Radar antenna structure damage and function damage mapping evaluation method and system based on subarray-level equivalent target

Abstract

The invention provides a radar antenna structure damage and function damage mapping evaluation method and system based on a subarray-level equivalent target, and relates to the technical field of radar antenna damage effect evaluation. According to the method, a baseline parameter set is formed through standardized filing of functional parameters of an equivalent target array structure, damage is loaded and damage parameters are collected under a preset working condition, main lobe gain, beam width, side lobe level and beam pointing are tested, functional variable quantity is calculated, array surface damage is divided and coded through a two-dimensional matrix and blocks, and standardized input is generated. A quantitative mapping model is established, and through cross validation, residual analysis optimization and adaptive updating, high-precision and traceable evaluation of structures and functions under multiple working conditions is realized, and support is provided for criterion optimization and anti-damage design.

Description

Technical Field

[0001] This invention relates to the field of radar antenna damage effect assessment technology, and in particular to a radar antenna structural damage and functional damage mapping assessment method and system based on subarray-level equivalent targets. Background Technology

[0002] As a core subsystem of advanced electronic systems, phased array radar antennas have a decisive impact on the overall performance of equipment due to their damage resistance and functional retention. Current antenna damage assessment mainly relies on equivalent target physical tests and finite parameter statistical analysis, typically judging the antenna's damage state by observing indicators such as structural deformation and the number of array element failures. However, with the increasing complexity of antenna structures and the continuous improvement of subarray integration, the distribution characteristics of physical damage in the array space have become increasingly diverse and complex. Damage assessment based solely on overall or single physical parameters can no longer accurately reflect the multidimensional impact of damage on key electrical performance aspects (such as array gain, beam directivity, and sidelobe suppression). Traditional experimental assessment methods have significant shortcomings in terms of spatial distribution details, parameter normalization processing, and the correlation modeling between array element failure modes and system functional performance, resulting in limitations in the resolution and scientific rigor of damage situation awareness.

[0003] Meanwhile, existing assessment methods for radar antenna damage mechanisms under extreme physical environments generally lack systematic collection, standardized processing, and batch management of spatial distribution data of subarray or element-level damage, making it difficult to achieve structured aggregation and feature extraction of large-scale damage samples. More critically, current quantitative mappings between physical damage parameters and functional damage indices largely rely on empirical relationships or low-dimensional models, lacking scientifically effective modeling methods to reveal the essential correlation and evolutionary laws between the two. This limits a deeper understanding and accurate assessment of antenna functional degradation mechanisms under complex damage conditions.

[0004] In summary, existing phased array radar antenna damage effect assessment systems still have significant shortcomings in terms of finely characterizing the spatial distribution features of subarray-level damage and achieving high-precision quantitative mapping between structural damage parameters and system functional performance indicators. These problems directly limit the scientific analysis and accurate assessment of radar antenna functional retention and damage effectiveness under complex damage environments. Therefore, there is an urgent need to develop a novel assessment method that integrates physical experiments, standardized acquisition of spatial distribution parameters, and data-driven modeling to achieve high-precision quantitative mapping between structural and functional damage and intelligent criterion optimization, thereby promoting the innovation and improvement of radar antenna damage effect assessment theory and technology. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, the purpose of this invention is to provide a radar antenna structural damage and functional damage mapping assessment method and system based on subarray-level equivalent targets. Using subarray-level equivalent targets and standardized data links as a foundation, the method expresses array surface damage through two-dimensional matrices and block partitioning, pairs damage characteristics with functional changes using unique codes, and employs a quantitative mapping model optimized through cross-validation and residual analysis. This achieves high-precision, traceable, and evolvable structural-functional assessment under multiple operating conditions, multiple spatial modes, and different array sizes, providing reliable data support and engineering pathways for antenna functional criterion optimization and damage-resistant design.

[0006] To achieve the above objectives, the present invention provides the following solution: A method for evaluating the structural and functional damage of radar antennas based on subarray-level equivalent targets includes: Select a subarray-level equivalent target with the same array structure and functional parameters as the target radar antenna, and standardize and archive the array structure and functional parameters of the equivalent target to form a baseline parameter set and object number for the experimental object. Damage determined by type, proportion, and spatial distribution is applied according to preset working conditions, and damage parameter sets are collected and archived. Test the main lobe gain, beamwidth, sidelobe level, and beam pointing; calculate the functional changes relative to the baseline parameter set of the experimental object; and archive the results by pairing them with the damage parameter set according to the operating condition number, object number, and timestamp. The array damage is encoded using a two-dimensional matrix and block division, and paired with the working condition number and subarray number to form a standardized set of input parameters. A quantitative mapping model is established using the set of damage parameters and changes in function. At least one method is selected from multiple linear regression, support vector machine, random forest, and artificial neural network. The model is optimized by k-fold cross-validation and residual analysis, and then updated adaptively.

[0007] Preferably, the damage parameter set includes the array element failure ratio, spatial distribution coding, maximum number of consecutively failed elements, distribution entropy, and damage concentration.

[0008] Preferably, a subarray-level equivalent target with array structure functional parameters consistent with the target radar antenna is selected, and the array structure functional parameters of the equivalent target are standardized and archived to form a baseline parameter set and object number for the experimental object, including: The experimental object was determined to be the equivalent target of a subarray-level radar antenna, and the array structure functional parameters were collected. The array structure functional parameters include: array type, array element arrangement, number of subarray elements, antenna key dimensions, module size, layout, operating frequency band, main lobe gain, beamwidth, radiation pattern, and sidelobe suppression ratio. The array structure functional parameters are stored in a database or spreadsheet system using a standardized data structure, and dual management of spreadsheets and databases is implemented. The array structure functional parameters are standardized and archived using unified fields, units, and codes. A unique numerical code is assigned to each experimental object, and this code is associated with the functional parameters of the array structure to form a baseline parameter set and an experimental object number.

[0009] Preferably, damage determined by type, proportion, and spatial distribution is applied according to preset working conditions, and a damage parameter set is collected and archived, including: Establish a set of damage conditions, and construct conditions with different damage ratios and spatial distributions by setting the failure state of array elements through physical loading or digital programming. Assign a condition number, record a timestamp and experimental object number to each condition. Collect a set of damage parameters; the set of damage parameters includes the failure rate of array elements, the maximum number of consecutively failed elements, the distribution entropy, the damage concentration, and the subarray number. Before being stored, the damage parameter set is normalized and outlier is removed. The damage parameter set is uniquely paired and archived with the working condition number, experimental object number, and timestamp using unified fields, units, and codes.

