Satellite composite material truss performance evaluation method, equipment, medium and product

Abstract

The invention discloses a satellite composite material truss performance evaluation method and device, a medium and a product, and relates to the field of large space structure composite material truss design and manufacture, and the method comprises the steps: building a composite material truss performance rapid evaluation model based on a physical information neural network; obtaining grid node coordinates of the to-be-tested satellite composite truss and geometric structure optimization variable parameters of the to-be-tested truss; and inputting the truss grid node coordinates of the to-be-tested satellite composite material truss and the geometric structure optimization variable parameters of the to-be-tested truss into the composite material truss performance rapid evaluation model to obtain a satellite composite material truss performance evaluation result. Through the composite material truss performance rapid evaluation model constructed based on the physical information neural grid, the mechanical performance of the satellite composite material truss can be calculated in real time, and high interpretability is achieved.

Description

Technical Field

[0001] The present application relates to the field of design and manufacture of composite trusses for large space structures, and in particular to a method, equipment, medium and product for evaluating the performance of composite trusses for satellites. Background Art

[0002] In recent years, with the increasing requirements for satellite signal strength and accuracy, it is necessary to design satellite antennas with larger sizes and higher structural strength. Composite trusses are a lightweight structural system made of materials such as PEEK (polyetheretherketone). They are usually based on triangular units and connect various rods through nodes to form a stable frame structure. They are an important supporting structure for satellite antennas.

[0003] Traditional satellite composite truss performance evaluation usually uses data-driven proxy models or traditional CAE (Computer Aided Engineering) software to drive parametric models through upper-level software to achieve rapid parametric modeling and perform mechanical performance analysis on the established model for further optimization.

[0004] However, the composite truss itself has a complex topological structure, numerous stress nodes, harsh working environment, and demanding performance requirements, which leads to the following problems in the design method based on the known performance evaluation method:

[0005] (1) Slow performance evaluation: When designing composite truss structures, the design needs to be modified hundreds or thousands of times. Finite element-based numerical simulation software needs to balance the contradiction between calculation speed and calculation accuracy. It is impossible to evaluate the mechanical properties of the new structure in real time, resulting in a long design cycle.

[0006] (2) A large amount of sample data is required: Data-driven proxy models can significantly improve simulation efficiency, but the model training process relies entirely on data, requiring a large amount of sample data to ensure the model's prediction accuracy. However, this makes it difficult to ensure that the model's prediction results conform to objective physical laws, and the interpretability is poor. Summary of the invention

[0007] The purpose of this application is to provide a satellite composite truss performance evaluation method, equipment, medium and product, which can determine the mechanical properties of the satellite composite truss in real time, ensure that the prediction results of the model conform to the objective physical laws, and have strong interpretability.

[0008] To achieve the above objectives, this application provides the following solutions:

[0009] In a first aspect, the present application provides a satellite composite truss performance evaluation method, the satellite composite truss performance evaluation method comprising:

[0010] Constructing a rapid performance evaluation model for composite trusses based on physical information neural networks;

[0011] Obtaining the grid node coordinates of the composite material truss of the satellite to be tested and the variable parameters of the geometric structure optimization of the truss to be tested; the variable parameters of the geometric structure optimization of the truss to be tested include the total length of the truss, the cross-sectional size of the truss, the thickness of the longitudinal beam, the width of the longitudinal beam, the pitch of the strip, the thickness of the strip and the width of the strip;

[0012] The truss grid node coordinates of the satellite composite truss to be tested and the optimized variable parameters of the geometric structure of the truss to be tested are input into the composite truss performance rapid evaluation model to obtain the satellite composite truss performance evaluation result.

