A digital twin model construction method for phased array antenna pattern reconstruction

Abstract

The application discloses a digital twin model construction method for phased array antenna pattern reconstruction, and relates to the technical field of phased array antenna testing, which comprises the following steps: full basic bit testing is performed on each T / R component of a phased array antenna to obtain actual phase shift values, actual attenuation values, phase shift parasitic attenuation and attenuation parasitic phase shift data, a T / R component non-ideal characteristic actual measurement database is constructed, a unit pattern database is obtained based on a coded aperture matrix excitation acquisition unit, and a digital twin model of the phased array antenna is constructed based on the T / R component non-ideal characteristic actual measurement database and the unit pattern database, so that the digital twin model is highly consistent with the actual performance of a physical antenna entity, high-precision construction and dynamic evolution of the digital twin model of the phased array antenna are realized, and the modeling efficiency and prediction accuracy are significantly improved. The model constructed by the application is excellent in beamforming prediction, fault diagnosis and positioning, and beam pre-distortion compensation, and effectively improves the prediction accuracy and efficiency.

Description

Technical Field

[0001] This invention relates to the field of phased array antenna testing technology, and more specifically to a method for constructing a digital twin model for reconstructing the radiation pattern of a phased array antenna. Background Technology

[0002] Phased array antennas possess outstanding advantages such as beam agility and flexible spatial scanning, making them core components in fields such as radar and wireless communication. The radiation performance of a phased array antenna is mainly determined by the coordinated operation of a large number of T / R components (including core components such as phase shifters and attenuators). Therefore, accurate and efficient antenna performance measurement and evaluation are the core prerequisites for ensuring the working accuracy of phased array antennas.

[0003] Currently, the mainstream measurement methods for phased array antennas include two categories: far-field measurement and near-field measurement. Near-field measurement technology calculates the far-field radiation pattern of the antenna through near-far-field transformation algorithms. Although it significantly improves testing efficiency compared to traditional far-field testing methods, in large-scale array testing scenarios, traditional uniform sampling technology based on the Nyquist sampling theorem still faces problems such as a large number of sampling points, a large amount of data collected, and a high computational load. Ultimately, this results in low overall testing efficiency and a long testing cycle, making it difficult to meet the rapid testing requirements of large-scale phased array antennas.

[0004] To address the low efficiency of traditional sampling tests, the industry has successively proposed digital models based on compressed sensing and coded aperture. Compressed sensing models leverage the sparsity of the spatial and angular domain distributions of phased array antennas, recovering the complete antenna element excitation distribution with far fewer measurements than the Nyquist sampling theorem through compressed sensing algorithms. Coding aperture models, on the other hand, introduce the concept of a coded mask into phased array excitation, reconstructing the excitation response of each element by acquiring and decoding probe response signals under different coding modes. However, existing compressed sensing and coded aperture models can only reconstruct the radiation pattern of radiating elements. In multi-wavelength radiation pattern calculations, they only substitute ideal quantization parameters of the T / R components for calculations, without considering the non-ideal characteristics of the T / R components in actual engineering conditions.

[0005] In practical engineering applications, the T / R components of phased array antennas generally exhibit inherent non-ideal error characteristics: phase shifters generate unexpected amplitude loss during phase adjustment, resulting in additional attenuation due to phase shift; attenuators generate additional phase shift during amplitude adjustment, resulting in additional phase shift due to attenuation. Therefore, there are varying degrees of deviation between the actual phase shift and attenuation values ​​of each T / R component and their theoretical quantized values. If the measurement model ignores these non-ideal characteristics, the calculated results for key performance indicators such as antenna beam pointing accuracy, sidelobe level, and radiation gain will deviate significantly, drastically reducing the accuracy and reliability of phased array antenna pattern reconstruction.

[0006] In summary, there is a lack of a reconstruction model that can combine the actual working characteristics of T / R components and predict the radiation characteristics of array antennas in real time with high accuracy. Summary of the Invention

[0007] The purpose of this invention is to propose a digital twin model construction method for phased array antenna pattern reconstruction. By constructing a digital twin model that can deeply integrate the non-ideal measured data of T / R components with the coupling data of the elements, the model can truly reflect the radiation characteristics of the physical array surface, thus solving the problems of low reconstruction accuracy and low measurement efficiency caused by not considering the actual phase shift / attenuation values ​​of the actual T / R components.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for constructing a digital twin model for pattern reconstruction of a phased array antenna includes: performing full-basic-bit testing using the physical entity of the target phased array antenna to obtain a measured database of non-ideal characteristics of the T / R components. A measured data model of the T / R component is constructed; near / mid field measurements are performed based on the coded aperture excitation matrix of the array surface to obtain near / mid field response datasets, and then the element coupling response is calculated to construct an element pattern database of the antenna array; a parameterized model of the phased array antenna is constructed, and the measured database of the non-ideal characteristics of the T / R component is used. The unit pattern database is integrated into the parameterized model to generate a digital twin model for reconstructing the phased array antenna pattern.

