A Method and System for On-Orbit Real-Time Active Adjustment of the Surface Accuracy of a Large Planar SAR Antenna

Abstract

The present invention discloses a method and system for real-time active adjustment of the surface accuracy of a large planar SAR antenna in orbit. Through in-orbit real-time temperature field training of the rod system errors, the coupling influence law between the orbital thermal environment, the rod system assembly and adjustment errors, and the surface accuracy of the SAR antenna is determined. Through the learning and training of the in-orbit real-time temperature field, the real-time prediction of the thermal deformation and surface accuracy of the SAR antenna at any angle and position during in-orbit operation is realized. At the same time, by studying the coupling law of the in-orbit temperature field and the rod system assembly and adjustment errors on the surface accuracy, the use of deep learning algorithms and optimization algorithms can achieve "prediction - adjustment - optimization - re-prediction - re-adjustment" of the surface accuracy of the SAR antenna at different times and positions during in-orbit operation, realizing the active prediction and adjustment of the in-orbit surface accuracy.

Description

Technical Field

[0001] The present invention belongs to the technical field of antennas, and relates to a method and system for actively adjusting the surface accuracy of a large planar SAR antenna in orbit in real time. Background Technique

[0002] During the in-orbit operation of a SAR antenna, its electrical performance is closely related to the mechanical structure accuracy and is directly affected by structural factors such as the shape, size, and flatness of the antenna panel. The change in the surface accuracy of the antenna panel during in-orbit operation will exacerbate the pointing instability and attitude instability of the spaceborne SAR antenna, seriously affecting the performance and imaging quality of the antenna, modulating the amplitude of the echo signal, resulting in imaging blurring, reduced resolution, and thus affecting the imaging result.

[0003] To obtain high-resolution SAR images, how to accurately ensure the accuracy and performance of the structure of a large planar SAR antenna during in-orbit operation is one of the key problems urgently to be solved in the field of satellite general assembly and integration testing. The factors affecting the surface accuracy of a SAR antenna during in-orbit operation are mainly in two aspects: 1) The surface adjustment error caused by the uncertainty problems such as geometric dimensions, material parameters, and gaps in the antenna rod system due to insufficient assembly and manufacturing processes when on the ground. 2) The thermal deformation of the surface under the complex space thermal environment during in-orbit operation. Therefore, it is very necessary to explore the relationship between the complex space thermal environment, adjustment error, and antenna surface accuracy and to control and improve the surface accuracy accordingly.

[0004] At present, the research on the surface accuracy control of SAR antennas mainly stays at the theoretical level and the on-the-ground analysis stage, and is adjusted by passive control methods. It relies on improving the structural design and processing technology, enhancing the structural reliability and adjustment stability, reducing error transmission and accumulation; selecting special materials and coatings to control the in-orbit heat absorption and radiation ratio of the antenna, reducing the temperature difference of the antenna panel to reduce deformation; performing ground pre-compensation, interacting with the error deformation in the in-orbit deployment state of the antenna for compensation and cancellation, so as to improve the surface accuracy; according to the experience of assembly and adjustment technicians, adopting the repeated trial-and-error mode of "detection - adjustment - detection", borrowing the camera and theodolite measurement system, and making the assembly accuracy reach the standard through repeated blind adjustment according to the measured errors.

[0005] The above passive control methods have the following deficiencies:

[0006] 1. Lack of pertinence and clarity: The traditional trial-and-error assembly mode mainly relies on the experience and "feeling" of technicians, lacking direction and pertinence. Since the coupling influence law of various factors on assembly accuracy / performance during the assembly process has not been accurately grasped at the theoretical level, it is difficult for traditional assembly technology to accurately quantify and control the influence of process factors on the deployment reliability and assembly accuracy of deployable mechanisms. Multiple disassembly, assembly, folding, and unfolding operations will cause damage to springs, coated parts, etc., leaving potential reliability hazards, and it is difficult to ensure the consistency of the assembly quality of deployable mechanisms.

[0007] 2. Low adjustment efficiency: Relying on structural optimization design, the process is complex and the engineering cost is huge. The assembly and adjustment technicians need to conduct trial-and-error adjustments many times, and the process is complex. Making special materials and coatings requires continuous experiments for proportioning and adjustment, all of which have the problem of long adjustment cycles.

[0008] 3. Unable to ensure in-orbit accuracy: The relationship between error and accuracy is not clear. Passive control stays at ground adjustment and pre-compensation. Relying solely on passive control means cannot keep the final in-orbit deployment and operation state of the space deployable mechanism unchanged in space, and the complex and changeable nature of the space orbit environment is not considered, and it is impossible to actively adjust according to the in-orbit situation. When the SAR antenna is in orbit, its position relative to the Earth and the Sun is constantly changing, and the shape error changes caused by the real-time change of the space thermal environment are not actively considered, predicted, and controlled. Summary of the Invention

[0009] The purpose of the present invention is to solve the problems in the prior art that the coupling influence law of various factors on assembly accuracy / performance during the assembly process has not been accurately grasped at the theoretical level, the adjustment efficiency is low, and the in-orbit accuracy of the SAR antenna cannot be guaranteed, and to provide a method and system for real-time active adjustment of the shape accuracy of a large planar SAR antenna in orbit.

[0010] To achieve the above purpose, the present invention adopts the following technical solutions:

[0011] A method for real-time active adjustment of the shape accuracy of a large planar SAR antenna in orbit, including:

[0012] Collect the geometric parameters and physical parameters of the components of the large spaceborne deployable antenna, and construct a theoretical simulation model;

[0013] Conduct in-orbit thermal analysis of the large planar SAR antenna based on the space orbit thermal environment, and collect the in-orbit temperature field in real time;

[0014] Based on the in-orbit temperature field, conduct thermo-mechanical coupling analysis of heat and rod system errors to obtain the thermo-mechanical coupling deformation results;

[0015] Based on the thermo-mechanical coupling deformation results, fit a plane and calculate the flatness of the SAR antenna panel;

[0016] Determine the influence law of temperature change and truss error on the surface accuracy of the antenna, and establish a deep learning model between the truss error and the surface accuracy based on the theoretical simulation model;

[0017] Predict the surface accuracy of the current SAR antenna in-orbit position based on the deep learning model, and establish a surface accuracy optimization model;

[0018] Based on the active adjustment of the antenna truss, continuously optimize the surface accuracy at different times and positions during the in-orbit operation of the SAR antenna until the optimal surface accuracy is achieved, and obtain the finite element prediction and optimization model of the SAR antenna with high surface accuracy.

