Simulation and Prediction Method of Electrical Performance of Integrated Mechano-Thermo-Electrical Radome

Abstract

This invention discloses a method for simulating and predicting the electrical performance of an integrated force-thermal-electric radome, belonging to the technical field of radome electrical performance research. Addressing the complexity of radome electrical performance research under force-thermal environments and the low efficiency of existing simulation methods, this invention establishes an efficient mapping relationship between force-thermal environment parameters and radome electrical performance, avoiding complex multi-physics iterative calculation processes. This provides a new technical approach for rapid performance evaluation and optimized design of radomes under force-thermal environments. The few-sample learning technique employed in this invention can achieve higher accuracy in electrical performance prediction under limited training data conditions, while reducing computation time from hours to minutes.

Description

Technical Field

[0001] This invention belongs to the technical field of radome electrical performance research, specifically relating to a method for simulating and predicting the electrical performance of an integrated mechanical-thermal-electrical radome. Background Technology

[0002] With the rapid development of high-speed aircraft, drones, and other high-end equipment, radomes, as key components protecting antenna systems from external environmental interference, directly impact the reliability of communication, navigation, and detection systems. Under complex force-thermal environments (such as aerodynamic pressure and high-temperature thermal loads), radomes undergo deformation and material parameter changes, leading to electrical performance issues such as decreased transmittance and increased aiming errors. Therefore, accurately predicting the electrical performance of radomes under force-thermal environments has become a core challenge in design and optimization. Traditional methods for analyzing the electrical performance of radomes mainly rely on multiphysics coupling simulation technology, which simulates performance changes under actual operating conditions through bidirectional coupling calculations of force, heat, and electromagnetic fields.

[0003] Traditional multiphysics coupled simulation methods require iterative calculations of the bidirectional coupling of force, heat, and electromagnetic fields, while maintaining extremely high mesh density, when analyzing the impact of force and thermal environments on the electrical performance of radomes. This results in a single complete simulation often taking tens of hours or even days. This high computational cost necessitates resimulating even minor changes in environmental parameters, severely limiting the efficiency of design optimization. Summary of the Invention

[0004] To address the aforementioned problems in the existing technology, this invention provides a method for simulating and predicting the electrical performance of an integrated thermoelectric radome. The technical problem to be solved by this invention is achieved through the following technical solution:

[0005] This invention provides a method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome, the method comprising:

[0006] The temperature field distribution of the radome is obtained by structural partitioning, and mechanical-thermal loads are applied to perform multi-field coupled simulation to obtain the radome model after deformation and temperature distribution.

[0007] Based on the radome model, the finite element mesh nodes are added to the displacement field to obtain the deformed radome solid model;

[0008] Material parameters are mapped to different zones based on the temperature field distribution to set electromagnetic simulation conditions for non-uniform material properties.

[0009] The deformed radome physical model is integrated with the antenna array to construct a radome antenna system; electromagnetic simulation is performed based on the radome antenna system and the set electromagnetic simulation conditions to output force-thermal-electric coupling simulation data.

[0010] A dataset based on force-thermal-electric coupling simulation data is used to train an electrical performance prediction model using a few-shot learning algorithm, resulting in a trained electrical performance prediction model. The trained electrical performance prediction model is then used to predict the electrical performance parameters of the radome under arbitrary force-thermal load combinations.

[0011] The beneficial effects of this invention are:

[0012] The solution provided by this invention establishes an efficient mapping relationship between mechanical and thermal environmental parameters and the electrical performance of the radome, avoiding the complex multi-physics iterative calculation process, and providing a new technical approach for rapid performance evaluation and optimization design of radomes under mechanical and thermal conditions. The small sample learning technology used in this invention can achieve higher accuracy in electrical performance prediction under limited training data conditions, while reducing the calculation time from hours to minutes. Attached Figure Description

[0013] Figure 1 This is a schematic diagram illustrating the steps of a method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome provided in an embodiment of the present invention.

[0014] Figure 2 This is a schematic diagram of the von Kármán radome in a method for simulating and predicting the electrical performance of a thermoelectric integrated radome provided in an embodiment of the present invention.

[0015] Figure 3 This is a layered schematic diagram of the radome in a method for simulating and predicting the electrical performance of an integrated thermoelectric radome provided in an embodiment of the present invention.

[0016] Figure 4 This is a schematic diagram of the deformation mesh in a method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome provided in an embodiment of the present invention.

[0017] Figure 5 This is a schematic diagram showing the deformation mesh being restored to a solid in a method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome provided in an embodiment of the present invention.

[0018] Figure 6 This is a schematic diagram of the antenna array in a method for simulating and predicting the electrical performance of a thermo-mechanical integrated radome provided in an embodiment of the present invention.

