Truss structure reliability analysis method and device, computer equipment and storage medium

Abstract

This invention discloses a method, apparatus, computer equipment, and storage medium for reliability analysis of truss structures. The method includes: setting parameters for the truss structure, including the distribution form, mean, and standard deviation of each random variable, and determining the center point of the HDMR formula expansion; setting various parameters and accuracy control parameters for the particle swarm optimization algorithm; establishing sub-Kriging surrogate models of each order based on the parameters; combining the sub-Kriging surrogate models of each order according to the HDMR formula as a substitute for the function, and generating multiple sample function values ​​using Monte Carlo simulation, statistically analyzing the proportion of samples with values ​​less than zero to obtain the predicted failure probability; and evaluating and predicting the performance and reliability of the structure or system under specific conditions based on the predicted failure probability, and identifying potential failure modes. This invention effectively reduces the dimensionality of the function, thereby improving the computational efficiency for high-dimensional reliability problems.

Description

Technical Field

[0001] This invention relates to a method, apparatus, computer equipment, and storage medium for reliability analysis of truss structures, belonging to the field of structural reliability analysis technology. Background Technology

[0002] In the field of reliability engineering, an active learning reliability method combining Kriging models and Monte Carlo simulations has been proposed and proven effective in conducting reliability analysis. However, as the number of random variables increases, the size of the design experiments required to build the model increases significantly, facing the challenge of the "curse of dimensionality." Summary of the Invention

[0003] In view of this, the present invention provides a method, apparatus, computer equipment and storage medium for reliability analysis of truss structures, which can effectively reduce the dimensionality of functions, thereby improving the computational efficiency for high-dimensional reliability problems.

[0004] The first objective of this invention is to provide a method for reliability analysis of truss structures.

[0005] The second objective of this invention is to provide a truss structure reliability analysis device.

[0006] A third objective of this invention is to provide a computer device.

[0007] A fourth objective of this invention is to provide a storage medium.

[0008] The first objective of this invention can be achieved by adopting the following technical solution:

[0009] A reliability analysis method for truss structures includes: setting parameters for the truss structure, the parameters including the distribution form, mean, and standard deviation of each random variable, and determining the center point of the HDMR formula expansion; establishing Kriging surrogate models of each order based on the parameters; combining the Kriging surrogate models of each order according to the HDMR formula as a substitute for the function, and using Monte Carlo simulation to generate function values ​​for multiple samples, statistically analyzing the proportion of samples with values ​​less than zero to obtain the predicted failure probability; and evaluating and predicting the performance and reliability of the structure or system under specific conditions based on the predicted failure probability, and identifying potential failure modes.

[0010] The second objective of this invention can be achieved by adopting the following technical solution:

[0011] A truss structure reliability analysis device includes: a setting module for setting parameters of the truss structure, the parameters including the distribution form, mean, and standard deviation of each random variable, and determining the center point of the HDMR formula expansion; a construction module for establishing Kriging surrogate models of each order based on the parameters; a prediction module for combining the Kriging surrogate models of each order according to the HDMR formula as a substitute for the function, and generating function values ​​of multiple samples using Monte Carlo simulation, statistically analyzing the proportion of samples with values ​​less than zero to obtain the predicted failure probability; and an analysis module for evaluating and predicting the performance and reliability of the structure or system under specific conditions based on the predicted failure probability, and identifying potential failure modes.

[0012] The third objective of this invention can be achieved by adopting the following technical solution:

[0013] A computer device includes a processor and a memory for storing a processor-executable program, wherein when the processor executes the program stored in the memory, it implements the above-described truss structure reliability analysis method.

[0014] The fourth objective of this invention can be achieved by adopting the following technical solution:

[0015] A storage medium storing a program, which, when executed by a processor, implements the aforementioned truss structure reliability analysis method.

[0016] The embodiments of the present invention have the following advantages over the prior art:

[0017] (1) This invention demonstrates high computational efficiency and accuracy in handling high-dimensional reliability problems. This indicates that the fusion of high-dimensional model representation and active learning-based Kriging surrogate model is an effective approach to developing high-performance surrogate model methods for solving high-dimensional reliability problems.

