence criterion is the loss value. .

[0024] S4, PCA dimensionality reduction:

[0025] Standardization preprocessing and principal component analysis (PCA) were performed on the 128-dimensional latent vectors obtained from S3 to reduce dimensionality. Principal components with a cumulative variance contribution of ≥90% were selected to extract 3-dimensional core feature vectors. Redundant features were removed to improve the training efficiency of subsequent surrogate models.

[0026] S5 and GPR proxy model construction and training:

[0027] Using a 3D core feature vector as input and microwave absorption performance and load-bearing performance indicators as outputs, a multi-output Gaussian process regression (GPR) surrogate model is constructed and trained. The model employs a squared exponential kernel function to control the prediction error below a set threshold, achieving accurate mapping between geometric features and structural performance. Microwave absorption performance includes effective absorption bandwidth and reflection loss, while load-bearing performance includes compressive strength and equivalent Young's modulus (stiffness).

[0028] S6, NSGA-II Multi-Objective Optimization:

[0029] Based on the trained GPR surrogate model, the NSGA-II non-dominated sorting genetic algorithm is used to carry out multi-objective iterative optimization to obtain the Pareto optimal solution set, from which the structural parameters with the best overall performance are selected.

[0030] S7, FDM integrated manufacturing:

[0031] Based on the optimal structural parameters, the FDM printing composite material and process parameters were selected, and the lattice structure was printed using fused deposition modeling technology. After printing, a secondary impregnation process was performed to load the microwave absorbing slurry, thereby achieving the integrated molding of the wave-transmitting layer, the wave-absorbing layer, and the reflective layer.

[0032] An integrated lattice structure optimization design system for wave absorption and load bearing

[0033] This system is used to implement the above-mentioned optimization design method and includes seven functional modules, each of which transmits data and works collaboratively in sequence:

[0034] Multi-topology screening module: Using wave absorption, load-bearing capacity, and lightweight as evaluation criteria, combined with multiphysics simulation and hierarchical analysis, the optimal lattice topology is screened and the results are output to the gradient configuration design module.

[0035] Gradient configuration design module: Designs a three-layer gradient lattice configuration based on optimal topology, determines design variables, constraints and optimization objectives, and inputs structural geometric data into the DeepSDF encoding module;

[0036] DeepSDF encoding module: Constructs an implicit neural network and incorporates a size-aware mechanism to encode the gradient structure and output the latent vector to the PCA dimensionality reduction module;

[0037] PCA dimensionality reduction module: performs dimensionality reduction on latent vectors, extracts core feature vectors, and transmits them to the GPR surrogate model module;

[0038] GPR Proxy Model Module: Completes the construction, training and validation of proxy models, establishes the mapping relationship between feature vectors and performance indicators, and provides performance prediction support for the optimization module;

[0039] NSGA-II Optimization Module: Calls the GPR surrogate model to complete multi-objective iterative optimization, outputs the Pareto optimal solution set and optimal structural parameters, and sends them to the FDM manufacturing control module;

[0040] FDM Manufacturing Control Module: Matches printing materials and process parameters according to optimal parameters, controls FDM equipment to complete structural printing and secondary impregnation, and achieves solid forming.

[0041] An electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement all the steps of the above-described integrated lattice structure optimization design method for absorbing and carrying waves.

[0042] Compared with the prior art, the present invention has the following advantages:

[0043] 1. Strong topology adaptability and excellent overall performance. By comparing multiple topologies laterally to select the optimal basic configuration, the limitations of traditional single-topology structures are overcome, ensuring from the source that the structure simultaneously meets the three major requirements of wave absorption, load-bearing capacity, and lightweight design. Tests show that the gradient honeycomb structure, compared to the uniform honeycomb structure, effectively increases the wave absorption bandwidth by approximately 22% and the structural stiffness by approximately 18%.