[0019] Figure 7 This is a flowchart illustrating the small-sample learning algorithm in a method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome provided in an embodiment of the present invention.

[0020] Figure 8 This is a schematic diagram illustrating the principle of meta-learning in the small-sample learning algorithm of the electromechanical integrated antenna radome simulation and prediction method provided in an embodiment of the present invention. Detailed Implementation

[0021] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0022] To address the complexity of studying the electrical performance of radomes under mechanical and thermal conditions and the low efficiency of existing simulation methods, this invention provides a method for simulating and predicting the electrical performance of an integrated mechanical, thermal, and electrical radome. Figure 1 As shown, it may include:

[0023] S1. The radome is structurally partitioned to obtain the temperature field distribution, and mechanical-thermal loads are applied to perform multi-field coupled simulation to obtain the radome model after deformation and temperature distribution.

[0024] Specifically, the process of structurally partitioning the radome to obtain the temperature field distribution can include:

[0025] Different forces are applied to different areas of the radome, and different ambient temperatures are set on the outer surface of the radome to obtain the temperature field distribution.

[0026] For radomes of different shapes, embodiments of this invention generate a von Kármán curve using a two-dimensional super-quadratic equation, which serves as the radome's generatrix. A rotationally symmetric radome is then generated using this generatrix, resulting in a von Kármán radome as shown below. Figure 2 As shown, the radome has a height of 480mm, a top thickness of 18.75mm, a bottom thickness of 16.7mm, and a bottom radius of 200mm.

[0027] The radome is structurally partitioned, resulting in a layered schematic diagram as shown below. Figure 3 As shown, it can be seen that Figure 3 In the figure, L represents the height of the radome, which is divided into layers according to the proportions shown in the figure.

[0028] In multi-field coupled simulations, the simulation can include: static load analysis and thermodynamic analysis; among which...

[0029] Static load analysis is used to simulate structural deformation under aerodynamic pressure and impact loads;

[0030] Thermodynamic analysis is used to calculate the thermal expansion effect and temperature gradient stress under a preset temperature environment.

[0031] Multi-field coupled simulation was performed by applying mechanical-thermal loads to obtain the radome model after deformation and temperature distribution. The corresponding deformed mesh in the radome model is shown below. Figure 4 As shown in the image. Specifically, the preset temperature can be set by the user according to the required high-temperature environment.

[0032] S2, based on the radome model, adds the finite element mesh nodes to the displacement field to obtain the deformed radome solid model, which may include:

[0033] The finite element mesh nodes are added to the displacement field in the radome model to obtain the deformed mesh file;

[0034] A surface mesh is generated from the finite element volume mesh in the deformed mesh file, a shell is generated from the surface mesh, and a solid is generated from the shell. This solid is then used as the solid model of the deformed radome. A schematic diagram of restoring the deformed mesh to a solid is shown below. Figure 5 As shown, Figure 5 The relatively higher outer portions of the image represent the solid entity before deformation, while the relatively lower inner portions represent the solid entity after deformation. This invention, through the superposition of finite element mesh nodes and displacement fields, maintains the mesh topology and rapidly generates the solid model of the deformed radome.

[0035] S3, based on the temperature field distribution, performs partitioned mapping of material parameters to set electromagnetic simulation conditions for non-uniform material properties, which may include:

[0036] The radome is divided into several material parameter zones based on the temperature field distribution. Each material parameter zone is assigned the dielectric constant and loss tangent at the corresponding temperature to set the electromagnetic simulation conditions for non-uniform material properties.

[0037] Specifically, in electromagnetic simulation software, the material parameters of the radome are mapped to different partitions based on the temperature field distribution to achieve electromagnetic simulation of non-uniform material properties. Therefore, before simulation, the material parameter variation curves of the radome material with temperature need to be imported into the electromagnetic software, along with the temperature field file. Finally, electromagnetic simulation of the radome antenna system is achieved. In this embodiment of the invention, the radome is divided into several material parameter partitions based on the temperature field distribution, and the dielectric constant and loss tangent values ​​are assigned to the corresponding temperatures to achieve high-precision full-wave electromagnetic simulation.

[0038] S4 integrates the deformed radome physical model with the antenna array to construct a radome antenna system; performs electromagnetic simulation based on the radome antenna system and the set electromagnetic simulation conditions, and outputs force-thermal-electric coupling simulation data.

[0039] Specifically, the deformed radome solid model obtained in step S2 and as shown in the figure... Figure 6 The antenna array shown forms a radome antenna system. Figure 6 The left side of the diagram shows the dimensions of the antenna elements in the antenna array, and an electromagnetic simulation of the system is performed using the finite element domain decomposition algorithm.