[0018] (2) When using the surrogate model method to handle high-dimensional reliability problems, a high-dimensional model representation is introduced to transform the construction of the high-dimensional surrogate model into the modeling of several low-dimensional sub-models, thereby achieving effective dimensionality reduction and avoiding the "curse of dimensionality". Based on the high-dimensional model representation, the method of this invention establishes a progressive modeling strategy according to the order of the sub-surrogate models, ensuring that the algorithm implementation has good systematicity and the model construction is convenient, and that all high-order sub-surrogate models involved in the modeling contribute fully to the overall modeling.

[0019] (3) In the active learning surrogate model method for reliability analysis, the strategy of introducing a PSO-based optimization algorithm to find highly representative sample points can overcome the shortcomings of CSP and achieve decoupling between the surrogate model construction and failure probability prediction process. Based on the differences between sub-surrogate models of different orders, this invention proposes a progressive modeling mode: for important first-order sub-surrogate models, sample points are selected with the goal of maximizing Kriging variance to ensure the accuracy of global surrogate; for higher-order sub-surrogate models, highly representative samples are selected to ensure the accuracy of the model in high probability density regions and extreme state regions. This invention demonstrates that this differentiated processing not only ensures the efficient implementation of the algorithm but also significantly improves sample quality and model accuracy. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0021] Figure 1 This is a simplified flowchart of a truss structure reliability analysis method according to an embodiment of the present invention.

[0022] Figure 2 This is a flowchart illustrating a truss structure reliability analysis method according to an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram of a load-bearing truss structure according to an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram showing the failure probability and relative error distribution under 30 runs according to an embodiment of the present invention.

[0025] Figure 5 This is a schematic diagram of the curve showing the change of the predicted failure probability as the number of sample points increases, according to an embodiment of the present invention.

[0026] Figure 6 This is a structural block diagram of a truss structure reliability analysis device according to an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0028] like Figure 1 and Figure 2 As shown in the figure, this embodiment provides a method for reliability analysis of truss structures, which includes:

[0029] Step 1. Parameter Definition: Set the parameters for this method and clarify the reliability issue to be analyzed.

[0030] Step 2. Construct the Kriging proxy model: Based on the parameters, establish sub-proxy models for each order.

[0031] Step 3. Model Validation: The constructed Kriging surrogate model is combined using the HDMR formula as a substitute for the function. Using the Monte Carlo simulation method, the function values ​​of a large number of samples are obtained, and the proportion of values ​​less than 0 is counted to obtain the predicted failure probability.

[0032] Step 4. Based on the reliability analysis results, evaluate and predict the performance and reliability of the structure or system under specific conditions, identify potential failure modes, assess risks, and propose improvement measures to enhance the reliability of the product or system.

[0033] Specifically, the details of each step are as follows:

[0034] Step 1. Parameter Definition:

[0035] Step 1.1: Determine the distribution form, mean and standard deviation of each variable, the center point of the HDMR expansion, the parameters of the particle swarm optimization algorithm: number of particles, number of iterations, particle movement speed, initial position, etc., and the stopping condition control parameters that control the update of the one-dimensional sub-surrogate model.

[0036] Step 2. Construct the Kriging proxy model:

[0037] Step 2.1: Establish an initial first-order sub-proxy model using the equidistant point sampling method. Utilize the PSO optimization algorithm to select high-value sample points and add them to the model, updating each model accordingly. The mathematical models involved are... This indicates that the DoE found by PSO is substituted into the Kriging model, where the model provides the Kriging variance. The stopping condition is...

[0038] Step 2.2: Identify the coupling relationships between variables and, using existing sample points, establish several second-order initial sub-surrogate models with significant coupling relationships. The specific identification formula is as follows: The formula for constructing a second-order sub-proxy model is as follows:

[0039] Step 2.3: Utilize the global optimization features of PSO and determine the most needed surrogate model and matching sample points based on Kriging's uncertainty variance. Update the surrogate model until convergence, and supplement with higher-dimensional sub-surrogate models until the requirements are met. The mathematical model for this stage is... u * It is the optimal solution found by PSO, u j ∈[-u lim ,u lim ] represents the search interval for the j-th design variable, δ is a preset minimum quantity used to describe the deviation between the search target and the limit state equation, and r s and r c These are the parameters for controlling the penalty range. The convergence condition is...