[0013] Optionally, a composite truss performance rapid evaluation model is constructed based on a physical information neural network, specifically including:

[0014] The satellite composite truss grid node coordinates and truss geometric structure optimization variables are used as input, and the stress, strain and load at the truss grid node coordinates are used as output to form a sample pair;

[0015] Constructing a training set and a test set based on the sample pairs;

[0016] Obtaining physical information from neural networks;

[0017] A meta-learning algorithm is used to train and test the physical information neural network based on the training set and the test set until the trained physical information neural network meets the set requirements, and the trained physical information neural network is used as a rapid performance evaluation model for composite material trusses.

[0018] Optionally, before obtaining the physical information neural network, the method further includes:

[0019] Obtaining initial physical information neural network;

[0020] The loss function of the initial physical information neural network is modified to obtain the physical information neural network.

[0021] Optionally, before taking the satellite composite material truss grid node coordinates and the truss geometric structure optimization variables as inputs and taking the stress, strain and load at the truss grid node coordinates as outputs to form sample pairs, the method further includes:

[0022] Construct a parametric design model for composite trusses;

[0023] Based on the composite material truss parametric design model, the composite material truss grid node coordinates, composite material truss geometric structure optimization variables, and the stress, strain and load at the composite material truss grid node coordinates are obtained.

[0024] Optionally, a meta-learning algorithm is used to train and test the physical information neural network based on the training set and the test set until the trained physical information neural network meets the set requirements, and the trained physical information neural network is used as a composite material truss performance rapid evaluation model, specifically including:

[0025] Using a meta-learning algorithm, offline training is performed on the physical information neural network based on the training set and the test set to obtain the physical information neural network after offline training;

[0026] The meta-learning algorithm is used to adjust the model parameters θ of the offline trained physical information neural network online to obtain a rapid performance evaluation model for composite trusses.

[0027] Optionally, the offline training process includes: a meta-training phase and a meta-testing phase;

[0028] In a meta-training stage, the physical information neural network is iteratively trained based on the training set to obtain a meta-trained physical information neural network;

[0029] In the meta-testing stage, the meta-trained physical information neural network is tested based on the test set to obtain the physical information neural network after offline training.

[0030] Optionally, the modified loss function is expressed as:

[0031] Loss=Loss data +Loss bc +Loss pde ;

[0032] Among them, Loss data is the data error loss, Loss bc is the boundary condition loss, Loss pde is the physical constraint loss.

[0033] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any one of the above-described satellite composite truss performance evaluation methods.

[0034] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of any of the above-described satellite composite truss performance evaluation methods.

[0035] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of any one of the above-mentioned satellite composite truss performance evaluation methods.

[0036] According to the specific embodiments provided in this application, this application has the following technical effects:

[0037] The present application provides a satellite composite truss performance evaluation method, equipment, medium and product. By constructing a composite truss performance rapid evaluation model based on physical information neural network, it can solve the problem of requiring a large amount of sample data to ensure the prediction accuracy of the model during the model training process, and can make the prediction results of the model conform to objective physical laws and have strong interpretability. By obtaining the satellite composite truss data to be tested and inputting the satellite composite truss data to be tested into the composite truss performance rapid evaluation model, the satellite composite truss performance evaluation result is obtained, which can solve the problem of slow performance evaluation speed, so that the mechanical properties of the satellite composite truss can be determined in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A schematic diagram of a process flow of a satellite composite material truss performance evaluation method provided in one embodiment of the present application;

[0040] Figure 2 A schematic diagram of a process for constructing a composite material truss performance rapid evaluation model provided in an embodiment of the present application;

[0041] Figure 3 A physical information neural network structure diagram provided for an embodiment of the present application;

[0042] Figure 4 A diagram of a parametric design interface for a composite material truss provided in an embodiment of the present application;

[0043] Figure 5 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0045] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0046] The present application provides a method for evaluating the performance of a satellite composite truss. Figure 1 As shown, the method includes:

[0047] Step 100: Construct a composite truss performance rapid evaluation model based on a physical information neural network.

[0048] In practical applications, the process of building a rapid performance evaluation model for composite trusses is as follows: Figure 2 As shown, the specific implementation steps are as follows:

[0049] Step 1001: construct a composite material truss parametric design model.

