Efficient identification method for connection rigidity of steel pipe pole of power transmission line

Abstract

The invention discloses an efficient identification method for connection rigidity of a steel pipe pole of a power transmission line, which relates to the field of structural health monitoring, and comprises the following steps: constructing a multi-scale finite element model of the steel pipe pole, acquiring displacement and frequency response of the structure to carry out parameter sensitivity analysis, and screening key correction parameters; constructing a key parameter-structure response sample data set, and training and optimizing an FEA-Net neural network model; and constructing a Bayesian correction framework, solving posterior probability distribution by adopting a TMCMC algorithm, and realizing the probabilistic identification of the connection stiffness and the systematic correction of the finite element model. By constructing a local refined contact model, a bolt pretightening force-connection rigidity corresponding relation is established, and an effective means is provided for verifying whether the rigidity of the corrected model is matched with the rigidity of an actual structure or not. The structure rigidity damage position can be accurately recognized, the rigidity degradation degree can be quantified, and a high-credibility model basis is provided for safety evaluation, digital twinborn construction and operation and maintenance decision making of engineering structures such as electric transmission line steel pipe poles and the like.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring technology, and in particular to an efficient method for identifying the connection stiffness of steel pipe poles in power transmission lines. Background Technology

[0002] With the continuous expansion of power grid construction and the increasing demand for long-distance power transmission, the safety and reliability of transmission lines, as key carriers of electrical energy transmission, are of paramount importance. Steel pipe pole structures, due to their excellent mechanical properties and high economic efficiency, have been widely used in high-voltage and ultra-high-voltage transmission lines. However, during long-term service, under the continuous influence of complex environmental loads (such as wind loads, ice loads, and temperature effects) and operational loads, critical components of steel pipe poles, such as flange bolt connections, are prone to stiffness degradation and damage accumulation, directly affecting the overall structural performance and safety status.

[0003] Currently, there are still significant shortcomings in the technology for identifying and correcting the connection stiffness of steel pipe poles, mainly in the following aspects: 1. Oversimplification in modeling: Traditional finite element modeling often simplifies bolted connections to ideal connection forms such as rigid or hinged connections, which cannot accurately reflect the semi-rigid connection characteristics and nonlinear contact behavior in actual structures, resulting in inherent biases in the model itself. Second, insufficient consideration of uncertainties: Traditional correction methods rely heavily on deterministic optimization frameworks and fail to fully consider the inherent uncertainties in model parameters and measured responses, resulting in limited credibility of correction results and difficulty in supporting probability-based reliability assessments. Third, low computational efficiency: Uncertainty quantification methods based on Bayesian theory require repeated calls to the finite element model for forward calculations. When large-scale sampling is performed in a high-dimensional parameter space, the computational cost is extremely high, making it difficult to meet the timeliness requirements of practical engineering applications. IV. Difficulty in solving the posterior probability distribution: In solving the Bayesian posterior probability distribution, traditional sampling algorithms have low sampling efficiency when dealing with complex posterior probability distributions such as high-dimensional and multi-peaked distributions. They are prone to problems such as slow chain convergence or result distortion, which limits the application of Bayesian methods in complex structures. Fifth, the verification methods are limited: the verification of the effect of traditional stiffness correction usually relies solely on the goodness of fit between the model's predicted response and the measured response. Since the true value of the connection stiffness in actual engineering structures is extremely difficult to measure directly, there is a lack of direct and effective verification methods to determine whether the corrected stiffness parameters match the actual physical state, which casts doubt on the physical reliability of the correction results. In summary, existing technologies cannot achieve accurate, reliable, and physically meaningful correction of the connection stiffness of steel pipe poles within a reasonable calculation cost. Summary of the Invention

[0004] To address the aforementioned challenges, this invention provides an efficient method for identifying the connection stiffness of steel pipe poles in transmission lines. Through multi-scale finite element modeling, parameter sensitivity analysis, a FEA-Net surrogate model, and a Bayesian-TMCMC joint correction framework, it achieves probabilistic identification of connection stiffness and systematic correction of the finite element model. This effectively solves problems such as insufficient consideration of uncertainties, low computational efficiency, difficulty in a posteriori solution, and limited verification methods in traditional methods, thereby improving the scientific rigor of safety assessments and operation and maintenance decisions for steel pipe pole structures in transmission lines.