[0044] 2. High accuracy in representing complex structures and good scale compatibility. By employing DeepSDF implicit encoding combined with a size-aware mechanism, it can accurately represent gradient lattice structures at different scales. The average chamfer distance between the reconstructed model and the original CAD model is only 0.012 mm. Combined with PCA dimensionality reduction and feature simplification, it significantly reduces data dimensionality and improves computational efficiency while retaining core geometric information.

[0045] 3. High optimization accuracy and fast iteration efficiency. The overall prediction error of the GPR surrogate model is less than 5%, which can accurately capture the nonlinear relationship between structural parameters and electromagnetic and mechanical properties; combined with the NSGA-II algorithm, multi-objective optimization is carried out, effectively avoiding local optima and shortening the overall optimization cycle.

[0046] 4. Complete design and manufacturing closed loop with strong engineering applicability. This invention deeply integrates simulation design, intelligent optimization, and 3D printing technology, clearly defining the matching relationship between material system, printing parameters, post-processing technology, and optimization parameters, forming a complete technical closed loop. The optimized solution can be directly implemented to prepare physical prototypes, and the technology has strong repeatability.

[0047] 5. Wide range of applications. The overall solution is compatible with various stealth equipment such as aerospace, drones, weaponry, and ships. The structural parameters and material system can be flexibly adjusted according to different working conditions, making it highly versatile. Attached Figure Description

[0048] Figure 1 is an overall flowchart of the integrated lattice structure optimization design method for absorbing and carrying waves according to the present invention;

[0049] Figure 2 shows a comparison of electromagnetic simulation results for nine typical lattice topologies in the 2–18 GHz frequency band, including the effective absorption bandwidth and average effective absorption bandwidth data for each structure. Figure 3 This invention provides a multi-objective optimization Pareto front distribution map, illustrating the trade-off between effective absorbing bandwidth and equivalent Young's modulus, including original data points, Pareto front points, and optimal data points. Figure 4 A visualization of the 3D reconstruction of the gradient honeycomb structure at different training stages of the DeepSDF neural network, showing the change in reconstruction accuracy during the model convergence process; Figure 5 This is a schematic diagram of the structure, dimensions, and layers of the FDM integrated manufacturing prototype of the present invention, with the wave-transmitting layer, the wave-absorbing core layer, the reflective layer, and the thickness dimensions of each region marked. Detailed Implementation

[0050] The present invention will be further described in conjunction with the accompanying drawings and embodiments.

[0051] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0052] As an embodiment of the present invention, such as Figure 1 As shown, an optimized design method for an integrated lattice structure that combines wave absorption and wave carrying includes the following specific steps:

[0053] Step S1, Multiple Topology Filtering:

[0054] This embodiment selects nine mainstream crystal lattice structures: honeycomb structure, rhombic lattice, tetrahedral lattice, octahedral lattice, Kagome lattice, pyramid lattice, cubic lattice, cross lattice, and star lattice, while also supplementing Gyroid, Kelvin and other structures as alternatives.

[0055] Evaluation weights were set as follows: absorption performance 0.4, load-bearing performance 0.4, and lightweight performance 0.2. CST MicrowaveStudio was used to conduct electromagnetic simulations across the 2–18 GHz frequency band to detect the effective absorption bandwidth and reflection loss. ANSYS was used to conduct compression mechanics simulations to detect compressive strength and structural stiffness. The structural volume fraction was used as the lightweight evaluation index.

[0056] Using the weighted comprehensive evaluation method of the analytic hierarchy process (AHP), the cellular structure exhibited the best overall performance: an effective absorption bandwidth of 12 GHz, reflection loss ≤-10 dB in the 8–18 GHz band, compressive strength of 6.5 MPa, stiffness of 1200 MPa, and a volume fraction of 0.35. The cellular structure was thus selected as the basic topology for subsequent designs. Electromagnetic simulation data and performance comparison results for nine typical lattice topologies in this step are as follows: Figure 2 As shown, the differences in effective absorption bandwidth and average absorption performance of each topology provide data support for the topology selection conclusions in this step.