[0040] Step 2.4: If necessary, identify and construct the corresponding higher-order sub-proxy models according to Steps 2.2 and 2.3.

[0041] Step 3. Failure probability prediction:

[0042] Step 3.1: Generate a batch of sample points with a standard normal distribution. The distribution of the sample points is determined based on the distribution of the DoE independent variables previously input into each sub-proxy model. In this embodiment, the distribution of the DoE independent variables is a standard normal distribution, so a batch of sample points that follow a standard normal distribution is generated.

[0043] Step 3.2: Use the combination of the pre-constructed Kriging surrogate models (combined according to the HDMR formula) to predict the function values ​​of the sample points.

[0044] Step 3.3: Calculate the proportion of the total number of sample points whose predicted function value is less than 0, thereby obtaining the predicted failure probability and calculating the coefficient of variation of the failure probability.

[0045] Step 4. Application of Results:

[0046] Step 4.1: Based on the failure probability, assess and predict the performance and reliability of the structure or system under specific conditions, identify potential failure modes, evaluate risks, and propose improvement measures to enhance the reliability of the structure or system. For example, for a specific structure, different shape parameters (such as the elastic modulus of a component, cross-sectional height, etc.) can be set for a key part. The overall failure probability of the structure can be predicted using a pre-built surrogate model, thus providing valuable reference for design.

[0047] This invention proposes a high-dimensional reliability analysis method based on the Kriging surrogate model, aiming to address the problems of high computational resource consumption and low efficiency in traditional uncertainty analysis methods when dealing with high-dimensional reliability problems. By introducing high-dimensional model representation (HDMR) technology, this method can effectively achieve function dimensionality reduction, thereby significantly improving the computational efficiency and accuracy of high-dimensional reliability problems. Specific advantages are as follows:

[0048] 1. High efficiency: By using the Kriging proxy model, the dependence on experimental samples is reduced, thus reducing the consumption of computing resources.

[0049] 2. Accuracy: Based on the Kriging method, high-precision prediction of the limit state surface is achieved, enabling accurate estimation of the failure probability.

[0050] 3. Versatility: Applicable to various function types, with broad application prospects.

[0051] 4. Ease of use: The model building process is clear and easy to understand and implement.

[0052] This invention provides an efficient analytical tool for complex structural design and reliability assessment, which can improve the scientific rigor and rationality of structural design.

[0053] For example, the truss structure with 30 variables is as follows: Figure 3 As shown in Table 1, the distribution types, means, and standard deviations of the 30 variables are as follows.

[0054] Table 1

[0055]

[0056]

[0057] In this example, the function expression is G(X) = v max -|Δ(x)|,v max The maximum displacement is set, Δ(x) is the structure in... Figure 3 The vertical displacement at the red position has a possible failure probability of 3.4941 × 10⁻⁶ under 30 variables. -5The initial DoE independent variables for each first-order sub-surrogate model are selected as [-3, 0, 3]. Using the above surrogate model method, the prediction accuracy of failure probability can be improved while reducing the number of function calls. Specific results are shown in Table 2.

[0058] Table 2

[0059]

[0060] In this example, the expression for the high-dimensional limit state function is: When N D When the values ​​are 20 and 60 respectively, a = 95000 and 830000 respectively. All random variables All values ​​follow a Gaussian distribution with a mean of 3.41 and a standard deviation of 0.2. The initial DoE independent variables for each first-order sub-surrogate model were selected as [-6.0, 0, 6.0]. The final calculation results and comparisons with other methods are shown in Table 3. Figure 4 As shown, this example also presents the distribution of failure probability and relative error over 30 runs. Figure 5 As shown in the example, this example also presents the curves of the predicted failure probability as the number of sample points increases during the first five runs. It is worth noting that the predicted failure probability is 0 before DoE = 245. This demonstrates the significant accuracy and efficiency advantages of this method in calculating high-dimensional reliability problems.