[0010] Preferably, after uniquely pairing and archiving the damage parameter set with the working condition number, experimental object number, and timestamp using unified fields, units, and codes, the method further includes: Visualize and save the archived damage parameter set; When preset conditions are met, the physical quantities of structural deformation and fracture are supplemented by digital image sensors and displacement sensors, and archived together with the damage parameter set.

[0011] Preferably, the main lobe gain, beamwidth, sidelobe level, and beam pointing are tested; the functional changes relative to the baseline parameter set of the experimental object are calculated; and these changes are paired with the damage parameter set according to the operating condition number, object number, and timestamp for archiving, including: The corresponding structural damage states are loaded sequentially according to the operating condition number, and the electrical performance test process is initiated. The four electrical performance parameters of main lobe gain, beamwidth, sidelobe level, and beam pointing were measured using a network analyzer, signal generator, and automated acquisition terminal, and were collected and recorded in real time according to a unified standard unit and data format. Based on the baseline parameter set of the experimental object, the functional changes of the electrical performance parameters relative to the baseline parameter set of the experimental object are calculated; the functional changes include: main lobe gain change, beamwidth change, sidelobe level change, and beam pointing offset; The functional changes are matched one-to-one with the damage parameter set under the corresponding working conditions according to the working condition number, experimental object number, and timestamp, and then archived in batches through an automated database interface to achieve full traceability. Before being stored, the damage parameter set and the electrical performance parameters are normalized, outlier removal and format verification are performed, and the equipment calibration and automatic pairing relationship verification are completed according to the standard operating procedure to obtain the paired dataset. After importing the paired datasets into the database, visualization is supported in the form of heatmaps, distribution maps, and parametric curves.

[0012] Preferably, the array surface damage is encoded using a two-dimensional matrix and block division, and paired with the working condition number and subarray number to form a standardized set of input parameters, including: A two-dimensional grid is established based on the actual arrangement of array elements to generate an array surface damage matrix; in the array surface damage matrix, "1" represents array element failure and "0" represents normal operation. The array damage matrix is ​​flattened into a one-dimensional encoded string in row order and stored as a spatial distribution encoded field. The array is divided into multiple functional blocks according to a preset area, and the number of failed array elements in each block is recorded. A standardized set of input parameters is constructed using the array damage matrix, one-dimensional spatial distribution code, number of failures in each block, and subarray number. These parameters are uniquely paired and archived with the working condition number, subarray number, experimental object number, and timestamp. Unified fields, units, and codes are used for database management.

[0013] Preferably, a quantitative mapping model is established based on the set of damage parameters and changes in function, using at least one method selected from multiple linear regression, support vector machine, random forest, and artificial neural network. This model is optimized through k-fold cross-validation and residual analysis, and then adaptively updated, including: A training sample set is constructed based on data that uniquely matches the working condition number, the experimental object number, and the timestamp; the training sample set consists of a set of damage parameters and corresponding functional changes and is archived in a standardized format. Perform sample normalization and feature filtering on the training sample set to improve modeling stability; At least one of the following methods—multivariate linear regression, support vector machine, random forest, and artificial neural network—is used to model the quantitative mapping relationship between structural damage parameters and functional changes. The generalization ability and prediction accuracy were evaluated by k-fold cross-validation and residual analysis, and the model structure and parameters obtained by modeling were optimized. Perform feature sensitivity analysis on key impairment features to identify input features that significantly affect the amount of functional change; Model parameters, evaluation indicators and modeling results are archived in batches using unified fields, units and codes to achieve version management, historical comparison and traceability; Incremental training and adaptive updates are performed when new operating conditions or samples are imported, and model parameters are optimized periodically to keep the model evolving with the data.

[0014] Preferably, the artificial neural network includes an input layer, at least two levels of gated residual feature layers, and an output layer. The input layer receives a set of impairment parameters, and the output layer regresses a vector of functional changes. The gated residual feature layers perform adaptive improvements based on feature sensitivity guidance, and the forward propagation satisfies: ; in, For the first The input feature vector of the layer, For the first The output feature vector of the layer, s, is a sensitivity vector obtained by normalization based on the feature sensitivity analysis results of claim 8. For the first Layer gate vector, For the first Layer non-negative gated strength scalar For the Sigmoid function, For element-wise activation functions, and For the first The layer weight matrix is ​​used to generate nonlinear branches and linear residual branches respectively. For element-wise multiplication, This is the vector of functional changes output by the network. This is the output layer weight matrix. This is the vector of changes in target function obtained by comparing it with the baseline parameter set of the experimental subjects. The training loss is calculated using the mean squared error.

[0015] A radar antenna structural and functional damage mapping and evaluation system based on a subarray-level equivalent target includes: The experimental object and baseline archiving unit is used to select a subarray-level equivalent target that is consistent with the array structure functional parameters of the target radar antenna, and to standardize and archive the array structure functional parameters of the equivalent target to form an experimental object baseline parameter set and object number. The damage condition loading and damage parameter archiving unit is used to load damage determined by type, proportion, and spatial distribution according to preset conditions, and to collect and archive the damage parameter set. The electrical performance testing and functional change pairing and archiving unit is used to test the main lobe gain, beamwidth, sidelobe level, and beam pointing, calculate the functional change relative to the baseline parameter set of the experimental object, and pair and archive it with the damage parameter set according to the operating condition number, object number, and timestamp. The spatial distribution coding and standardized input generation unit is used to encode the array surface damage by dividing it into two-dimensional matrices and blocks, and pair it with the working condition number and subarray number to form a standardized input parameter set. The quantitative mapping modeling and adaptive optimization unit is used to establish a quantitative mapping model based on the set of damage parameters and functional changes. It selects at least one method from multiple linear regression, support vector machine, random forest, and artificial neural network, optimizes it through k-fold cross-validation and residual analysis, and adaptively updates it.

[0016] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: (1) This invention uses the subarray-level equivalent target as a carrier, first establishes the baseline parameter set of the experimental object, then collects damage parameters in a standardized manner under various damage ratios and spatial distributions, and encodes the array damage in a two-dimensional matrix and block, simultaneously tests the main lobe gain, beamwidth, sidelobe level, beam pointing and calculates the functional change, and finally trains a quantitative mapping model with data paired with samples to achieve high resolution, interpretability and quantifiable correlation between structural damage and functional degradation, which significantly overcomes the distortion and instability caused by existing judgments based on a single statistical quantity.