[0019] A further improvement of the present invention lies in:

[0020] Furthermore, the theoretical simulation model includes: a high-voltage power supply, a PZT actuator, and three-section truss materials; the high-voltage power supply is connected to the PZT actuator to provide power to the PZT actuator; the PZT actuator is connected to the three-section truss materials; the three-section truss materials are connected in sequence; the PZT actuator is driven by the high-voltage power supply, and under the piezoelectric effect, the PZT actuator generates an axial displacement, and the PZT actuator is directly connected to the truss material, indirectly realizing the adjustment of the truss length;

[0021]

[0022] ΔL = ΔL1 + ΔL2 + ΔL3

[0023] Where: l1, l2, and l3 are the lengths of the three sections of materials respectively; E1, E2, and E3 are the elastic moduli of the three sections of materials; A1, A2, and A3 are the cross-sectional areas of the three sections of materials; ΔL is the adjustment amount of the rod length.

[0024] Furthermore, conduct in-orbit thermal analysis of the large planar SAR antenna based on the in-orbit thermal environment, and collect the in-orbit temperature field in real time; specifically: conduct in-orbit thermal analysis based on the finite element analysis Simens Ug / Nx software, and extract the in-orbit real-time temperature field; the moments of the extracted temperature field are: when the sun is at the vernal equinox, summer solstice, autumnal equinox, and winter solstice positions, 24 interpolation calculation positions are selected for each orbit, including: direct sun illumination, side sun illumination, and the spacecraft entering and leaving the earth's shadow, obtain the temperature field change law, and obtain the temperature field change of the SAR antenna at any moment and position on the orbit based on the interpolation method.

[0025] Furthermore, based on the in-orbit temperature field, conduct a thermal-mechanical coupling analysis of heat and truss error to obtain the thermal-mechanical coupling deformation result, specifically:

[0026] Import the real-time on-orbit temperature field into the theoretical simulation model, and set the constraints and boundary conditions; set the rod length error, process it based on the temperature equivalence method, and use thermal expansion to convert the error change of the rod system into the temperature change of the rod system to achieve the adjustment of the rod length error:

[0027]

[0028] In the formula: L is the rod length, ΔL is the rod length adjustment amount, ΔT is the temperature change amount, and α is the linear thermal expansion coefficient of the rod;

[0029] Based on the on-orbit real-time temperature field and the rod system alignment error, use the finite element analysis software Abaqus to perform thermal-mechanical coupling analysis to obtain the deformation field of the SAR antenna panel.

[0030] Furthermore, based on the thermal-mechanical coupling deformation results, fit the plane and calculate the flatness of the SAR antenna panel. Specifically:

[0031] Extract the thermal-mechanical coupling analysis results to obtain the deformation field of the SAR antenna panel. Fit the three-dimensional space scatter points into a plane based on the least squares method and calculate the flatness of the SAR antenna panel;

[0032] The plane equation is obtained by fitting n space points. Take this plane as the reference surface of the spaceborne SAR antenna shape, regard each space point as a point on the actual plane, and obtain the deviation of each point from this fitted plane; that is:

[0033]

[0034] In the formula: z ij is the coordinate of the original coordinate point in the Z direction; z is the coordinate of the projection of this point on the reference plane; is the normalization factor;

[0035] According to the definition, the flatness error is:

[0036] E = max(e i ) - min(e i )

[0037] Among them, E is the flatness error.

[0038] Furthermore, a deep learning model between the truss error and the surface accuracy is established, specifically: taking the truss error as the input and the surface accuracy as the output, based on the simulation model, a number of training samples are generated; the samples are shuffled and divided into m training sets and n test sets; the inputs and outputs of the training set and the test set are normalized respectively; an RBF neural network is established, the number of neurons in the hidden layer is the number m of the training set, and the initial expansion speed spread(0) is set; after learning and training, the outputs of the training set and the test set are obtained, the training effect is evaluated by the root mean square error RMSE, and the inverse normalization process is carried out to obtain the predicted output and the prediction accuracy of the surface accuracy.

[0039] Furthermore, a surface accuracy optimization model is established, specifically:

[0040] Taking the surface accuracy of the simulation model as the objective function, taking the highest surface accuracy of the spaceborne SAR antenna as the optimization goal, and the implicit function as the deep learning model; taking the reduction of flatness error, the adjustment amount of each support rod size, and the difference between the optimization results of the previous and the current generations as the constraint conditions, the surface accuracy is optimized and designed based on the particle swarm optimization algorithm to obtain the optimal truss adjustment amount;

[0041] E = E[(a1 + Δa1), (a2 + Δa2), (a3 + Δa3), (a4 + Δa4), (a5 + Δa5), (a6 + Δa6), (a7 + Δa7)]

[0042] In the formula: E is the flatness error value of the antenna surface after applying the truss adjustment; a1, a2, …, a7 are the original errors of each rod of the SAR antenna; Δa1, Δa2, …, Δa7 are the adjustment amounts of each support rod;

[0043] The constraint conditions are:

[0044]

[0045] where K is the number of iterations.

[0046] Furthermore, based on the particle swarm optimization algorithm, the surface accuracy is optimized and designed to obtain the optimal truss adjustment amount, specifically:

[0047] Initialize the particle swarm, and use the random number method to give the initial position and velocity of the particles;

[0048] Set the flatness error of the spaceborne SAR antenna array surface as the optimization goal, take the flatness of the finite element simulation model as the objective function, evaluate the fitness of each particle in the initial state according to the objective function, and use it as the position of the optimal surface accuracy in the initial state of the particle. The position of the global optimal surface accuracy is obtained by comparison;

[0049] Determine whether the fitness of each particle at the initial position meets the requirements. If it meets, end; if it does not meet the end condition, enter the next step, and update the velocity and position of each particle according to the current state;

[0050] Re-evaluate the surface accuracy of each particle according to the objective function; compare the surface accuracy of each particle in the current state with the individual optimum, and take the better position as the new individual extreme value;

[0051] Compare the surface accuracy of each particle with the surface accuracy corresponding to the global optimum. The highest surface accuracy among all particles and higher than the fitness value at the current state global optimum position is used as the new global extreme value;

[0052] Judge whether the position meets the surface accuracy requirements of the large planar SAR antenna. If it does not meet, continue the optimization calculation; until it meets the requirements, usually the algorithm ends the optimization calculation when it reaches the set maximum iteration number or the optimal fitness value meets the minimum limit.