[0057] Step S2, Gradient Configuration Design:

[0058] Based on the honeycomb structure design, a three-layer gradient honeycomb lattice configuration is designed, with a gradual change in wall thickness along the thickness direction to form a dielectric constant gradient transition layer, thereby improving the broadband absorption capability.

[0059] Design variables and their value ranges:

[0060] Cellular height of each layer 3–8 mm; Cell diameter of each layer of honeycomb 5–15 mm; honeycomb wall thickness of each layer 0.5–1.0 mm; impregnation thickness of each layer of microwave absorbing slurry : 0.1~0.5mm.

[0061] Constraints:

[0062] Total height of absorbing core layer ; cell wall thickness Single cell cycle The lattice-equivalent Young's modulus is not lower than the lower limit of the design domain.

[0063] Optimization goal:

[0064] Maximize the effective absorption bandwidth of 2–18 GHz, the average reflection loss of the key frequency band of 6–12 GHz ≤ -10 dB, minimize the structural mass, and maximize the compressive strength.

[0065] 200 sets of parameter samples were generated using Latin hypercube sampling, and electromagnetic and mechanical simulations were performed one by one to build a dataset for model training.

[0066] Step S3: DeepSDF implicit neural network encoding

[0067] A DeepSDF implicit neural network was constructed with the following structure: input layer → 4 fully connected layers (with 256, 128, 64, and 32 neurons respectively) → output layer. The input data included the 3D coordinates of the structure surface and normalized size parameters. A size-aware mechanism was implemented to achieve multi-scale structural collaborative encoding. 600,000 structure coordinate points were extracted as training samples. The Adam optimizer was used with a learning rate of 0.001 and a loss function of mean squared error (MSE). Iterative training was performed until the loss value was reached. The model converges. All samples are encoded into 128-dimensional latent vectors to complete geometric feature representation. This step, including the model training process and the 3D reconstruction effect of the gradient honeycomb structure, is shown below. Figure 4 As shown, the structural reconstruction accuracy in the early stage of model training and after convergence intuitively verifies the accurate representation capability of DeepSDF implicit encoding.

[0068] Step S4, PCA dimensionality reduction:

[0069] The 128-dimensional latent vectors were standardized, and principal component analysis was performed. Principal components with a cumulative variance contribution of ≥90% were selected, ultimately reducing the vector dimension to 3 dimensions, preserving the core geometric features of the structure, and eliminating redundant information.

[0070] Step S5, GPR surrogate model construction and training:

[0071] A multi-output GPR surrogate model is constructed, using a squared exponential kernel as the kernel function, with a kernel length scale of 0.3 and a noise variance set to... The model takes a 3D core feature vector as input and outputs effective absorption bandwidth, average reflection loss of 6–12 GHz, compression strength, and structural quality. The 200 datasets are divided into a training set (160 sets) and a test set (40 sets) in an 8:2 ratio. After model training, the overall prediction error on the test set is 3.8%, which is lower than the preset threshold of 5%, and the mapping accuracy meets the requirements.

[0072] Step S6, NSGA-II multi-objective optimization:

[0073] The NSGA-II algorithm parameters are set as follows: population size 100, number of iterations 500, crossover probability 0.8, mutation probability 0.05. Using the GPR surrogate model as the performance prediction vehicle, multi-objective iterative optimization is performed within the constraints to obtain the Pareto optimal solution set. Based on engineering requirements, the set of structural parameters with the best overall performance is selected from the solution set. The multi-objective optimization results of this step are as follows: Figure 3As shown, the optimized Pareto front distribution map clearly presents the performance trade-off between the effective absorption bandwidth of the structure and the equivalent Young's modulus, as well as the selection location of the optimal parameter point.

