Spatio-temporal Evolution Prediction Method and Device for Service Performance of Bridge Structures Based on Graph Networks

Abstract

The present invention provides a method and device for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network, which relates to the technical field of predicting the service performance of a bridge structure. The method includes: constructing a bridge graph model; constructing a basic prediction model for the evolution of the service performance of the bridge structure based on the bridge graph model and a spatio-temporal attention graph convolutional neural network; embedding a parameterized graph learning module for adaptively updating the spatio-temporal dependence relationship of the bridge structure components, obtaining a sample data set of the time series of the service performance evolution of the bridge structure components to be predicted, training the embedded prediction model for the evolution of the service performance of the bridge structure, and obtaining a trained prediction model for the evolution of the service performance of the bridge structure; and obtaining the spatio-temporal evolution prediction result of the service performance of the bridge structure according to the model. The purpose of the present invention is to effectively capture the spatio-temporal dependence relationship between the same type of structural components in the bridge and between different types of structural components, and improve the accuracy of the spatio-temporal evolution prediction model of the service performance of the bridge structure.

Description

Technical Field

[0001] The present invention relates to the technical fields of intelligent operation and maintenance of bridges and prediction of service performance of bridge structures, and particularly to a method and device for predicting spatio-temporal evolution of service performance of bridge structures based on graph networks. Background Art

[0002] As a key structure for crossing natural obstacles, bridges are an important part of the transportation network and play a significant role in promoting national economic development and improving people's quality of life. The service performance of bridge structures directly affects the service level and safety of the transportation network. During the long-term service process of bridge structures, affected by service environment corrosion, vehicle loads and their coupling effects, etc., the service performance will degrade as the service time increases. At the same time, the evolution of the service performance of bridge structures shows obvious spatio-temporal dependence. Specifically, among the same type of structural components in a bridge, as well as between different types of structural components, their service environments, material properties and mechanical properties are interrelated and jointly affect the service performance of the bridge. Therefore, predicting the spatio-temporal evolution law of the service performance of bridge structures is crucial for making intelligent maintenance decisions throughout the life cycle of bridges.

[0003] Traditional methods for predicting the evolution of the service performance of bridge structures mainly include regression analysis and probability statistics. To a certain extent, these methods can predict the evolution of the service performance of bridge structures, but they often require a large number of assumptions and simplifications, are difficult to handle the influence of multi-factor coupling effects, and ignore the spatio-temporal dependence relationships between the same type of structural components and between different types of structural components in bridges. Regression analysis processes the time series data of bridge structures through techniques such as polynomial fitting, characterizes the non-linear relationship between the service performance of bridge structures and time, and constructs a model for its evolution over time. Existing research has determined the influence of multiple variables such as loads, environments, and design features on the service performance of bridge structures through multiple regression analysis, and established and verified the service performance evolution models for the superstructure and substructure of prestressed concrete bridges (Srikanth I, Arockiasamy M. Remaining service life prediction of aging concrete bridges based on multiple relevant explanatory variables[J]. Practice Periodical on Structural Design and Construction, 2021, 26(4):04021036.). However, such methods do not consider the randomness and uncertainty in the evolution process of service performance, and ignore the spatio-temporal dependence relationships between the same type of structural components and between different types of structural components in bridges. The prediction models based on probability statistics capture the randomness in the evolution process of bridge structures by simulating the changes in random variables such as the service performance of bridge structures or the duration of the structure in a specific service performance. They are mainly divided into two categories: one is the state-based model, which is based on the probability of transitioning from one state to another at discrete time intervals. For example, the Markov model captures the randomness in its evolution process by regarding the evolution of the service performance of bridge structures as a random process, and realizes the prediction of bridge performance. Existing research has established a semi-Markov continuous-time process bridge service performance evolution prediction model based on bridge historical inspection data and verified its performance (Furtado F, Ribeiro D. Railway Bridge Management System Based on Visual Inspections with Semi-Markov Continuous Time Process[J]. KSCE Journal of Civil Engineering, 2023, 27(1): 233-250.).Another type is the time-based model, which regards the duration of the bridge structure in a certain specific state as a random variable and uses a probability distribution function to describe the degradation process. The disadvantage of this type of method is that it highly depends on the orderliness and standardization of data, and ignores the spatio-temporal dependence relationships among the same type of structural components and different types of structural components in the bridge. Due to the inability to handle the influence of the multi-factor coupling effect during the service performance evolution process and the spatio-temporal dependence relationships among structural components, the above methods have certain limitations in application.

[0004] In recent years, with the development of big data and artificial intelligence, especially the breakthroughs in deep learning technology in image recognition and complex pattern analysis, the prediction method of bridge service performance evolution based on machine learning has gradually attracted attention. Machine learning methods can handle high-dimensional, non-linear and noisy data, and capture the potential patterns and trends of complex data. The current prediction models based on machine learning can be roughly divided into artificial neural network models, fuzzy set theory models, deep learning models, etc., and the prediction performance of these models has been verified through a large number of data sets. The bridge service performance evolution prediction model based on traditional machine learning, represented by the artificial neural network model, can accurately predict the overall service performance of the bridge. Deep learning methods have stronger capabilities than traditional machine learning methods in processing large-scale data and obtaining the evolution laws under corrosion, vehicle loads and their coupling effects. The bridge service performance evolution prediction model based on convolutional neural network learns tens of thousands of historical data from different bridges and predicts the evolution of bridge structural service performance under the coupling action of multiple factors (Liu H, Zhang Y. Bridge condition rating data modeling using deep learning algorithm[J]. Structure and Infrastructure Engineering, 2020, 16(10): 1447-1460.; Liu H, Nehme J, Lu P. An application of convolutional neural network for deterioration modeling of highway bridge components in the United States[J]. Structure and Infrastructure Engineering, 2023, 19(6): 731-744.). The bridge service performance evolution prediction model based on recurrent neural network learns the evolution laws from the integrated data of several years and better completes the prediction of bridge structural service performance evolution (Liu K, El-Gohary N. Deep learning –based analysis of multisource heterogeneous bridge data for enhanced data-driven bridge deterioration prediction[J]. Journal of Computing in Civil Engineering, 2022, 36(5): 04022023.).However, the above model does not consider the spatio-temporal dependence relationships between structural components of the same type and between different types of structural components in the bridge, making it difficult to accurately predict the service performance of the bridge structure and having certain limitations.