[0053] A large planar SAR antenna surface accuracy on-orbit real-time active adjustment system, comprising:

[0054] An acquisition module, which acquires the geometric parameters and physical parameters of the components of the large spaceborne deployable antenna and constructs a theoretical simulation model;

[0055] An analysis module, which conducts on-orbit thermal analysis of the large planar SAR antenna based on the space orbit thermal environment and acquires the on-orbit temperature field in real time;

[0056] An acquisition module, which conducts thermo-mechanical coupling analysis of heat and rod system errors based on the on-orbit temperature field and acquires the thermo-mechanical coupling deformation result;

[0057] A fitting module, which fits a plane based on the thermo-mechanical coupling deformation result and calculates the flatness of the SAR antenna panel;

[0058] A first construction module, which determines the influence law of temperature change and rod system alignment error on the antenna surface accuracy, and based on the theoretical simulation model, establishes a deep learning model between the rod system alignment error and the surface accuracy;

[0059] A second construction module, which predicts the surface accuracy of the current SAR antenna on-orbit position based on the deep learning model and establishes a surface accuracy optimization model;

[0060] An optimization module, which conducts active adjustment based on the antenna rod system, continuously optimizes the surface accuracy of the SAR antenna at different times and positions during on-orbit operation until the optimal surface accuracy is achieved, and obtains a finite element prediction and optimization model of the high surface accuracy SAR antenna.

[0061] Compared with the prior art, the present invention has the following beneficial effects:

[0062] Through in-orbit real-time temperature field for truss error training, the present invention determines the coupling influence law among the orbital thermal environment, truss alignment error and SAR antenna surface accuracy. Through the learning and training of the in-orbit real-time temperature field, the real-time prediction of thermal deformation and surface accuracy of the SAR antenna at any angle and position during in-orbit operation is realized. At the same time, by studying the coupling law of the in-orbit temperature field and truss alignment error on the surface accuracy, the deep learning algorithm and optimization algorithm can realize the "prediction - adjustment - optimization - re-prediction - re-adjustment" of the surface accuracy of the SAR antenna at different moments and positions during in-orbit operation, and realize the active prediction and adjustment of the in-orbit surface accuracy.

[0063] Furthermore, the present invention conducts data regression training and prediction between errors and surface accuracy based on the RBF neural network. The RBF neural network can approximate any nonlinear function with arbitrary precision in a compact set, has a fast learning speed and high training accuracy. By inputting truss errors, a large amount of data training can ensure the accuracy of the surface error prediction of large planar SAR antennas, realize the rapid prediction of the surface accuracy of SAR antennas, and improve the alignment efficiency.

[0064] Furthermore, the present invention uses the particle swarm algorithm for optimal design. Based on the surface error obtained from in-orbit real-time prediction, the optimal adjustment amount scheme of the truss for assembly adjustment can be quickly solved, avoiding the limitations of passive adjustment methods such as the existing trial-and-error alignment method, and greatly improving the antenna assembly accuracy and alignment efficiency.

[0065] Furthermore, the present invention directly adjusts the truss based on the PZT actuator. When the surface accuracy of the antenna does not meet the working requirements, the required truss adjustment amount is obtained through optimization calculation. An actuation voltage is applied to the actuator by the drive device. Under the piezoelectric effect, the actuator generates an axial displacement, and the actuator is directly connected to the truss to realize truss adjustment, optimize and improve the surface accuracy of the SAR antenna. The piezoelectric actuator has the characteristics of fast response speed, high positioning accuracy, low power consumption and good linearity, which greatly improves the surface accuracy optimization effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0067] Figure 1 It is a schematic flow chart of the method for on-orbit real-time active adjustment of the surface accuracy of a large planar SAR antenna of the present invention;

[0068] Figure 2 This is a schematic structural diagram of the in-orbit real-time active adjustment system for the surface accuracy of the large planar SAR antenna of the present invention;

[0069] Figure 3 This is a schematic structural diagram of the large planar SAR antenna according to an embodiment of the present invention;

[0070] Figure 4 This is a simplified schematic diagram of the actuator arrangement according to an embodiment of the present invention;

[0071] Figure 5 This is a schematic diagram of the technical solution flow according to an embodiment of the present invention;

[0072] Figure 6 This is a schematic diagram of the space orbit thermal analysis process according to an embodiment of the present invention;

[0073] Figure 7 This is a structure diagram of the RBF neural network of the deep learning model for error factors and surface accuracy according to an embodiment of the present invention;

[0074] Figure 8 This is a schematic diagram of the optimization algorithm process according to an embodiment of the present invention; Detailed implementation manners

[0075] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0076] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0077] It should be noted that: like reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0078] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper", "lower", "horizontal", "inner", etc. are used to indicate the orientation or positional relationship, it is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the invention product is usually placed during use. This is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, terms such as "first", "second", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0079] In addition, if the term "horizontal" appears, it does not mean that the component is required to be absolutely horizontal, but it can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that the structure must be completely horizontal, but it can be slightly inclined.

[0080] In the description of the embodiments of the present invention, it should also be noted that unless otherwise clearly specified and limited, if terms such as "set", "installed", "connected", "connected" are used, they should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0081] The following further describes the present invention in detail with reference to the drawings:

[0082] See Figure 1 , the present invention discloses a method for real-time active adjustment of the surface accuracy of a large planar SAR antenna in orbit, including:

[0083] S101, collect the geometric parameters and physical parameters of the components of the large spaceborne deployable antenna, and construct a theoretical simulation model.

[0084] The theoretical simulation model includes a high-voltage power supply, a PZT actuator, and a three-section rod system material; the high-voltage power supply is connected to the PZT actuator to provide power to the PZT actuator; the PZT actuator is connected to the three-section rod system material; the three-section rod system materials are connected in sequence; the PZT actuator is driven by the high-voltage power supply. Under the piezoelectric effect, the PZT actuator generates an axial displacement, and the PZT actuator is directly connected to the rod system material, indirectly realizing the adjustment of the rod system length;

[0085]

[0086] ΔL = ΔL1 + ΔL2 + ΔL3

[0087] Where: l1, l2, and l3 are the lengths of the three segments of materials respectively; E1, E2, and E3 are the elastic moduli of the three segments of materials; A1, A2, and A3 are the cross-sectional areas of the three segments of materials; ΔL is the adjustment amount of the rod length.

[0088] S102, conduct on-orbit thermal analysis of a large planar SAR antenna based on the space orbit thermal environment, and collect the on-orbit temperature field in real time.