[0010] Furthermore, the measured database of non-ideal characteristics of the T / R component. This includes: actual phase shift value lookup table, actual attenuation value lookup table, phase shift parasitic attenuation data, and attenuation parasitic phase shift data.

[0011] The full basic bit test includes: for each T / R component in the phased array antenna, traversing all basic bit commands of the attenuator, measuring the actual amplitude response and actual phase shift of the T / R component under each basic bit command, and establishing an actual attenuation value lookup table and attenuation parasitic phase shift data; traversing all basic bit commands of the phase shifter, measuring the actual phase response and actual amplitude shift of the T / R component under each basic bit command, and establishing an actual phase shift value lookup table and phase shift parasitic attenuation data.

[0012] Furthermore, the phased array antenna is equipped with a near / field measurement system, including a probe; the coded aperture excitation matrix is ​​constructed according to a coding algorithm, and each row in the coded aperture excitation matrix corresponds to a type of coded excitation.

[0013] The near / field measurement performed by the array-based coded aperture excitation matrix, and the subsequent calculation of the element coupling response, includes: sequentially loading the coded excitation from the coded aperture excitation matrix onto all T / R components of the array; controlling the probe to acquire near / field complex signals point by point on the scanning plane to obtain a measurement dataset containing near / field complex signals from all scanning points; and processing the measurement dataset according to the decoding algorithm corresponding to the encoding algorithm. Decoding operations are performed to obtain the element coupling response of each element, resulting in the far-field radiation pattern of the corresponding element under all scan points, and an element radiation pattern database is constructed.

[0014] Furthermore, the parameterized model of the phased array antenna includes: a radiating element model, an array layout, a feeding network, and an interface for interfacing with actual TR components.

[0015] Furthermore, the measured database of the non-ideal characteristics of the T / R component will be used. Integrating the unit pattern database into the parameterized model includes: establishing a measured T / R component model in the digital twin model, and configuring the measured T / R component model to take beam control commands as input and the actual excitation complex coefficients of each antenna element as output. Its calculation process includes:

[0016] The theoretical phase shift and theoretical amplitude weighting are calculated based on the target beam direction and quantized into phase shifter command values ​​and attenuator theoretical command values;

[0017] Based on the measured database of non-ideal characteristics of T / R components The actual phase shift value and parasitic phase shift attenuation data are queried according to the phase shifter command value, and the actual attenuation value and attenuation parasitic phase shift data are queried according to the attenuator theoretical command value, thereby obtaining the actual excitation complex coefficients; a pattern reconstruction module is established in the digital twin model, and the pattern reconstruction module is configured to synthesize a far-field pattern based on the element pattern database of the phased array antenna.

[0018] Furthermore, the digital twin model is also configured with a model verification module for performing a rapid verification process to verify the reconstruction accuracy of the digital twin model.

[0019] The rapid verification process includes: taking the measured radiation pattern and the predicted radiation pattern under a preset typical beam state as input, extracting key indicators from the measured radiation pattern and the predicted radiation pattern respectively; if the error of each key indicator is within a preset range, the model verification is passed; if the error of any key indicator exceeds the preset range, the dynamic update process of the digital twin model is triggered; wherein, the measured radiation pattern is obtained by planar near-field scanning or far-field testing under the current state of the phased array antenna, and the predicted radiation pattern is generated by the current digital twin model.

[0020] Furthermore, the digital twin model is also configured with an error corrector to obtain the correction amount of the amplitude and phase of the unit pattern in the unit pattern database based on the errors of each key indicator.

[0021] The error corrector is built on a lightweight neural network. Its input is the error of each key indicator, and its output is the correction amount of the amplitude and phase of the unit radiation pattern in the unit radiation pattern database. Its configuration is as follows: if the error of each key indicator is within the preset range, the unit radiation pattern in the unit radiation pattern database is fine-tuned by the correction amount output by the error corrector after rapid verification. If the cumulative error of a key indicator exceeds the preset threshold, the dynamic update process of the digital twin model is executed.

[0022] The dynamic update process of the digital twin model includes: re-performing the complete coded measurement, updating the cell pattern database, and retraining the error corrector.

[0023] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0024] (1) This invention is the first to deeply integrate the array unit radiation pattern database based on the physical entity of the phased array antenna with the measured data of the non-ideal characteristics of the T / R component into the digital twin modeling process. It completely abandons the ideal device assumption and ideal unit radiation assumption commonly used in traditional array simulation, eliminates the deviation between theoretical simulation and physical working conditions from the root of modeling, and makes the beam prediction and performance deduction results of the digital twin model highly consistent with the actual performance of the physical antenna entity. It effectively breaks through the traditional simulation error bottleneck and realizes high-precision mapping between the model and the physical entity.

