Intelligent pre-pressing control method and system for high-speed railway continuous rigid frame beam

Abstract

The application provides an intelligent pre-pressing control method and system for a high-speed railway continuous rigid frame beam, and relates to the technical field of pre-pressing control. The method comprises the following steps: acquiring beam body structure design data of the high-speed railway continuous rigid frame beam, a three-dimensional real scene image of a rigid frame beam hanging basket and cast-in-place section construction load parameters; determining multiple key structure response points of the rigid frame beam hanging basket and test pressures of each key structure response point based on the beam body structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real scene image of the rigid frame beam hanging basket; and determining a construction access monitoring result of the continuous rigid frame beam hanging basket based on multiple full-field deformation point information corresponding to each advanced test structure response point and multiple full-field deformation point information corresponding to each key structure response point. The method can accurately simulate the overall form evolution of the rigid frame beam hanging basket under complex loads and determine the construction access monitoring result.

Description

Technical Field

[0001] This invention relates to the field of prestressing control technology, specifically to an intelligent prestressing control method and system for continuous rigid frame beams of high-speed railways. Background Technology

[0002] Continuous rigid frame beams for high-speed railways are the core structure of long-span bridges, and their construction quality directly affects the operational safety and durability of the line. Cantilever casting with hanging baskets is the mainstream construction technique for continuous rigid frame beams in high-speed railways, and preloading with reaction frames is a crucial step before construction. However, existing preloading control methods for reaction frames in continuous rigid frame beams for high-speed railways have significant limitations. Traditional preloading often employs a fixed-point, graded loading method, conducting pressure tests only on a few pre-set key points such as the main truss nodes of the hanging basket and the anchorage points of the suspension belts. This makes it difficult to cover the full-range stress characteristics of the hanging basket structure. Data acquisition focuses only on single-point deformation or stress values, failing to capture the overall deformation correlation of the hanging basket under load. This limitation not only results in one-sided structural response data from preloading tests but also weakens the ability of test results to characterize the overall performance of the hanging basket. Consequently, traditional preloading tests can only rely on data from a few initially pre-set fixed points for evaluation, making it difficult to identify hidden high-risk stress areas in the hanging basket structure where no measuring points are deployed, and lacking a basis for optimizing and adjusting the test plan. Meanwhile, traditional methods cannot dynamically simulate and predict the deformation trend of the formwork during the casting process, which ultimately results in insufficient matching between the pre-stress test and the actual construction conditions, making it difficult to form accurate construction access monitoring results.

[0003] Therefore, how to accurately simulate the overall morphological evolution of the rigid frame beam hanging basket under complex loads and determine the construction access monitoring results is an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem solved by this invention is how to accurately simulate the overall morphological evolution of a rigid frame beam formwork under complex loads and determine the construction access monitoring results.

[0005] According to a first aspect, the present invention provides an intelligent preloading control method for a continuous rigid frame beam of a high-speed railway, comprising: acquiring beam structure design data of the continuous rigid frame beam of the high-speed railway, a three-dimensional real-scene image of the formwork for the rigid frame beam, and construction load parameters of the cast-in-place section; based on the beam structure design data of the continuous rigid frame beam of the high-speed railway and the three-dimensional real-scene image of the formwork for the rigid frame beam, determining multiple key structural response points of the formwork for the rigid frame beam and the test pressure of each key structural response point; applying the test pressure of each key structural response point to the corresponding key structural response point, and acquiring multiple full-field deformation points corresponding to each key structural response point. Information; Based on the information of multiple full-field deformation points corresponding to each key structural response point, determine multiple advanced test structural response points and the test pressure of each advanced test structural response point; Apply the test pressure of each advanced test structural response point to the corresponding advanced test structural response point for pressure testing, and obtain the information of multiple full-field deformation points corresponding to each advanced test structural response point; Based on the information of multiple full-field deformation points corresponding to each advanced test structural response point and the information of multiple full-field deformation points corresponding to each key structural response point, determine the construction access monitoring results of the continuous rigid frame beam hanging basket.

[0006] In one possible implementation, determining multiple advanced test structure response points and the test pressure of each advanced test structure response point based on the information of multiple full-field deformation points corresponding to each key structural response point includes: clustering based on the information of multiple full-field deformation points corresponding to each key structural response point to determine multiple deformation feature clusters corresponding to each key structural response point; determining multiple high-order deformation regions and multiple low-order deformation regions corresponding to each key structural response point based on the multiple deformation feature clusters corresponding to each key structural response point; and determining multiple advanced test structure response points and the test pressure of each advanced test structure response point based on the multiple high-order deformation regions and the multiple low-order deformation regions corresponding to each key structural response point.