[0005] To achieve the above objectives, the present invention provides an efficient method for identifying the connection stiffness of steel pipe poles in transmission lines, comprising the following steps: S1: Construct a multi-scale finite element model of the steel pipe pole and obtain model response data under different structural parameters. The model response data includes the displacement and modal frequency data of the structure. S2: Based on the model's response data, perform sensitivity analysis on the model's structural parameters to determine key correction parameters; S3: Based on the key correction parameters, construct a key parameter-structural response sample dataset; and use the sample dataset to train the FEA-Net neural network model to obtain the trained FEA-Net neural network model; S4: Construct a Bayesian correction framework that integrates the prior information of the key correction parameters; based on the trained FEA-Net neural network model, use the TMCMC (Transitional Markov Chain Monte Carlo) algorithm to sample and solve the posterior probability distribution of the key correction parameters; S5: Based on the posterior probability distribution of key correction parameters, determine the correction value of connection stiffness to achieve probabilistic identification of stiffness damage location and degradation degree.

[0006] Preferably, the construction of a multi-scale finite element model of the steel pipe pole in S1 includes: A simplified global-scale model is established based on the material and structural parameters of the actual steel pipe pole structure, including modeling of the tower body, crossarm, and conductor, as well as defining the connection stiffness through matrix elements; A detailed local contact model is established for the flange bolt connection, including detailed modeling of the bolt and flange, setting of contact and friction behavior between various parts, and application of bolt preload. Stiffness calculations are performed based on the local refined contact model to establish the correspondence between bolt preload and connection stiffness.

[0007] Preferably, determining the key correction parameters includes: Static and modal analyses were performed on a simplified model of a steel pipe pole at the global scale using finite element software to obtain displacement and modal frequency data of the structure. Calculate the normalized sensitivity of each response index to each structural parameter; Based on the sensitivity calculation results and sensitivity threshold, key correction parameters are determined.

[0008] Preferably, the normalized sensitivity of each response index to each structural parameter is expressed as: ; in, For structural response indicators, For structural parameters, the partial derivatives are approximated using the finite difference method: ;in, This represents a small positive change in the structural parameters.

[0009] Preferably, S3 specifically includes: Random sampling is performed within the range of key correction parameters to generate different parameter combinations. These combinations are then input into the finite element model to obtain the corresponding structural responses, thus forming a key parameter-structural response sample dataset. Construct a FEA-Net neural network model, with key correction parameters and structural response as inputs and outputs, respectively. The FEA-Net neural network model is trained using the sample dataset until the model prediction error is less than a preset threshold, thus obtaining the trained FEA-Net neural network model.

[0010] Preferably, constructing a Bayesian correction framework that integrates prior information of the key correction parameters specifically includes: Based on historical operation and maintenance data of steel pipe poles and relevant literature experience, the prior probability distribution of key correction parameters is determined. A likelihood function is constructed to measure the fit between the model's predicted response and the measured response under different structural parameters. The model's predicted response is calculated by the trained FEA-Net neural network model.

[0011] Preferably, the likelihood function of a single structural response parameter is expressed as: ; Where: ω p To predict the response, ω e For the measured response, The variance of the measured response is given by , and the total likelihood function is the product of the likelihood functions corresponding to each structural response index.

[0012] Preferably, the TMCMC algorithm is used to sample and solve the posterior probability distribution of the key correction parameters; specifically, it includes: An initial sample set is drawn from the prior probability distribution of the key correction parameters; Calculate the weights and weight variation coefficients of each sample in the current stage, and adaptively select the index for the next stage so that the weight variation coefficients reach a preset threshold. The current sample is resampled based on the normalized weights to obtain the guide sample for the next stage; Perform the MCMC move step on the guide sample, perform Metropolis-Hastings move with Gaussian proposal distribution to generate new candidate samples, and determine whether to accept the new candidate samples based on the acceptance / rejection criterion; Repeat the above steps until the stage exponent reaches 1, at which point the iteration stops.

[0013] Preferably, the correction value for the connection stiffness is determined based on the posterior probability distribution of the key correction parameters, specifically including: Statistical analysis was performed on the sample set of the final output of the TMCMC algorithm to extract the posterior mean, variance and confidence interval of each key correction parameter; and the posterior mean was used as the correction value for the connection stiffness.

[0014] Preferably, the following steps are also included: S6: Substitute the connection stiffness correction value determined by S5 into the finite element model, compare the model response with the measured structural response, and verify the correction effect; and / or calculate the theoretical stiffness value by combining the bolt preload in the measured structure with the corresponding relationship between the bolt preload and connection stiffness, and compare it with the correction value to verify the correction result.

