Satellite solar wing unfolding test key parameter prediction system and method
Through deep learning algorithms, the prediction model of key parameter of satellite solar wing expansion test is constructed, combined with simulation and actual measured data, high-precision online prediction and abnormal alarm of satellite solar wing expansion process is realized, and the inaccurate simulation guidance and data acquisition in the existing technology is solved. The problem of incompleteness improves the success rate and efficiency of the experiment.
Patent Information
- Application Number
- CN202410859199.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-28
- Publication Date
- 2024-11-19
AI Technical Summary
The existing satellite solar wing expansion process simulation is difficult to conduct high confidence simulation guidance on the ground, it takes a long time and cannot achieve online prediction. The traditional testing process fails to comprehensively collect and analyze key data such as hinge angle and tooling parameters, resulting in insufficient quality control of the deployment test.
A key parameter prediction system for satellite solar wing expansion experiments was designed, including data processing and management, prediction model construction, online prediction and decision support modules, and a deep learning algorithm was used to build a prediction model, combining simulation and measured data to perform data preprocessing and Optimize to realize online prediction and abnormal alarm for key parameters such as hinge angle, posture, and speed.
It improves the success rate and efficiency of the solar wing expansion test, realizes high-precision, full-process simulation guidance and data management, and supports real-time management and risk control of the expansion process.
Smart Images

Figure CN118981830A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of spacecraft test technology, and in particular to a system and method for predicting key parameters of a satellite solar wing deployment test. Background Art
[0002] With the rapid development of space technologies such as satellite earth observation, communication, space science and deep space exploration, the application of space deployable mechanisms such as solar arrays, synthetic aperture radar antennas, and space extension arms is becoming more and more widespread. Subject to the limitation of the carrier envelope, the space deployable mechanism completes the space deployment through a series of mechanical movements such as unlocking, releasing, unfolding, and locking. Its deployment reliability and deployment accuracy will directly affect the service performance of the satellite. The ground deployment test of the space mechanism is an important way to verify whether it can be successfully deployed in orbit. The deployment of the space mechanism can be divided into active drive, passive drive, and active-passive hybrid drive according to the driving mode; according to the rigidity and flexibility of the array surface deployed by the mechanism, it can be divided into rigid, semi-rigid, and flexible; according to the number of array surfaces of the mechanism, it can be divided into single-body and multi-body. The most commonly used passive multi-plate deployment mechanism for solar wings. At present, the design and simulation of the solar wing deployment test plan only consider the product's on-orbit deployment conditions, without considering the influence of ground factors. The confidence of the ground simulation is not high, the test criteria and boundary conditions of the deployment test are inaccurate, and the factors such as tooling, equipment, and environment in the actual process of the deployment test are not considered, and the accuracy of the prediction of the deployment test process cannot be guaranteed.
[0003] At present, both at home and abroad, the application of deep learning technology has been initially explored in the prediction of test parameters, hoping to predict test results through deep learning technology. The invention patent with publication number CN110057537A discloses a method for predicting the impact of aircraft aerodynamic performance. This method provides a method for predicting the impact of aerodynamic performance on simulated aircraft wind tunnel tests, and realizes the prediction of airway on startup performance parameters in wind tunnel tests.
[0004] The existing systems and methods are difficult to meet the needs of predicting key parameters of satellite solar wing tests. The main problems are as follows: ① The existing satellite solar wing deployment process simulation is mainly based on the design model, which makes it difficult to conduct high-confidence simulation to guide the optimization of process solutions for ground deployment tests; ② The existing solar wing deployment process simulation is time-consuming, and the deployment test simulation of each satellite model needs to be carried out separately, which makes it difficult to achieve online prediction of key parameters of solar wing deployment based on actual test conditions; ③ In the traditional deployment test process, the collection and record of key data such as hinge angle, tooling parameters, real-time posture and speed of the mechanism are single, and the product status data, tooling shape and position, support force unloading, friction, wind resistance and other data are not fully collected. The inherent coupling relationship between the data is not deeply explored, and the actual deployment state of the mechanism cannot be fully reflected, and there are deficiencies in the quality control and evaluation of the deployment test process.
[0005] Therefore, the present invention has developed a key parameter prediction system and method for a satellite solar wing deployment test, which realizes the test data processing and management of the satellite solar wing deployment process, the prediction model construction, and the online prediction and decision support modules, providing technical support and decision support for the satellite solar wing high-efficiency, high-precision, and high-reliability deployment test and test adjustment, thereby improving the one-time success rate of the solar wing deployment test and the deployment test efficiency. Summary of the invention
[0006] In view of the defects in the prior art, the object of the present invention is to provide a system and method for predicting key parameters of a satellite solar wing deployment test.