[0050] The purpose of this step is to be able to quickly generate a CAD (Computer-Aided Design) model given specified geometric parameters.

[0051] In the initial stage of establishing the parametric design model of composite trusses, since the various parameters of the truss structure need to be changed many times during the optimization process of the truss structure, it is necessary to use Creo (CAD design software package) 3D modeling software for parametric modeling, and design each structural parameter as a variable in the modeling process. Corresponding constraints should be added between related design variables, and the range of design variables related to the composite truss design should be clarified, and finally the parametric design model of the composite truss is obtained. In the design, the required truss geometric structure optimization variables are shown in Table 1.

[0052] Table 1 Truss geometry optimization variable description

[0053] serial number Variable Name <![CDATA[k1]]> Truss total length <![CDATA[k2]]> Truss section size <![CDATA[k3]]> Stringer thickness <![CDATA[k4]]> Width of longitudinal beam <![CDATA[k5]]> Strip pitch <![CDATA[k6]]> Strip thickness <![CDATA[k7]]> Strip width

[0054] The interface for parametric design of composite trusses is as follows: Figure 4 As shown, during the design process, the values ​​of some variables can be changed, and the truss scheme after the changed parameters will be generated in real time on the design interface, providing a simple and user-friendly design interface to support the rapid design of satellite antennas.

[0055] Step 1002: Based on the composite material truss parametric design model, the composite material truss grid node coordinates, the composite material truss geometric structure optimization variables, and the load at the composite material truss grid node coordinates are p, the stress at the truss grid node coordinates are σ, and the strain at the truss grid node coordinates are ε, wherein the truss elastic mechanics control equation in the present application is:

[0056] σ=E·ε;

[0057]

[0058]

[0059] in, is the gradient operator, F is the body force, u is the displacement, E is the elastic modulus, is the transpose of the product of the gradient operator and the shift.

[0060] Step 1003: Taking the satellite composite material truss grid node coordinates and the truss geometric structure optimization variables as inputs, and taking the stress, strain and load at the truss grid node coordinates as outputs, a sample pair is formed.

[0061] The data set is a collection of complete physical field data, which stores the data of the physical field that can be measured for the entire truss. In the actual application process, according to the data obtained in step 1002, we stipulate that the three coordinates of the composite material truss grid node are x, y, and z, the load at the truss grid node coordinate is p, the stress at the truss grid node coordinate is σ, and the strain at the truss grid node coordinate is ε. Then, in this embodiment, this part of the data set sample pair is defined as Dataset:

[0062] Dataset={x,y,z,k1,k2,k3,k4,k5,k6,k7,p,σ,ε}.

[0063] Step 1004: construct a training set and a test set based on the sample pairs.

[0064] In actual application, the data set sample pairs obtained in step 1003 are divided into a training set and a test set according to a certain ratio.

[0065] Step 1005: Obtain an initial physical information neural network.

[0066] In this application, a physical information neural network method is used to evaluate the mechanical properties of composite trusses. This method combines deep learning and physics knowledge, uses a neural network driven by data and mechanism fusion, and combines actual measurement data and physical mechanisms. It can not only use the rich information in the data to capture the behavior of complex systems, but also ensure that the output of the model conforms to physical laws, thereby improving the accuracy and generalization ability of the model.

[0067] The satellite composite truss grid node coordinates and truss geometric structure optimization variables are used as the input of the initial physical information neural network, and the stress, strain and load at the truss grid node coordinates are used as the output of the initial physical information neural network. The number of hidden layers, activation function and other hyperparameters are further set. The structure diagram of the physical information neural network is shown in the figure. Figure 3 The hyper parameters of the physical information neural network are shown in Table 2.

[0068] Table 2 Physical information neural network hyperparameters

[0069] Hyperparameters Configuration Number of hidden layers 10 Number of nodes per layer 64 Activation Function Sigmoid(x)

[0070] Step 1006: modify the loss function of the initial physical information neural network to obtain the physical information neural network.