[0061] Table 3

[0062]

[0063]

[0064] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware, and the corresponding program can be stored in a computer-readable storage medium.

[0065] It should be noted that although the method operations of the above embodiments are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the order of execution of the described steps may be changed. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0066] like Figure 6 As shown, this embodiment provides a truss structure reliability analysis device, which includes:

[0067] The setting module 601 is used to set the parameters of the truss structure, including the distribution form, mean and standard deviation of each random variable, and to determine the center point of the HDMR formula expansion.

[0068] Module 602 is used to build sub-Kriging proxy models of each order based on the parameters.

[0069] The prediction module 603 is used to combine the Kriging surrogate models of each order according to the HDMR formula as a substitute for the function, and use the Monte Carlo simulation method to generate function values ​​of multiple samples, count the proportion of samples with values ​​less than zero, and obtain the predicted failure probability.

[0070] Analysis module 604 is used to evaluate and predict the performance and reliability of a structure or system under specific conditions based on the predicted failure probability, and to identify potential failure modes.

[0071] In one embodiment, the parameters include the cross-sectional area of ​​the truss members, the elastic modulus of the truss members, the load on the truss structure, and the parameters of the particle swarm optimization algorithm.

[0072] In one embodiment, establishing the sub-Kriging proxy model for each order based on the parameters includes:

[0073] A first-order initial sub-agent model is constructed using the equidistant point selection method, and high-value sample points are selected and added to the model using the particle swarm optimization algorithm to achieve model update.

[0074] Identify the coupling relationships between variables and construct several second-order initial sub-proxy models with significant coupling relationships based on existing sample points;

[0075] By combining the global optimization characteristics of the particle swarm optimization algorithm with the uncertainty variance of the Kriging model, the surrogate model and corresponding sample points that most need to be updated are determined, and the updates are continued until the model converges.

[0076] In one embodiment, the step of using the particle swarm optimization algorithm to select high-value sample points and add them to the model to update the model includes:

[0077] Using the particle swarm optimization algorithm, with the goal of maximizing the Kriging prediction variance, we search for the optimal solution within the constraint interval to obtain the single-dimensional sample point u* that maximizes the prediction variance.

[0078] Add the sample point u* to the DoE sample set of the current sub-agent model and update the first-order initial sub-agent model until the stopping condition is met. This means that the DoE found by the particle swarm optimization algorithm is substituted into the Kriging model, which provides the Kriging variance. This represents the set of true function response values ​​corresponding to the DoE used when constructing the sub-model for the i-th variable. This indicates the preset precision coefficient.

[0079] In one embodiment, the combination of the global optimization features of the particle swarm optimization algorithm and the uncertainty variance of the Kriging model determines the surrogate model and corresponding sample points that most need updating, and continues to update until the model converges, as shown in the following formula:

[0080] u * This represents the optimal solution found by the particle swarm optimization algorithm. F represents the optimal point across all dimensions obtained by the particle swarm optimization algorithm. obj (u) represents the constraints of the constructed mathematical model. This represents the overall predicted value of a function formed by combining the predicted mean values ​​of each sub-proxy model according to the HDMR formula, with the optimal point being substituted into each sub-proxy model. u represents the optimal point found by the particle swarm optimization algorithm in the current iteration under the constraints. N represents the largest uncertainty variance among the uncertain variances provided by each sub-surrogate model after decomposing u and substituting it into each sub-surrogate model. D This represents the total number of variables in the limit state function that needs to be solved, P represents the penalty coefficient, which is usually taken as 1, and u j ∈[-u lim ,u lim ] represents the search interval for the j-th design variable, δ represents the preset value used to describe the deviation between the search target and the limit state equation, and r s and r c This indicates the configured penalty range control parameters;

[0081] Convergence condition is This means that after the global optimum obtained through the particle swarm optimization algorithm iterations is broken down and input into each sub-proxy model, the sub-proxy model with the largest uncertainty variance output by each sub-proxy model is selected as the current sub-proxy model that most needs to be updated. The predicted mean output by this model. This is the variance of the prediction uncertainty output by the model.