[0005] The evolution of the service performance of the bridge structure is affected by service environment corrosion, vehicle loads, and their coupling effects, etc., and exhibits obvious spatio-temporal dependence relationships. Specifically, between structural components of the same type and between different types of structural components in the bridge, their service environments, material properties, and mechanical properties are interrelated and jointly affect the service performance of the bridge structure. The prediction model for the evolution of the service performance of the bridge structure based on the regression analysis method can characterize the non-linear relationship between each variable affecting the evolution of the service performance and the service performance, but it cannot predict the evolution of the service performance under the coupling action of multiple factors and ignores the spatio-temporal dependence relationships between structural components of the same type and between different types of structural components in the bridge. The prediction model for the evolution of the service performance of the bridge structure based on the probability statistics method, such as the Markov chain model, has a high dependence on the orderliness and normality of the data, and the obtained prediction results are very sensitive to the model parameters. At the same time, this method ignores the spatio-temporal dependence relationships between structural components of the same type and between different types of structural components in the bridge. The prediction model for the evolution of the service performance of the bridge structure based on machine learning can process high-dimensional, non-linear, and noisy data, capture the potential patterns and trends of complex data, and the prediction effect of the model is better than the above traditional prediction models. Among them, the prediction model for the evolution of the service performance of the bridge structure based on deep learning can process large-scale bridge structure data, and its performance is better than other traditional machine learning models. However, the existing prediction models for the evolution of the service performance of the bridge structure based on deep learning are all based on time-series data for training, and do not consider the spatio-temporal dependence relationships between structural components of the same type and between different types of structural components in the bridge, making it difficult to accurately predict the service performance of the bridge structure. Therefore, the spatio-temporal evolution prediction model for the service performance of the bridge structure based on deep learning needs to be further developed.

[0006] There are still deficiencies in the existing machine learning methods for predicting the evolution of the service performance of the bridge structure. The research mainly focuses on the evolution prediction in the time dimension and ignores the mutual relationships in the space dimension between structural components of the same type and between different types of structural components in the bridge. It is necessary to develop a prediction model for the service performance of the bridge structure that can consider the spatio-temporal dependence relationships of different structural components of the bridge to improve the prediction accuracy and efficiency of the service performance, so as to better support the intelligent maintenance decision-making of the bridge. Summary of the Invention

[0007] To address the deficiencies still existing in the existing machine learning methods for predicting the service performance evolution of bridge structures. The research mainly focuses on the evolution prediction in the time dimension, ignoring the mutual relationships in the spatial dimension among the same type of structural components and different types of structural components in bridges. It is necessary to develop a prediction model for the service performance of bridge structures that can consider the spatio-temporal dependencies of different structural components of bridges, so as to improve the prediction accuracy and efficiency of service performance and better support the intelligent maintenance decision-making of bridges. The embodiments of the present invention provide a spatio-temporal evolution prediction method and device for the service performance of bridge structures based on graph networks. The technical solutions are as follows:

[0008] On the one hand, a spatio-temporal evolution prediction method for the service performance of bridge structures based on graph networks is provided. This method is implemented by a spatio-temporal evolution prediction device for the service performance of bridge structures, and the method includes:

[0009] S1. Construct a bridge graph model.

[0010] S2. Based on the bridge graph model and the spatio-temporal attention graph convolutional neural network, construct a basic evolution prediction model for the service performance of bridge structures.

[0011] S3. For the basic evolution prediction model for the service performance of bridge structures, embed a parametric graph learning module that adaptively updates the spatio-temporal dependencies of bridge structural components to obtain an embedded evolution prediction model for the service performance of bridge structures.

[0012] S4. Obtain a sample data set of the service performance evolution time series of the structural components of the bridge to be predicted, train the embedded evolution prediction model for the service performance of bridge structures, obtain the optimal network parameters of the bridge graph model and the spatio-temporal attention graph convolutional neural network, and obtain a trained evolution prediction model for the service performance of bridge structures.

[0013] S5. Obtain the service performance evolution time series data of the structural components of the bridge to be predicted, input it into the trained evolution prediction model for the service performance of bridge structures, and obtain the spatio-temporal evolution prediction result of the service performance of bridge structures.

[0014] Optionally, constructing the bridge graph model in S1 includes:

[0015] S11. Based on the existing theory, simulate the service performance evolution of important structural components of the bridge, construct a finite element model of the bridge structure, perform static analysis on the bridge structure with service performance evolution, obtain the spatial dependencies among the structural components in the bridge, map the spatial dependencies to the topological relationships of the graph neural network, and construct a bridge graph structure according to the topological relationships.

[0016] Among them, the bridge graph structure is represented as , the node set The nodes in Represents a bridge structural component, represents the number of nodes; edge set The edges in represent the connection relationships between bridge structural components; represents the adjacency matrix.

[0017] S12. Parametrize the material property parameters of the bridge structural components in the bridge graph structure, and then obtain the bridge graph model.

[0018] Optionally, based on the bridge graph model and the spatio-temporal attention graph convolutional neural network in S2, construct a basic bridge structural service performance evolution prediction model, including:

[0019] S21. Obtain the time series sample data set of the bridge structural component service performance evolution.

[0020] S22. Construct a spatio-temporal attention graph convolutional neural network; among them, the spatio-temporal attention graph convolutional neural network includes: a spatio-temporal attention module and a spatio-temporal graph convolutional module.

[0021] The spatio-temporal attention module includes: a time attention layer and a space attention layer. The time attention layer is used to extract the time attention matrix in the data, and the space attention layer is used to extract the space attention matrix in the data.

[0022] The spatio-temporal graph convolutional module includes: a time convolutional layer and a space convolutional layer. The time convolutional layer is used to identify the time dependence of adjacent time points, and the space convolutional layer is used to capture the spatial dependence of the neighborhood.

[0023] S23. Based on the time series sample data set of the bridge structural component service performance evolution, train the network parameters of the spatio-temporal attention graph convolutional neural network to obtain a basic bridge structural service performance evolution prediction model.

[0024] Optionally, the process of obtaining the time attention matrix in S22 includes:

[0025] Obtain the time attention matrix according to the following formula (1) :

[0026] (1)

[0027] Normalize the time attention matrix through the softmax function to obtain the time attention matrix output by the time attention layer , as shown in the following formula (2):

[0028] (2)

[0029] In the formula, , , , , represents learnable parameters, represents the activation function, represents the temporal data information of the input to the r-th layer spatio-temporal block of the model at time step h, represents the transpose of a matrix, represents the temporal attention matrix element in, The value of represents the temporal attention matrix in time and the degree of dependence between, represents the length of the temporal dimension of the r-th layer spatio-temporal block.

[0030] Optionally, the process of obtaining the spatial attention matrix in S22 includes:

[0031] Obtain the spatial attention matrix according to the following formula (3) :

[0032] (3)

[0033] Normalize the spatial attention matrix through the softmax function to obtain the spatial attention matrix output by the spatial attention layer, as shown in the following formula (4):

[0034] (4)

[0035] In the formula, , , , , represent learnable parameters, represents the transpose of a matrix, represents the activation function, represents the temporal data information of the input to the r-th layer spatio-temporal block of the model at time step h, represents the spatial attention matrix element in, The value of represents the spatial attention matrix at position and the degree of dependence between represents, represents the size of the spatial dimension of the r-th layer spatio-temporal block.

[0036] Optionally, the parametric graph learning module in S3 for adaptively updating the spatio-temporal dependence relationship of bridge structural components is used to extract the bidirectional dependence relationship between the nodes of the structural components in the bridge.

[0037] ​​Extract the bidirectional dependency relationships between the nodes of the structural components in the bridge, including:

[0038] Generate an affinity matrix through the following formula (5) :

[0039] (5)

[0040] In the formula, 、 represent learnable parameters, represents the transpose of the matrix, is used to generate the weights at the diagonal positions.