[0071] Add physical information terms to the loss function of the neural network, that is, the physical laws followed in reality. In this application, they are the control equations followed by the composite truss. The loss function in this application includes three parts: "data error term", "boundary condition term" and "physical constraint term". Among them, the "data error term" is used to measure the error between the output result of the neural network and a small amount of actual measured data values, usually expressed by mean square error (MSE); the "boundary condition term" is used to measure the consistency between the calculation results of the neural network and the physical boundary, also expressed by mean square error; "physical constraint" is used to measure the consistency between the calculation results of the neural network and the physical mechanism, usually expressed by the residual of the physical constraint. In the process of model training, not only the data error should be minimized, but also the boundary error and physical constraint error should be minimized.

[0072] The "data error term" in the loss function is defined as follows:

[0073]

[0074] Where i is the i-th node of the satellite composite truss grid, (x i ,y i ,z i ) is the coordinate of the ith node of the satellite composite truss grid, u(x i ,yi ,z i ) is the prediction result of the model, u data (x i ,y i ,z i ) is the real data value, N data is the number of data points.

[0075] The "boundary condition term" in the loss function is defined as follows:

[0076]

[0077] Among them, u bc (x i ,y i ,z i ) is the real data value on the boundary, N bc is the number of data points on the boundary.

[0078] The "physical constraint term" in the loss function is defined as the accumulation of the residuals of the control equation, and the expression is as follows:

[0079]

[0080] Among them, N pde is the number of data points.

[0081] The loss function Loss of the physical information neural network in this application is:

[0082] Loss=Loss data +Loss bc +Loss pde .

[0083] In practical applications, it is generally necessary to adjust the parameters of the physical information neural network according to the gradient information of the loss function through the optimizer, and minimize the loss function by gradually updating the weights, so that the neural network can gradually improve its performance and accuracy. It is also necessary to set the learning rate reasonably to ensure the stability and convergence speed of the training process. It is also necessary to set the parameters of the optimizer. The detailed parameters of the optimization are shown in Table 3.

[0084] Table 3 Optimized hyperparameters

[0085] Hyperparameters Configuration Learning Rate 0.001 Optimizer AdamOptimizer

[0086] Step 1007: adopt a meta-learning algorithm to train and test the physical information neural network based on the training set and the test set until the trained physical information neural network meets the set requirements, and use the trained physical information neural network as a composite material truss performance rapid evaluation model.

[0087] As we all know, meta-learning aims to design models that can quickly adapt to new tasks from a small amount of experience. This application uses the method of meta-learning to perform offline training and online fine-tuning of physical information neural networks. First, based on the constructed physical information neural network, the control equations and boundary conditions under different working conditions are integrated into the meta-learning framework, and offline training is performed through the set loss function to obtain a set of initial model parameters with strong generalization performance. Secondly, the composite truss performance evaluation system will face different working conditions during the application process. When facing new working conditions, the control equations, boundary conditions and initial conditions in the real physics followed by the model may change. At this time, it is necessary to quickly fine-tune the parameters of the meta-learning model online to obtain a meta-learning model that can quickly adapt to new working conditions, and finally realize the training of the model.

[0088] The offline training process goes through two phases: meta-training and meta-testing. The meta-training phase involves training the physical neural network to learn the ability to extract and generalize knowledge from multiple training tasks, and usually includes multiple iterations, each of which evaluates the performance of the truss model under a set of specific working conditions to optimize the parameters of the physical neural network and enhance the generalization of the model so that it can quickly learn on new tasks. The meta-testing phase is performed after meta-training and is used to evaluate the performance of the physical neural network on new tasks. Through this process, the meta-model is further optimized so that it can quickly adapt to new tasks.