[0089] Conduct space orbit thermal analysis based on the finite element analysis software Simens Ug / Nx, and extract the on-orbit real-time temperature field; the moments of the extracted temperature fields are: when the sun is at the vernal equinox, summer solstice, autumnal equinox, and winter solstice positions, 24 interpolation calculation positions are selected for each orbit, including: direct sun illumination, side sun illumination, and the spacecraft entering and leaving the earth's shadow, to obtain the variation law of the temperature field, and obtain the temperature field variation of the SAR antenna at any moment and position on the orbit based on the interpolation method.

[0090] S103, conduct thermo-mechanical coupling analysis of heat and rod system errors based on the on-orbit temperature field, and obtain the thermo-mechanical coupling deformation results.

[0091] Import the on-orbit real-time temperature field into the theoretical simulation model, and set the constraints and boundary conditions; set the rod length error, and process it based on the method of temperature equivalence. Use thermal expansion to convert the error change of the rod system into the temperature change of the rod system, so as to realize the adjustment of the rod length error:

[0092]

[0093] Where: L is the rod length, ΔL is the adjustment amount of the rod length, ΔT is the temperature change amount, and α is the linear thermal expansion coefficient of the rod;

[0094] Based on the on-orbit real-time temperature field and the rod system alignment error, conduct thermo-mechanical coupling analysis using the finite element analysis software Abaqus to obtain the deformation field of the SAR antenna panel.

[0095] S104, fit a plane based on the thermo-mechanical coupling deformation results and calculate the flatness of the SAR antenna panel.

[0096] Extract the thermo-mechanical coupling analysis results to obtain the deformation field of the SAR antenna panel. Fit the three-dimensional space scatter points into a plane based on the least squares method and calculate the flatness of the SAR antenna panel;

[0097] The plane equation is obtained by fitting n space points. Take this plane as the reference plane of the spaceborne SAR antenna surface, and regard each space point as a point on the actual plane to obtain the deviation amount of each point from this fitted plane; that is:

[0098]

[0099] Where: z ijis the Z - direction coordinate of the original coordinate point; z is the coordinate of the projection of this point on the reference plane; is the normalization factor;

[0100] From the definition, the flatness error is:

[0101] E = max(e i ) - min(e i )

[0102] where E is the flatness error.

[0103] S105. Determine the influence laws of temperature change and truss error on the antenna surface accuracy, and based on the theoretical simulation model, establish a deep - learning model between the truss error and the surface accuracy.

[0104] Taking the truss error as the input and the surface accuracy as the output, based on the simulation model, generate a number of training samples; shuffle the samples and divide them into m training sets and n test sets; normalize the inputs and outputs of the training sets and test sets respectively; establish an RBF neural network, the number of neurons in the hidden layer is the number m of the training set, and set the initial spread rate spread(0); after learning and training, obtain the outputs of the training set and test set, evaluate the training effect with the root - mean - square error RMSE, and perform anti - normalization processing to obtain the predicted output and prediction accuracy of the surface accuracy.

[0105] S106. Based on the deep - learning model, predict the surface accuracy of the current SAR antenna in the on - orbit position and establish a surface accuracy optimization model.

[0106] Taking the surface accuracy of the simulation model as the objective function, with the highest surface accuracy of the space - borne SAR antenna as the optimization goal, and the implicit function as the deep - learning model; taking the reduction of the flatness error, the adjustment amount of each support rod size, and the difference between the optimization results of the previous and current generations as the constraint conditions, optimize the surface accuracy based on the particle swarm optimization algorithm to obtain the optimal truss adjustment amount;

[0107] E = E[(a1 + Δa1),(a2 + Δa2),(a3 + Δa3),(a4 + Δa4),(a5 + Δa5),(a6 + Δa6),(a7 + Δa7)]

[0108] In the formula: E is the flatness error value of the antenna surface after applying truss adjustment; a1, a2, …, a7 are the original errors of each rod of the SAR antenna; Δa1, Δa2, …, Δa7 are the adjustment amounts of each support rod;

[0109] The constraint conditions are:

[0110]

[0111] where K is the number of iterations.

[0112] Optimize the surface accuracy based on the particle swarm optimization algorithm to obtain the optimal adjustment amount of the rod system. Specifically:

[0113] Initialize the particle swarm, and use the method of random numbers to give the initial positions and velocities of the particles;

[0114] Set the flatness error of the spaceborne SAR antenna array surface as the optimization objective, take the flatness of the finite element simulation model as the objective function, evaluate the fitness of each particle in the initial state according to the objective function, as the position of the optimal surface accuracy in the initial state of the particle, and obtain the position of the global optimal surface accuracy by comparison;

[0115] Judge whether the fitness of each particle at the initial position meets the requirements. If it meets, end; if it does not meet the end condition, enter the next step, and update the velocity and position of each particle according to the current state;

[0116] Evaluate the surface accuracy of each particle again according to the objective function; compare the surface accuracy of each particle in the current state with the individual optimum, and take the better position as the new individual extreme value;

[0117] Compare the surface accuracy of each particle with the surface accuracy corresponding to the global optimum. The highest among all particles and higher than the fitness value at the current state of the global optimum position is used as the new global extreme value;

[0118] Judge whether this position meets the requirements of the surface accuracy of the large planar SAR antenna. If it does not meet, continue the optimization calculation; until it meets the requirements, usually the algorithm ends the optimization calculation when it reaches the set maximum number of iterations or the optimal fitness value meets the minimum limit.

[0119] S107. Based on the antenna rod system, make active adjustments, continuously optimize the surface accuracy at different times and positions during the on-orbit operation of the SAR antenna until the optimal surface accuracy is achieved, and obtain a finite element prediction and optimization model of the SAR antenna with high surface accuracy.

[0120] See Figure 2 , the present invention discloses a large planar SAR antenna surface accuracy on-orbit real-time active adjustment system, including:

[0121] A collection module, which collects the geometric parameters and physical parameters of the components of the large spaceborne deployable antenna and constructs a theoretical simulation model;

[0122] An analysis module, which conducts on-orbit thermal analysis of the large planar SAR antenna based on the space orbit thermal environment and collects the on-orbit temperature field in real time;

[0123] An acquisition module, which conducts thermal-mechanical coupling analysis of heat and rod system errors based on the on-orbit temperature field and obtains the thermal-mechanical coupling deformation result;

[0124] A fitting module, which fits a plane based on the thermo-mechanical coupling deformation result and calculates the flatness of the SAR antenna panel plane;

[0125] A first construction module, which determines the influence law of temperature change and rod system alignment error on the antenna surface accuracy, and based on the theoretical simulation model, establishes a deep learning model between the rod system alignment error and the surface accuracy;

[0126] A second construction module, which predicts the surface accuracy of the current SAR antenna in-orbit position based on the deep learning model and establishes a surface accuracy optimization model;

[0127] An optimization module, which actively adjusts based on the antenna rod system, continuously optimizes the surface accuracy at different times and positions during the in-orbit operation of the SAR antenna until the optimal surface accuracy is achieved, and obtains a finite element prediction and optimization model of the SAR antenna with high surface accuracy.