[0025] (2) This invention establishes a closed-loop mechanism for dynamic model updates, consisting of "actual measurement comparison - error correction - parameter iteration," which can track in real time the performance changes of physical antennas caused by device aging, environmental interference, and parameter drift during long-term operation. By continuously iterating and optimizing the parameters of the digital twin model through the error corrector, the digital twin model can continuously evolve along with the working state of the physical entity, achieving accurate mapping throughout its entire life cycle. This effectively solves the problems of accuracy decay and poor adaptability of traditional static simulation models over long-term use, and significantly improves the stability and availability of the system during long-term operation.

[0026] (3) The digital twin model constructed by the present invention is not only a high-precision simulator, but also equipped with a fault database and a beam predistortion compensation module. It has a powerful tool platform that can be used for performance prediction, fault diagnosis, beam compensation optimization, and realize the capability leap from "design verification" to "state management". Attached Figure Description

[0027] Figure 1 This is a flowchart of the digital twin model construction process provided in Embodiment 1 of the present invention;

[0028] Figure 2 This is a schematic diagram of the measured data identification of the T / R component provided in Embodiment 2 of the present invention. Detailed Implementation

[0029] To better understand this invention, the following description, in conjunction with the accompanying drawings of the embodiments of this invention, will provide further details, but these descriptions are not intended to limit the scope of the invention. Various modifications and improvements made to the technical solutions of this invention by those skilled in the art without departing from the inventive concept should fall within the protection scope of this invention.

[0030] refer to Figure 1 This embodiment takes an X-band phased array antenna containing 256 elements as an example to disclose a method for constructing a digital twin model for reconstructing the radiation pattern of a phased array antenna, including the following steps:

[0031] Step S100: For the physical entity of the phased array antenna to be modeled, perform full-basic-bit testing on each of the 256 T / R components in its array surface to obtain a measured database of non-ideal characteristics of the T / R components. , and construct a measured data model of the T / R component.

[0032] Among them, the measured database of non-ideal characteristics of T / R components At least including:

[0033] Actual phase shift value lookup table : at the theoretical command value of the phase shifter The actual phase shift amount;

[0034] Actual attenuation value lookup table At the attenuator's theoretical command value The actual attenuation value;

[0035] Phase-shifting parasitic attenuation data : Depending on the actual phase shift value The varying additional attenuation value;

[0036] Attenuation parasitic phase shift data : with actual decay value The changing additional phase shift value.

[0037] This embodiment uses an automated T / R component testing platform, which consists of a vector network analyzer, a multi-channel switch matrix, a temperature control box, automated test fixtures, and host computer control software.

[0038] Before testing, 256 T / R modules were sequentially mounted on a dedicated test fixture. Each T / R module contained a 6-bit phase shifter (theoretical phase shift step 5.625°, range 0~360°) and a 6-bit attenuator (theoretical attenuation step 0.5dB, range 0~31.5dB). Scattering parameters of each T / R module were recorded by stepping samples across the full code range of the phase shifter and attenuator.

[0039] The test environment temperature was controlled at 25℃±0.5℃ to eliminate the influence of temperature on the performance of TR components.

[0040] The full basic bit test includes: actual phase shift value and phase-shift parasitic attenuation value test, as well as actual attenuation value and attenuation parasitic phase shift value test.

[0041] Specifically, the testing process for the actual phase shift value and the parasitic phase shift attenuation value is as follows:

[0042] For each T / R component, the attenuator is set to the reference state, i.e., the minimum attenuation is 0dB. The phase shifter is iterated through one reference state and six basic bit instructions. Corresponding theoretical phase shift value At each command value, the S21 parameters of the T / R component are measured using a vector network analyzer to extract the actual phase response. and phase-shifting parasitic attenuation value .

[0043] The test results of the basic bits are superimposed and combined to obtain the actual phase shift value and the phase shift parasitic attenuation value under the corresponding arbitrary phase shift command, and then organized into a lookup table. and And stored in the T / R component non-ideal characteristic measured database. It is used to construct the measured data model of T / R components.

[0044] The testing process for the actual attenuation value and the attenuation parasitic phase shift value is as follows:

[0045] For each T / R component, the phase shifter is set to the reference state, i.e., 0°, and the attenuator's one ground state and six basic bit instructions are traversed. Corresponding theoretical attenuation value At each command value, the actual attenuation response was measured. and attenuation parasitic phase shift value .

[0046] The actual attenuation test results of the basic bits are superimposed and combined to obtain the actual attenuation value and the attenuation parasitic phase shift value of the corresponding arbitrary attenuation value, and then organized into a lookup table. and Stored in the T / R component non-ideal characteristic measured database It is used to construct the measured data model of T / R components.