[0041] Obtain the old affinity matrix through the parameterized graph learning module , and combine the old affinity matrix and the generated affinity matrix through the adaptive aggregation module in the parameterized graph learning module:

[0042] (6)

[0043] In the formula, represents the adaptive score matrix obtained by the adaptive aggregation module, which is used to determine the combination weights between the newly learned old and new affinity matrices, represents a non-linear activation function, represents one or more 1×1 convolutional layers.

[0044] Fuse the old affinity matrix and the affinity matrix to generate a new affinity matrix :

[0045] (7)

[0046] In the formula, represents element-wise multiplication.

[0047] According to the new affinity matrix , obtain the sparsified affinity matrix through the following formula (8):

[0048] (8)

[0049] In the formula, the element of the matrix , represents the row of the matrix, represents the column of the matrix, represents the threshold.

[0050] The sparsified affinity matrix Normalize to obtain the affinity matrix :

[0051] (9)

[0052] In the formula, for the elements of matrix .

[0053] Optionally, obtaining the structural component service performance evolution time series sample dataset of the bridge to be predicted in S4, training the embedded bridge structural service performance evolution prediction model, obtaining the optimal bridge graph model and the network parameters of the spatio-temporal attention graph convolutional neural network, and obtaining the trained bridge structural service performance evolution prediction model, including:

[0054] S41. Obtain the structural component service performance evolution time series sample dataset of the bridge, and perform preprocessing to obtain the preprocessed sample dataset. Train and optimize the final bridge structural service performance evolution prediction model according to the preprocessed sample dataset to obtain the trained final bridge structural service performance evolution prediction model.

[0055] S42. Based on the optimal affinity matrix obtained in step S44, train the final bridge structural service performance evolution prediction model to obtain the network parameters of the optimized bridge structural service performance evolution prediction model .

[0056] S43. Fix the network parameters of the optimized bridge structural service performance evolution prediction model , and train the parameterized graph learning module to learn and update the spatial dependence relationship between nodes.

[0057] S44. Update the optimal affinity matrix according to the result output by the trained parameterized graph learning module;

[0058] S45. Determine whether the preset number of iterations is reached. If so, output the optimal model configuration of the final bridge structural service performance evolution prediction model; if not, go to step S41 to execute.

[0059] On the other hand, a spatio-temporal evolution prediction device for the service performance of a bridge structure based on a graph network is provided. The device is applied to the spatio-temporal evolution prediction method for the service performance of a bridge structure based on a graph network. The device includes:

[0060] A bridge graph model construction module for constructing a bridge graph model.

[0061] A prediction model construction module for constructing a basic bridge structural service performance evolution prediction model based on the bridge graph model and the spatio-temporal attention graph convolutional neural network. ​

[0062] An embedding module is used to embed a parametric graph learning module that adaptively updates the spatio-temporal dependency relationship of bridge structure components into a basic bridge structure service performance evolution prediction model, so as to obtain an embedded bridge structure service performance evolution prediction model.

[0063] A training module is used to obtain a sample data set of the service performance evolution time series of the structural components of the bridge to be predicted, train the embedded bridge structure service performance evolution prediction model, obtain the optimal network parameters of the bridge graph model and the spatio-temporal attention graph convolutional neural network, and obtain a trained bridge structure service performance evolution prediction model.

[0064] An output module obtains the service performance evolution time series data of the structural components of the bridge to be predicted, inputs it into the trained bridge structure service performance evolution prediction model, and obtains the spatio-temporal evolution prediction result of the bridge structure service performance.

[0065] On the other hand, a device for predicting the spatio-temporal evolution of the service performance of a bridge structure is provided. The device for predicting the spatio-temporal evolution of the service performance of a bridge structure includes: a processor; a memory, and a computer-readable instruction is stored on the memory. When the computer-readable instruction is executed by the processor, any one of the methods in the above-mentioned method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network is implemented.

[0066] On the other hand, a computer-readable storage medium is provided. At least one instruction is stored in the storage medium, and the at least one instruction is loaded and executed by a processor to implement any one of the methods in the above-mentioned method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network.

[0067] The beneficial effects brought by the technical solutions provided in the embodiments of the present invention at least include:

[0068] In the present invention, a method for predicting the evolution of the service performance of a bridge structure based on a graph network is proposed, aiming to effectively capture the spatio-temporal dependency relationships between structural components of the same type and between different types of structural components in the bridge, and improve the accuracy of the spatio-temporal evolution prediction model of the bridge structure service performance. The technical effects are as follows:

[0069] 1. Improving prediction accuracy: Different types of structural components in a bridge (such as main girders, piers, stay cables) are interrelated in terms of space and load effects, and the same type of structural components are interrelated in terms of space and mechanical properties. Traditional prediction models mainly focus on the analysis in the time dimension, ignoring the spatial dependence relationships between bridge structural components. The present invention constructs a bridge graph model to accurately represent the spatial dependence relationships between bridge structural components. First, map the spatial dependence relationships between structural components in the bridge into the topological relationships of a graph neural network, and represent the connection relationships between components through the edges in the graph structure. Second, parameterize the material property of the bridge structural component to represent the spatial dependence relationships between the same type of structural components. This graph model can reflect the spatial dependence relationships between different types of structural components and the same type of structural components in the bridge at the same time. Based on the bridge graph model and the spatio-temporal attention graph convolutional neural network, construct a prediction model for the evolution of the service performance of the bridge structure. The model takes the time series data of the service performance of the bridge structure as input, and uses the spatio-temporal attention module and the graph convolutional module to capture the spatio-temporal dependence relationships between structural components hidden in the data. Therefore, the prediction results obtained by the model are more in line with the real state.

[0070] 2. Self - adaptability and model optimization: As the service time increases, the spatio-temporal dependence relationships between bridge structural components will change. Based on the prediction model for the evolution of service performance, the present invention embeds a parametric graph learning algorithm for adaptively updating the spatio-temporal dependence relationships of bridge structural components. This algorithm dynamically captures and updates the spatio-temporal dependence relationships between structural components during the service process, improves the bridge graph model in the graph convolutional neural network, and adapts to the dynamically changing bridge structural characteristics during the service process. This dynamic adjustment mechanism enables the model to adapt to the changes in the spatio-temporal dependence relationships of structural components caused by factors such as environment and load during the long-term service of the bridge. Combining the spatio-temporal attention graph convolutional neural network and the parametric graph learning algorithm, construct a prediction model for the evolution of the service performance of the bridge structure. The training of the prediction model, parametric graph learning, and affinity matrix update in the model together constitute a single optimization cycle. By setting multiple iterative cycles to continuously adjust and improve the model parameters, the optimal model configuration for predicting the service performance of the bridge structure can be obtained, further improving the model performance. Description of the Drawings

[0071] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0072] Figure 1 It is a flowchart of a spatio-temporal evolution prediction method for the service performance of a bridge structure based on a graph network provided by an embodiment of the present invention;