[0007] The key parameter prediction system for satellite solar wing deployment test provided by the present invention comprises:
[0008] Solar wing deployment test data processing and management module: used to read solar wing deployment test parameters, perform satellite attitude, installation accuracy, hanging force, torque deployment test simulation and preprocess historical test data, build key data models and output data sets for data management;
[0009] Solar wing deployment process prediction model construction module: used to read the solar wing deployment process simulation and test data sets from the solar wing deployment test data processing and management module, use deep learning algorithms to build a solar wing deployment process prediction model, and optimize the prediction model based on the deployment time and node posture key parameter indicators, and store and manage the prediction model;
[0010] Online prediction module for solar wing deployment test process: used to read the solar wing deployment test process data collected in real time from the solar wing deployment test data processing and management module, call the prediction model of the corresponding product from the solar wing deployment process prediction model construction module to input the measured data, and output the hinge angle, posture, speed, acceleration and deployment time prediction data set of the key nodes;
[0011] Solar wing deployment test process decision support module: used to read the prediction data set of key nodes from the online prediction module of the solar wing deployment test process, merge and analyze it, compare and analyze the prediction data with the deployment test target value, issue alarms for abnormal situations and guide the adjustment of the deployment test plan.
[0012] Preferably, the solar wing deployment test data processing and management module includes:
[0013] Solar wing deployment test data preprocessing module: used to read the simulation data and historical data of the solar wing deployment test, and complete the solar wing deployment test data preprocessing by using data cleaning, feature analysis, data normalization, and data set division steps to form a solar wing deployment test training data set;
[0014] Solar wing deployment test data model building module: used to call the solar wing deployment test data set, build the solar wing deployment process data model according to the needs of solar wing deployment process prediction, combined with the key parameters, key tooling and equipment of the solar wing deployment process;
[0015] Solar wing deployment test data management module: used to combine the solar wing deployment process data model, and complete the storage and management of measured data, simulation data, training data, and output data in the solar wing deployment test according to the satellite model, solar wing type, test time, and data type.
[0016] Preferably, the solar wing deployment process prediction model building module includes:
[0017] Solar wing deployment process prediction model construction and training module: used to call the training set in the solar wing deployment test training data set in the solar wing deployment test data management module, use bp neural network for deep learning and training, and pre-train in the form of multi-layer autoencoder neural network, which is divided into input layer, output layer and hidden layer; the input layer input data includes gravity unloading force, ambient temperature and air floating platform flatness; the hidden layer is linked by weights and biases to complete the dimensional unification of input and output; the output layer directly outputs the key parameters of the deployment process, including the solar wing deployment time, hinge degree and the position, speed and acceleration of the key nodes; establish a prediction model for key parameters, and solve the parameters and of the model;
[0018] Solar wing deployment process prediction model optimization module: used to call the test set in the solar wing deployment test training data set in the solar wing deployment test data management module to perform parameter optimization and model correction of the prediction model. The parameter optimization is judged by defining the loss function LP, and the expression is: LP = |PP R | 1 , where |*| 1 is the 1 norm, P is the prediction result, P R To collect data in practice, the prediction model is iterated and calculated until LP meets the requirements, completing parameter optimization and model correction;
[0019] Solar wing deployment process prediction model management module: used to manage various optimized solar wing deployment test prediction models, and classify and store prediction models based on solar wing configuration, number of panels, and size parameters.
[0020] Preferably, the online prediction module for the solar wing deployment test process includes:
[0021] Solar wing deployment parameter online prediction module: used to read the actual deployment test data collected in real time, call the solar wing prediction model of the corresponding configuration, and complete the online prediction of the solar wing deployment time, hinge degree, and the position, speed, and acceleration of key nodes;
[0022] Prediction data management module: used to save the predicted solar wing deployment test data, which is saved according to the solar wing configuration and test time.
[0023] Preferably, the solar wing deployment test process decision support module includes:
[0024] Abnormal alarm module for solar wing deployment process: used to read the hinge angle, displacement, speed, acceleration, and deployment time prediction data of key nodes from the online prediction module of the solar wing deployment test process, merge and analyze them, compare and analyze the prediction data with the deployment test target value, and issue an alarm and handle abnormal situations, including:
[0025] For the hinge angle of the key node, the predicted angle at each moment is compared with the target parameter to determine whether the unfolding angle exceeds the preset threshold. If so, the corresponding parameter is adjusted;
[0026] For the displacement, velocity, and acceleration motion parameters, the prediction curve is drawn and compared with the target curve to determine whether there is an out-of-tolerance or mutation. If so, the corresponding parameters are adjusted;
[0027] For the deployment time, the predicted time is compared with the design index time to determine whether the deployment time meets the design index. If it deviates from the design index, the corresponding parameters are adjusted;
[0028] Solar wing deployment process prediction result visualization module: used to read the hinge angle, posture, speed, acceleration, and deployment time prediction data sets of key nodes from the online prediction module of the solar wing deployment test process, merge and analyze them, and visualize the prediction results of the solar wing deployment process through two-dimensional charts and three-dimensional models, and support the rapid export of prediction results.