[0082] In one embodiment, the random variables are distributed in a log-normal form; the step of combining the Kriging surrogate models of each order according to the HDMR formula as a substitute for the function, and using the Monte Carlo simulation method to generate function values ​​for multiple samples, and statistically analyzing the proportion of samples with values ​​less than zero to obtain the predicted failure probability, includes:

[0083] Generate a batch of sample points that follow a standard normal distribution;

[0084] Using the pre-constructed Kriging surrogate models of various orders, combined according to the HDMR formula of the high-dimensional model representation, the function values ​​of the sample points are predicted using the Monte Carlo simulation method;

[0085] The proportion of sample points with function values ​​less than zero is statistically analyzed to determine the predicted failure probability.

[0086] In one embodiment, the HDMR formula is a decomposition representation of a high-dimensional limit state function.

[0087] This embodiment provides a computer device including a processor, a memory, an input device, a display device, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. When the computer programs are executed by the processor, they implement the truss structure reliability analysis method described in the above embodiment.

[0088] This embodiment provides a storage medium, which is a computer-readable storage medium, storing a computer program. When the computer program is executed by a processor, it implements the truss structure reliability analysis method of the above embodiment.

[0089] It should be noted that the computer-readable storage medium in this embodiment can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0090] In this embodiment, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this embodiment, the computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable storage medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0091] The computer-readable storage medium described above can be used to write computer programs for executing this embodiment in one or more programming languages ​​or combinations thereof. These programming languages ​​include object-oriented programming languages—such as Java, Python, and C++—and conventional procedural programming languages—such as C or similar programming languages. The program can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0092] In summary, this invention proposes a surrogate model method that integrates an active learning mechanism with high-dimensional model representation. In this method, based on the rules of high-dimensional model representation, the high-dimensional limit state function is approximated as a combination of several low-dimensional sub-surrogate models, which are constructed through an active learning mechanism. The specific construction process of this surrogate model includes the following key operations: construction of univariate sub-models, identification of the requirements for coupled variable sub-models, and construction of coupled variable sub-models. During the construction of coupled variable sub-models, highly representative samples are selected based on a learning function and using a particle swarm optimization algorithm. Subsequently, the maximum contribution sub-model is updated. The performance of the proposed method is evaluated through numerical examples. The results show that this method not only ensures the representativeness of the experimental samples but also effectively overcomes the difficulties encountered by existing methods in handling high-dimensional problems, thus verifying the feasibility and efficiency of this method in dealing with high-dimensional reliability problems.

[0093] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope disclosed in the present invention, based on the technical solution and inventive concept of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. A reliability analysis method for truss structures, characterized in that, include: Set the parameters of the truss structure, including the distribution form, mean and standard deviation of each random variable, and determine the center point of the HDMR formula expansion; Based on the parameters, establish Kriging proxy models for each order; The Kriging surrogate models of each order are combined according to the HDMR formula as a substitute for the function, and the function values ​​of multiple samples are generated using the Monte Carlo simulation method. The proportion of samples with values ​​less than zero is counted to obtain the predicted failure probability. Based on the predicted failure probability, assess and predict the performance and reliability of the structure or system under specific conditions, and identify potential failure modes. The step of establishing sub-Kriging proxy models of each order based on the parameters includes: A first-order initial sub-agent model is constructed using the equidistant point selection method, and high-value sample points are selected and added to the model using the particle swarm optimization algorithm to achieve model update. Identify the coupling relationships between variables and construct several second-order initial sub-proxy models with significant coupling relationships based on existing sample points; By combining the global optimization characteristics of the particle swarm optimization algorithm with the uncertainty variance of the Kriging model, the surrogate model and corresponding sample points that most need to be updated are determined, and the updates are continued until the model converges. The process of using particle swarm optimization to select high-value sample points and add them to the model for model updates includes: Using the particle swarm optimization algorithm, with the goal of maximizing the Kriging prediction variance, the algorithm seeks the optimal solution within the constraint interval to obtain the single-dimensional sample points that maximize the prediction variance. u* ; The sample points u* Add the DoE sample set of the current sub-agent model and update the first-order initial sub-agent model until the stopping condition is met.