[0128] Embodiment:

[0129] See Figure 3 , the large planar SAR antenna is a symmetric structure, and the large planar SAR antenna includes: a satellite body 1, a satellite connecting rod 2, an outer support rod 3, a middle support rod 4, an inner support rod 5, a 90° locking hinge 6, an actuator 7, an SAR antenna panel 8, a 180° locking hinge 9, and a driving component.

[0130] Among them, the satellite body 1 is connected to the two middlemost antenna panels 8 through the 90° locking hinge 6; there are two antenna panels 8 on each side of the satellite body 1, and the antenna panels 8 on each side are connected through the 180° hinge 9; the antenna panel 8 is connected to the satellite body 1 through the 90° hinge 6; the outer support rod 3 is connected to the outermost antenna panel 8, one end of the satellite connecting rod 2 is fixed on the satellite body 1, and the other end of the satellite connecting rod 2 is connected to the outer support rod 3; one of the two middlemost antenna panels 8 is respectively connected to two middle support rods 4 and two inner support rods 5, and the other end of the satellite connecting rod 2 is simultaneously connected to the intersection of the outer support rod 3, the middle support rod 4, and the inner support rod 5; PZT actuators 7 are arranged on the satellite connecting rod 2, the outer support rod 3, the middle support rod 4, and the inner support rod 5, and during modeling, its structure is simplified through parameter equivalent calculation and is regarded as a beam when connected to the rod system as a whole. The simplified schematic diagram is as Figure 4 shown. The driving component is installed at the 90° hinge 6 and the 180° hinge 9, and the deployable antenna is symmetrically distributed on both sides of the satellite body 1.

[0131] See Figure 5, the method for on-orbit real-time active adjustment of the large planar SAR antenna surface accuracy in this embodiment includes: carrying out the simulation of the large planar SAR antenna surface accuracy in the space environment, establishing a deep learning model between the rod system error and the surface accuracy, and on-orbit real-time adjustment and optimization of the large planar SAR antenna surface accuracy.

[0132] (1) Carry out the simulation of the large planar SAR antenna surface accuracy in the space environment

[0133] (1-1) Establish a simplified simulation model

[0134] As Figure 4 shown in the simplified schematic diagram of the rod system actuator layout, the actual rod is a "link-actuator-link" structure, and the theoretical simulation model includes a high-voltage power supply, a PZT actuator, and three sections of rod system materials; the high-voltage power supply is connected to the PZT actuator to provide power to the PZT actuator; the PZT actuator is connected to the three sections of rod system materials; the three sections of rod system materials are connected in sequence; the PZT actuator is driven by the high-voltage power supply, and under the piezoelectric effect, the PZT actuator generates an axial displacement, and the PZT actuator is directly connected to the rod system material, indirectly realizing the adjustment of the rod system length.

[0135]

[0136] ΔL = ΔL1 + ΔL2 + ΔL3

[0137] In the formula: l1, l2, and l3 are the lengths of the three sections of materials respectively; E1, E2, and E3 are the elastic moduli of the three sections of materials; A1, A2, and A3 are the cross-sectional areas of the three sections of materials; ΔL is the adjustment amount of the rod length.

[0138] Based on the geometric parameters and physical parameters of the large planar SAR antenna, parametric modeling is carried out in the general finite element analysis software Abaqus using Python. An equivalent analysis model of the spatial hinge clearance is established based on the contact theory and the spring model, an equivalent shell element model of the antenna panel is established based on the equivalent plate theory, an equivalent beam element of the rod system is established based on the elasticity theory, and a finite element assembly model is established based on the coupling and constraint relationships.

[0139] (1-2) Spatial orbit thermal analysis

[0140] As Figure 6 shown, in the spatial orbit thermal analysis process, it is necessary to define the orbit type and parameters, the space thermal environment, the SAR antenna attitude and position, the sun position and parameters.

[0141] Space is a vacuum with almost no air convection. Therefore, the heat transfer between the SAR antenna and the external environment occurs only through radiation. The heat transfer between the internal equipment and structures of the SAR antenna is only conduction and radiation between solids, as well as thermal coupling between different grid cells. At the same time, the shadow occlusion effect between the SAR antenna structures and the occlusion effect of the Earth on the Sun are considered. The space heat sources are mainly the thermal radiation of the Sun, the Earth, other planets, and their reflections of the solar radiation. Since the SAR satellite is an Earth-facing satellite, ignoring the influence of other celestial bodies in space, the thermal radiation mainly considers solar thermal radiation, Earth infrared thermal radiation, and Earth albedo solar thermal radiation. The space is equivalent to a black body at 4K, and the Earth is a black body at 250K. The influence of solar radiation pressure and space microgravity is ignored.

[0142] Based on the general finite element analysis Simens Ug / Nx software, space orbit thermal analysis is carried out to extract the on-orbit real-time temperature field. Further, the moments of the extracted temperature field are: when the Sun is at the vernal equinox, summer solstice, autumnal equinox, and winter solstice positions, 24 interpolation calculation positions are selected for each orbit, mainly including: direct solar illumination, side solar illumination, the spacecraft entering and leaving the Earth's shadow, etc., to find the variation law of the temperature field, and through interpolation to obtain the temperature field variation of the SAR antenna at any moment and position on the orbit.

[0143] (1-3) Thermal-mechanical coupling analysis based on heat and error

[0144] Based on the finite element model established in step (1-1), the real-time on-orbit temperature field obtained in step (1-2) is imported, and the constraints and boundary conditions are set. The rod length error is set and processed based on the temperature equivalence method. The error change of the rod system is transformed into the temperature change of the rod system by thermal expansion to realize the adjustment of the rod length error:

[0145]

[0146] In the formula: L is the rod length, ΔL is the rod length adjustment amount, ΔT is the temperature change amount, and α is the linear thermal expansion coefficient of the rod.