[0029] The method for predicting key parameters of a satellite solar wing deployment test provided by the present invention includes:
[0030] Solar wing deployment test data processing and management steps: read solar wing deployment test parameters, perform satellite attitude, installation accuracy, hanging force, torque deployment test simulation and preprocess historical test data, build key data models and output data sets for data management;
[0031] Steps for building a prediction model for the solar wing deployment process: read the simulation and test data sets of the solar wing deployment process, use a deep learning algorithm to build a prediction model for the solar wing deployment process, optimize the prediction model based on the deployment time and key parameter indicators of the node position, and store and manage the prediction model;
[0032] Online prediction steps of the solar wing deployment test process: read the real-time collected solar wing deployment test process data, call the prediction model of the corresponding product to input the measured data, and output the hinge angle, posture, speed, acceleration and deployment time prediction data set of the key nodes;
[0033] Decision support steps for the solar wing deployment test process: read the predicted data sets of key nodes for combined analysis, compare and analyze the predicted data with the deployment test target values, issue alarms for abnormal situations and guide adjustments to the deployment test plan.
[0034] Preferably, the solar wing deployment test data processing and management steps include:
[0035] Solar wing deployment test data preprocessing steps: read the simulation data and historical data of the solar wing deployment test, perform data cleaning, feature analysis, data normalization and data set division, complete the solar wing deployment test data preprocessing, and form a solar wing deployment test training data set;
[0036] Steps for constructing the solar wing deployment test data model: call the solar wing deployment test data set, and build the solar wing deployment process data model based on the needs of solar wing deployment process prediction, combined with the key parameters, key tooling and equipment of the solar wing deployment process;
[0037] Solar wing deployment test data management steps: Combined with the solar wing deployment process data model, the storage and management of measured data, simulation data, training data, and output data in the solar wing deployment test are completed according to the satellite model, solar wing type, test time, and data type.
[0038] Preferably, the solar wing deployment process prediction model building step includes:
[0039] The steps of building and training the prediction model of the solar wing deployment process are as follows: calling the training set in the solar wing deployment test training data set, using the BP neural network for deep learning and training, and pre-training in the form of a multi-layer autoencoder neural network, which is divided into an input layer, an output layer, and a hidden layer; the input layer inputs data including gravity unloading force, ambient temperature, and the flatness of the air-floating platform; the hidden layer is linked by weights and biases to unify the dimensions of input and output; the output layer directly outputs the key parameters of the deployment process, including the solar wing deployment time, hinge degree, and the position, velocity, and acceleration of the key nodes; a prediction model for key parameters is established, and the parameters and of the model are solved;
[0040] Optimization steps of the prediction model for the solar wing deployment process: call the test set in the solar wing deployment test training data set to perform parameter optimization and model correction of the prediction model. The parameter optimization is judged by defining the loss function LP, and the expression is: LP = |PP R | 1, where |*| 1 is the 1 norm, P is the prediction result, P R To collect data in practice, the prediction model is iterated and calculated until LP meets the requirements, completing parameter optimization and model correction;
[0041] Solar wing deployment process prediction model management steps: manage various optimized solar wing deployment test prediction models, and classify and store the prediction models based on solar wing configuration, number of panels, and size parameters.
[0042] Preferably, the online prediction step of the solar wing deployment test process includes:
[0043] Online prediction steps for solar wing deployment parameters: read the actual deployment test data collected in real time, call the solar wing prediction model of the corresponding configuration, and complete the online prediction of the solar wing deployment time, hinge degree, and the position, speed, and acceleration of key nodes;
[0044] Prediction data management steps: Save the predicted solar wing deployment test data according to the solar wing configuration and test time.
[0045] Preferably, the solar wing deployment test process decision support steps include:
[0046] Abnormal alarm steps for the solar wing deployment process: read the hinge angle, displacement, speed, acceleration, and deployment time prediction data of key nodes, merge and analyze them, compare and analyze the prediction data with the deployment test target value, and alarm and handle abnormal situations, including:
[0047] For the hinge angle of the key node, the predicted angle at each moment is compared with the target parameter to determine whether the unfolding angle exceeds the preset threshold. If so, the corresponding parameter is adjusted;
[0048] For the displacement, velocity, and acceleration motion parameters, the prediction curve is drawn and compared with the target curve to determine whether there is an out-of-tolerance or mutation. If so, the corresponding parameters are adjusted;
[0049] For the deployment time, the predicted time is compared with the design index time to determine whether the deployment time meets the design index. If it deviates from the design index, the corresponding parameters are adjusted;
[0050] Steps for visualizing the prediction results of the solar wing deployment process: read the hinge angle, posture, speed, acceleration, and deployment time prediction data sets of key nodes, merge and analyze them, and visualize the prediction results of the solar wing deployment process through two-dimensional charts and three-dimensional models, and support the rapid export of prediction results.