[0047] T / R component non-ideal characteristics measured database Subsequently, it serves as the core input to the digital twin model, replacing the traditional ideal TR component model. Its data identification diagram is as follows: Figure 2 As shown.

[0048] The full-base-level testing revealed that the characteristics of different T / R components exhibit a certain degree of dispersion, confirming the necessity of measuring each component individually in this embodiment.

[0049] Step S200: Obtain the element pattern database of the antenna array through coding measurement.

[0050] It should be noted that the near / mid field measurement system of the phased array antenna physical entity is used for coded measurement to obtain the element radiation pattern database of the phased array antenna. The definition standard of the near / mid field measurement area is: the measurement distance meets the far field test conditions relative to the element antenna inside the array, but it is still in the radiation near field region for the large-aperture array antenna. The near / mid field has both the far field measurement of the element antenna and the near field scanning characteristics of the array.

[0051] The planar near / field measurement system mainly includes: an open waveguide probe, a two-dimensional scanning frame, a vector network analyzer, a controller, and data acquisition software.

[0052] Specifically, the open-aperture waveguide probe is used for non-contact reception of near-field / field electromagnetic wave signals radiated by the phased array antenna under different coded excitation states; the two-dimensional scanning frame is used to support the open-aperture waveguide probe to achieve controllable and precise scanning in a two-dimensional plane; the vector network analyzer is used to measure the near-field / field electromagnetic wave signals picked up by the open-aperture waveguide probe and output the corresponding complex signal data for each excitation state; the controller is used to control the motion trajectory, sampling timing and sampling points of the two-dimensional scanning frame, synchronously trigger the data acquisition action of the vector network analyzer, and coordinate the two-dimensional scanning frame, the vector network analyzer and the antenna T / R assembly to work together.

[0053] In a specific implementation, the planar near / midfield measurement system is either a single-probe measurement system or a multi-probe measurement system.

[0054] Before measurement, in a dark room, the antenna array to be measured is vertically mounted in front of the scanning frame, with the antenna array parallel to the probe aperture.

[0055] In one embodiment, the coding measurement specifically includes:

[0056] A set of beam state codes based on a coding algorithm is defined, and a coded excitation matrix for a 256-element array is generated according to the coding algorithm to control the array to generate 256 coded beams. Each row of the coded aperture excitation matrix corresponds to a type of coded excitation.

[0057] The encoding stimulus matrix here can be either a random matrix or a deterministic matrix. Random matrices include Gaussian random matrices, Bernoulli random matrices, etc., while deterministic matrices include Hadamard matrices, uniform redundancy matrices, non-redundancy matrices, etc.

[0058] In a specific implementation, the Hadamard matrix is ​​an excitation matrix composed of ±1 elements of order 256, the uniform redundancy matrix is ​​an excitation matrix of order 256 based on quadratic remainders, and the non-redundancy matrix is ​​an excitation matrix of order 256 with the lowest possible autocorrelation function outside the central main peak.

[0059] All T / R components on the array are set to the reference state, the phase shifter is 0°, and the attenuator is at its minimum attenuation, i.e., 0dB. The scanning frame traverses all scanning points in the moving order. For each scanning point, the array changes the coded excitation state in sequence according to the encoding. The vector network analyzer records the S21 parameters (amplitude and phase) to obtain the near / field complex signal (amplitude and phase) of that scanning point under the switching of 256 coded beams.

[0060] Near / midfield complex signal formation measurement dataset for all scan points .

[0061] Based on measurement dataset The unit coupling response is calculated by the corresponding decoding algorithm of the encoding algorithm.

[0062] In one embodiment, the measurement dataset is processed by a decoding algorithm corresponding to the encoding matrix. Decoding operations were performed to obtain the field distribution of each of the 256 elements on the far-field plane of the element. This distribution integrates the radiation characteristics of the unit itself, the mutual coupling effect between units, and the combined influence of the test environment.

[0063] The active radiation patterns of 256 sparsely sampled elements are interpolated using an interpolation algorithm. The reconstructed radiation patterns are then normalized and converted to spherical coordinates for subsequent synthesis calculations, ultimately yielding radiation patterns for all 256 elements, forming an element radiation pattern database. .

[0064] In specific implementations, Hadamard coding is solved by its transpose matrix, uniform redundancy coding is solved by the decoding matrix constructed by the balanced correlation method, and non-redundant array coding is solved by deconvolution and frequency filtering.

[0065] In this embodiment, the element coupling response serves as the transfer vector from the element excitation to the near / field response, and its elements reflect the radiation characteristics of the antenna element considering the coupling effect.

[0066] Step S300: Based on the antenna element distribution and size data of the phased array antenna, establish a parameterized model of the phased array antenna. Furthermore, a data interface program was developed to integrate the measured data model of the T / R component with the radiation pattern database of the unit. The data is integrated into the parameterized model to generate a digital twin model of the phased array antenna pattern reconstruction.