[0074] Step S7, FDM Integrated Manufacturing:

[0075] Material preparation:

[0076] The printing substrate is PLA (polylactic acid), and the microwave-absorbing filler is carbon black. They are mixed at a mass ratio of 85%:15% and the composite printing filament is prepared by a twin-screw extruder. The wave-transparent layer is made of pure PLA material.

[0077] Printing parameter settings:

[0078] The printing temperature is 220℃, the printing speed is 50 mm / s, and the printing layer thickness is 0.1 mm, matching the honeycomb wall thickness and impregnation thickness. Integrated printing is completed in the order of wave-transmitting layer → wave-absorbing layer → reflective layer.

[0079] Secondary impregnation process:

[0080] Prepare a carbon-based microwave absorbing paste with a solid content of 30%; immerse the printed sample in the paste and let it stand for 5 minutes, then remove it and dry it at 60℃ for 2 hours. Repeat this process twice to complete the microwave absorbing paste loading. The overall dimensional error of the molded sample is ≤±0.05 mm, meeting the processing accuracy requirements. The layered structure, dimensional parameters, and printing sequence of the integrated molded sample in this step are as follows: Figure 5 As shown, the layered structure and thickness dimensions of the wave-transmitting layer, the wave-absorbing core layer, and the reflective layer of the sample are clearly defined, fully demonstrating the final formed structure of the integrated FDM manufacturing process.

[0081] Performance testing and verification

[0082] The absorption performance was tested in the 2–18 GHz frequency band using the bow-shaped method and a vector network analyzer; the structural compressive performance was tested using an electronic universal testing machine at a loading speed of 1 mm / min. Test results: The effective absorption bandwidth of the sample reached 14.2 GHz (2–16.2 ​​GHz), with an average reflection loss ≤-20 dB in the key frequency band of 6–12 GHz; the compressive strength was 7.8 MPa, and the surface density was 425 g / m². Compared with a uniform honeycomb structure with the same parameters, the effective absorption bandwidth was increased by 22%, and the structural stiffness was increased by 18%, with all indicators meeting the design requirements.

Claims

1. A method for optimizing the design of an integrated lattice structure for both wave absorption and wave carrying, characterized in that, Includes the following steps: S1. Multi-topology screening: Using wave absorption, load-bearing capacity, and lightweight performance as evaluation criteria, and combining multiphysics simulation and hierarchical analysis, the topology with the best comprehensive performance is screened from various typical lattice topologies. S2. Gradient configuration design: Based on the optimal topology, a multi-layer gradient lattice configuration is designed, and design variables, constraints, and multi-objective optimization functions are determined. S3. DeepSDF encoding: A DeepSDF implicit neural network with a size-aware mechanism is constructed to discretize and encode the gradient lattice structure, outputting a latent vector. S4. PCA dimensionality reduction: Principal component analysis is performed on the latent vector. S5. Dimensionality reduction analysis to extract core feature vectors; S6. GPR surrogate model construction and training: Using the core feature vectors as input and structural performance indicators as output, train a multi-output GPR surrogate model to ensure that the prediction error is below a preset threshold; S7. NSGA-II optimization: Based on the GPR surrogate model, use the NSGA-II algorithm to perform multi-objective iterative optimization, outputting the Pareto optimal solution set and optimal structural parameters; S8. FDM integrated manufacturing: According to the optimal structural parameters, match the printing material and process parameters, and complete the integrated molding of the lattice structure through FDM fused deposition modeling combined with a secondary impregnation process.

2. The method for optimizing the design of an integrated lattice structure for absorbing and supporting waves according to claim 1, characterized in that, The various typical lattice topologies include honeycomb structures, rhombic lattices, tetrahedral lattices, octahedral lattices, Kagome lattices, pyramid lattices, cubic lattices, cross lattices, and star lattices. The weights for the selection and evaluation criteria are: wave absorption performance 0.4, load-bearing performance 0.4, and lightweight performance 0.