[0051] Compared with the prior art, the present invention has the following beneficial effects:
[0052] 1. In view of the traditional method, most of the simulations before the deployment of satellite solar wings are simulations of ideal conditions on orbit, and the test criteria and boundary conditions of the deployment test are inaccurate. The present invention comprehensively considers factors such as tooling, equipment, and environment during the actual deployment test process, and conducts high-precision, full-process, high-efficiency, and high-confidence simulations that take into account ground deployment test conditions to guide process optimization;
[0053] 2. In view of the problem that the amount of data from traditional satellite solar wing deployment tests is small and difficult to support the training of prediction models, the present invention adopts a method of combining simulation data with actual data to train prediction models, thereby ensuring the efficiency and accuracy of prediction model training and optimization;
[0054] 3. The traditional satellite solar wing deployment test lacks effective, real-time and complete data collection of key parameters such as hinge angle and tooling parameters. The present invention constructs online prediction of key parameters of solar wing deployment test through deep learning technology, and realizes full-process and multi-factor online control of key parameters to support abnormal warning and risk control of the deployment process;
[0055] 4. Compared with the traditional method that requires three-dimensional modeling and simulation analysis for each model, the present invention adopts deep learning technology to construct a prediction model for key parameters of satellite solar wing deployment tests according to the solar wing configuration classification, which realizes the rapid prediction of key parameters of solar wing deployment tests of various configurations such as deployment time, hinge degree, key point posture, speed, etc., supports the rapid adjustment of the deployment test process plan and the test site in advance, and effectively improves the test efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:
[0057] Figure 1 It is a composition diagram of a key parameter prediction system for a satellite solar wing deployment test provided by an embodiment of the present invention;
[0058] Figure 2 It is a flow chart of a method for predicting key parameters of a satellite solar wing deployment test provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0059] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.
[0060] Example
[0061] Figure 1 1 is a diagram showing the composition of a key parameter prediction system for a satellite solar wing deployment test provided by an embodiment of the present invention. Figure 1 As shown, the satellite deployable mechanism deployment test control system based on digital twin includes a solar wing deployment test data processing and management module, a solar wing deployment process prediction model construction module, a solar wing deployment test process online prediction module, and a solar wing deployment test process decision support module.
[0062] The solar wing deployment test data processing and management module is used to read the solar wing deployment test parameters, perform deployment test simulation and historical test data preprocessing such as the attitude of the star body, adjustment accuracy, hanging force, torque, etc., build key data models and output data sets for data management. It further includes a solar wing deployment test data preprocessing module, a solar wing deployment test data model building module, and a solar wing deployment test data management module. The solar wing deployment test data preprocessing module is used to read the simulation data and historical data of the solar wing deployment test, and use data cleaning, feature analysis, data normalization, and data set division steps to complete the solar wing deployment test data preprocessing to form a solar wing deployment test training data set; the solar wing deployment test data model building module is used to call the solar wing deployment test data set, and build a solar wing deployment process data model M according to the needs of the solar wing deployment process prediction, combined with the key parameters, key tooling and equipment of the solar wing deployment process. z , the expression is: M z ={M P ,M F ,M D}, where M P represents the key parameters of the solar wing deployment process, M F Indicates key tooling parameters, M D Represents key equipment parameters; the solar wing deployment test data management module is used to combine the solar wing deployment process data model to complete the storage and management of measured data, simulation data, training data, and output data in the solar wing deployment test according to the satellite model, solar wing type, test time, data type, etc.