[0073] Figure 2 It is an architecture diagram of a method for predicting spatiotemporal evolution of service performance of a bridge structure based on a graph network provided in an embodiment of the present invention;

[0074] Figure 3 is a schematic diagram of a bridge graph model provided by an embodiment of the present invention;

[0075] Figure 4 It is an architecture diagram of a bridge structure service performance prediction model based on a spatiotemporal attention graph convolutional neural network provided by an embodiment of the present invention;

[0076] Figure 5 It is an architecture diagram of a prediction model embedded with a parameterized graph learning algorithm provided by an embodiment of the present invention;

[0077] Figure 6 It is a diagram of the model training optimization process of the graph network-based bridge structure service performance spatiotemporal evolution prediction method provided in an embodiment of the present invention;

[0078] Figure 7 It is a block diagram of a device for predicting the spatiotemporal evolution of service performance of a bridge structure based on a graph network provided by an embodiment of the present invention;

[0079] Figure 8 It is a structural schematic diagram of a device for predicting the spatiotemporal evolution of service performance of a bridge structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0080] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0081] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0082] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0083] In the embodiments of the present invention, sometimes subscripts such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meanings they express are the same.

[0084] To make the technical problems, technical solutions, and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0085] The embodiments of the present invention provide a method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network. This method can be implemented by a device for predicting the spatio-temporal evolution of the service performance of a bridge structure, and this device for predicting the spatio-temporal evolution of the service performance of a bridge structure can be a terminal or a server. As Figure 1 、 2 shown in the flowchart of the method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network, the processing flow of this method can include the following steps:

[0086] S1. Construct a bridge graph model.

[0087] Optionally, the above step S1 can include the following steps S11 - S12:

[0088] S11. Based on existing theories, simulate the service performance evolution of important structural components of the bridge, construct a finite element model of the bridge structure, perform static analysis on the bridge structure with service performance evolution, obtain the spatial dependence relationship between structural components in the bridge, map the spatial dependence relationship to the topological relationship of the graph neural network, and construct a bridge graph structure according to the topological relationship.

[0089] Among them, the bridge graph structure is represented as , the nodes in the node set represent the structural components of the bridge, represents the number of nodes; the edges in the edge set represent the connection relationship between the structural components of the bridge; represents the adjacency matrix.

[0090] S12. Parameterize the material properties of the structural components in the bridge graph structure by means of one-hot encoding, etc., assign eigenvalue to the nodes, and thus obtain the bridge graph model.

[0091] In a feasible implementation, different types of structural components in the bridge (such as main girders, bridge piers, stay cables) are spatially and load-related, and the same type of structural components are spatially and mechanically related.

[0092] Specifically, a bridge graph model is constructed to represent the spatial dependence relationships among bridge structural components. First, the spatial dependence relationships among the structural components in the bridge are mapped to the topological relationships of a graph neural network, and the connection relationships among the components are represented by the edges in the graph structure. The bridge graph structure consists of a node set and an edge set and is denoted as . Among them, the nodes represent the bridge structural components, and the edges represent the connection relationships among the components. denotes the adjacency matrix of the graph . If there is an edge between nodes and , then , otherwise . Second, the material property parameters of the bridge structural components are parameterized to represent the spatial dependence relationships among the structural components of the same type. If the material properties of nodes and are the same, then , otherwise . The schematic diagram of the bridge graph model is shown in Appendix Figure 3 . In the figure, if there is a spatial dependence relationship of service performance evolution between nodes and , then the nodes and are connected. Further, the intra-class dependence relationship between components of the same type is represented by a dashed line, and the inter-class dependence relationship between components of different types is represented by a solid line.

[0093] S2. Based on the bridge graph model and the spatio-temporal attention graph convolutional neural network, a basic bridge structural service performance evolution prediction model is constructed.

[0094] In a feasible implementation, based on the bridge graph model and the spatio-temporal attention graph convolutional neural network, the spatio-temporal attention module and the graph convolutional module are used to capture the spatio-temporal dependence relationships among the structural components in the data, and a bridge structural service performance evolution prediction model is constructed. The overall architecture diagram of the model is shown in Appendix Figure 4 . The model takes the time series data of the bridge structural service performance as input, uses the spatio-temporal attention module and the graph convolutional module to capture the spatio-temporal dependence relationships among the structural components in the data, and outputs the service performance evolution prediction results for a future period of time. The spatio-temporal attention module consists of a time attention layer and a space attention layer, which can extract the time attention matrix and the space attention matrix in the data. The information processed by the attention module is input into the spatio-temporal graph convolutional module, which combines the graph convolution in the spatial dimension and the standard convolution in the time dimension. The graph convolution in the spatial dimension is used to capture the spatial dependence of the neighborhood, and the standard convolution in the time dimension is used to identify the time dependence of adjacent time points.

[0095] Optionally, the above step S2 may include the following steps S21 - S23:

[0096] S21. Obtain a sample data set of the service performance evolution time series of bridge structural components.

[0097] S22. Construct a spatio - temporal attention graph convolutional neural network; wherein, the spatio - temporal attention graph convolutional neural network includes: a spatio - temporal attention module and a spatio - temporal graph convolutional module.

[0098] The spatio - temporal attention module includes: a time attention layer and a space attention layer. The time attention layer is used to extract the time attention matrix in the data, and the space attention layer is used to extract the space attention matrix in the data.

[0099] The spatio - temporal graph convolutional module includes: a time convolutional layer and a space convolutional layer. The time convolutional layer is used to identify the time dependence of adjacent time points, and the space convolutional layer is used to capture the spatial dependence of the neighborhood.

[0100] Optionally, the process of obtaining the time attention matrix in S22 includes:

[0101] Obtain the time attention matrix according to the following formula (1) :

[0102] (1)

[0103] Normalize the time attention matrix through the softmax function to obtain the time attention matrix output by the time attention layer , as shown in the following formula (2):

[0104] (2)

[0105] In the formula, , , , , represent learnable parameters, represents the activation function, represents the time - series data information of the input of the r - th spatio - temporal block of the model at time step h, represents the transpose of the matrix, represents the time attention matrix in the elements, the value of represents the time and the degree of dependence between them in the time attention matrix represents the length of the time dimension of the r - th spatio - temporal block.

[0106] Optionally, the process of obtaining the spatial attention matrix in S22 includes:

[0107] Obtain the spatial attention matrix according to the following formula (3) :

[0108] (3)

[0109] Normalize the spatial attention matrix through the softmax function to obtain the spatial attention matrix output by the spatial attention layer , as shown in the following formula (4):

[0110] (4)

[0111] In the formula, , , , , represent learnable parameters, represents the transpose of the matrix, represents the activation function, represents the temporal data information of the input of the model to the r-th spatio-temporal block at time step h, represents the spatial attention matrix in the elements, The value of represents the spatial attention matrix in the position and between the degree of dependence, represents the size of the spatial dimension of the r-th spatio-temporal block.

[0112] Furthermore, the information processed by the attention module is input into the spatio-temporal graph convolution module, which combines graph convolution in the spatial dimension and standard convolution in the temporal dimension. Graph convolution in the spatial dimension is used to capture the spatial dependence of neighborhoods, and standard convolution in the temporal dimension is used to identify the temporal dependence of neighboring time points.