[0067] Specifically, the parametric model includes: array layout, feed network, radiation pattern reconstruction module, and interface for interfacing with the actual T / R components. It primarily provides array geometry information, cell position coordinates, and radiation characteristics under ideal conditions, serving as a benchmark for subsequent data fusion.

[0068] In this embodiment, the interface that interfaces with the actual T / R component includes:

[0069] The set_beam interface is used to set the target beam direction;

[0070] The `get_excitation` interface is used to return the excitation values ​​of each unit. The calculated actual excitation complex coefficients;

[0071] The `get_farfield` interface is used to return data based on... Synthesized far-field radiation pattern at a specified angle;

[0072] The get_metrics interface is used to return key performance metrics (gain, beamwidth, sidelobe level, pointing accuracy, etc.).

[0073] In one embodiment, the measured data model of the T / R component and the radiation pattern database of the unit are used. Integration into the parametric model includes: establishing a measured model of the T / R component within the parametric model to replace the ideal device model.

[0074] The T / R module's measured model is configured as follows: Beam control commands are input, and the actual excitation complex coefficients of each antenna element are output. The calculation process is as follows: First, the ideal excitation amplitude / phase is calculated based on the target beam direction; then, the ideal excitation amplitude / phase is quantized into theoretical command values ​​for the phase shifter and attenuator; finally, based on the measured database of the T / R module's non-ideal characteristics... The actual phase shift value and the parasitic phase shift attenuation value are queried according to the phase shifter command value, and the actual attenuation value and the attenuation parasitic phase shift value are queried according to the attenuator theoretical command value, thereby obtaining the actual excitation complex coefficient.

[0075] A pattern reconstruction module is established in the parameterized model. The pattern reconstruction module is configured as follows: synthesizing far-field patterns based on the element pattern database of phased array antennas.

[0076] In this embodiment, a lookup table from the measured data model of the T / R component is used instead of the ideal model of the T / R component. When the digital twin model needs to calculate the value of a certain unit in the instruction... The excitation is first queried from the actual phase shift value lookup table. Lookup table for actual attenuation values Obtain the base value, then add the query phase-shifting parasitic attenuation value. and attenuation parasitic phase shift value After correction, the final excitation complex coefficients that best approximate the theoretical phase shift and attenuation are obtained. .

[0077] Through the above processing, the excitation complex coefficients in the digital twin model fully reflect the influence of the measured non-ideal characteristics of the T / R component.

[0078] Next, we will introduce the process of rapid verification and dynamic updating of digital twin models.

[0079] In this embodiment, the digital twin model is configured with a model verification module to perform a rapid verification process to verify the accuracy of the digital twin model in reconstructing the direction map.

[0080] The rapid verification process includes: selecting typical beam states as verification states; and for each verification state, calling the get_farfield interface of the digital twin model to calculate its synthetic far-field radiation pattern. As predicted values, key indicators such as gain, beamwidth, and sidelobe level are extracted. On the other hand, for each verification state, the measured far-field radiation pattern is obtained through a preset planar near-field scanning method or far-field testing method under the current state of the phased array antenna. The target value is used to extract key indicators such as gain, beamwidth, and sidelobe level.

[0081] Calculate the gain error under each verification state. Beamwidth error Sidelobe level error .

[0082] The errors of each key indicator are observed. If the errors of each key indicator are within the preset range, it indicates that the digital twin model has high accuracy and the model verification is successful. If the errors of any key indicator exceed the preset range, the dynamic update process of the digital twin model is triggered.

[0083] It should be noted that the rapid verification process can be set to be triggered periodically, such as once a week; or triggered when a critical event occurs, such as when the device is powered on or when the ambient temperature changes by more than 5°C.

[0084] To address the mismatch between the characteristics of digital models and physical entities after long-term operation, this embodiment sets up a dynamic iterative update mechanism for the model to continuously ensure the model's prediction accuracy and control reliability.

[0085] Specifically, the digital twin model is also equipped with an error corrector, which is used to fine-tune the digital twin model and execute the dynamic update process of the digital twin model.

[0086] In one embodiment, a lightweight neural network is established as an error corrector, comprising: a 3-layer multilayer perceptron (MLP), with the input being a rapid verification measurement. and The differences are analyzed by backpropagation to optimize the network parameters, and the output is a correction amount for the amplitude and phase of the cell radiation pattern in the cell radiation pattern database.

[0087] In this embodiment, rapid verification measurement will be performed. and The differences are used as monitoring signals to optimize network parameters through backpropagation. When the errors of each key indicator are small and within the preset range, the cell pattern database is fine-tuned based on the correction amount output by the error corrector after each rapid verification. When the errors of key indicators exceed the preset range, such as sidelobe errors greater than 2dB, a complete encoding measurement is re-executed to update the cell pattern database. This is to match the digital twin model with the physical entity and retrain the error corrector.