[0089] In practical applications, the specific process of offline training is as follows:

[0090] First, the meta-learning model parameters θ of the physical information neural network are randomly initialized. The meta-training phase uses the data of the training set to optimize the base learner through the physical information neural network. The model parameters θ' are continuously updated through the gradient descent algorithm, and the above steps are repeated until the number of tasks reaches the batch size. The gradient descent algorithm is as follows:

[0091]

[0092] Among them, α represents the step size of each update, f θ is the parameterized model before parameter update, L is the loss function value of the model, is the gradient operator when the meta-learning model parameter is θ.

[0093] Then, the parameterized model f is updated using the parameters θ' Calculate the loss function on the corresponding test set, sum the loss function values ​​of all tasks, and update the parameters based on the gradient of the sum of the loss function values ​​according to the following logic:

[0094]

[0095] Among them, β represents the meta-step size of each update.

[0096] Repeat the above steps until the number of training iterations of meta-learning is reached.

[0097] Finally, after the offline training is completed, the physical information neural network model after offline training is obtained.

[0098] However, the composite truss performance evaluation system will face different working conditions during application. When facing new working conditions, the control equations, boundary conditions, and initial conditions in the real physics followed by the model may change. Therefore, it is necessary to quickly fine-tune the parameters of the meta-learning model online to obtain a meta-learning model that can quickly adapt to new working conditions.

[0099] When evaluating the mechanical properties of the truss under new working conditions, the offline trained neural network model is first used as the initial model. The parameters of this model reflect the stress and strain distribution of the truss structure and other properties. In the fine-tuning stage, a small number of new working condition sample sets are required to train a small amount of data on the loads and boundary conditions under the new working conditions. The model parameters θ are quickly fine-tuned online so that the model can learn the mechanical characteristics of the truss under different loads and boundary conditions under the new working conditions, so as to ensure that the truss can achieve rapid performance evaluation under the new working conditions. Finally, the trained physical information neural network is used as a rapid performance evaluation model for composite trusses.

[0100] Step 101: Obtain the grid node coordinates of the composite material truss of the satellite to be tested and the variable parameters of the geometric structure optimization of the truss to be tested; wherein the variable parameters of the geometric structure optimization of the truss to be tested include the total length of the truss, the cross-sectional size of the truss, the thickness of the longitudinal beam, the width of the longitudinal beam, the pitch of the strip, the thickness of the strip and the width of the strip.

[0101] Step 102: input the truss grid node coordinates of the satellite composite truss to be tested and the geometric structure optimization variable parameters of the truss to be tested into the composite truss performance rapid evaluation model to obtain the satellite composite truss performance evaluation result.

[0102] Based on the above description, in this application, a parametric design model of a composite truss is obtained by secondary development of CAD (Computer-Aided Design) software, and then the design parameters required for the parametric design of the truss are extracted, and the relationship between the parameters is established, so that the truss design scheme can be generated in real time. The rapid evaluation model of the composite truss performance uses the residual of the elastic mechanics control equation that the composite truss obeys when running in space as the regularization term of the loss function, and can rely only on a smaller scale of simulation data or not rely on simulation data for training. In the rapid update stage of the composite truss evaluation model driven by meta-learning, the analysis of the truss model by the physical information neural network is accelerated by two means: offline model pre-training and online rapid fine-tuning of the evaluation model under new working conditions.

[0103] Furthermore, based on the physical information neural network method, the present application integrates the mechanism equation of the composite truss and a small amount of sample data into a rapid performance evaluation model of the composite truss, providing an efficient method for solving the mechanical properties of the composite truss.

[0104] The beneficial effects of this application mainly include: First, a truss mechanical performance rapid evaluation model is constructed based on a physical information neural grid to replace the CAE (Computer Aided Engineering) simulation software, and a truss parametric design system is constructed, which shortens the design time of satellite antenna trusses and speeds up design efficiency. Second, by introducing physical mechanisms in the neural network model training process, the model's demand for sample data volume is reduced, while improving interpretability. Third, through a pre-training-fine-tuning mechanism based on meta-learning, a performance evaluation model for new working conditions is quickly generated, which speeds up performance evaluation and has great potential engineering value.