[0113] S23. Based on the service performance evolution time series sample data set of bridge structural components, train the network parameters of the spatio-temporal attention graph convolutional neural network to obtain a bridge structural service performance evolution prediction model.

[0114] S3. For the basic bridge structural service performance evolution prediction model, embed a parametric graph learning module that adaptively updates the spatio-temporal dependence relationship of bridge structural components to obtain an embedded bridge structural service performance evolution prediction model.

[0115] In a feasible implementation manner, based on the service performance evolution prediction model, a parametric graph learning algorithm for adaptively updating the spatio-temporal dependence relationship of bridge structural components is embedded. The overall architecture diagram of the model is as shown in the appendix Figure 5 As shown. After embedding the parametric graph learning module, the entire model is divided into a prediction module and a graph learning module. Through training with input data, the optimal graph model output by the graph learning module and the optimal network weight parameters of the prediction module are obtained. This algorithm dynamically captures and updates the spatio-temporal dependence relationship between structural components during the service process, improves the bridge graph model in the graph convolutional neural network, adapts to the dynamically changing bridge structural characteristics during the service process, and further improves the accuracy and robustness of the spatio-temporal evolution prediction model of the bridge structural service performance.

[0116] Optionally, the parametric graph learning module for adaptively updating the spatio-temporal dependence relationship of bridge structural components in S3 is used to extract the bidirectional dependence relationship between the nodes of the structural components in the bridge.

[0117] Extracting the bidirectional dependence relationship between the nodes of the structural components in the bridge may include:

[0118] Generating a rough affinity matrix through the following formula (5) :

[0119] (5)

[0120] In the formula, , represent learnable parameters, represents the transpose of the matrix, is used to generate the weights at the diagonal positions.

[0121] Using the adaptive aggregation module to combine the old spatial dependence relationship and the new spatial dependence relationship:

[0122] (6)

[0123] In the formula, represents the adaptive score matrix obtained by the adaptive aggregation module, which is used to determine the combination weight between the newly learned old and new affinity matrices, represents the non-linear activation function, represents one or more 1×1 convolutional layers.

[0124] Using the adaptive aggregation module to fuse the old and new affinity matrices to generate a new affinity matrix :

[0125] (7)

[0126] In the formula, represents element-wise multiplication.

[0127] To enhance the sparsity of the generation matrix, a sparsified affinity matrix is obtained through the following formula (8) :

[0128] (8)

[0129] In the formula, for the matrix element , represents the row of the matrix, represents the column of the matrix, represents the threshold value, which is used to filter out weak relationships.

[0130] Normalize the sparsified affinity matrix to obtain the affinity matrix :

[0131] (9)

[0132] In the formula, for the matrix element .

[0133] This algorithm can dynamically capture and update the spatio-temporal dependence relationship of bridge structural components during the service process, thereby improving the initial bridge graph model defined in the graph convolutional neural network and adapting to the dynamically changing bridge structural characteristics during the service process.

[0134] Furthermore, combining the spatio-temporal attention graph convolutional neural network and the parameterized graph learning algorithm, a prediction model for the evolution of the service performance of bridge structures is constructed. The training and optimization process of the model is divided into three steps. First, based on the preprocessed data set, the prediction model is trained and optimized to ensure that the model can effectively capture the spatio-temporal dependence relationship between structural components in the data and output accurate prediction results. Second, fix the network parameters of the current prediction model and train the parameterized graph learning module to accurately learn and update the spatial dependence relationship between nodes through an adaptive algorithm. Third, update the current optimal affinity matrix based on the results output by the parameterized graph learning module to ensure that the model adapts to the dynamically changing bridge structural characteristics during the service process. The above three steps together constitute a single optimization loop of the model. To further improve the model performance, multiple iterative loops are carried out to continuously adjust and improve the model parameters until the optimal model configuration for predicting the service performance of bridge structures is found. The training and optimization process of the model is as shown in the appendix Figure 6 as shown, and specifically may include the following steps S41 - S45:

[0135] S41. Obtain the service performance evolution time series sample data set of bridge structural components, and perform preprocessing to obtain the preprocessed sample data set. Train and optimize the final bridge structural service performance evolution prediction model according to the preprocessed sample data set to obtain the trained final bridge structural service performance evolution prediction model.

[0136] In a feasible implementation manner, train and optimize the prediction model based on the preprocessed data set to ensure that the model can effectively capture the spatio-temporal dependence relationship between structural components hidden in the data and output accurate prediction results.

[0137] S42. Obtain the optimal affinity matrix through the parametric graph learning module, and optimize the trained final bridge structural service performance evolution prediction model according to the optimal affinity matrix to obtain the network parameters of the optimized bridge structural service performance evolution prediction model. .

[0138] In a feasible implementation manner, obtain the optimal affinity matrix from the set of affinity matrices , and use it to optimize the prediction model to obtain the network parameters . As shown in the following formula (10), select the prediction value and the true value to minimize the L1 loss between them as the training objective of the prediction model:

[0139] (10)

[0140] In the formula: is the true value; is the prediction value.

[0141] S43. Fix the network parameters of the optimized bridge structural service performance evolution prediction model , and train the parametric graph learning module to learn and update the spatial dependence relationship between nodes.

[0142] In a feasible implementation manner, fix the network parameters of the current prediction model as the estimated value, train the parametric graph learning module, and accurately learn and update the spatial dependence relationship between nodes through an adaptive algorithm. To ensure the sparsity of the generated dependence relationship, optimize based on the loss function shown in formula (11) :

[0143] (11)

[0144] In the formula, is a hyperparameter used to control the sparsity rate. For at and set the value to 1 at the position and 0 at other positions, indicating the dimension of the affinity matrix.

[0145] S44. Update the optimal affinity matrix according to the result output by the trained parametric graph learning module.

[0146] In a feasible implementation, update the current optimal affinity matrix based on the result output by the parametric graph learning module , ensuring that the model adapts to the dynamically changing bridge structure characteristics during service. First, incorporate the generated adjacency matrix into the affinity matrix set A. Then, input the graphs in the set and the validation set into the prediction model respectively, and according to the loss function , calculate all sub-graph losses according to Equation (12):

[0147] (12)

[0148] In the formula, is the prediction function, represents the k-th matrix in the affinity matrix set, are the learnable parameters in the prediction model.

[0149] Let be the vector composed of all prediction losses, be the maximum loss value, and the weight vector is as shown in the following formula (13):

[0150] (13)

[0151] represents the weight vector.

[0152] The current affinity matrix can be obtained by weighting all sub-graphs:

[0153] (14)

[0154] S45. Judge whether the preset number of iterations is reached. If so, output the optimal model configuration of the final bridge structure service performance evolution prediction model; if not, go to step S41. Continuously adjust and improve the model parameters through multiple iteration loops to obtain the optimal model configuration for predicting the bridge structure service performance and improve the model performance. The training process of the entire model is as shown in the appendix Figure 6 as shown.