[0017] (2) The data of this invention are stored in the database with unified fields, units and codes. The damage parameter set, spatial distribution code and functional change quantity are uniquely matched according to the working condition number, experimental object number, subarray number and timestamp. Normalization and outlier removal are performed before storage. After storage, heat map, matrix distribution map and parameter curve visualization are supported. Closed-loop management from damage loading, index testing, sample matching to modeling and evaluation is realized, which improves data quality and traceability of reproducible experiments and reduces the deviation caused by human operation differences.

[0018] (3) The input of this invention simultaneously covers statistical damage features and two-dimensional spatial distribution coding, and the output is the functional change compared with the baseline. In the modeling stage, at least one of the following methods can be selected: multiple linear regression, support vector machine, random forest, artificial neural network. The structure and parameters are optimized by k-fold cross-validation and residual analysis to ensure that it still has good generalization ability and prediction accuracy under different array sizes, subarray layouts and damage modes.

[0019] (4) The present invention automates the electrical performance testing and data pairing, and after forming a standardized set of input parameters, it can directly drive model inference, thereby achieving rapid evaluation of the degree and level of functional degradation. This process is suitable for batch samples and multi-batch test scenarios, shortens the evaluation cycle, reduces test costs, and supports quantitative optimization of functional criteria and test schemes based on evaluation results, thereby improving the efficiency of engineering applications.

[0020] (5) This invention supports incremental training and adaptive updates, and automatically optimizes the model as new operating conditions and new samples are added; it identifies damage features that have a significant impact on function through feature sensitivity analysis, and provides quantitative basis for subarray layout, device redundancy, criterion threshold and maintenance strategy, thereby enhancing the functional retention and damage resistance design capability under extreme operating conditions. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart of the method provided in an embodiment of the present invention; Figure 2 A flowchart illustrating the technical route provided in the embodiments of the present invention; Figure 3 This is a schematic diagram of the equivalent target structure of a subarray-level radar antenna provided in an embodiment of the present invention; Figure 4 A schematic diagram of a standard module structure provided for an embodiment of the present invention; Figure 5 This is a schematic diagram of the antenna unit structure provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the transceiver chip structure provided in an embodiment of the present invention; Figure 7 This is a schematic diagram showing the arrangement of the strain gauge interface on the equivalent target according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure and connection of a single strain gauge interface provided in an embodiment of the present invention; Figure 9 This is a schematic diagram of the "zero-damage" working condition spatial distribution provided in an embodiment of the present invention. All elements of the array are functioning normally, and there are no failed elements. Figure 10 This is a schematic diagram of the spatial distribution of the "10% damage" working condition provided in an embodiment of the present invention, where 10% of the elements within the array fail. Figure 11This is a schematic diagram of the spatial distribution of the "20% damage" working condition provided in an embodiment of the present invention, where 20% of the elements within the array fail. Figure 12 This is a schematic diagram of the spatial distribution of the "30% damage" working condition provided in an embodiment of the present invention, where 30% of the elements within the array fail. Figure 13 This is a schematic diagram of the spatial distribution of the "40% damage" working condition provided in an embodiment of the present invention, showing that 40% of the elements within the array fail. Figure 14 This is a schematic diagram of the spatial distribution of the "50% damage" working condition provided in an embodiment of the present invention, where 50% of the elements within the array fail. Figure 15 This is a schematic diagram of the array current distribution under the condition of edge-concentrated distribution provided in an embodiment of the present invention; Figure 16 This is a schematic diagram of the array current distribution under the condition of centralized distribution in an embodiment of the present invention; Figure 17 This is a schematic diagram of the array current distribution under random distribution conditions provided in an embodiment of the present invention; Figure 18 This is a schematic diagram of the subarray-level radar antenna damage assessment method provided in an embodiment of the present invention; Figure 19 This is a schematic diagram of the structural damage-functional impairment mapping modeling provided in an embodiment of the present invention.

[0023] Explanation of reference numerals in the attached figures: 1. Radome; 2. Heat sink housing; 3. Active subarray; 4. Waveguide circuit board; 5. Centrifugal fan board; 6. Standard module; 7. Antenna unit; 8. Transceiver chip; 9. Strain gauge reserved interface. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] Figure 1 The method flowchart provided in the embodiments of the present invention is as follows: Figure 1 As shown, this invention provides a method for evaluating the structural and functional damage of radar antennas based on subarray-level equivalent targets, including: Step 100: Select a subarray-level equivalent target that matches the array structure functional parameters of the target radar antenna, and standardize and archive the array structure functional parameters of the equivalent target to form a baseline parameter set and object number for the experimental object. Step 200: Load damage determined by type, proportion, and spatial distribution according to preset working conditions, and collect and archive the damage parameter set; Step 300: Test the main lobe gain, beamwidth, sidelobe level, and beam pointing; calculate the functional change relative to the baseline parameter set of the experimental object; and archive the test results with the damage parameter set according to the operating condition number, object number, and timestamp. Step 400: Encode the array surface damage using a two-dimensional matrix and block division, and pair it with the working condition number and subarray number to form a standardized set of input parameters; Step 500: Establish a quantitative mapping model based on the set of damage parameters and changes in function. Select at least one method from multiple linear regression, support vector machine, random forest, and artificial neural network. Optimize the model through k-fold cross-validation and residual analysis, and then update it adaptively.

[0027] Reference Figure 2 The flowchart below illustrates a method for quantitatively mapping structural and functional damage of radar antennas based on subarray-level spatial distribution parameters, comprising the following steps: S1: Determine the experimental object and the spatial distribution parameter system of the subarray level. In this embodiment, the experimental object is an equivalent target of a subarray-level modular radar antenna (such as...). Figure 3 , Figure 4 (As shown). This equivalent target is modularly designed based on the array structure of an actual phased array radar antenna, possessing highly replicated physical dimensions and functional partitions. The equivalent target consists of multiple pluggable active subarrays, each containing standard modules (such as...). Figure 4 Each module further includes several antenna elements and transceiver chips (such as...). Figure 5 , Figure 6 It also reserves multiple standardized interfaces to facilitate rapid assembly under different experimental requirements.

[0028] See Figures 3 to 8 Furthermore, the radome 1 covers and protects the entire subarray-level radar antenna equivalent target array surface, and forms a complete external protection structure with the heat dissipation housing 2 through a fixed connection, thereby achieving environmental isolation and wave transmission protection for the internal array and devices.

[0029] Furthermore, the heat dissipation housing 2 is located outside the equivalent target of the subarray-level radar antenna, and is used to support and enclose the internal core components, including the active subarray 3, the wave control circuit board 4, and the centrifugal fan board 5. The heat dissipation housing not only provides structural support, but also works in conjunction with the centrifugal fan board 5 to achieve heat dissipation management of the core components.