[0088] It should be noted that the weights of the error corrector are initialized to 0.

[0089] The digital twin model constructed in this embodiment deeply integrates two types of core measured databases to achieve high-fidelity replication of physical phased array antennas: one is a cell pattern database containing the coupling radiation characteristics of all array elements. First, it accurately characterizes the true active radiation characteristics of each unit under array mutual coupling, structural shielding, and scattering interference; second, it provides a database of non-ideal characteristics of T / R components. The system records non-ideal operating characteristics of each channel, such as phase shift drift, attenuation deviation, and amplitude-phase nonlinearity. Driven by two types of real-world databases, the digital twin model abandons the assumptions of traditional ideal array simulations, fully mirroring the actual operating state of the physical array surface. It possesses high-precision beam prediction, state awareness, and adaptive control capabilities. Furthermore, when the phased array antenna experiences device aging and parameter drift during long-term operation, leading to a slight decrease in beam performance at a specific scanning angle, the system periodically and automatically triggers a rapid verification measurement process. After the process is initiated, the model acquires the latest measured radiation pattern data of the physical array surface, compares it with the deviation characteristics of historical model prediction data, and trains an error corrector through deep learning to adaptively correct the unit radiation patterns in the unit radiation pattern database. This compensates for characteristic deviations caused by device aging and environmental interference, enabling the digital twin model to achieve high-precision matching with the physical antenna entity again, ensuring long-term operational stability.

[0090] Next, we will introduce the application of digital twin models in accurate beamforming prediction, intelligent diagnosis and location of antenna faults, and adaptive compensation for beam predistortion, which solves the technical problems of low beam control accuracy, delayed fault diagnosis, and difficulty in offsetting the effects of non-ideal characteristics and inter-array coupling in traditional phased array antennas.

[0091] Firstly, digital twin models are used for beamforming and pattern prediction.

[0092] In practice, when it is necessary to switch to a completely new beam pointing or beam pattern, the operator inputs the target beam parameters into the human-machine interface of the digital twin model. The model then uses its built-in beamforming algorithm to fuse the beams. The corresponding non-ideal characteristics of the T / R component and The corresponding unit coupling radiation characteristics are used to quickly iterate and solve for the optimal T / R component control command set, and the predicted far-field radiation pattern is synthesized and visualized in real time.

[0093] After the operator confirms that the beam performance meets the standards, the model can directly send the optimized control commands to the physical phased array to achieve precise beam control. The measured beam pattern is highly consistent with the model prediction results, and all beam indicators, especially the sidelobe level, fully meet the design requirements.

[0094] It is evident that the digital twin model possesses multi-scenario beamforming and accurate far-field pattern prediction capabilities, and can be adapted to three mainstream working scenarios: conventional scanning beam, low sidelobe beam, and anti-interference zero-dip beam. It fully couples the non-ideal characteristics of the T / R component and the mutual coupling characteristics of the unit, and the prediction results are highly consistent with the physical measurements.

[0095] Firstly, conventional scanning beam calculation and prediction. Staff input the target beam direction through a digital twin model's human-computer interface. The model calls the built-in beamforming algorithm, combined with Compensation for non-ideal amplitude and phase deviations in each T / R component, while matching It corrects beam distortion caused by unit mutual coupling, automatically iterates and calculates the optimal excitation coefficients of all units in the array, generates corresponding excitation weight vectors, completes array beam synthesis, and finally accurately outputs far-field radiation patterns and core performance indicators such as beam pointing accuracy, main lobe gain, side lobe level, and beamwidth, realizing rapid modeling and performance prediction of conventional beams.

[0096] Secondly, low sidelobe beam optimization generation. To meet the low sidelobe operation requirements of the array antenna, the operator inputs the target beam direction... With a preset desired sidelobe level threshold, the model calls the built-in beamforming algorithm to suppress sidelobes and ensure main lobe performance as the optimization objective. It fully considers the amplitude and phase non-ideal errors of the T / R components and the coupling interference between array elements, and adaptively iteratively optimizes the amplitude weighted distribution of each element. Under the premise of minimizing the main lobe gain loss, the sidelobe level of the synthesized beam pattern meets the preset index requirements, thereby achieving high-precision low-sidelobe beamforming.

[0097] Third, anti-interference adaptive null generation. For space electromagnetic interference scenarios, staff input the target communication beam direction... relative to the incident direction of the interference signal The model calls the Linear Constrained Minimum Variance (LCMV) adaptive beamforming algorithm. Under the constraint that the main lobe beam gain and pointing accuracy meet the working requirements, it adaptively generates a radiation null with a depth greater than 40dB in the direction of interference incident, effectively suppressing spatial interference signals, while minimizing the main lobe gain loss and ensuring the working stability of the array antenna in complex electromagnetic environments.

[0098] Secondly, digital twin models are used for fault diagnosis and location.