[0155] S5. Obtain the time series data of the service performance evolution of the structural components of the bridge to be predicted, input it into the trained bridge structure service performance evolution prediction model, and obtain the spatio-temporal evolution prediction result of the bridge structure service performance.

[0156] The present invention proposes a method for predicting the service performance evolution of a bridge structure based on a graph network, including the following two parts: a prediction model for the service performance evolution of a bridge structure based on a spatio-temporal attention graph neural network, and a parametric graph learning algorithm for adaptively updating the spatio-temporal dependence relationship of bridge structure components. The technical problems to be solved include: describing the topological relationship between bridge structure components through a graph structure, constructing a bridge graph model, and accurately characterizing the spatial dependence relationship between components; constructing a prediction model for the service performance evolution of a bridge structure based on the bridge graph model and a spatio-temporal attention graph convolutional neural network; and embedding a parametric graph learning algorithm for adaptively updating the spatio-temporal dependence relationship of bridge structure components on the basis of the service performance evolution prediction model. This algorithm dynamically captures and updates the spatio-temporal dependence relationship of bridge structure components during the service process, improves the initial bridge graph model defined in the graph convolutional neural network, adapts to the dynamically changing bridge structure characteristics during the service process, and further enhances the prediction performance of the model.

[0157] In an embodiment of the present invention, a method for predicting the service performance evolution of a bridge structure based on a graph network is proposed, aiming to effectively capture the spatio-temporal dependence relationship between the same type of structural components and different types of structural components in a bridge, and improve the accuracy of the spatio-temporal evolution prediction model of the bridge structure service performance. The technical effects are as follows:

[0158] 1. Improve prediction accuracy: Different types of structural components in a bridge (such as main girders, piers, and stay cables) are spatially and load-related, and the same type of structural components are spatially and mechanically related. Traditional prediction models mainly focus on the analysis of the time dimension and ignore the spatial dependence relationship between bridge structure components. The present invention constructs a bridge graph model to accurately characterize the spatial dependence relationship between bridge structure components. First, the spatial dependence relationship between bridge structure components is mapped into the topological relationship of a graph neural network, and the connection relationship between components is characterized by the edges in the graph structure. Second, the material attribute parameters of bridge structure components are parameterized to characterize the spatial dependence relationship between the same type of structural components. This graph model can simultaneously reflect the spatial dependence relationship between different types of structural components and the same type of structural components in a bridge. Based on the bridge graph model and a spatio-temporal attention graph convolutional neural network, a prediction model for the service performance evolution of a bridge structure is constructed. The model takes the time series data of the bridge structure service performance as input, and uses the spatio-temporal attention module and the graph convolutional module to capture the spatio-temporal dependence relationship hidden in the data between structural components. Therefore, the prediction results obtained by the model are more in line with the real state.

[0159] 2. Adaptability and Model Optimization: As the service time increases, the spatio-temporal dependence relationship between bridge structural components will change. Based on the service performance evolution prediction model, this invention embeds a parametric graph learning algorithm that adaptively updates the spatio-temporal dependence relationship of bridge structural components. This algorithm dynamically captures and updates the spatio-temporal dependence relationship between structural components during service, improves the bridge graph model in the graph convolutional neural network, and adapts to the dynamically changing bridge structural characteristics during service. This dynamic adjustment mechanism enables the model to adapt to the changes in the spatio-temporal dependence relationship of structural components caused by factors such as environment and load during the long-term service of the bridge. Combining the spatio-temporal attention graph convolutional neural network and the parametric graph learning algorithm, a prediction model for the evolution of the service performance of bridge structures is constructed. The prediction model training, parametric graph learning, and affinity matrix update in the model together constitute a single optimization cycle. By setting multiple iterative cycles to continuously adjust and improve the model parameters, the optimal model configuration for predicting the service performance of bridge structures can be obtained, further enhancing the model performance.

[0160] Figure 7 is a block diagram of a device for spatio-temporal evolution prediction of the service performance of a bridge structure based on a graph network shown according to an exemplary embodiment. This device is used for the method of spatio-temporal evolution prediction of the service performance of a bridge structure based on a graph network. Referring to Figure 7 this, the device includes a bridge graph model construction module 310, a prediction model construction module 320, an embedding module 330, a training module 340, and an output module 350. Among them:

[0161] The bridge graph model construction module 310 is used to construct a bridge graph model.

[0162] The prediction model construction module 320 is used to construct a basic prediction model for the evolution of the service performance of a bridge structure based on the bridge graph model and the spatio-temporal attention graph convolutional neural network.

[0163] The embedding module 330 is used to embed a parametric graph learning module that adaptively updates the spatio-temporal dependence relationship of bridge structural components into the basic prediction model for the evolution of the service performance of a bridge structure, obtaining an embedded prediction model for the evolution of the service performance of a bridge structure.

[0164] The training module 340 is used to obtain a sample data set of the time series of the evolution of the service performance of the structural components of the bridge to be predicted, train the embedded prediction model for the evolution of the service performance of a bridge structure, obtain the network parameters of the optimal bridge graph model and the spatio-temporal attention graph convolutional neural network, and obtain a trained prediction model for the evolution of the service performance of a bridge structure.

[0165] The output module 350 obtains the time series data of the evolution of the service performance of the structural components of the bridge to be predicted, inputs it into the trained prediction model for the evolution of the service performance of a bridge structure, and obtains the spatio-temporal evolution prediction result of the service performance of the bridge structure.

[0166] In an embodiment of the present invention, a method for predicting the service performance evolution of a bridge structure based on a graph network is proposed, aiming to effectively capture the spatio-temporal dependence relationships between structural components of the same type and between different types of structural components in the bridge, and improve the accuracy of the spatio-temporal evolution prediction model of the bridge structure service performance. The technical effects are as follows:

[0167] 1. Improve prediction accuracy: Different types of structural components in the bridge (such as main girders, piers, stay cables) are spatially and load-related, and structural components of the same type are spatially and mechanically related. Traditional prediction models mainly focus on the analysis in the time dimension and ignore the spatial dependence relationships between bridge structural components. The present invention constructs a bridge graph model to accurately represent the spatial dependence relationships between bridge structural components. First, map the spatial dependence relationships between structural components in the bridge into the topological relationships of a graph neural network, and represent the connection relationships between components through the edges in the graph structure. Second, parameterize the material properties of bridge structural components to represent the spatial dependence relationships between structural components of the same type. This graph model can reflect both the spatial dependence relationships between different types of structural components and between structural components of the same type in the bridge. Based on the bridge graph model and the spatio-temporal attention graph convolutional neural network, a prediction model for the service performance evolution of the bridge structure is constructed. The model takes the time series data of the bridge structure service performance as input, and uses the spatio-temporal attention module and the graph convolutional module to capture the spatio-temporal dependence relationships hidden in the data. Therefore, the prediction results obtained by the model are more in line with the real state.