[0030] Furthermore, the active subarray 3 is installed inside the heat dissipation housing 2, with multiple active subarrays arranged in a planar manner to form a complete array. Each active subarray is connected to the wave control circuit board 4 through standardized electrical connections to realize signal transmission and reception, power distribution and control functions, while forming a heat dissipation channel with the centrifugal fan board 5 to improve the system's heat dissipation efficiency.

[0031] Furthermore, the wave control circuit board 4 and the active subarray 3 are connected by signal lines and power lines and are uniformly set on the back or side edge of the array to realize the regulation and power management of the radio frequency signals of each active subarray, ensuring the consistency and controllability of the electromagnetic performance of the entire array.

[0032] Furthermore, the centrifugal fan plate 5 is installed on the back of the equivalent target of the subarray radar antenna, forming a heat dissipation channel with the active subarray 3 for active ventilation and heat dissipation, effectively controlling the system operating temperature and improving equipment stability.

[0033] Furthermore, each active subarray 3 contains several standard modules 6, which are connected by a plug-in structure. This facilitates maintenance and replacement, and ensures stable distribution of electrical signals and power, thereby achieving modularity and scalability of the array structure.

[0034] Furthermore, the standard module 6 integrates multiple antenna units 7 and transceiver chips 8, wherein the antenna units 7 are electrically connected to the transceiver chips 8 through printed circuits or feed lines to realize the transmission, reception and preliminary modulation / demodulation of radar signals.

[0035] Furthermore, the transceiver chip 8 not only achieves direct connection with the antenna unit 7, but also exchanges signals with the wave control circuit board 4 through a data line to complete the switching of radio frequency signal transmission and reception and intelligent control.

[0036] Furthermore, to facilitate structural health monitoring and physical damage archiving, a strain gauge reserved interface 9 is provided at a key location on the wave control circuit board 4. This interface can be plugged into and connected to an external strain testing system for real-time acquisition and archiving of structural strain signals, supporting accurate data acquisition for damage assessment experiments.

[0037] Specifically, the spatial structural parameters of the equivalent target are first systematically archived, including array type (e.g., linear array, area array), array element arrangement, number of subarray elements, module size, and layout. All of the above structural parameters are managed using both spreadsheets and a database to ensure data standardization and traceability for subsequent experiments.

[0038] Furthermore, the key functional parameters of the equivalent target (such as operating frequency band, main lobe gain, beamwidth, radiation pattern, side lobe suppression ratio, etc.) are standardized and archived, and all data are in a unified format and unit to improve the comparability of experimental objects of different models and batches.

[0039] Furthermore, a unique digital code is assigned to each type of equivalent target and its experimental object. Through the organic association of structural parameters and functional parameters, a complete experimental object description system is formed, realizing full-process digital traceability and batch management.

[0040] In addition, the equivalent target adopts a modular and open architecture, with standardized interfaces covering multiple levels of structure such as subarrays, antenna units and transceiver chips, which facilitates loading of various damage conditions and acquisition of multiple parameters, and supports damage assessment experiments in multiple scenarios and batches.

[0041] Finally, the equivalent target can be integrated with the host computer software and data acquisition platform to support array element-level physical state monitoring, damage loading, and automatic archiving of multiple parameters, ensuring a closed-loop data chain throughout the experiment and improving the scientific validity and engineering reproducibility of the evaluation results.

[0042] The above implementation methods ensure that the experimental subjects are representative of engineering and the methods are universal, laying a solid foundation for the systematic loading of damage conditions and the acquisition of high-resolution spatial parameters, and providing reliable data support for the subsequent establishment of a structural damage-functional damage mapping model.

[0043] S2: Set damage conditions and collect standardized physical parameters In this embodiment, step S2 is based on the subarray-level radar antenna equivalent target. To meet the needs of actual engineering applications or experiments, diverse damage scenarios are constructed, and multi-level, standardized archiving and management of structural damage parameters are carried out. Specifically, the following operational procedures are included: First, based on the pre-set test plan, various damage conditions are determined. Damage conditions can be obtained through physical loading (such as fragments, shock waves, and their coupling effects), or, when experimental conditions are limited, through digital programming to control the functional state of equivalent target array elements, achieving damage simulation at multiple scales and in multiple modes (such as edge-concentrated, center-concentrated, and random distribution). Each damage condition is assigned a unique condition number for easy data traceability and management. For visualization representations of damage conditions with different damage scales and spatial distributions, see [link to relevant documentation]. Figures 9 to 14 This demonstrates the typical spatial distribution of array elements from "zero damage" to "50% damage". Exemplary, each image is a heatmap of array current distribution, with the horizontal and vertical axes representing the row and column numbers of the array elements, respectively; color levels or shades reflect the current magnitude or failure state of each element (failed elements correspond to dark / low current); the spatial distribution of failed elements differs under different damage ratios, intuitively reflecting the randomness and cumulative effect of damage.

[0044] Furthermore, to demonstrate the impact of different spatial distribution types under the same damage ratio, this embodiment sets three spatial distributions for 20% array element failure: edge concentration, center concentration, and random distribution. Their array surface current distributions are as follows: Figure 15(Marginal concentration) Figure 16 (Centralized) Figure 17 As shown in the (random distribution) diagram, the above distribution pattern helps to deeply analyze the mechanism by which spatial damage characteristics affect the changes in radar antenna electrical performance, and provides diverse data support for subsequent high-precision mapping modeling.

[0045] Furthermore, structural damage parameters are collected and archived. To systematically characterize the spatial distribution features of frontal surface damage, this embodiment establishes a multi-layered spatial distribution parameter system, specifically including but not limited to: (1) Array element failure ratio: The ratio of the number of failed array elements to the total number of array elements, used to measure the overall damage level; (2) Spatial distribution coding: using a two-dimensional matrix The method represents the damage state of each element on the array surface. express Array element failure express The array element failure is normal. For example, the damage state of an 8×8 array can be represented as:

[0046] The matrix can be stored directly as a two-dimensional table, or it can be flattened into a one-dimensional encoded string (such as 00100000|01100000|……) and archived in batches in the database.

[0047] (3) Partition coding and statistics: Divide the array into several functional blocks (such as A, B, C, D), and record the number of failed array elements in each block to support regional damage analysis; (4) Maximum number of consecutive failure units: For each row, column or block, scan and count the maximum number of consecutive "1"s to reflect the damage concentration trend; (5) Spatial distribution entropy: The distribution entropy (H) is calculated based on the failure ratio of each block to quantify the dispersion or aggregation characteristics of the damage distribution.

[0048] (6) Subarray numbering: Damage parameters correspond one-to-one with subarray structures, enabling regional and local fine analysis.