[0099] In this embodiment, the digital twin model is equipped with a fault database and an intelligent matching and diagnosis module. Relying on the real-time measured unit radiation pattern database, it realizes automatic identification, accurate positioning and visual annotation of phased array T / R components and antenna channels, solving the problems of low efficiency, large error and inability to predict hidden faults in traditional manual fault diagnosis.

[0100] The fault database stores a variety of common fault types, including:

[0101] A single unit fails completely, meaning that the single T / R component fails to function and has no radiation output;

[0102] Multi-unit cluster failure, i.e., multiple array units fail simultaneously in succession;

[0103] Phase shifter drift fault, i.e., the actual phase shift value of the unit deviates from the specified value. The reference phase shift value exceeds the preset error threshold;

[0104] Attenuator drift fault, i.e., the actual attenuation of the unit deviates from the specified value. The reference attenuation amount exceeds the preset error threshold.

[0105] During the long-term operation of the phased array antenna, the digital twin model collects and updates the measured unit radiation pattern data in real time, and continuously compares the performance differences between the model's predicted standard radiation pattern and the physically measured radiation pattern.

[0106] When physical measured radiation patterns show performance degradation phenomena such as abnormally high sidelobe levels, beam pointing deviation, and main lobe gain attenuation, for example, an abnormally high sidelobe level of 5dB, the digital twin model automatically retrieves the fault database to complete feature matching, accurately determines the fault type, locks the faulty T / R unit number and the corresponding faulty channel, and at the same time completes the visual annotation of fault location, fault type, and abnormal parameters on the human-machine interface, providing accurate data support for equipment operation and maintenance and fault repair.

[0107] Thirdly, digital twin models are used for beam predistortion compensation.

[0108] To address the issues of beam distortion, pointing offset, and sidelobe lift caused by the non-ideal characteristics of T / R components and the mutual coupling effect of array elements, this embodiment uses a digital twin model to achieve adaptive beam predistortion compensation, thereby canceling non-ideal interference from the command source and significantly improving the array beamforming accuracy.

[0109] The core principle of beam predistortion compensation is to abandon the traditional ideal beam command generation logic and introduce predistortion excitation correction in advance during the generation of T / R component beam control commands. The optimal predistortion excitation is solved so that the physical radiation field after non-ideal transmission and mutual coupling radiation of T / R components can approximate the preset ideal target radiation field to the greatest extent, thereby suppressing beam distortion problems from the root.

[0110] The specific compensation process is as follows: Staff input the desired ideal beam field. The digital twin model constructs an optimization objective function with regularization constraints:

[0111] ;

[0112] in, The array radiation pattern predicted by the digital twin model. The regularization coefficient is . It is a pre-distortion excitation used to constrain the excitation weight amplitude and avoid problems such as excitation overload and noise amplification.

[0113] In one embodiment, the digital twin model uses a gradient descent algorithm to iteratively solve the objective function to obtain the optimal array excitation. Then through The inverse mapping relationship of the database is used to solve for the corrected phase shift commands for each channel. With decay command Finally, the predistortion correction command is sent to each T / R component to complete the predistortion compensation of the entire channel beam, so that the physical array synthesized beam height is in line with the ideal beam index.

[0114] Through the detailed description of the embodiments, those skilled in the art can clearly understand that the core of the present invention lies in: deeply fusing the measured data of the non-ideal characteristics of the T / R component with the unit pattern data based on the coding matrix reconstruction, constructing a high-precision, dynamically evolving digital twin model for phased array antenna pattern reconstruction, solving the core problems of low reconstruction accuracy caused by not considering the actual T / R value, as well as low measurement efficiency and inability to dynamically update compensation.

[0115] The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made in accordance with the spirit and essence of the present invention should be covered by the present invention.

Claims

1. A method for constructing a digital twin model for pattern reconstruction of a phased array antenna, characterized in that, include: Full-basic-bit testing was conducted using the physical entity of the target phased array antenna to obtain a measured database of non-ideal characteristics of the T / R component. Construct a measured data model for the T / R component; Near / midfield measurements are performed based on the coded aperture excitation matrix of the array surface to obtain near / midfield response datasets, and then the element coupling response is calculated to construct an element pattern database of the antenna array. A parameterized model of the phased array antenna is constructed, and the measured database of the non-ideal characteristics of the T / R components is used. The unit pattern database is integrated into the parameterized model to generate a digital twin model for reconstructing the phased array antenna pattern.

2. The method for constructing a digital twin model for reconstructing the radiation pattern of a phased array antenna according to claim 1, characterized in that, The measured database of non-ideal characteristics of the T / R component Includes: actual phase shift value lookup table, actual attenuation value lookup table, phase shift parasitic attenuation data, and attenuation parasitic phase shift data; The full fundamental bit test includes: For each T / R component in a phased array antenna: Iterate through all the basic bit instructions of the attenuator, measure the actual amplitude response and actual phase shift of the T / R component under each basic bit instruction, and establish a lookup table of actual attenuation value and attenuation parasitic phase shift data; Iterate through all the basic bit commands of the phase shifter, measure the actual phase response and actual amplitude offset of the T / R component under each basic bit command, and establish a lookup table of actual phase shift values ​​and phase shift parasitic attenuation data.