[0168] 2. Self-adaptability and model optimization: As the service time increases, the spatio-temporal dependence relationships between bridge structural components will change. Based on the service performance evolution prediction model, the present invention embeds a parametric graph learning algorithm for adaptively updating the spatio-temporal dependence relationships of bridge structural components. This algorithm dynamically captures and updates the spatio-temporal dependence relationships between structural components during the service process, improves the bridge graph model in the graph convolutional neural network, and adapts to the dynamically changing bridge structure characteristics during the service process. This dynamic adjustment mechanism enables the model to adapt to the changes in the spatio-temporal dependence relationships of structural components caused by factors such as the environment and load during the long-term service of the bridge. Combining the spatio-temporal attention graph convolutional neural network and the parametric graph learning algorithm, a prediction model for the service performance evolution of the bridge structure is constructed. The prediction model training, parametric graph learning, and affinity matrix update in the model together constitute a single optimization cycle. By setting multiple iterative cycles to continuously adjust and improve the model parameters, the optimal model configuration for predicting the bridge structure service performance can be obtained, further improving the model performance.

[0169] Figure 8 is a schematic structural diagram of a device for predicting the spatio-temporal evolution of the service performance of a bridge structure provided by an embodiment of the present invention, as Figure 8As shown, the spatio-temporal evolution prediction device for the service performance of a bridge structure may include the above-mentioned Figure 7 spatio-temporal evolution prediction device for the service performance of a bridge structure based on a graph network as shown. Optionally, the spatio-temporal evolution prediction device 410 for the service performance of a bridge structure may include a first processor 2001.

[0170] Optionally, the spatio-temporal evolution prediction device 410 for the service performance of a bridge structure may further include a memory 2002 and a transceiver 2003.

[0171] Among them, the first processor 2001 is connected to the memory 2002 and the transceiver 2003, for example, through a communication bus.

[0172] Next, in combination with Figure 8 each component of the spatio-temporal evolution prediction device 410 for the service performance of a bridge structure will be specifically introduced:

[0173] Among them, the first processor 2001 is the control center of the spatio-temporal evolution prediction device 410 for the service performance of a bridge structure, which can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPUs), or can be an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention, for example: one or more digital signal processors (DSPs), or one or more field programmable gate arrays (FPGAs).

[0174] Optionally, the first processor 2001 can execute various functions of the spatio-temporal evolution prediction device 410 for the service performance of a bridge structure by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0175] In a specific implementation, as an embodiment, the first processor 2001 may include one or more CPUs, for example Figure 8 CPU0 and CPU1 shown in

[0176] In a specific implementation, as an embodiment, the spatio-temporal evolution prediction device 410 for the service performance of a bridge structure may also include multiple processors, for example Figure 8The first processor 2001 and the second processor 2004 shown in []. Each of these processors can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The processor here can refer to one or more devices, circuits, and / or processing cores for processing data (such as computer program instructions).

[0177] Among them, the memory 2002 is used to store the software program for implementing the solution of the present invention and is controlled by the first processor 2001 for execution. The specific implementation manner can refer to the above method embodiments and will not be elaborated here.

[0178] Optionally, the memory 2002 can be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or can also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 can be integrated with the first processor 2001 or exist independently and is coupled to the first processor 2001 through the interface circuit of the bridge structure service performance spatio-temporal evolution prediction device 410 ( Figure 8 not shown in []). The embodiments of the present invention do not make specific limitations on this.

[0179] The transceiver 2003 is used to communicate with network devices or communicate with terminal devices.

[0180] Optionally, the transceiver 2003 can include a receiver and a transmitter ( Figure 8 not shown separately in []). Among them, the receiver is used to implement the receiving function, and the transmitter is used to implement the sending function.

[0181] Optionally, the transceiver 2003 can be integrated with the first processor 2001 or exist independently and is coupled to the first processor 2001 through the interface circuit of the bridge structure service performance spatio-temporal evolution prediction device 410 ( Figure 8 not shown in []). The embodiments of the present invention do not make specific limitations on this.

[0182] It should be noted that Figure 8 the structure of the bridge structure service performance spatio-temporal evolution prediction device 410 shown in does not constitute a limitation on the router. The actual knowledge structure recognition device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0183] In addition, the technical effects of the bridge structure service performance spatio-temporal evolution prediction device 410 can refer to the technical effects of the method for predicting the spatio-temporal evolution of the bridge structure service performance based on the graph network described in the above method embodiments, which will not be elaborated here.

[0184] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), and this processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or this processor may also be any conventional processor, etc.

[0185] It should also be understood that the memory in the embodiments of the present invention can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchlink DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0186] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0187] It should be understood that the term "and / or" in this document is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after, but it may also represent an "and / or" relationship, which can be specifically understood by referring to the context.

[0188] In the present invention, "at least one" means one or more, and "a plurality" means two or more. "At least one of the following" or its similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, and c can be single or multiple.

[0189] It should be understood that in various embodiments of the present invention, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0190] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in connection with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled professionals may use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0191] Those skilled in the art can clearly understand that for the sake of convenience and brevity of description, the specific working processes of the devices, apparatuses, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0192] In several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be in electrical, mechanical, or other forms.

[0193] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0194] In addition, the functional units in various embodiments of the present invention can be integrated in one processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit.

[0195] When the above-mentioned functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0196] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network, characterized in that, The method includes: S1. Construct a bridge graph model; S2. Based on the bridge graph model and the spatio-temporal attention graph convolutional neural network, construct a basic prediction model for the service performance evolution of the bridge structure; S3. For the basic prediction model for the service performance evolution of the bridge structure, embed a parametric graph learning module that adaptively updates the spatio-temporal dependence relationship of the bridge structure components to obtain an embedded prediction model for the service performance evolution of the bridge structure; S4. Obtain a sample data set of the service performance evolution time series of the structural components of the bridge to be predicted, train the embedded prediction model for the service performance evolution of the bridge structure, obtain the network parameters of the optimal bridge graph model and the spatio-temporal attention graph convolutional neural network, and obtain a trained prediction model for the service performance evolution of the bridge structure; S5. Obtain the service performance evolution time series data of the structural components of the bridge to be predicted, input it into the trained prediction model for the service performance evolution of the bridge structure, and obtain the spatio-temporal evolution prediction result of the bridge structure service performance.

2. The method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network according to claim 1, wherein The construction of the bridge graph model in S1 includes: S11. Based on existing theories, simulate the service performance evolution of important structural components of the bridge, construct a finite element model of the bridge structure, perform static analysis on the bridge structure with service performance evolution, obtain the spatial dependence relationship between the structural components in the bridge, map the spatial dependence relationship to the topological relationship of the graph neural network, and construct a bridge graph structure according to the topological relationship; Among them, the bridge diagram structure is represented as G=(V, E, A), where the nodes (v1, v2, …, v i , …, v j , …, v n ) in the node set V represent bridge structure components, n represents the number of nodes; the edges in the edge set E represent the connection relationships between bridge structure components; A represents the adjacency matrix; S12. Parametrize the material property parameters of the bridge structural components in the bridge graph structure to obtain a bridge graph model.