[0063] The solar wing deployment process prediction model construction module is used to read the solar wing deployment process simulation and test data set from the solar wing deployment test data processing and management module, use the deep learning algorithm to build the solar wing deployment process prediction model, and optimize the prediction model based on key parameter indicators such as deployment time and node posture, and store and manage the prediction model. It further includes a solar wing deployment process prediction model construction and training module, a solar wing deployment process prediction model optimization module and a solar wing deployment process prediction model management module. The solar wing deployment process prediction model construction and training module is used to call the training set in the solar wing deployment test training data set in the solar wing deployment test data management module, use the bp neural network for deep learning and training, and form a prediction model for the solar wing deployment time, hinge degree, and the posture, speed, and acceleration of key nodes;
[0064] The BP neural network is pre-trained in the form of a multi-layer autoencoder neural network, which is divided into an input layer, an output layer and a hidden layer. SI Including gravity unloading force Fg, ambient temperature T and air-floating platform flatness PP, it is regarded as a 1*i matrix, i represents the number of input parameter types, and the expression is:
[0065] X SI ={X SFg , X ST , X SPP}
[0066]
[0067] The hidden layer is connected through weight W and bias b to unify the dimensions of input and output;
[0068] The output layer directly outputs the key prediction parameters EP of the deployment process. The key parameter set EP includes the solar wing deployment time Tp, hinge degree θhp, and the position PiLp (Xpi, Ypi, Zpi), speed PVpi, and acceleration PApi of the key node Pi;
[0069] The prediction model MP of the key parameter EP is established, and the parameters β and ε of the model are solved, namely:
[0070] EP=Xβ+ε
[0071]
[0072] The solar wing deployment process prediction model optimization module is used to call the test set in the solar wing deployment test training data set in the solar wing deployment test data management module to perform parameter optimization and model correction of the prediction model. The parameter optimization is judged by defining the loss function LP, that is:
[0073] LP=|PP R | 1
[0074] In the formula, |*| 1 is the 1 norm, P is the prediction result, P R To collect data in practice, the prediction model is iterated and calculated until LP meets the requirements, completing parameter optimization and model correction;
[0075] The solar wing deployment process prediction model management module is used to manage various optimized solar wing deployment test prediction models, and to classify and store the prediction models based on parameters such as solar wing configuration, number of panels, and size.
[0076] The online prediction module of the solar wing deployment test process is used to read the solar wing deployment test process data collected in real time from the solar wing deployment test data processing and management module, call the prediction model of the corresponding product from the solar wing deployment process prediction model construction module to input the measured data, and output the prediction data set such as the hinge angle, posture, speed, acceleration and deployment time of the key nodes. It further includes a solar wing deployment parameter online prediction module and a prediction data management module. Among them, the solar wing deployment parameter online prediction module is used to read the actual deployment test data collected in real time, call the solar wing prediction model of the corresponding configuration, and complete the online prediction of key parameters such as the solar wing deployment time, hinge degree, and the posture, speed, acceleration of the key nodes; the prediction data management module is used to save the predicted solar wing deployment test data, and save it according to the solar wing configuration, test time, etc.
[0077] Solar wing deployment parameter online prediction module: used to read the actual deployment test data collected in real time RI , call the solar wing prediction model MP of the corresponding configuration (configuration number is c) c , complete the online prediction of key parameters such as the solar wing deployment time Tp, hinge degree θhp, and the position PiLp (Xpi, Ypi, Zpi), speed PVpi, acceleration PApi of the key node Pi, that is:
[0078] EP=X RI β c +ε c
[0079]
[0080] The solar wing deployment test process decision support module is used to read the prediction data sets of force, position, speed, acceleration, deployment time, etc. of key nodes from the online prediction module of the solar wing deployment test process, and merge and analyze them, compare and analyze the prediction data with the deployment test target value, give an alarm for abnormal situations and guide the adjustment of the deployment test plan, which further includes the solar wing deployment process abnormal alarm module and the solar wing deployment process prediction result visualization module. Among them, the solar wing deployment process abnormal alarm module is used to read the prediction data sets of hinge angle, position, speed, acceleration, deployment time, etc. of key nodes from the online prediction module of the solar wing deployment test process, and merge and analyze them, compare and analyze the prediction data with the deployment test target value, and give an alarm for abnormal situations; the solar wing deployment process prediction result visualization module is used to read the prediction data sets of hinge angle, position, speed, acceleration, deployment time, etc. of key nodes from the online prediction module of the solar wing deployment test process, and merge and analyze them, and realize the visualization of the prediction results of the solar wing deployment process through two-dimensional charts and three-dimensional models, and support the rapid export of prediction results.
[0081] When the present invention is working, business and data integration is realized among the modules: the solar wing deployment process prediction model construction module reads the training set composed of simulation data and historical test data of the satellite solar wing deployment test from the solar wing deployment test data processing and management module, constructs prediction models of the deployment time, hinge degree and posture, speed and acceleration of various solar wing types, and calls the test set in the solar wing deployment test training data set to optimize the parameters of the prediction model and correct the model, so as to form the deployment process prediction model of various solar wings, and save it to the solar wing deployment process prediction model management module; the solar wing deployment parameter online prediction module reads the real-time acquired data from the solar wing deployment test data processing and management module The actual deployment test data of the set is collected, and the deployment process prediction models of various solar wings are called from the solar wing deployment process prediction model construction module to output the key parameters such as solar wing deployment time, hinge degree, and position, speed, acceleration of key nodes, and transmit them to the solar wing deployment test process decision support module; the solar wing deployment test process decision support module reads the prediction data set such as hinge angle, position, speed, acceleration, deployment time, etc. of key nodes from the solar wing deployment parameter online prediction module, compares and analyzes the prediction data with the deployment test target value, outputs the prediction data through two-dimensional charts and three-dimensional models, and passes the prediction and alarm data to the solar wing deployment test data processing and management module.