[0040] The finite element domain decomposition method facilitates simulation of corresponding models with limited resources. The process is similar to conventional simulation methods: importing the model, setting electromagnetic parameters, configuring the finite element method, determining the number of partitions, and performing simulation calculations are all part of the process. Mechano-thermal-electric coupling simulation data can include aiming error and transmittance. After completing these steps, the results are post-processed to calculate the aiming error and transmittance at different scanning angles.

[0041] The antenna array in this embodiment of the invention is a phased array antenna. Its beam direction is not achieved through mechanical rotation, but rather by adjusting the phase of the feed current to each antenna element to change the pointing angle. The radiation field of the array antenna can be considered as a vector superposition of the fields generated by each antenna element. Therefore, when the amplitude and phase of the feed current to each antenna element satisfy a specific relationship, the desired radiation characteristics can be achieved.

[0042] Scripts can be written to automate multiphysics simulations and resolve file transfer and compatibility issues between software programs. These scripts can include: calling mechanical-thermal simulation software, exporting mesh files, processing mesh files in modeling software, and calling electromagnetic simulation software. This invention automates multiphysics simulations (mechanical-thermal-electromagnetic) using scripts, resolving file transfer and compatibility issues between software programs and improving data generation efficiency.

[0043] S5 is a dataset constructed based on force-thermal-electric coupling simulation data. A few-shot learning algorithm is used to train an electrical performance prediction model to obtain a trained electrical performance prediction model. The trained electrical performance prediction model is used to predict the electrical performance parameters of the radome under arbitrary force-thermal load combinations.

[0044] S51, a dataset constructed based on force-thermal-electric coupling simulation data, uses a few-shot learning algorithm to train an electrical performance prediction model, resulting in a trained electrical performance prediction model, which may include:

[0045] S511: A dataset is constructed based on multiple sets of force-thermal-electric coupling simulation data. The dataset is then augmented and normalized to obtain the processed dataset.

[0046] S512 divides the processed dataset into training and test sets through random sampling;

[0047] S513, initialize the neural network;

[0048] S514, based on the meta-learning framework, uses the training set to train the neural network and obtain the trained electrical performance prediction model.

[0049] S515 uses a test set to evaluate the electrical performance prediction model and obtain the error;

[0050] S516, repeat steps S511-S515 several times until the average error meets the preset error requirement or the number of iterations reaches the preset number of iterations, and obtain the trained electrical performance prediction model.

[0051] Specifically, initializing a neural network can include:

[0052] Initialize the number of nodes and weights in the input, hidden, and output layers of the neural network.

[0053] Based on the meta-learning framework, a neural network is trained using a training set to obtain a trained electrical performance prediction model, which may include:

[0054] The neurons in the neural network are delinearized using a nonlinear activation function to obtain the processed neural network.

[0055] The meta-learning framework is used to pre-train the processed neural network based on the training set to obtain global parameters.

[0056] The processed neural network is trained based on global parameters to obtain a trained electrical performance prediction model.

[0057] The mechanical-thermal-electric coupling simulation data is used as the training set for training the electrical performance prediction model. A flowchart of the few-shot learning algorithm is shown below. Figure 7 As shown, firstly, the data is normalized and data augmented. Then, the data is partitioned using Bootstrapping random sampling, dividing the samples into training and test sets. During the training phase, a meta-learning framework is used to pre-train the neural network processed by the non-linear activation function based on the training set, obtaining global parameters. These global parameters will be further optimized during subsequent neural network training. After the model training is complete, it is evaluated using the initially partitioned test set, and the error is calculated. Subsequently, the training and test sets are re-partitioned, and the training and evaluation process is repeated to obtain multiple sets of different error values. Finally, the overall predictive ability of the model is evaluated by calculating the average of these errors.

[0058] Understandably, embodiments of the present invention employ few-shot learning algorithms (such as residual neural networks of transfer learning, integrated models of support vector regression and Gaussian processes, and prototype networks under the meta-learning framework) to establish an efficient mapping relationship between mechanical and thermal environmental parameters and the electrical performance of the radome, significantly reducing the dependence on a large amount of training data.

[0059] A schematic diagram illustrating the principle of meta-learning in few-shot learning algorithms, as shown below. Figure 8As shown, when the problem to be learned is a single-task problem, the meta-learning training set consists of multiple pseudo-tasks, each further divided into a support set and a query set. The inner loop starts with the initial parameters (usually provided by the outer loop) and performs several gradient updates using data from the support set in the pseudo-task, adjusting the model parameters to adapt to the current pseudo-task. The parameter updates in the inner loop occur locally within the task and are not shared with other tasks.

[0060] The outer loop doesn't select an "optimal set of parameters" from all tasks. Instead, it calculates gradients and updates the global parameters based on the adaptation results of each inner loop task and the loss on the pseudo-task query set. The goal of the outer loop is to find a global set of initial parameters from which the model can quickly adapt to all tasks (including future new tasks). In other words, the outer loop's optimization objective is to improve the model's adaptability and generalization ability, rather than optimizing parameters individually for each task.