2. The truss structure reliability analysis method according to claim 1, characterized in that, The parameters include the cross-sectional area of ​​truss members, the elastic modulus of truss members, the load on the truss structure, and the parameters of the particle swarm optimization algorithm.

3. The truss structure reliability analysis method according to claim 1, characterized in that, The stopping condition is , This means that the DoE found by the particle swarm optimization algorithm is substituted into the Kriging model, which provides the Kriging variance. Indicates the first i When constructing a sub-model using variables, the set of actual function response values ​​corresponding to the DoE used. φ This indicates the preset precision coefficient.

4. The truss structure reliability analysis method according to claim 1, characterized in that, The process combines the global optimization features of the particle swarm optimization algorithm with the uncertainty variance of the Kriging model to determine the surrogate model and corresponding sample points that most need updating. This process continues until the model converges, as shown in the following equation: , This represents the optimal solution found by the particle swarm optimization algorithm. This represents the optimal point across all dimensions obtained by the particle swarm optimization algorithm. This represents the constraints of the constructed mathematical model. This represents the overall predicted value, which is a function formed by combining the predicted means of each sub-proxy model using the HDMR formula, after substituting the optimal point into each sub-proxy model. This represents the optimal point found by the particle swarm optimization algorithm in the current iteration under the constraints. Indicates will After decomposing and substituting into each sub-proxy model, the largest uncertainty variance among the uncertainty variances provided by each sub-proxy model is... N D This represents the total number of variables in the limit state function that needs to be solved. P Indicates the penalty coefficient. Indicates the first The search range for each design variable. This represents a preset value used to describe the deviation between the search target and the limit state equation. and This indicates the configured penalty range control parameters; Convergence condition is , This means that after the global optimum obtained through the particle swarm optimization algorithm iterations is broken down and input into each sub-proxy model, the sub-proxy model with the largest uncertainty variance output by each sub-proxy model is selected as the current sub-proxy model that most needs to be updated. The predicted mean output by this model. This is the variance of the prediction uncertainty output by the model.

5. The truss structure reliability analysis method according to claim 1, characterized in that, The random variables are distributed in a log-normal manner; the Kriging surrogate models of each order are combined according to the HDMR formula as a substitute for the function, and the Monte Carlo simulation method is used to generate function values ​​for multiple samples. The proportion of samples with values ​​less than zero is counted to obtain the predicted failure probability, including: Generate a batch of sample points that follow a standard normal distribution; Using the pre-constructed Kriging surrogate models of various orders, combined according to the HDMR formula of the high-dimensional model representation, the function values ​​of the sample points are predicted using the Monte Carlo simulation method; The proportion of sample points with function values ​​less than zero is statistically analyzed to determine the predicted failure probability.

6. The truss structure reliability analysis method according to claim 1, characterized in that, The HDMR formula is a decomposition representation of the high-dimensional limit state function.

7. A truss structure reliability analysis device, characterized in that, The method for reliability analysis of truss structures according to any one of claims 1-6 includes: The setting module is used to set the parameters of the truss structure, including the distribution form, mean and standard deviation of each random variable, and to determine the center point of the HDMR formula expansion; A building module is used to establish sub-Kriging proxy models of each order based on the parameters. The prediction module is used to combine the Kriging surrogate models of each order according to the HDMR formula as a substitute for the function, and use the Monte Carlo simulation method to generate function values ​​of multiple samples, count the proportion of samples with values ​​less than zero, and obtain the predicted failure probability. The analysis module is used to evaluate and predict the performance and reliability of a structure or system under specific conditions based on the predicted failure probability, and to identify potential failure modes.

8. A computer device comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the truss structure reliability analysis method according to any one of claims 1-6.

9. A storage medium storing a program, characterized in that, When the program is executed by the processor, it implements the truss structure reliability analysis method according to any one of claims 1-6.

Images