[0049] Furthermore, all damage parameters are automatically and digitally archived in a structured database in a standardized format. Each parameter group records the operating condition number, timestamp, and unique identifier of the experimental object, supporting batch retrieval and automated pairing. The archiving process automatically performs normalization and outlier removal to ensure the comparability and engineering applicability of data from different batches. The relevant parameter system and archiving format are shown in Table 1.

[0050] In practical applications, if physical loading is not used, digital programming can be used to parameterize the failure states of array elements and systematically simulate different damage ratios and spatial distributions (see...). Figures 15 to 17 This method can significantly improve the efficiency and flexibility of data acquisition, support the acquisition of large-sample, high-throughput data, and provide a solid data foundation for complex damage-function mapping modeling and batch evaluation.

[0051] In this embodiment, all damage spatial distribution parameters can not only be archived using digital encoding, but also support visualization output (such as heat maps, matrix distribution maps, etc.), enhancing the ability to intuitively analyze the damage state. For experiments with physical loading conditions, high-precision measurement methods such as three-dimensional digital image correlation (DIC) and displacement sensors can be combined to collect multi-dimensional physical quantities such as structural deformation and fracture, further enriching the damage parameter system.

[0052] Through the aforementioned multi-level, standardized, automated archiving, and batch management process, this embodiment ensures high-precision recording of structural damage parameters and engineering traceability, significantly improving the scientific rigor, efficiency, and scalability of subsequent evaluation processes.

[0053] S3: Electrical performance testing and synchronous data acquisition In this embodiment, standardized field measurements of key electrical performance indicators are conducted for various damage conditions set in S2, and high-precision automatic pairing and batch archiving management of structural damage parameters and electrical performance data are achieved. The specific implementation process is as follows: First, a standardized procedure for experimental loading and electrical performance measurement was established. Based on the damage condition numbers archived in step S2, the corresponding structural damage states were sequentially applied to the equivalent target. Under each condition, key electrical performance parameters of the target, such as main lobe gain, beamwidth, sidelobe level, and beam pointing, were measured using a professional testing platform including a network analyzer, signal generator, and automated data acquisition terminal. All test parameters were acquired and recorded in real time using standardized units and data formats to ensure the experimental data had engineering comparability and standardized management capabilities. The testing system and data acquisition structure can be found in [reference needed]. Figure 18 ...

[0054] Furthermore, the system automates the pairing and archiving of damage parameters and electrical performance data. After each damage condition test is completed, the system automatically pairs the structural damage parameters under the current condition (such as the failure rate of array elements, spatial distribution coding, maximum number of consecutively failed elements, damage concentration, subarray number, etc.) with all synchronously measured electrical performance indicators. The pairing process is based on multi-dimensional information such as the condition number, the unique identifier of the experimental object, and the timestamp, and uses an automated database interface for batch archiving, achieving full-process traceability management of the data flow.

[0055] Furthermore, standardization and outlier handling processes are implemented. Before archiving, all structural damage parameters and electrical performance indicators undergo automatic data normalization, outlier removal, and format verification to ensure direct comparison and statistical analysis of data from different experimental batches and operating conditions. Testing equipment is regularly calibrated using standardized procedures, and the entire testing process follows standard operating protocols to improve the authenticity, completeness, and batch reproducibility of experimental results.

[0056] Furthermore, the process offers visualization and engineering support. After the paired dataset is imported into the database in real time, it supports data visualization in various ways, such as heatmaps, distribution maps, and parameter curves, effectively assisting subsequent structure-function mapping modeling (S4 step). Test data, damage parameters, and the pairing process are all digitally archived and batch-managed in the database, facilitating subsequent model training, engineering evaluation, and process optimization. For related data flows and pairing processes, please refer to [link to relevant documentation / documentation]. Figure 18 The flowchart shown is for testing and data archiving.

[0057] Through the standardized, automated, and batch-managed electrical performance testing and data acquisition process described above, this embodiment achieves high-precision synchronous acquisition and automatic pairing of structural damage parameters and key functional indicators, significantly improving the scientific nature, engineering adaptability, and data closed-loop capability of the evaluation process.

[0058] S4: Mapping Modeling of Physical Parameters and Functional Impairment In this embodiment, step S4 focuses on establishing a high-precision quantitative mapping model between the pre-archived structural damage parameters and key functional damage indicators, thereby improving the scientific rigor and engineering applicability of the evaluation scheme. Specifically, it includes the following process: First, the structural and functional parameters are batch archived and uniquely paired. Normalized and outlier-removed data from S2 and S3 stages are retrieved, and each set of structural damage parameters is automatically paired with its corresponding functional damage index using operating condition numbers, unique object identifiers, and timestamps. Each pairing includes structural parameters such as array element failure ratio, spatial distribution code, maximum number of consecutively failed elements, damage concentration, and subarray number, as well as functional indicators such as main lobe gain variation, beamwidth variation, sidelobe level variation, and beam pointing offset. The paired data is stored in batches in a database, supporting batch retrieval and analysis. See [link to database]. Figure 19 The modeling flowchart shown.

[0059] Furthermore, spatial distribution parameters are archived and coded at multiple levels. The spatial distribution state of each damage condition is represented by a two-dimensional matrix. The data is digitally represented and archived in a database. For example, the spatial distribution of surface damage under a certain working condition can be represented as:

[0060] in, =1 indicates that the corresponding array element is invalid. =0 indicates normal operation. This matrix can be stored as a two-dimensional array or flattened into a one-dimensional encoded string for batch archiving. In the database parameter table, the surface damage status is saved as a "spatial distribution code" field.

[0061] Meanwhile, the array is further divided into several functional blocks (such as A, B, C, and D), recording the number of failed array elements and the maximum number of consecutively failed units in each block, and calculating statistical characteristics such as distribution entropy H and damage concentration. Each parameter corresponds one-to-one with the operating condition number, realizing the standardization, batch processing, and automatic pairing of input parameters (see Table 1).

[0062] Furthermore, a standardized parametric modeling process is implemented. Using multidimensional structural damage parameters as input and key functional impairment indicators as output, algorithms such as multiple linear regression, support vector machine (SVM), random forest (RF), and artificial neural network (ANN) are selected to automatically perform parametric modeling. The specific operations are as follows: (1) Before modeling, the system automatically performs feature screening (such as principal component analysis and stepwise regression) and sample normalization to improve modeling accuracy and stability.

[0063] (2) During the modeling process, k-fold cross-validation, residual analysis and other methods are used to systematically evaluate the generalization ability and prediction accuracy of the model, and the best performing model is selected and retained.