[0015] Therefore, this invention adopts the above-mentioned efficient identification method for the connection stiffness of steel pipe poles in transmission lines. First, it accurately captures the semi-rigid characteristics of flange bolt connections through multi-scale finite element modeling, and selects key correction parameters by combining parameter sensitivity analysis. Then, it uses the FEA-Net proxy model to replace the expensive finite element simulation, which greatly improves the calculation efficiency. Finally, it realizes the probabilistic identification of stiffness parameters based on the Bayesian-TMCMC framework, which fully considers the uncertainty of the model and measurement. It has the following advantages: (1) It uses parameter sensitivity analysis to reduce the correction dimension, focuses on key parameters, and improves the correction efficiency; (2) It uses the FEA-Net proxy model to replace the traditional finite element calculation, avoids repeated calls to the finite element model, and significantly reduces the calculation cost; (3) It combines Bayesian theory and TMCMC algorithm to efficiently solve the posterior probability distribution on the basis of fully considering the uncertainty, and realizes the accurate location of stiffness damage and the quantification of the degree of degradation; (4) It provides the correspondence between "bolt preload and connection stiffness" as a verification means, enhances the physical credibility of the correction results, and provides a reliable model basis for the safety assessment and operation and maintenance decision of engineering structures.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] Figure 1 This is a flowchart of an efficient method for identifying the connection stiffness of steel pipe poles in power transmission lines, as described in this invention. Detailed Implementation

[0018] The following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0019] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0020] The terms "comprising" or "including" as used in this invention mean that the element preceding the term encompasses the element listed after the term, and do not exclude the possibility of encompassing other elements. Terms such as "inner," "outer," "upper," and "lower" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. In this invention, unless otherwise explicitly specified and limited, the term "attached" and similar terms should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can refer to a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication of two elements or the interaction relationship between two elements. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0021] Example An efficient method for identifying the connection stiffness of steel pipe poles in transmission lines, such as... Figure 1 As shown, it includes the following steps: S1: Construct a multi-scale finite element model of the steel pipe pole and obtain model response data under different structural parameters. The model response data includes the displacement and modal frequency data of the structure. Constructing a multi-scale finite element model of the steel pipe pole, specifically including: A simplified global-scale model is established based on the material and structural parameters of the actual steel pipe pole structure, including modeling of the tower body, crossarm, and conductor, as well as defining the connection stiffness through matrix elements, thereby simplifying the treatment of flange bolt connections. A detailed local contact model is established for the flange bolt connection, including detailed modeling of the bolt and flange, setting of contact and friction behavior between various parts, and application of bolt preload. Stiffness calculations are performed based on a locally refined contact model, establishing the correspondence between bolt preload and connection stiffness.

[0022] S2: Based on the model's response data, perform sensitivity analysis on the model's structural parameters to determine key correction parameters; Static and modal analyses were performed on a simplified model of a steel pipe pole at the global scale using finite element software to obtain displacement and modal frequency data of the structure. Calculate the normalized sensitivity of each response index to each structural parameter; response For structural parameters The normalized sensitivity (S) is defined as: ; The partial derivatives are approximated using the finite difference method: .

[0023] Based on the sensitivity calculation results and sensitivity threshold, key correction parameters are determined.

[0024] Specifically, by calculating the sensitivity of each response index to each structural parameter, parameters with minimal impact on the response are eliminated, and key stiffness parameters strongly correlated with the response are retained, thereby reducing the dimensionality of subsequent model corrections. S3: Based on the key correction parameters, construct a key parameter-structural response sample dataset; and use the sample dataset to train the FEA-Net neural network model to obtain the trained FEA-Net neural network model; Random sampling is performed within the range of key correction parameters to generate different parameter combinations. These combinations are then input into the finite element model to obtain the corresponding structural responses, thus forming a key parameter-structural response sample dataset. Specifically, Pyansys is used to randomly sample within a reasonable range of structural stiffness parameters using appropriate sampling methods to generate different parameter combinations, which are then input into the finite element model to obtain a "key parameter-structural response" sample dataset, laying the data foundation for subsequent model training. Construct a FEA-Net neural network model, with key correction parameters and structural response as inputs and outputs, respectively. The FEA-Net neural network model is trained using the sample dataset until the model prediction error is less than a preset threshold, thus obtaining the trained FEA-Net neural network model.