[0105] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store satellite composite truss performance evaluation data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a satellite composite truss performance evaluation method is implemented.

[0106] Those skilled in the art will understand that Figure 5The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0107] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0108] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0109] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0110] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0111] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0112] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0113] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A satellite composite truss performance evaluation method, characterized in that: The satellite composite material truss performance evaluation method comprises: Constructing a rapid performance evaluation model for composite trusses based on physical information neural networks; Obtaining the grid node coordinates of the composite material truss of the satellite to be tested and the variable parameters of the geometric structure optimization of the truss to be tested; the variable parameters of the geometric structure optimization of the truss to be tested include the total length of the truss, the cross-sectional size of the truss, the thickness of the longitudinal beam, the width of the longitudinal beam, the pitch of the strip, the thickness of the strip and the width of the strip; The truss grid node coordinates of the satellite composite truss to be tested and the optimized variable parameters of the geometric structure of the truss to be tested are input into the composite truss performance rapid evaluation model to obtain the satellite composite truss performance evaluation result.

2. The satellite composite truss performance evaluation method according to claim 1, characterized in that: A rapid performance evaluation model for composite trusses is constructed based on physical information neural networks, including: The satellite composite truss grid node coordinates and truss geometric structure optimization variables are used as input, and the stress, strain and load at the truss grid node coordinates are used as output to form a sample pair; Constructing a training set and a test set based on the sample pairs; Obtaining physical information from neural networks; A meta-learning algorithm is used to train and test the physical information neural network based on the training set and the test set until the trained physical information neural network meets the set requirements, and the trained physical information neural network is used as a rapid performance evaluation model for composite material trusses.

3. The satellite composite truss performance evaluation method according to claim 2, characterized in that: Before acquiring physical information neural network, it also includes: Obtaining initial physical information neural network; The loss function of the initial physical information neural network is modified to obtain the physical information neural network.

4. The satellite composite truss performance evaluation method according to claim 2, characterized in that: Before forming a sample pair with the satellite composite truss grid node coordinates and truss geometric structure optimization variables as input and the stress, strain and load at the truss grid node coordinates as output, it also includes: Construct a parametric design model for composite trusses; Based on the composite material truss parametric design model, the composite material truss grid node coordinates, composite material truss geometric structure optimization variables, and the stress, strain and load at the composite material truss grid node coordinates are obtained.

5. The satellite composite truss performance evaluation method according to claim 2, characterized in that: The physical information neural network is trained and tested based on the training set and the test set by using a meta-learning algorithm until the trained physical information neural network meets the set requirements, and the trained physical information neural network is used as a composite material truss performance rapid evaluation model, specifically including: Using a meta-learning algorithm, offline training is performed on the physical information neural network based on the training set and the test set to obtain the physical information neural network after offline training; The meta-learning algorithm is used to adjust the model parameters θ of the offline trained physical information neural network online to obtain a rapid performance evaluation model for composite trusses.

6. The satellite composite truss performance evaluation method according to claim 5, characterized in that: The offline training process includes: a meta-training phase and a meta-testing phase; In a meta-training stage, the physical information neural network is iteratively trained based on the training set to obtain a meta-trained physical information neural network; In the meta-testing stage, the meta-trained physical information neural network is tested based on the test set to obtain the physical information neural network after offline training.

7. The satellite composite truss performance evaluation method according to claim 3, characterized in that: The modified loss function is expressed as: Loss=Loss data +Loss bc +Loss pde ; Among them, Loss data is the data error loss, Loss bc is the boundary condition loss, Loss pde is the physical constraint loss.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the satellite composite truss performance evaluation method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the satellite composite truss performance evaluation method according to any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the satellite composite truss performance evaluation method according to any one of claims 1 to 7 is implemented.

Images