[0147] Based on the on-orbit real-time temperature field and the rod system error, the deformation field of the SAR antenna panel is obtained by thermal-mechanical coupling analysis using the general finite element analysis software Abaqus.

[0148] (2) Establish a deep learning model between rod system error and surface accuracy

[0149] The establishment of the deep learning model between the rod system error and the surface accuracy described in this embodiment includes: clarifying the influence law of the on-orbit temperature field and the rod system alignment error on the antenna surface accuracy; considering the dimensional error, generating multiple samples for training based on the simulation model, with the input of the learning model being the rod system error and the output being the surface accuracy.

[0150] The specific solution steps are as follows:

[0151] (2-1) Least squares method for fitting a plane to obtain flatness

[0152] According to the thermo-mechanical coupling analysis results in step (1-3), the deformation field of the SAR antenna panel is extracted, and the three-dimensional space scatter points are fitted to a plane by the least squares method to calculate the flatness of the SAR antenna panel.

[0153] Flatness error refers to the distance between the actual surface of an object and the ideal reference plane. In this simulation analysis, the plane equation is obtained by fitting n space points, so this plane can be used as the reference plane for the shape of the spaceborne SAR antenna. Regarding each space point as a point on the actual plane, the deviation of each point from this fitted plane can be calculated. That is:

[0154]

[0155] In the formula: z ij is the coordinate of the original coordinate point in the Z direction; z is the coordinate of the projection of this point on the reference plane; is the normalization factor, which approaches 1 when the coordinates of the point in the X and Y directions are large.

[0156] From the definition, the flatness error can be obtained as:

[0157] E = max(e i ) - min(e i )

[0158] Among them, E is the flatness error.

[0159] (2-2) Clarify the influence laws of temperature change and alignment error on the antenna surface accuracy

[0160] Taking the on-orbit real-time temperature field as the predefined initial field condition, generating multiple sets of truss errors, conducting simulation training to obtain flatness, clarifying the coupling influence laws between the orbital thermal environment, truss alignment error and the SAR antenna surface accuracy, and obtaining the thermal deformation and surface accuracy of the SAR antenna at a specific moment during on-orbit operation;

[0161] According to the SPSS experimental analysis platform, design orthogonal experiments, conduct range analysis and multi-factor variance analysis to investigate the influence relationship between 14 trusses such as 2, 3, 4, and 6 in the SAR antenna Figure 3 and the SAR antenna surface accuracy;

[0162] (2-3) Establish a deep learning model between truss error and surface accuracy

[0163] Considering dimensional errors, based on the simulation model, generate multiple samples for training, and conduct data regression training and prediction between error and surface accuracy based on the RBF neural network.

[0164] As Figure 7 shown, it is the structure diagram of the RBF neural network. The topological structure of the RBF neural network is a three-layer forward network:

[0165] The first layer is the input layer, which is composed of signal source nodes and only plays the role of transmitting data information without any transformation of the input information;

[0166] The second layer is the hidden layer, and the number of nodes is determined as needed. The kernel function (activation function) of the hidden layer neurons is the Gaussian function, which performs a spatial mapping transformation on the input information;

[0167] The third layer is the output layer, which responds to the input pattern. The activation function of the output layer neurons is a linear function, which linearly weights the information output by the hidden layer neurons and then outputs it as the output result of the entire neural network.

[0168] The RBF neural network can approximate any nonlinear function with arbitrary precision in a compact set, with fast learning speed and high training accuracy. Based on the RBF neural network, the learning and prediction of the surface accuracy are realized. As Figure 7 shown, the basic scheme process is as follows:

[0169] The input is the truss error, and the output is the surface accuracy. The truss error is trained based on the on-orbit real-time temperature field

[0170] (2-3-1) Taking the truss error as the input and the surface accuracy as the output, multiple training samples are generated based on the simulation model;

[0171] (2-3-2) Shuffle the samples and divide them into m training sets and n test sets;

[0172] (2-3-3) Normalize the input and output of the training set and the test set respectively;

[0173] (2-3-4) Establish an RBF neural network, with the number of hidden layer neurons being the number m of the training set, and set the initial spread rate spread(0);

[0174] (2-3-5) After learning and training, the outputs of the training set and the test set are obtained. The training effect is evaluated by the root mean square error RMSE, and the denormalization process is performed to obtain the predicted output and predicted accuracy of the surface accuracy.

[0175] (3) On-orbit real-time adjustment and optimization of the surface accuracy of the large planar SAR antenna

[0176] (3-1) Establish a surface accuracy optimization model

[0177] In this adjustment and optimization model, the form accuracy of the simulation model is selected as the objective function, and the highest form accuracy of the spaceborne SAR antenna is taken as the ultimate optimization goal. The implicit function is the learning model in step (2-3). Taking the reduction in flatness error, the adjustment amounts of each support rod, and the difference between the optimization results of the previous and current generations as the constraint conditions, the form accuracy is optimized by the particle swarm optimization algorithm to obtain the optimal adjustment amount of the rod system. E = E[(a1 + Δa1), (a2 + Δa2), (a3 + Δa3), (a4 + Δa4), (a5 + Δa5), (a6 + Δa6), (a7 + Δa7)]

[0178] In the formula: E is the flatness error value of the antenna form after applying the rod system adjustment; a1, a2,..., a7 are the original errors of each rod of the SAR antenna; Δa1, Δa2,..., Δa7 are the adjustment amounts of each support rod.

[0179] The constraint conditions are:

[0180]

[0181] where K is the number of iterations.

[0182] Based on the particle swarm optimization algorithm, solve the rod system adjustment scheme. As Figure 8 shown, the basic process includes:

[0183] (3-1-1) First, it is necessary to initialize the particle swarm, and use the method of random numbers to give the initial positions and velocities of the particles;

[0184] (3-1-2) Set the flatness error of the spaceborne SAR antenna array surface as the optimization goal, take the flatness of the finite element simulation model as the objective function, evaluate the fitness of each particle in the initial state according to the objective function, and use it as the position of the optimal form accuracy in the initial state of the particle. Obtain the position of the global optimal form accuracy by comparison;

[0185] (3-1-3) Judge whether the fitness of each particle at the initial position meets the requirements. If it meets, end;

[0186] (3-1-4) If the end condition is not met, enter the next step, and update the velocity and position of each particle according to the current state;

[0187] (3-1-5) Re-evaluate the form accuracy of each particle according to the objective function;

[0188] (3-1-6) Compare the form accuracy of each particle in the current state with the individual optimum, and take the better position as the new individual extreme value;

[0189] (3-1-7) Then, compare the surface accuracy of each particle with the surface accuracy corresponding to the global optimum. The highest surface accuracy among all particles that is higher than the fitness value at the global optimum position of the current state is used as the new global extreme value.