[0025] Specifically, a FEA-Net neural network model was built based on Python, with the number of neurons in its input and output layers matched to the number of key correction parameters and structural response indices, respectively. The FEA-Net model was then trained using the aforementioned simulation dataset, and the network hyperparameters were adjusted until the model prediction error was less than 1%, thereby ensuring that the model could replace traditional finite element direct calculations.

[0026] S4: Construct a Bayesian correction framework that integrates the prior information of the key correction parameters; based on the trained FEA-Net neural network model, use the TMCMC algorithm to sample and solve the posterior probability distribution of the key correction parameters; Constructing a Bayesian correction framework that integrates prior information of the key correction parameters specifically includes: Based on historical operation and maintenance data of steel pipe poles and relevant literature experience, the prior probability distribution of key correction parameters is determined. A likelihood function is constructed to measure the fit between the model's predicted response and the measured response under different structural parameters. The model's predicted response is calculated by the trained FEA-Net neural network model.

[0027] The likelihood function of a single structural response parameter is expressed as: ; Where: ω p To predict the response, ω e For the measured response, The variance of the measured response is given by , and the total likelihood function is the product of the likelihood functions corresponding to each structural response index.

[0028] After obtaining the likelihood function for each response parameter, the total likelihood function is the product of all likelihood functions, and the posterior distribution of the parameters is then calculated based on this product. However, since analytical solutions are extremely difficult to find in this process, sampling is often used as a substitute. The Transitional Markov Chain Monte Carlo (TMCMC) algorithm was written in Python. It uses a multi-stage sampling method to iteratively update samples and achieve a smooth transition from the prior distribution to the posterior distribution. During this process, a pre-trained FEA-Net model can be called to calculate the structural response, replacing the time-consuming finite element calculation and significantly improving computational efficiency. Specifically, it is generated by sampling from the prior distribution of structural stiffness parameters. N 0 samples { θ o,k}, k =1, ..., N 0; The first stage was calculated based on the initial sample. j =0, pWeights and evidence for 0=0: , ; ( j For the number of stages, j =0, 1, ..., m-1; p j (This refers to the index for the current stage).

[0029] Select the next stage index p j+1 Make the coefficient of variation of weights (COV) (coefficient of variation of weights = standard deviation of weights / mean of weights) reach a preset threshold of 100%, and calculate the updated weights; Use the updated version p j+1 Recalculate the weights and normalize them; Resample the current samples according to the normalized weights to obtain the guide samples for the new stage; The MCMC move step is performed on the guide sample, and the Metropolis-Hastings move is performed with the Gaussian proposal distribution. Then, the acceptance / rejection criterion is used to determine whether to accept the new sample, thereby updating the sample set at this stage. Repeat the above steps, when the stage index p m When =1, the iteration stops; S5: Based on the posterior probability distribution of key correction parameters, determine the correction value of connection stiffness to achieve probabilistic identification of stiffness damage location and degradation degree.

[0030] Statistical analysis was performed on the sample set of the final output of the TMCMC algorithm to extract the posterior mean, variance and confidence interval of each key correction parameter; and the posterior mean was used as the correction value for the connection stiffness.

[0031] It also includes the following steps: S6: Substitute the connection stiffness correction value determined by S5 into the finite element model, compare the model response with the measured structural response, and verify the correction effect; and / or calculate the theoretical stiffness value by combining the bolt preload in the measured structure with the corresponding relationship between the bolt preload and connection stiffness, and compare it with the correction value to verify the correction result.

[0032] Example 1 This embodiment provides a more detailed description of the method provided in this application, with reference to a specific structure, as follows: Construction of finite element model and parameter sensitivity analysis of steel pipe pole: A global-scale simplified finite element model of the sample steel pipe pole structure was established. The tower body is divided into three sections, including two flange connection structures. The stiffness properties of each connection in six degrees of freedom are considered, for a total of 12 stiffness parameters. Static and modal analyses were performed on the model to obtain the static displacements of two nodes and the first four modal frequencies of the structure under static load. The normalized sensitivity of each response index to each stiffness parameter was calculated. The results showed that the sensitivity values ​​of four stiffness parameters to the structural response were almost 0. Therefore, these were treated as fixed values ​​in the subsequent correction process, and only the remaining eight stiffness parameters with higher sensitivity were corrected, thereby effectively reducing the problem dimension. Furthermore, a localized, refined model of the flange bolt connection was established, and the correspondence between "bolt preload and connection stiffness" was obtained through stiffness calculation. In actual structures, the theoretical stiffness value of the connection can be calculated by measuring the readily available bolt preload, providing a new approach to verifying the consistency between the corrected model and the actual structure.