[0082] The present invention also provides a satellite deployable mechanism deployment test control method based on digital twin, referring to Figure 2 As shown, the specific steps of this method are as follows:
[0083] Step 1, using the solar wing deployment test data processing and management module to read the simulation data and historical test data of the satellite solar wing deployment test, and preprocessing the test data such as friction coefficient μ, driving torque T, friction resistance torque Mf, hinge degree θh, key point Pi pose PiL (Xi, Yi, Zi), speed PVi, acceleration PAi, deployment time T, etc., including data cleaning, feature analysis, data normalization and data set division, to form a solar wing deployment test data set;
[0084] Step 2: Use the mechanism deployment virtual-real fusion simulation model construction module to call the solar wing deployment test data set, and build a solar wing deployment process data model based on the needs of solar wing deployment process prediction, combined with solar wing type, deployment process key parameters, key tooling and equipment, including measured data, simulation data and output data;
[0085] Step 3: Encapsulate various solar wing deployment process test data sets according to solar wing configurations to form a solar wing deployment test data collection TD, including a training set and a test set, and manage them according to the solar wing configuration in a solar wing deployment test data management module;
[0086] Step 4, using the solar wing deployment process prediction model construction and training module, calling the training set in the solar wing deployment test data set in the solar wing deployment test data management module, using the BP neural network for deep learning and training, building prediction models for the deployment time, hinge degree, and posture, speed, and acceleration of various solar wing types, and calling the test set in the solar wing deployment test training data set to optimize the parameters of the prediction model and correct the model, so as to form a deployment process prediction model for various solar wings;
[0087] Step 5, classifying various solar wing deployment process prediction models according to the solar wing configuration, and storing the prediction models in the solar wing deployment process prediction model management module with parameters such as solar wing configuration, number of panels, and size;
[0088] Step 6, using the solar wing deployment parameter online prediction module, read the actual solar wing deployment test data collected in real time, call the solar wing prediction model of the corresponding configuration from the solar wing deployment process prediction model management module according to the solar wing configuration, complete the online prediction of key parameters such as the solar wing deployment time Tp, hinge degree θhp, and the position PiLp (Xpi, Ypi, Zpi), speed PVpi, acceleration PApi of the key node Pi, form the solar wing deployment test prediction data collection PD, and save and manage the solar wing deployment test data in the prediction data management module according to the solar wing configuration, test time, etc.;
[0089] Step 7, using the abnormal alarm module of the solar wing deployment process, read and analyze the prediction data set PD of the hinge angle, posture, speed, acceleration, deployment time and other key nodes from the online prediction module of the solar wing deployment test process, compare and analyze the prediction data PD with the deployment test target value, and issue an alarm for abnormal situations;
[0090] Step 8, using the solar wing deployment process prediction result visualization module, reads the prediction data sets such as hinge angle, position, speed, acceleration, deployment time, etc. of key nodes from the solar wing deployment test process online prediction module, and merges and analyzes them. Through two-dimensional charts and three-dimensional models, the prediction results of the solar wing deployment process are visualized, and the prediction results can be quickly exported.
[0091] Those skilled in the art know that, in addition to implementing the system, device and its various modules provided by the present invention in a purely computer-readable program code, it is entirely possible to implement the same program in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers and embedded microcontrollers by logically programming the method steps. Therefore, the system, device and its various modules provided by the present invention can be considered as a hardware component, and the modules included therein for implementing various programs can also be considered as structures within the hardware component; the modules for implementing various functions can also be considered as both software programs for implementing the method and structures within the hardware component.
[0092] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.
Claims
1. A satellite solar wing deployment test key parameter prediction system, characterized in that: include: Solar wing deployment test data processing and management module: used to read solar wing deployment test parameters, perform satellite attitude, installation accuracy, hanging force, torque deployment test simulation and preprocess historical test data, build key data models and output data sets for data management; Solar wing deployment process prediction model construction module: used to read the solar wing deployment process simulation and test data sets from the solar wing deployment test data processing and management module, use deep learning algorithms to build a solar wing deployment process prediction model, and optimize the prediction model based on the deployment time and node posture key parameter indicators, and store and manage the prediction model; Online prediction module for solar wing deployment test process: used to read the solar wing deployment test process data collected in real time from the solar wing deployment test data processing and management module, call the prediction model of the corresponding product from the solar wing deployment process prediction model construction module to input the measured data, and output the hinge angle, posture, speed, acceleration and deployment time prediction data set of the key nodes; Solar wing deployment test process decision support module: used to read the prediction data set of key nodes from the online prediction module of the solar wing deployment test process, merge and analyze it, compare and analyze the prediction data with the deployment test target value, issue alarms for abnormal situations and guide the adjustment of the deployment test plan.