[0190] (3-1-8) Again, determine whether this position meets the surface accuracy requirements of the large planar SAR antenna. If it does not meet the requirements, continue the optimization calculation. Usually, the algorithm ends the optimization calculation when it reaches the set maximum number of iterations or the optimal fitness value meets the minimum limit.

[0191] (3-2) On-orbit real-time adjustment of actuators

[0192] By studying the coupling law of the on-orbit temperature field and the assembly error of the rod system on the surface accuracy, the "prediction-adjustment-optimization-re-prediction-re-adjustment" of the surface accuracy at different times and positions during the on-orbit operation of the SAR antenna can be realized by using deep learning algorithms and optimization algorithms, achieving the active prediction and adjustment of the on-orbit surface accuracy.

[0193] Predict the surface accuracy of the current on-orbit position of the SAR antenna through deep learning. Based on the particle swarm optimization algorithm, output the adjustment amount of the rod system that makes the surface accuracy optimal. After obtaining the adjustment amount of the rod system, adjust the rod system error through the actuator to make the surface accuracy optimal: Use PZT piezoelectric actuators arranged on the rod system. Apply an actuator voltage to the actuator through the drive device. Under the piezoelectric effect, the actuator generates an axial displacement to produce a given adjustment amount. The actuator is directly connected to the rod system. The rod system adjustment is realized through the actuator to make the surface accuracy optimal, turning the rod system error to good use, and finally obtaining a finite element prediction and optimization model of the SAR antenna with high surface accuracy.

[0194] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, various changes and modifications can be made to the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for on-orbit real-time active adjustment of the surface accuracy of a large planar SAR antenna, characterized in that, Including: Collecting the geometric and physical parameters of the components of a large spaceborne deployable antenna and constructing a theoretical simulation model; Conducting on-orbit thermal analysis of a large planar SAR antenna based on the space orbit thermal environment and collecting the on-orbit temperature field in real time; Performing thermo-mechanical coupling analysis of heat and rod system errors based on the on-orbit temperature field to obtain the thermo-mechanical coupling deformation results; Fitting a plane based on the thermo-mechanical coupling deformation results and calculating the flatness of the SAR antenna panel; Determining the influence law of temperature change and rod system errors on the surface accuracy of the antenna, and establishing a deep learning model between the rod system errors and the surface accuracy based on the theoretical simulation model; Predicting the surface accuracy of the current SAR antenna on-orbit position based on the deep learning model and establishing a surface accuracy optimization model; Based on the active adjustment of the antenna rod system, continuously optimizing the surface accuracy at different times and positions during the on-orbit operation of the SAR antenna until the optimal surface accuracy is achieved, and obtaining a finite element prediction and optimization model of a high-surface-accuracy SAR antenna; The step of fitting a plane based on the thermo-mechanical coupling deformation results and calculating the flatness of the SAR antenna panel is specifically as follows: Extracting the thermo-mechanical coupling analysis results to obtain the deformation field of the SAR antenna panel, and fitting the three-dimensional space scatter points into a plane based on the least squares method and calculating the flatness of the SAR antenna panel; The plane equation is obtained by fitting n spatial points, and this plane is used as the surface reference plane of the spaceborne SAR antenna. Regarding each spatial point as a point on the actual plane, the deviation of each point from this fitted plane is obtained; that is: In the formula: is the coordinate in the Z direction of the original coordinate point; z is the coordinate of the projection of this point on the reference plane; is the normalization factor; The flatness error is obtained by definition as: where E is the flatness error; The step of establishing the surface accuracy optimization model is specifically as follows: Taking the surface accuracy of the simulation model as the objective function, taking the highest surface accuracy of the spaceborne SAR antenna as the optimization goal, and the implicit function as the deep learning model; taking the reduction in flatness error, the adjustment amount of each support rod size, and the difference between the optimization results of the previous and current generations as the constraint conditions, and optimizing the design of the surface accuracy based on the particle swarm optimization algorithm to obtain the optimal rod system adjustment amount; Where: E is the flatness error value of the antenna surface after adjusting the rod system; a1, a2, …, a7 are the original errors of each rod of the SAR antenna; are the adjustment amounts of each strut; The constraint conditions are: where K is the number of iterations; the step of optimizing the design of the surface accuracy based on the particle swarm optimization algorithm to obtain the optimal rod system adjustment amount is specifically as follows: Initializing the particle swarm and using the method of random numbers to give the initial positions and velocities of the particles; Setting the flatness error of the spaceborne SAR antenna array surface as the optimization goal, taking the flatness of the finite element simulation model as the objective function, evaluating the fitness of each particle in the initial state according to the objective function, taking it as the position of the optimal surface accuracy in the initial state of the particle, and obtaining the position of the global optimal surface accuracy through comparison; Judging whether the fitness of each particle at the initial position meets the requirements. If it meets, end; if it does not meet the end condition, enter the next step, and update the velocity and position of each particle according to the current state; Evaluating the surface accuracy of each particle again according to the objective function; comparing the surface accuracy of each particle in the current state with the individual optimum, and taking the better position as the new individual extreme value; Comparing the surface accuracy of each particle with the surface accuracy corresponding to the global optimum, and taking the highest among all particles and higher than the fitness value at the current state global optimum position as the new global extreme value; Judge whether the position meets the surface accuracy requirements of the large planar SAR antenna. If not, continue the optimization calculation until the requirements are met, the algorithm reaches the set maximum number of iterations, or the optimal fitness value meets the minimum limit, then end the optimization calculation.

2. The method for on-orbit real-time active adjustment of the surface accuracy of a large planar SAR antenna according to claim 1, characterized in that, The theoretical simulation model includes: a high-voltage power supply, a PZT actuator, and a three-section rod system material. The high-voltage power supply is connected to the PZT actuator to provide power to the PZT actuator. The PZT actuator is connected to the three-section rod system material, and the three-section rod system materials are connected in sequence. The PZT actuator is driven by the high-voltage power supply. Under the piezoelectric effect, the PZT actuator generates an axial displacement. Since the PZT actuator is directly connected to the rod system material, the adjustment of the rod system length is indirectly realized. In the formula: is the length of each of the three segments of material; is the elastic modulus of the three segments of material; is the cross-sectional area of the three segments of material; is the adjustment amount of the length of the rod.