3. The method for constructing a digital twin model for reconstructing the radiation pattern of a phased array antenna according to claim 1, characterized in that, The phased array antenna is equipped with a near / field measurement system, including a probe; The coding aperture excitation matrix is ​​constructed according to the coding algorithm, and each row in the coding aperture excitation matrix corresponds to a coding excitation. The near / field measurement performed by the array-based coded aperture excitation matrix, and then the unit coupling response was calculated, includes: sequentially loading the coded excitation in the coded aperture excitation matrix onto all T / R components of the array, controlling the waveguide probe to collect near / field complex signals point by point on the scanning plane, and obtaining a measurement dataset containing near / field complex signals of all scanning points; The measurement dataset is processed according to the decoding algorithm corresponding to the encoding algorithm. Decoding operations are performed to obtain the element coupling response of each element, resulting in the far-field radiation pattern of the corresponding element under all scan points, and an element radiation pattern database is constructed.

4. The method for constructing a digital twin model for phased array antenna pattern reconstruction according to claim 1, characterized in that, The parameterized model of the phased array antenna includes: radiating element model, array layout, feed network, and interface for interfacing with actual TR components.

5. The method for constructing a digital twin model for phased array antenna pattern reconstruction according to claim 1, characterized in that, The measured database of non-ideal characteristics of the T / R component Integrating the unit pattern database into the parameterized model includes: establishing a measured T / R component model in the digital twin model, and configuring the measured T / R component model to take beam control commands as input and the actual excitation complex coefficients of each antenna element as output. Its calculation process includes: The theoretical phase shift and theoretical amplitude weighting are calculated based on the target beam direction and quantized into phase shifter command values ​​and attenuator theoretical command values; Based on the measured database of non-ideal characteristics of T / R components The actual phase shift value and phase shift parasitic attenuation data are queried according to the phase shifter command value, and the actual attenuation value and attenuation parasitic phase shift data are queried according to the attenuator theoretical command value, thereby obtaining the actual excitation complex coefficients; A pattern reconstruction module is established in the digital twin model, and the pattern reconstruction module is configured to synthesize a far-field pattern based on the element pattern database of the phased array antenna.

6. The method for constructing a digital twin model for phased array antenna pattern reconstruction according to claim 1, characterized in that, The digital twin model is also configured with a model verification module for performing a rapid verification process to verify the reconstruction accuracy of the digital twin model. The rapid verification process includes: taking the measured radiation pattern and the predicted radiation pattern under the preset typical beam state as input, extracting key indicators from the measured radiation pattern and the predicted radiation pattern respectively; if the error of each key indicator is within the preset range, the model verification is passed; if the error of any key indicator exceeds the preset range, the dynamic update process of the digital twin model is triggered. The measured radiation pattern is obtained by planar near-field scanning or far-field testing under the current state of the phased array antenna, while the predicted radiation pattern is generated by the current digital twin model.

7. The method for constructing a digital twin model for phased array antenna pattern reconstruction according to claim 6, characterized in that, The digital twin model is also equipped with an error corrector, which is used to obtain the correction amount of the amplitude and phase of the unit pattern in the unit pattern database based on the error of each key indicator. The error corrector is built on a lightweight neural network. Its input is the error of each key indicator, and its output is the correction amount of the amplitude and phase of the unit radiation pattern in the unit radiation pattern database. Its configuration is as follows: If the error of each key indicator is within the preset range, the unit radiation pattern in the unit radiation pattern database is fine-tuned by the correction amount output by the error corrector after rapid verification. If the cumulative error of a key indicator exceeds the preset threshold, the dynamic update process of the digital twin model is executed. The dynamic update process of the digital twin model includes: re-performing the complete coded measurement, updating the cell pattern database, and retraining the error corrector.

8. The method for constructing a digital twin model for reconstructing the radiation pattern of a phased array antenna according to claim 1, characterized in that, The digital twin model is also equipped with a beam predistortion compensation module for compensating for the non-ideal characteristics and mutual coupling effects of the T / R components; The compensation for the non-ideal characteristics and mutual coupling effects of the T / R component includes: Based on the input target radiation pattern, define the optimization objective function and solve for the optimal amplitude and phase excitation using a preset optimization algorithm; By inversely mapping the measured data model of the T / R components, the optimal amplitude and phase excitation is converted into phase shifter and attenuator commands for each T / R component, and then sent to each T / R component so that the actual radiation field of the phased array antenna approximates the target radiation pattern.

Images