[0064] (3) All model parameters, modeling results and evaluation indicators are archived in batches in a standardized format to realize model iteration management, historical comparison and automatic traceability.

[0065] As an optional implementation, this embodiment provides an improved artificial neural network, including an input layer, at least two levels of gated residual feature layers, and an output layer. The input layer receives a set of impairment parameters, the output layer regresses the function change vector, and the gated residual feature layers perform adaptive improvements based on feature sensitivity guidance. Forward propagation satisfies: ; in, For the first The input feature vector of the layer, For the first The output feature vector of the layer, s, is a sensitivity vector obtained by normalization based on the feature sensitivity analysis results of claim 8. For the first Layer gate vector, For the first Layer non-negative gated strength scalar For the Sigmoid function, For element-wise activation functions, and For the first The layer weight matrix is ​​used to generate nonlinear branches and linear residual branches respectively. For element-wise multiplication, This is the vector of functional changes output by the network. This is the output layer weight matrix. This is the vector of changes in target function obtained by comparing it with the baseline parameter set of the experimental subjects. The training loss is calculated using the mean squared error.

[0066] In this embodiment, based on the unique pairing relationship between the working condition number, the experimental object number, and the timestamp, the damage parameter set of each sample is selected as input, and the functional change is used as the supervised output to construct a training sample set and an independent validation set. The damage parameter set includes at least the array element failure ratio, the maximum number of consecutively failed units, the distribution entropy, the damage concentration, and the subarray number. These parameters are first linearly normalized according to the upper and lower limits of the fields, and outliers are pruned by percentile before being stored. The functional change is calculated according to the baseline parameter set of the experimental object, resulting in four components: main lobe gain change, beamwidth change, sidelobe level change, and beam pointing offset. These components are standardized to form a target vector that corresponds one-to-one with the input. Cross-validation is used during the training phase to ensure that the training, validation, and test sets of each fold do not overlap. After each fold of training is completed, residual statistics are performed, and model snapshots, metrics, and metadata are saved for subsequent comparison and backtracking.

[0067] The artificial neural network consists of an input layer, at least two levels of gated residual feature layers, and an output layer. The input layer dimension equals the number of fields in the damage parameter set, and the output layer dimension is four, corresponding to the four functional variation components. Each level of gated residual feature layer contains one nonlinear main branch and one linear residual branch: the main branch performs a fully connected mapping and activation on the features of the previous level; the linear residual branch performs a linear mapping on the same input and adds it to the result of the main branch. A gate vector is set within the layer to modulate the main branch output element-wise. The gate vector value ranges from zero to one, and is obtained by applying a gate strength to a normalized feature sensitivity vector and then compressing it, achieving adaptive emphasis on key damage features and suppression of non-key features. The multi-level gated residual feature layers are concatenated and then connected to a fully connected regression head to directly output the four-dimensional functional variation prediction.

[0068] Training targets mean squared error, employing a gradient-based optimization strategy combined with learning rate annealing. Each training fold is repeated several times under a fixed random seed. Training is terminated early if the validation set metrics do not improve within a certain number of rounds to prevent overfitting. After training, the relationship between residual distribution, prediction interval, and error variation with damage intensity is statistically analyzed. If systematic biases are found, the gating strength range, hidden layer width, and number of layers are adjusted slightly while keeping the input fields unchanged, and cross-validation is performed again to determine the updated structure and parameters. When new working conditions or samples are added to the database, incremental training is performed based on the existing weights, and the version number and metric records are updated to form a traceable, evolving model library.

[0069] The exemplary layer, "Gated Residual Feature Layer," refers to a network layer that simultaneously contains a nonlinear main branch and a linear residual branch. The output of the main branch is scaled element-wise by the gate vector and then added to the residual branch to maintain gradient stability while emphasizing key features. The "Sensitivity Vector" refers to the weight sequence obtained and normalized based on feature sensitivity analysis, used to characterize the relative strength of each damage feature on the target. The "Gating Strength" refers to the scaling factor of the sensitivity vector, used to adjust the strength of the gating effect. The "Functional Variation Vector" refers to the ordered set of four functional variation components calculated relative to the baseline parameter set of the experimental subject. The "Training Sample Set" refers to the paired data set consisting of the damage parameter set and the corresponding functional variation, archived according to a unified field, unit, and code.

[0070] Furthermore, the input dimension of the artificial neural network is derived from the number of fields in the damage parameter set, and the output dimension is fixed at 4; the length of the gating vector is consistent with the input dimension, and the value range is 0-1; the gating strength is a non-negative real number, initially set to 1, and can be automatically updated during training; the hidden layer width is an integer between 128 and 256, and the number of layers is ≥2; the activation function is a rectified linear unit; the objective function is mean squared error; the cross-validation fold is 5; the numerical range of the damage parameter set is set with upper and lower limits based on the percentile statistics of historical samples, and out-of-bounds data is truncated according to the boundaries; the maximum number of training epochs is 200, the batch size is 64, and the early termination patience value is 15; the model archive includes weights, configuration, training and validation metrics, residual statistics, and data version number. Sample size ≥ 1200; input fields include at least 5 items: array element failure rate, maximum number of consecutively failed units, distribution entropy, damage concentration, and subarray number; initial gating strength of 1, which typically converges to 0.5-1 after training; the units of the four components of functional variation are dB, °, dB, and ° respectively; single inference latency not exceeding 10ms on a single sample; model version and data version correspond 1:1, and archive records include version number, timestamp, and mean and standard deviation of cross-validation metrics.

[0071] Furthermore, this embodiment supports data archiving and modeling for various damage conditions and different equivalent target structures, featuring adaptive optimization and open expansion. For new conditions or complex spatial distributions, the system automatically imports new sample data through an interface, incorporating it into dynamic model training and adaptive optimization to achieve regular updates and upgrades of model parameters. For different types of damage, new spatial distribution patterns, etc., the system can automatically expand the modeling algorithm library to improve the engineering adaptability and innovation capabilities of the evaluation scheme.

[0072] Through the above-mentioned multi-layer spatial parameter encoding, unique pairing, standardized modeling and automatic optimization closed-loop process, this embodiment achieves high-resolution, batch-capable, automated quantitative mapping between structural damage and functional damage, effectively breaking through the bottlenecks of existing technologies in batch archiving, spatial feature expression and intelligent modeling, and providing strong engineering support for the scientific assessment of radar antenna damage effects.

[0073] The modeling and pairing process can be found in [reference needed]. Figure 19 As shown in Table 1, spatial distribution parameters, pairing examples, and key archiving fields are detailed below.