3. The method for on-orbit real-time active adjustment of the surface accuracy of a large planar SAR antenna according to claim 2, characterized in that, The on-orbit thermal analysis of the large planar SAR antenna based on the space orbit thermal environment is to collect the on-orbit temperature field in real time. Specifically, based on the finite element analysis software Siemens Ug / Nx, the space orbit thermal analysis is carried out to extract the on-orbit real-time temperature field. The moments of the extracted temperature field are: when the sun is at the vernal equinox, summer solstice, autumnal equinox, and winter solstice positions, 24 interpolation calculation positions are selected for each orbit, including: solar direct illumination, solar side illumination, and the spacecraft entering and leaving the earth's shadow, to obtain the temperature field change law, and based on the interpolation method, the temperature field change of the SAR antenna at any moment and position on the orbit is obtained.

4. The on-orbit real-time active adjustment method for the surface accuracy of a large planar SAR antenna according to claim 3, wherein, Based on the on-orbit temperature field, the thermo-mechanical coupling analysis of heat and rod system error is carried out to obtain the thermo-mechanical coupling deformation result. Specifically: Import the real-time on-orbit temperature field into the theoretical simulation model, and set the constraints and boundary conditions. Set the rod length error, and process it based on the temperature equivalence method. The error change of the rod system is transformed into the temperature change of the rod system by thermal expansion to realize the adjustment of the rod length error: In the formula: is the rod length, is the rod length adjustment amount, is the temperature change amount, is the linear thermal expansion coefficient of the rod; Based on the on-orbit real-time temperature field and the rod system alignment error, use the finite element analysis software Abaqus to carry out thermo-mechanical coupling analysis to obtain the deformation field of the SAR antenna panel.

5. The on-orbit real-time active adjustment method for the surface accuracy of a large planar SAR antenna according to claim 4, wherein, The deep learning model between the rod system error and the surface accuracy is established as follows: taking the rod system error as the input and the surface accuracy as the output, based on the simulation model, several training samples are generated. Shuffle the samples and divide them into m training sets and n test sets. Normalize the inputs and outputs of the training sets and test sets respectively. Establish an RBF neural network, the number of neurons in the hidden layer is the number m of the training set, and set the initial expansion speed spread(0); After learning and training, the outputs of the training set and the test set are obtained. The training effect is evaluated by the root mean square error RMSE, and the anti-normalization process is carried out to obtain the predicted output and prediction accuracy of the surface accuracy.

6. A on-orbit real-time active adjustment system for the surface accuracy of a large planar SAR antenna, wherein, It includes: A collection module that collects the geometric parameters and physical parameters of the components of the large spaceborne deployable antenna and constructs a theoretical simulation model; An analysis module that conducts on-orbit thermal analysis of the large planar SAR antenna based on the space orbit thermal environment and collects the on-orbit temperature field in real time; An acquisition module that conducts thermo-mechanical coupling analysis of heat and rod system error based on the on-orbit temperature field to obtain the thermo-mechanical coupling deformation result; A fitting module that fits a plane based on the thermo-mechanical coupling deformation result and calculates the flatness of the SAR antenna panel; The first construction module determines the influence law of temperature change and rod system alignment error on the antenna surface accuracy, and based on the theoretical simulation model, establishes a deep learning model between the rod system alignment error and the surface accuracy; The second construction module predicts the surface accuracy of the current SAR antenna on-orbit position based on the deep learning model, and establishes a surface accuracy optimization model; The optimization module actively adjusts based on the antenna rod system, continuously optimizes the surface accuracy at different times and positions during the on-orbit operation of the SAR antenna until the optimal surface accuracy is achieved, and obtains a finite element prediction and optimization model of the high surface accuracy SAR antenna; Among them, the method of fitting a plane and calculating the flatness of the SAR antenna panel based on the thermo-mechanical coupling deformation result is specifically as follows: Extract the thermo-mechanical coupling analysis result to obtain the deformation field of the SAR antenna panel, and fit the three-dimensional space scatter points into a plane based on the least squares method and calculate the flatness of the SAR antenna panel; The plane equation is obtained by fitting n spatial points, and this plane is used as the surface reference plane of the spaceborne SAR antenna. Regarding each spatial point as a point on the actual plane, obtain the deviation of each point from this fitted plane; that is: In the formula: is the coordinate in the Z direction of the original coordinate point; z is the coordinate of the projection of this point on the reference plane; is the normalization factor; The flatness error is obtained by definition as: Among them, E is the flatness error; The establishment of the surface accuracy optimization model is specifically as follows: Taking the surface accuracy of the simulation model as the objective function, taking the highest surface accuracy of the spaceborne SAR antenna as the optimization goal, and the implicit function as the deep learning model; taking the reduction in flatness error, the adjustment amount of each support rod size, and the difference between the optimization results of the previous and current generations as constraints, optimize the design of the surface accuracy based on the particle swarm optimization algorithm to obtain the optimal rod system adjustment amount; Where: E is the flatness error value of the antenna surface after adjusting the rod system; a1, a2, …, a7 are the original errors of each rod of the SAR antenna; are the adjustment amounts of each strut; The constraint conditions are: Among them, K is the number of iterations; the method of optimizing the design of the surface accuracy based on the particle swarm optimization algorithm to obtain the optimal rod system adjustment amount is specifically as follows: Initialize the particle swarm, and use the method of random numbers to give the initial position and velocity of the particles; Set the flatness error of the spaceborne SAR antenna array surface as the optimization goal, take the flatness of the finite element simulation model as the objective function, evaluate the fitness of each particle in the initial state according to the objective function, and use it as the position of the optimal surface accuracy in the initial state of the particle. Obtain the position of the global optimal surface accuracy by comparison; Judge whether the fitness of each particle at the initial position meets the requirements. If it meets, end; if it does not meet the end condition, enter the next step, and update the velocity and position of each particle according to the current state; Evaluate the surface accuracy of each particle again according to the objective function; compare the surface accuracy of each particle in the current state with the individual optimum, and take the better position as the new individual extreme value; Compare the surface accuracy of each particle with the surface accuracy corresponding to the global optimum. The highest among all particles and higher than the fitness value at the current state global optimum position is used as the new global extreme value; Judge whether this position meets the surface accuracy requirements of the large planar SAR antenna. If it does not meet, continue the optimization calculation; until it meets the requirements, the algorithm reaches the set maximum number of iterations or the optimal fitness value meets the minimum limit, then end the optimization calculation.

Images