3. The method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network according to claim 1, wherein The construction of the basic prediction model for the service performance evolution of the bridge structure based on the bridge graph model and the spatio-temporal attention graph convolutional neural network in S2 includes: S21. Obtain a sample data set of the service performance evolution time series of the bridge structural components; S22. Construct a spatio-temporal attention graph convolutional neural network; wherein, the spatio-temporal attention graph convolutional neural network includes: a spatio-temporal attention module and a spatio-temporal graph convolutional module; The spatio-temporal attention module includes: a time attention layer and a spatial attention layer. The time attention layer is used to extract the time attention matrix in the data, and the spatial attention layer is used to extract the spatial attention matrix in the data; The spatio-temporal graph convolutional module includes: a time convolutional layer and a spatial convolutional layer. The time convolutional layer is used to identify the time dependence of adjacent time points, and the spatial convolutional layer is used to capture the spatial dependence of the neighborhood; S23. Based on the sample data set of the service performance evolution time series of the bridge structural components, train the network parameters of the spatio-temporal attention graph convolutional neural network to obtain a basic prediction model for the service performance evolution of the bridge structure.

4. The method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network according to claim 3, wherein The process of obtaining the time attention matrix in S22 includes: Obtain the time attention matrix E according to the following formula (1): Normalize the time attention matrix E through the softmax function to obtain the time attention matrix E' output by the time attention layer i,j , as shown in the following formula (2): Where, V e , U1, U2, U3, b e represent learnable parameters, σ represents an activation function, represents the temporal data information of the input of the model to the r-th layer spatio-temporal block at time step h, T represents the transpose of a matrix, E i,j represents an element in the temporal attention matrix E, E i,j value represents the degree of dependence between time i and j in the temporal attention matrix E, T r-1 represents the length of the temporal dimension of the r-th layer spatio-temporal block.

5. The method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network according to claim 3, wherein The process of obtaining the spatial attention matrix in S22 includes: Obtain the spatial attention matrix S according to the following formula (3): Normalize the spatial attention matrix S through the softmax function to obtain the spatial attention matrix S′ output by the spatial attention layer i,j , as shown in the following equation (4): Wherein, V s , W1, W2, W3, b s represent learnable parameters, T represents the transpose of a matrix, σ represents an activation function, represents the temporal data information of the input of the model to the r-th layer spatio-temporal block at time step h, S i,j represents an element in the spatial attention matrix S, and the value of S i,j represents the degree of dependence between positions i and j in the spatial attention matrix S, and N represents the size of the spatial dimension of the r-th layer spatio-temporal block.

6. The method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network according to claim 1, wherein The parametric graph learning module that adaptively updates the spatio-temporal dependence relationship of the bridge structure components in S3 is used to extract the bidirectional dependence relationship between the nodes of the structural components in the bridge; The extraction of the bidirectional dependence relationship between the nodes of the structural components in the bridge includes: Generate the affinity matrix A1 through the following formula (5): In the formula, M1 and M2 represent learnable parameters, T represents the transpose of the matrix, and Diag(A) is used to generate the weights at the diagonal positions; Obtain the old affinity matrix A through the parametric graph learning module old , and combine the old affinity matrix A old with the generated affinity matrix A1 through the adaptive aggregation module in the parametric graph learning module: S = g(h([A1,A old )) (6) In the formula, S represents the adaptive score matrix obtained by the adaptive aggregation module, which is used to determine the combination weights between the newly learned old and new affinity matrices, g represents a non-linear activation function, and h represents one or more 1×1 convolutional layers; Fuse the old affinity matrix A old with the affinity matrix A1 to generate a new affinity matrix A2: A2 = S ⊙ A1+(1 - S) ⊙ A old (7) In the formula, ⊙ represents element-wise multiplication; According to the new affinity matrix A2, obtain the sparsified affinity matrix A3 through the following formula (8): wherein, the elements of matrix D2 i represents the row of the matrix, and j represents the column of the matrix, ∈(0, 1) represents a threshold value; Normalize the sparse affinity matrix A3 to obtain the affinity matrix A new : In the formula, the elements of matrix D3 7. The method for predicting the spatio-temporal evolution of the service performance of a bridge structure based on a graph network according to claim 1, wherein The obtaining of the sample data set of the service performance evolution time series of the structural components of the bridge to be predicted in S4, training the embedded prediction model for the service performance evolution of the bridge structure, obtaining the optimal network parameters of the bridge graph model and the spatio-temporal attention graph convolutional neural network, and obtaining the trained prediction model for the service performance evolution of the bridge structure includes: S41. Obtain the sample data set of the service performance evolution time series of the structural components of the bridge, and perform preprocessing to obtain the preprocessed sample data set. Train and optimize the final prediction model for the service performance evolution of the bridge structure according to the preprocessed sample data set to obtain the trained final prediction model for the service performance evolution of the bridge structure; S42. Obtain the optimal affinity matrix through the parametric graph learning module, train the final prediction model for the evolution of the service performance of the bridge structure based on the optimal affinity matrix, and obtain the network parameters Θ of the optimized prediction model for the evolution of the service performance of the bridge structure * ; S43. Fix the network parameters Θ of the optimized prediction model for the service performance evolution of the bridge structure * , train the parameterized graph learning module, and learn and update the spatial dependence relationship between nodes; S44. Update the optimal affinity matrix according to the result output by the trained parameterized graph learning module; S45. Determine whether the preset number of iterations is reached. If so, output the optimal model configuration of the final prediction model for the service performance evolution of the bridge structure; if not, go to step S42.

8. A spatio-temporal evolution prediction device for the service performance of a bridge structure based on a graph network, the spatio-temporal evolution prediction device for the service performance of a bridge structure based on a graph network is used to implement the spatio-temporal evolution prediction method for the service performance of a bridge structure based on a graph network according to any one of claims 1-7, characterized in that, The device includes: A bridge graph model construction module for constructing a bridge graph model; A prediction model construction module for constructing a basic prediction model for the service performance evolution of the bridge structure based on the bridge graph model and the spatio-temporal attention graph convolutional neural network; An embedding module for embedding a parameterized graph learning module that adaptively updates the spatio-temporal dependence relationship of the structural components of the bridge into the basic prediction model for the service performance evolution of the bridge structure to obtain the embedded prediction model for the service performance evolution of the bridge structure; A training module for obtaining the sample data set of the service performance evolution time series of the structural components of the bridge to be predicted, training the embedded prediction model for the service performance evolution of the bridge structure, obtaining the optimal network parameters of the bridge graph model and the spatio-temporal attention graph convolutional neural network, and obtaining the trained prediction model for the service performance evolution of the bridge structure; An output module for obtaining the service performance evolution time series data of the structural components of the bridge to be predicted, inputting it into the trained prediction model for the service performance evolution of the bridge structure, and obtaining the spatio-temporal evolution prediction result of the service performance of the bridge structure.

9. A prediction device for the spatio-temporal evolution of the service performance of a bridge structure, characterized in that, The spatio-temporal evolution prediction device for the service performance of the bridge structure includes: A processor; A memory, on which computer-readable instructions are stored. When the computer-readable instructions are executed by the processor, the method described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that, Program code is stored in the computer-readable storage medium, and the program code can be called by the processor to execute the method described in any one of claims 1 to 7.

Images