2. The satellite solar wing deployment test key parameter prediction system according to claim 1 is characterized in that: The solar wing deployment test data processing and management module includes: Solar wing deployment test data preprocessing module: used to read the simulation data and historical data of the solar wing deployment test, and complete the solar wing deployment test data preprocessing by using data cleaning, feature analysis, data normalization, and data set division steps to form a solar wing deployment test training data set; Solar wing deployment test data model building module: used to call the solar wing deployment test data set, build the solar wing deployment process data model according to the needs of solar wing deployment process prediction, combined with the key parameters, key tooling and equipment of the solar wing deployment process; Solar wing deployment test data management module: used to combine the solar wing deployment process data model, and complete the storage and management of measured data, simulation data, training data, and output data in the solar wing deployment test according to the satellite model, solar wing type, test time, and data type.
3. The satellite solar wing deployment test key parameter prediction system according to claim 1 is characterized in that: The solar wing deployment process prediction model building module includes: Solar wing deployment process prediction model construction and training module: used to call the training set in the solar wing deployment test training data set in the solar wing deployment test data management module, use bp neural network for deep learning and training, and pre-train in the form of multi-layer autoencoder neural network, which is divided into input layer, output layer and hidden layer; the input layer input data includes gravity unloading force, ambient temperature and air floating platform flatness; the hidden layer is linked by weights and biases to complete the dimensional unification of input and output; the output layer directly outputs the key parameters of the deployment process, including the solar wing deployment time, hinge degree and the position, speed and acceleration of the key nodes; establish a prediction model for key parameters, and solve the parameters and of the model; Solar wing deployment process prediction model optimization module: used to call the test set in the solar wing deployment test training data set in the solar wing deployment test data management module to perform parameter optimization and model correction of the prediction model. The parameter optimization is judged by defining the loss function LP, and the expression is: LP = |PP R |1, where |*|1 is the 1 norm, P is the prediction result, P R To collect data in practice, the prediction model is iterated and calculated until LP meets the requirements, completing parameter optimization and model correction; Solar wing deployment process prediction model management module: used to manage various optimized solar wing deployment test prediction models, and classify and store prediction models based on solar wing configuration, number of panels, and size parameters.
4. The satellite solar wing deployment test key parameter prediction system according to claim 1 is characterized in that: The online prediction module of the solar wing deployment test process includes: Solar wing deployment parameter online prediction module: used to read the actual deployment test data collected in real time, call the solar wing prediction model of the corresponding configuration, and complete the online prediction of the solar wing deployment time, hinge degree, and the position, speed, and acceleration of key nodes; Prediction data management module: used to save the predicted solar wing deployment test data, which is saved according to the solar wing configuration and test time.
5. The satellite solar wing deployment test key parameter prediction system according to claim 1 is characterized in that: The solar wing deployment test process decision support module includes: Abnormal alarm module for solar wing deployment process: used to read the hinge angle, displacement, speed, acceleration, and deployment time prediction data of key nodes from the online prediction module of the solar wing deployment test process, merge and analyze them, compare and analyze the prediction data with the deployment test target value, and issue an alarm and handle abnormal situations, including: For the hinge angle of the key node, the predicted angle at each moment is compared with the target parameter to determine whether the unfolding angle exceeds the preset threshold. If so, the corresponding parameter is adjusted; For the displacement, velocity, and acceleration motion parameters, the prediction curve is drawn and compared with the target curve to determine whether there is an out-of-tolerance or mutation. If so, the corresponding parameters are adjusted; For the deployment time, the predicted time is compared with the design index time to determine whether the deployment time meets the design index. If it deviates from the design index, the corresponding parameters are adjusted; Solar wing deployment process prediction result visualization module: used to read the hinge angle, posture, speed, acceleration, and deployment time prediction data sets of key nodes from the online prediction module of the solar wing deployment test process, merge and analyze them, and visualize the prediction results of the solar wing deployment process through two-dimensional charts and three-dimensional models, and support the rapid export of prediction results.
6. A method for predicting key parameters of a satellite solar wing deployment test, characterized in that: include: Solar wing deployment test data processing and management steps: read solar wing deployment test parameters, perform satellite attitude, installation accuracy, hanging force, torque deployment test simulation and preprocess historical test data, build key data models and output data sets for data management; Steps for building a prediction model for the solar wing deployment process: read the simulation and test data sets of the solar wing deployment process, use a deep learning algorithm to build a prediction model for the solar wing deployment process, optimize the prediction model based on the deployment time and key parameter indicators of the node position, and store and manage the prediction model; Online prediction steps of the solar wing deployment test process: read the real-time collected solar wing deployment test process data, call the prediction model of the corresponding product to input the measured data, and output the hinge angle, posture, speed, acceleration and deployment time prediction data set of the key nodes; Decision support steps for the solar wing deployment test process: read the predicted data sets of key nodes for combined analysis, compare and analyze the predicted data with the deployment test target values, issue alarms for abnormal situations and guide adjustments to the deployment test plan.