[0074] Table 1

[0075] S5: Damage Assessment and Severity Determination In this embodiment, step S5, based on the structure-function mapping model established in S4, performs quantitative evaluation and multi-level automatic determination of the radar antenna functional status under different damage conditions. The specific process is as follows: First, import the S4 mapping model and archived damage parameter data. For each experimental condition, input the structural damage parameters (such as the failure rate of array elements, spatial distribution coding, maximum number of consecutively failed elements, damage concentration, etc.) into the mapping model, and automatically output the predicted or measured results of key functional indicators (such as main lobe gain change, beamwidth change, sidelobe level change, etc.).

[0076] Furthermore, based on the established multi-dimensional parameter system, a quantitative analysis of antenna damage status is conducted by comprehensively considering various structural and functional damage indicators. The evaluation grading criteria can preferentially reference existing industry standards, engineering experience, and historical cases. For example, by combining key criteria such as "main lobe gain reduction," "beamwidth increase," and "damage ratio," multiple threshold levels can be set to achieve segmented determination of antenna functional status. For different application scenarios, the criterion system can be flexibly adjusted according to actual needs to adapt to diverse engineering requirements.

[0077] Furthermore, an automated data processing workflow is adopted to digitally archive all judgment results and related process data. Each evaluation process records input parameters, model judgment results, grading criteria, and historical data comparisons, ensuring the traceability and continuous improvement capability of the evaluation system.

[0078] Through the above steps, this embodiment realizes the automated damage classification assessment of the structural damage and functional damage mapping model in actual engineering, improves the objectivity, scientificity and engineering applicability of the assessment process, and provides strong support for the digital management of radar antenna damage status and subsequent decision-making.

[0079] S6 Model Feedback and Continuous Optimization In this embodiment, step S6 designs a full-process management mechanism for continuous feedback and adaptive optimization based on the established structure-function mapping model. The specific operation is as follows: First, in the practical application of the radar antenna damage assessment system, the structural damage parameters and corresponding functional degradation data acquired in each round are automatically archived. All new data are digitally stored in a unified format and form a traceable database with historical assessment samples.

[0080] Furthermore, the system periodically analyzes and evaluates the model performance of the data in the archived database. Methods such as model residual statistics and prediction error analysis are used to quantify the model's prediction accuracy under new operating conditions and complex samples. For data with significant prediction deviations, the system automatically triggers model retraining or parameter optimization processes to promptly correct model coefficients or update core feature parameters, ensuring that the structure-function mapping relationship continues to evolve with data accumulation.

[0081] Furthermore, the system achieves data interoperability and closed-loop management with the testing and evaluation platform. New damage-functional data collected by the testing platform can be directly transmitted to the model optimization module, and the optimized model parameters and criteria are transmitted back to the testing platform in real time, realizing automatic updates and result synchronization throughout the entire process, thereby improving the system's engineering adaptability and intelligence level.

[0082] Through the above process, this embodiment realizes the dynamic optimization and self-learning capability of the structure-function mapping model, ensuring the long-term effectiveness and accuracy of the evaluation method, and providing solid support for the scientific analysis and engineering management of the damage effects of complex radar antennas.

[0083] Corresponding to the above method, this embodiment also provides a radar antenna structural damage and functional damage mapping evaluation system based on subarray-level equivalent targets, characterized in that it includes: The experimental object and baseline archiving unit is used to select a subarray-level equivalent target that is consistent with the array structure functional parameters of the target radar antenna, and to standardize and archive the array structure functional parameters of the equivalent target to form an experimental object baseline parameter set and object number. The damage condition loading and damage parameter archiving unit is used to load damage determined by type, proportion, and spatial distribution according to preset conditions, and to collect and archive the damage parameter set. The electrical performance testing and functional change pairing and archiving unit is used to test the main lobe gain, beamwidth, sidelobe level, and beam pointing, calculate the functional change relative to the baseline parameter set of the experimental object, and pair and archive it with the damage parameter set according to the operating condition number, object number, and timestamp. The spatial distribution coding and standardized input generation unit is used to encode the array surface damage by dividing it into two-dimensional matrices and blocks, and pair it with the working condition number and subarray number to form a standardized input parameter set. The quantitative mapping modeling and adaptive optimization unit is used to establish a quantitative mapping model based on the set of damage parameters and functional changes. It selects at least one method from multiple linear regression, support vector machine, random forest, and artificial neural network, optimizes it through k-fold cross-validation and residual analysis, and adaptively updates it.

[0084] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0085] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for evaluating the mapping between structural and functional damage of radar antennas based on subarray-level equivalent targets, characterized in that, include: Select a subarray-level equivalent target with the same array structure and functional parameters as the target radar antenna, and standardize and archive the array structure and functional parameters of the equivalent target to form a baseline parameter set and object number for the experimental object. Damage determined by type, proportion, and spatial distribution is applied according to preset working conditions, and damage parameter sets are collected and archived. Test the main lobe gain, beamwidth, sidelobe level, and beam pointing; calculate the functional changes relative to the baseline parameter set of the experimental object; and archive the results by pairing them with the damage parameter set according to the operating condition number, object number, and timestamp. The array damage is encoded using a two-dimensional matrix and block division, and paired with the working condition number and subarray number to form a standardized set of input parameters. A quantitative mapping model is established using the set of damage parameters and changes in function. At least one method is selected from multiple linear regression, support vector machine, random forest, and artificial neural network. The model is optimized by k-fold cross-validation and residual analysis, and then updated adaptively.

2. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 1, characterized in that, The damage parameter set includes the array element failure ratio, spatial distribution coding, maximum number of consecutively failed elements, distribution entropy, and damage concentration.

3. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 1, characterized in that, A subarray-level equivalent target with array structure functional parameters consistent with the target radar antenna is selected, and the array structure functional parameters of the equivalent target are standardized and archived to form a baseline parameter set and object number for the experimental object, including: The experimental object was determined to be the equivalent target of a subarray-level radar antenna, and the array structure functional parameters were collected. The array structure functional parameters include: array type, array element arrangement, number of subarray elements, antenna key dimensions, module size, layout, operating frequency band, main lobe gain, beamwidth, radiation pattern, and sidelobe suppression ratio. The array structure functional parameters are stored in a database or spreadsheet system using a standardized data structure, and dual management of spreadsheets and databases is implemented. The array structure functional parameters are standardized and archived using unified fields, units, and codes. A unique numerical code is assigned to each experimental object, and this code is associated with the functional parameters of the array structure to form a baseline parameter set and an experimental object number.

4. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 1, characterized in that, Damage is applied according to preset working conditions, determined by type, proportion, and spatial distribution. Damage parameter sets are collected and archived, including: Establish a set of damage conditions, and construct conditions with different damage ratios and spatial distributions by setting the failure state of array elements through physical loading or digital programming. Assign a condition number, record a timestamp and experimental object number to each condition. Collect a set of damage parameters; the set of damage parameters includes the failure rate of array elements, the maximum number of consecutively failed elements, the distribution entropy, the damage concentration, and the subarray number. Before being stored, the damage parameter set is normalized and outlier is removed. The damage parameter set is uniquely paired and archived with the working condition number, experimental object number, and timestamp using unified fields, units, and codes.

5. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 4, characterized in that, After uniquely pairing and archiving the damage parameter set with the working condition number, experimental object number, and timestamp using unified fields, units, and codes, the following is also included: Visualize and save the archived damage parameter set; When preset conditions are met, the physical quantities of structural deformation and fracture are supplemented by digital image sensors and displacement sensors, and archived together with the damage parameter set.

6. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 1, characterized in that, Test the main lobe gain, beamwidth, sidelobe level, and beam pointing; calculate the functional changes relative to the baseline parameter set of the experimental object; and archive these changes with the damage parameter set according to the operating condition number, object number, and timestamp, including: The corresponding structural damage states are loaded sequentially according to the operating condition number, and the electrical performance test process is initiated. The four electrical performance parameters of main lobe gain, beamwidth, sidelobe level, and beam pointing were measured using a network analyzer, signal generator, and automated acquisition terminal, and were collected and recorded in real time according to a unified standard unit and data format. Based on the baseline parameter set of the experimental object, the functional changes of the electrical performance parameters relative to the baseline parameter set of the experimental object are calculated; the functional changes include: main lobe gain change, beamwidth change, sidelobe level change, and beam pointing offset; The functional changes are matched one-to-one with the damage parameter set under the corresponding working conditions according to the working condition number, experimental object number, and timestamp, and then archived in batches through an automated database interface to achieve full traceability. Before being stored, the damage parameter set and the electrical performance parameters are normalized, outlier removal and format verification are performed, and the equipment calibration and automatic pairing relationship verification are completed according to the standard operating procedure to obtain the paired dataset. After importing the paired datasets into the database, visualization is supported in the form of heatmaps, distribution maps, and parametric curves.

7. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 1, characterized in that, The array damage is encoded using a two-dimensional matrix and block division, paired with the working condition number and subarray number to form a standardized set of input parameters, including: A two-dimensional grid is established based on the actual arrangement of array elements to generate an array surface damage matrix; in the array surface damage matrix, "1" represents array element failure and "0" represents normal operation. The array damage matrix is ​​flattened into a one-dimensional encoded string in row order and stored as a spatial distribution encoded field. The array is divided into multiple functional blocks according to a preset area, and the number of failed array elements in each block is recorded. A standardized set of input parameters is constructed using the array damage matrix, one-dimensional spatial distribution code, number of failures in each block, and subarray number. These parameters are uniquely paired and archived with the working condition number, subarray number, experimental object number, and timestamp. Unified fields, units, and codes are used for database management.

8. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 1, characterized in that, A quantitative mapping model is established using a set of damage parameters and changes in function. At least one method from multiple linear regression, support vector machine, random forest, and artificial neural network is selected. The model is optimized through k-fold cross-validation and residual analysis, and then adaptively updated. This includes: A training sample set is constructed based on data that uniquely matches the working condition number, the experimental object number, and the timestamp; the training sample set consists of a set of damage parameters and corresponding functional changes and is archived in a standardized format. Perform sample normalization and feature filtering on the training sample set to improve modeling stability; At least one of the following methods—multivariate linear regression, support vector machine, random forest, and artificial neural network—is used to model the quantitative mapping relationship between structural damage parameters and functional changes. The generalization ability and prediction accuracy were evaluated by k-fold cross-validation and residual analysis, and the model structure and parameters obtained by modeling were optimized. Perform feature sensitivity analysis on key impairment features to identify input features that significantly affect the amount of functional change; Model parameters, evaluation indicators and modeling results are archived in batches using unified fields, units and codes to achieve version management, historical comparison and traceability; Incremental training and adaptive updates are performed when new operating conditions or samples are imported, and model parameters are optimized periodically to keep the model evolving with the data.

9. The radar antenna structural damage and functional damage mapping assessment method based on subarray-level equivalent targets according to claim 1, characterized in that, The artificial neural network includes an input layer, at least two levels of gated residual feature layers, and an output layer. The input layer receives a set of impairment parameters, and the output layer regresses a vector of functional changes. The gated residual feature layers perform adaptive improvements based on feature sensitivity guidance, and the forward propagation satisfies: ； in, For the first The input feature vector of the layer, For the first The output feature vector of the layer, s, is a sensitivity vector obtained by normalization based on the feature sensitivity analysis results of claim 8. For the first Layer gate vector, For the first Layer non-negative gated strength scalar For the Sigmoid function, For element-wise activation functions, and For the first The layer weight matrix is ​​used to generate nonlinear branches and linear residual branches respectively. For element-wise multiplication, This is the vector of functional changes output by the network. This is the output layer weight matrix. This is the vector of changes in target function obtained by comparing it with the baseline parameter set of the experimental subjects. The training loss is calculated using the mean squared error.

10. A radar antenna structural damage and functional damage mapping and evaluation system based on a subarray-level equivalent target, characterized in that, include: The experimental object and baseline archiving unit is used to select a subarray-level equivalent target that is consistent with the array structure functional parameters of the target radar antenna, and to standardize and archive the array structure functional parameters of the equivalent target to form an experimental object baseline parameter set and object number. The damage condition loading and damage parameter archiving unit is used to load damage determined by type, proportion, and spatial distribution according to preset conditions, and to collect and archive the damage parameter set. The electrical performance testing and functional change pairing and archiving unit is used to test the main lobe gain, beamwidth, sidelobe level, and beam pointing, calculate the functional change relative to the baseline parameter set of the experimental object, and pair and archive it with the damage parameter set according to the operating condition number, object number, and timestamp. The spatial distribution coding and standardized input generation unit is used to encode the array surface damage by dividing it into two-dimensional matrices and blocks, and pair it with the working condition number and subarray number to form a standardized input parameter set. The quantitative mapping modeling and adaptive optimization unit is used to establish a quantitative mapping model based on the set of damage parameters and functional changes. It selects at least one method from multiple linear regression, support vector machine, random forest, and artificial neural network, optimizes it through k-fold cross-validation and residual analysis, and adaptively updates it.

Images