[0061] This invention establishes an efficient mapping relationship between mechanical and thermal environmental parameters and the electrical performance of the radome, avoiding complex multiphysics iterative calculation processes. This provides a new technical approach for rapid performance evaluation and optimized design of radomes under mechanical and thermal conditions. The few-sample learning technique used in this invention can achieve higher accuracy in electrical performance prediction under limited training data conditions, while reducing the calculation time from hours to minutes, thus achieving higher accuracy in electrical performance prediction.

[0062] It should be noted that, in the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0063] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for simulating and predicting the electrical performance of a thermo-mechanical integrated radome, characterized in that, include: The temperature field distribution of the radome is obtained by structural partitioning, and mechanical-thermal loads are applied to perform multi-field coupled simulation to obtain the radome model after deformation and temperature distribution. Based on the radome model, the finite element mesh nodes are added to the displacement field to obtain the deformed radome solid model; Material parameters are mapped to different zones based on the temperature field distribution to set electromagnetic simulation conditions for non-uniform material properties. The deformed radome physical model is integrated with the antenna array to construct a radome antenna system. Electromagnetic simulation is performed based on the radome antenna system and the set electromagnetic simulation conditions, and force-thermal-electric coupling simulation data is output. A dataset based on force-thermal-electric coupling simulation data is used to train an electrical performance prediction model using a few-shot learning algorithm, resulting in a trained electrical performance prediction model. The trained electrical performance prediction model is then used to predict the electrical performance parameters of the radome under arbitrary force-thermal load combinations.

2. The method for simulating and predicting the electrical performance of a thermo-mechanical integrated radome according to claim 1, characterized in that, The process of structurally partitioning the radome to obtain the temperature field distribution includes: Different forces are applied to different areas of the radome, and different ambient temperatures are set on the outer surface of the radome to obtain the temperature field distribution.

3. The method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome according to claim 1, characterized in that, The multi-field coupled simulation includes: static load analysis and thermodynamic analysis; wherein, The static load analysis is used to simulate structural deformation under aerodynamic pressure and impact loads; The thermodynamic analysis is used to calculate the thermal expansion effect and temperature gradient stress under a preset temperature environment.

4. The method for simulating and predicting the electrical performance of a thermo-mechanical integrated radome according to claim 1, characterized in that, The process of adding the finite element mesh nodes and the displacement field based on the radome model to obtain the deformed radome solid model includes: The finite element mesh nodes are added to the displacement field in the radome model to obtain the deformed mesh file; A surface mesh is generated from the finite element volume mesh in the deformed mesh file, a shell is generated from the surface mesh, a solid is generated from the shell, and the solid is used as the solid model of the deformed radome.

5. The method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome according to claim 1, characterized in that, The step of mapping material parameters to different temperature fields to set electromagnetic simulation conditions for non-uniform material properties includes: The radome is divided into several material parameter zones based on the temperature field distribution. Each material parameter zone is assigned the dielectric constant and loss tangent at the corresponding temperature to set the electromagnetic simulation conditions for non-uniform material properties.

6. The method for simulating and predicting the electrical performance of an integrated thermoelectric radome according to claim 1, characterized in that, The dataset constructed based on the mechanical-thermal-electric coupling simulation data is used to train an electrical performance prediction model using a few-shot learning algorithm, resulting in a trained electrical performance prediction model, including: S511: A dataset is constructed based on multiple sets of force-thermal-electric coupling simulation data. The dataset is then augmented and normalized to obtain the processed dataset. S512 divides the processed dataset into training and test sets through random sampling; S513, initialize the neural network; S514, based on the meta-learning framework, uses the training set to train the neural network and obtain the trained electrical performance prediction model. S515 uses a test set to evaluate the electrical performance prediction model and obtain the error; S516, repeat steps S511-S515 several times until the average error meets the preset error requirement or the number of iterations reaches the preset number of iterations, and obtain the trained electrical performance prediction model.

7. The method for simulating and predicting the electrical performance of an integrated thermoelectric radome according to claim 6, characterized in that, The initialization of the neural network includes: Initialize the number of nodes and weights in the input, hidden, and output layers of the neural network.

8. The method for simulating and predicting the electrical performance of a thermoelectric integrated antenna radome according to claim 6, characterized in that, The meta-learning framework, which trains a neural network using a training set to obtain a trained electrical performance prediction model, includes: The neurons in the neural network are delinearized using a nonlinear activation function to obtain the processed neural network. The meta-learning framework is used to pre-train the processed neural network based on the training set to obtain global parameters. The processed neural network is trained based on global parameters to obtain a trained electrical performance prediction model.

Images