7. The method for predicting key parameters of a satellite solar wing deployment test according to claim 6, characterized in that: The solar wing deployment test data processing and management steps include: Solar wing deployment test data preprocessing steps: read the simulation data and historical data of the solar wing deployment test, perform data cleaning, feature analysis, data normalization and data set division, complete the solar wing deployment test data preprocessing, and form a solar wing deployment test training data set; Steps for constructing the solar wing deployment test data model: call the solar wing deployment test data set, and build the solar wing deployment process data model based on the needs of solar wing deployment process prediction, combined with the key parameters, key tooling and equipment of the solar wing deployment process; Solar wing deployment test data management steps: Combined with the solar wing deployment process data model, the storage and management of measured data, simulation data, training data, and output data in the solar wing deployment test are completed according to the satellite model, solar wing type, test time, and data type.
8. The method for predicting key parameters of a satellite solar wing deployment test according to claim 6, characterized in that: The solar wing deployment process prediction model construction step includes: The steps of building and training the prediction model of the solar wing deployment process are as follows: calling the training set in the solar wing deployment test training data set, using the BP neural network for deep learning and training, and pre-training in the form of a multi-layer autoencoder neural network, which is divided into an input layer, an output layer, and a hidden layer; the input layer inputs data including gravity unloading force, ambient temperature, and the flatness of the air-floating platform; the hidden layer is linked by weights and biases to unify the dimensions of input and output; the output layer directly outputs the key parameters of the deployment process, including the solar wing deployment time, hinge degree, and the position, velocity, and acceleration of the key nodes; a prediction model for key parameters is established, and the parameters and of the model are solved; Optimization steps of the prediction model for the solar wing deployment process: call the test set in the solar wing deployment test training data set to perform parameter optimization and model correction of the prediction model. The parameter optimization is judged by defining the loss function LP, and the expression is: LP = |PP R |1, where |*|1 is the 1 norm, P is the prediction result, P R To collect data in practice, the prediction model is iterated and calculated until LP meets the requirements, completing parameter optimization and model correction; Solar wing deployment process prediction model management steps: manage various optimized solar wing deployment test prediction models, and classify and store the prediction models based on solar wing configuration, number of panels, and size parameters.
9. The method for predicting key parameters of a satellite solar wing deployment test according to claim 6, characterized in that: The online prediction step of the solar wing deployment test process includes: Online prediction steps for solar wing deployment parameters: read the actual deployment test data collected in real time, call the solar wing prediction model of the corresponding configuration, and complete the online prediction of the solar wing deployment time, hinge degree, and the position, speed, and acceleration of key nodes; Prediction data management steps: Save the predicted solar wing deployment test data according to the solar wing configuration and test time.
10. The method for predicting key parameters of a satellite solar wing deployment test according to claim 6, characterized in that: The solar wing deployment test process decision support steps include: Abnormal alarm steps for the solar wing deployment process: read the hinge angle, displacement, speed, acceleration, and deployment time prediction data of key nodes, merge and analyze them, compare and analyze the prediction data with the deployment test target value, and alarm and handle abnormal situations, including: For the hinge angle of the key node, the predicted angle at each moment is compared with the target parameter to determine whether the unfolding angle exceeds the preset threshold. If so, the corresponding parameter is adjusted; For the displacement, velocity, and acceleration motion parameters, the prediction curve is drawn and compared with the target curve to determine whether there is an out-of-tolerance or mutation. If so, the corresponding parameters are adjusted; For the deployment time, the predicted time is compared with the design index time to determine whether the deployment time meets the design index. If it deviates from the design index, the corresponding parameters are adjusted; Steps for visualizing the prediction results of the solar wing deployment process: read the hinge angle, posture, speed, acceleration, and deployment time prediction data sets of key nodes, merge and analyze them, and visualize the prediction results of the solar wing deployment process through two-dimensional charts and three-dimensional models, and support the rapid export of prediction results.
Citation Information
Cited By
Solar wing unfolding state detection method and device, electronic equipment and storage medium
CN119887704A
Solar wing battery selecting and matching method and system based on three-dimensional layout
CN120509112A
A method and system for selecting solar array cells based on three-dimensional layout
CN120509112B
Unfolding and folding attitude measurement system and method for space flexible film deployable structure
CN121252754A