2. The optimal topology obtained from the selection is the cellular structure.

3. The method for optimizing the design of an integrated lattice structure for absorbing and supporting waves according to claim 1, characterized in that, The design variables include the lattice height, aperture, wall thickness, and impregnation thickness of the absorbing paste for each layer; the constraint is the total height of the absorbing core layer. The cell wall is thick single cell cycle The lattice equivalent Young's modulus is not lower than the lower bound of the design domain; the multi-objective optimization function is: to maximize the effective absorption bandwidth of 2-18 GHz wideband, optimize the reflection loss of the key frequency band of 6-12 GHz, minimize the structural mass, and maximize the compressive load-bearing strength.

4. The method for optimizing the design of an integrated lattice structure for absorbing and supporting waves according to claim 1, characterized in that, The DeepSDF implicit neural network includes an input layer, four fully connected layers, and an output layer. The size-aware mechanism normalizes the size parameters of the gradient lattice structure and co-encodes them using the three-dimensional coordinates of the structure's surface. The network training loss function is the mean squared error (MSE), and the convergence criterion is the loss value. The initial dimension of the latent vector output by encoding is 128.

5. The method for optimizing the design of an integrated lattice structure for absorbing and supporting waves according to claim 1, characterized in that, The core feature vector obtained after PCA dimensionality reduction has a dimension of 3.

6. The method for optimizing the design of an integrated lattice structure for absorbing and supporting waves according to claim 1, characterized in that, The GPR surrogate model uses a squared exponential kernel function with a length scale of 0.1–0.5 and a noise variance of . The structural performance indicators are divided into wave absorption performance and load-bearing performance. Wave absorption performance includes effective absorption bandwidth and reflection loss, while load-bearing performance includes compressive strength and equivalent Young's modulus.

7. The method for optimizing the design of an integrated lattice structure for absorbing and supporting waves according to claim 1, characterized in that, The NSGA-II algorithm parameters are set as follows: population size 100, number of iterations 500, crossover probability 0.8, and mutation probability 0.

05.

8. The method for optimizing the design of an integrated lattice structure for absorbing and supporting waves according to claim 1, characterized in that, The FDM printing material is a composite of a microwave absorbing functional component and a carrier matrix component; the microwave absorbing functional component is at least one of carbon black and graphene, and the carrier matrix component is at least one of PLA and PEEK; the printing parameters are: printing temperature 180~260℃, printing speed 30~80 mm / s, and layer thickness 0.05~0.2 mm; the secondary impregnation process is as follows: the printed part is immersed in a carbon-based microwave absorbing slurry with a solid content of 30% and left to stand for 5 minutes, then taken out and dried at 60℃ for 2 hours, and the operation is repeated twice.

9. A lattice structure optimization design system integrating wave absorption and wave carrying, characterized in that, The method for optimizing the integrated lattice structure for absorbing and supporting waves as described in any one of claims 1-8 includes: a multi-topology screening module for screening the optimal lattice topology and outputting the result to a gradient configuration design module; a gradient configuration design module for designing multi-layer gradient lattice configurations, determining design variables, constraints, and optimization objectives, and outputting structural data to a DeepSDF encoding module; a DeepSDF encoding module for encoding the gradient lattice structure and outputting the latent vector to a PCA dimensionality reduction module; a PCA dimensionality reduction module for reducing the dimensionality of the latent vector and extracting core feature vectors, and transmitting them to a GPR surrogate model module; a GPR surrogate model module for constructing and training a surrogate model and establishing a mapping relationship between feature vectors and performance indicators; an NSGA-II optimization module for performing multi-objective optimization based on the GPR surrogate model and outputting the optimal structural parameters to an FDM manufacturing control module; and an FDM manufacturing control module for controlling an FDM device to complete structural printing and secondary impregnation molding according to the optimal parameters.

10. An electronic device, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the integrated lattice structure optimization design method for absorbing and carrying waves as described in any one of claims 1-8.

Images