[0033] Constructing the FEA-Net proxy model: Based on the key stiffness parameters obtained from the above screening, Pyansys was used to randomly sample within their reasonable range to generate 10,000 sets of parameter combinations. These were then input into the finite element model to extract the first four modal frequencies of the structure and the displacement data of two measuring points, thereby constructing a "key parameter-structural response" sample dataset. Subsequently, a FEA-Net neural network model based on an MLP architecture was built using Python. Eight key stiffness parameters were used as inputs, and six structural responses were used as outputs. The model was trained using the aforementioned sample dataset. By optimizing the network hyperparameters, the error between the model's predicted responses and the finite element simulation responses was reduced to less than 1%. This ensured that the FEA-Net model could effectively replace direct finite element calculations, significantly improving the computational efficiency of subsequent correction processes without sacrificing accuracy.

[0034] Establish a Bayesian-TMCMC joint correction framework: First, by reviewing historical operation and maintenance data and relevant literature and experience of steel pipe poles, the prior distribution information of 8 key stiffness parameters was determined; A likelihood function is established to measure the goodness of fit between the model's predicted response and the measured response under different structural parameters. After obtaining the likelihood functions for each of the six response parameters, the total likelihood function is the product of the individual likelihood functions, from which the posterior distribution of the parameters can be calculated. However, since analytical solutions are extremely difficult in this process, sampling is used instead in this method. The specific steps are as follows: 2000 initial samples were generated by sampling from the prior distributions of 8 key stiffness parameters. θ o,k},k =1, ..., 2000; the number of stages in the initial stage. j =0, stage index p 0 = 0.

[0035] Iteratively execute the following steps (stages) j =0, 1, ..., m -1): 1) Adaptive selection of the next stage index p j+1 Make the weight of the current sample The coefficient of variation (COV, weighted coefficient of variation = weighted standard deviation / weighted mean) reaches the preset threshold of 100%; 2) Calculate the weight of each sample. and stage evidence Based on normalized weights Resampling is performed on the current sample to obtain 2000 "guide samples" for the next stage. For each resampled guide sample, a Metropolis-Hastings shift is performed based on a Gaussian proposal distribution, centered on its current value. The covariance matrix of the proposal distribution is obtained by multiplying the weighted covariance matrix of the current stage samples by a scaling factor β (usually β=0.2). Subsequently, the acceptance / rejection criterion (Metropolis-Hastings Criterion) is used to determine whether to accept the new sample, thereby updating the sample set for this stage.

[0036] Until the stage index p m When the value equals 1, the final stage is reached, the iteration stops, and statistical analysis is performed on the samples in the final stage to extract their posterior mean, variance, and confidence interval.

[0037] Verification of the corrected results: The mean values ​​of each stiffness parameter in the final stage sample of the TMCMC algorithm are used as correction values ​​and substituted into the finite element model. By comparing the model response with the measured structural response, the correction effect of this method can be preliminarily verified.

[0038] However, the above verification is based solely on the degree of matching of response data and lacks clear physical meaning. Therefore, by measuring the bolt preload of the structure and combining it with a local fine contact model of the connection points, the theoretical stiffness value of the structure can be calculated. Comparing this theoretical stiffness with the TMCMC correction result allows for verification of the rationality of the correction result from a physical mechanism perspective, significantly improving its engineering reliability.

[0039] Therefore, the present invention adopts the above-mentioned efficient identification method for the connection stiffness of steel pipe poles in transmission lines, which can effectively solve the problems of insufficient consideration of uncertainties, low computational efficiency, difficulty in a posteriori solution and single verification means in traditional methods. The correction efficiency is significantly improved compared with traditional methods. It can accurately identify the location of structural stiffness damage and quantify the degree of stiffness degradation, providing a highly reliable model foundation for the safety assessment, digital twin construction and operation and maintenance decision-making of engineering structures such as steel pipe poles in transmission lines.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A highly efficient method for identifying the connection stiffness of steel pipe poles in transmission lines, characterized in that, Includes the following steps: S1: Construct a multi-scale finite element model of the steel pipe pole and obtain model response data under different structural parameters. The model response data includes the displacement and modal frequency data of the structure. S2: Based on the model's response data, perform sensitivity analysis on the model's structural parameters to determine key correction parameters; S3: Construct a key parameter-structural response sample dataset based on key correction parameters; The FEA-Net neural network model was trained using the sample dataset to obtain the trained FEA-Net neural network model. S4: Construct a Bayesian correction framework that integrates the prior information of the key correction parameters; based on the trained FEA-Net neural network model, use the TMCMC algorithm to sample and solve the posterior probability distribution of the key correction parameters; S5: Based on the posterior probability distribution of key correction parameters, determine the correction value of connection stiffness to achieve probabilistic identification of stiffness damage location and degradation degree.

2. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 1, characterized in that, In S1, a multi-scale finite element model of the steel pipe pole is constructed, specifically including: A simplified global-scale model is established based on the material and structural parameters of the actual steel pipe pole structure, including modeling of the tower body, crossarm, and conductor, as well as defining the connection stiffness through matrix elements; A detailed local contact model is established for the flange bolt connection, including detailed modeling of the bolt and flange, setting of contact and friction behavior between various parts, and application of bolt preload. Stiffness calculations are performed based on the local refined contact model to establish the correspondence between bolt preload and connection stiffness.

3. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 2, characterized in that, The determination of key correction parameters includes: Static and modal analyses were performed on a simplified model of a steel pipe pole at the global scale using finite element software to obtain displacement and modal frequency data of the structure. Calculate the normalized sensitivity of each response index to each structural parameter; Based on the sensitivity calculation results and sensitivity threshold, key correction parameters are determined.

4. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 3, characterized in that, The normalized sensitivity of each response index to each structural parameter is expressed as: ； in, For structural response indicators, For structural parameters, the partial derivatives are approximated using the finite difference method: ;in, This represents a small positive change in the structural parameters.

5. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 4, characterized in that, S3 specifically includes: Random sampling is performed within the range of key correction parameters to generate different parameter combinations. These combinations are then input into the finite element model to obtain the corresponding structural responses, thus forming a key parameter-structural response sample dataset. Construct a FEA-Net neural network model, with key correction parameters and structural response as inputs and outputs, respectively. The FEA-Net neural network model is trained using the sample dataset until the model prediction error is less than a preset threshold, thus obtaining the trained FEA-Net neural network model.

6. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 2, characterized in that, Constructing a Bayesian correction framework that integrates prior information of the key correction parameters specifically includes: Based on historical operation and maintenance data of steel pipe poles and relevant literature experience, the prior probability distribution of key correction parameters is determined. A likelihood function is constructed to measure the fit between the model's predicted response and the measured response under different structural parameters. The model's predicted response is calculated by the trained FEA-Net neural network model.

7. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 6, characterized in that, The likelihood function of a single structural response parameter is expressed as: ； Where: ω p To predict the response, ω e For the measured response, The variance of the measured response is given by , and the total likelihood function is the product of the likelihood functions corresponding to each structural response index.

8. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 7, characterized in that, The TMCMC algorithm is used to sample and solve the posterior probability distribution of the key correction parameters; specifically, it includes: An initial sample set is drawn from the prior probability distribution of the key correction parameters; Calculate the weights and weight variation coefficients of each sample in the current stage, and adaptively select the index for the next stage so that the weight variation coefficients reach a preset threshold. The current sample is resampled based on the normalized weights to obtain the guide sample for the next stage; The MCMC move step is performed on the guide sample, and the Metropolis-Hastings move is performed with the Gaussian proposal distribution to generate new candidate samples. The acceptance / rejection criterion is used to determine whether to accept the new candidate samples. Repeat the above steps until the stage exponent reaches 1, at which point the iteration stops.

9. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 8, characterized in that, Based on the posterior probability distribution of key correction parameters, the correction value for connection stiffness is determined, specifically including: Statistical analysis was performed on the sample set of the final output of the TMCMC algorithm to extract the posterior mean, variance and confidence interval of each key correction parameter; and the posterior mean was used as the correction value of the connection stiffness.

10. The efficient identification method for the connection stiffness of steel pipe poles in transmission lines according to claim 9, characterized in that, It also includes the following steps: S6: Substitute the connection stiffness correction value determined by S5 into the finite element model, compare the model response with the measured structural response, and verify the correction effect; and / or calculate the theoretical stiffness value by combining the bolt preload in the measured structure with the corresponding relationship between the bolt preload and connection stiffness, and compare it with the correction value to verify the correction result.

Images