[0007] In one possible implementation, determining the construction access monitoring results of the continuous rigid frame beam formwork based on the multiple full-field deformation point information corresponding to each advanced test structural response point and the multiple full-field deformation point information corresponding to each key structural response point includes: generating multiple full-field deformation point information corresponding to multiple simulated structural response points based on the multiple full-field deformation point information corresponding to each advanced test structural response point and the multiple full-field deformation point information corresponding to each key structural response point; constructing a structural response map, wherein the structural response map includes multiple key structural response point nodes, multiple advanced test structural response point nodes, and multiple simulated structural response point nodes, the node feature of each key structural response point node is the multiple full-field deformation point information corresponding to each key structural response point, and the node feature of each advanced test structural response point node is the multiple full-field deformation point information corresponding to each advanced test structural response point. Information on multiple full-field deformation points corresponding to each simulated structural response point is provided. The node features of each simulated structural response point node are information on multiple full-field deformation points corresponding to each simulated structural response point. The structural response map is processed based on a graph neural network to determine the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient. Based on the construction load parameters of the cast-in-place section and the three-dimensional real-scene image of the rigid frame beam formwork, multiple construction stress points of the rigid frame beam formwork and the load intensity change sequence corresponding to each construction stress point are determined. Based on the multiple construction stress points of the rigid frame beam formwork, the load intensity change sequence corresponding to each construction stress point, and the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient, a dynamic morphological simulation video of the continuous rigid frame beam formwork during the pouring process is generated. Based on the dynamic morphological simulation video of the continuous rigid frame beam formwork during the pouring process, the construction access monitoring results of the continuous rigid frame beam formwork are determined.

[0008] In one possible implementation, the input of the graph neural network is the structural response map, and the output of the graph neural network is the overall morphological evolution sequence information of the rigid beam hanging basket under load gradient.

[0009] According to a second aspect, the present invention provides an intelligent preloading control system for a continuous rigid frame beam of a high-speed railway, comprising: a data acquisition module for acquiring beam structure design data of the continuous rigid frame beam of the high-speed railway, a three-dimensional real-scene image of the rigid frame beam formwork, and construction load parameters of the cast-in-place section; a key point determination module for determining multiple key structural response points of the rigid frame beam formwork and the test pressure of each key structural response point based on the beam structure design data of the continuous rigid frame beam of the high-speed railway and the three-dimensional real-scene image of the rigid frame beam formwork; and a key point testing module for applying the test pressure of each key structural response point to the corresponding key structural response point and acquiring multiple full-field deformation point information corresponding to each key structural response point. The system includes: an advanced point determination module, used to determine multiple advanced test structural response points and test pressures for each advanced test structural response point based on multiple full-field deformation point information corresponding to each key structural response point; an advanced point testing module, used to apply the test pressure of each advanced test structural response point to the corresponding advanced test structural response point for pressure testing, and to obtain multiple full-field deformation point information corresponding to each advanced test structural response point; and a monitoring result determination module, used to determine the construction access monitoring results of the continuous rigid frame beam hanging basket based on the multiple full-field deformation point information corresponding to each advanced test structural response point and the multiple full-field deformation point information corresponding to each key structural response point.

[0010] In one possible implementation, the advanced point determination module is further configured to: cluster and determine multiple deformation feature clusters corresponding to each key structural response point based on the multiple full-field deformation point information corresponding to each key structural response point; determine multiple high-level deformation regions and multiple low-level deformation regions corresponding to each key structural response point based on the multiple deformation feature clusters corresponding to each key structural response point; and determine multiple advanced test structural response points and the test pressure of each advanced test structural response point based on the multiple high-level deformation regions and the multiple low-level deformation regions corresponding to each key structural response point.

[0011] In one possible implementation, the monitoring result determination module is further configured to: generate multiple full-field deformation point information corresponding to multiple simulated structural response points based on the multiple full-field deformation point information corresponding to each advanced test structural response point and the multiple full-field deformation point information corresponding to each key structural response point; construct a structural response map, wherein the structural response map includes multiple key structural response point nodes, multiple advanced test structural response point nodes, and multiple simulated structural response point nodes, wherein the node feature of each key structural response point node is the multiple full-field deformation point information corresponding to each key structural response point, the node feature of each advanced test structural response point node is the multiple full-field deformation point information corresponding to each advanced test structural response point, and the node feature of each simulated structural response point node is... Information on multiple full-field deformation points corresponding to each simulated structural response point; processing the structural response spectrum based on a graph neural network to determine the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient; determining multiple construction stress points of the rigid frame beam formwork and the load intensity change sequence corresponding to each construction stress point based on the construction load parameters of the cast-in-place section and the three-dimensional real-scene image of the rigid frame beam formwork; generating a dynamic morphological simulation video of the continuous rigid frame beam formwork during the pouring process based on the multiple construction stress points of the rigid frame beam formwork, the load intensity change sequence corresponding to each construction stress point, and the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient; and determining the construction access monitoring results of the continuous rigid frame beam formwork based on the dynamic morphological simulation video of the continuous rigid frame beam formwork during the pouring process.

[0012] In one possible implementation, the input of the graph neural network is the structural response map, and the output of the graph neural network is the overall morphological evolution sequence information of the rigid beam hanging basket under load gradient.

[0013] According to a third aspect, embodiments of the present invention provide an electronic device, including: a processor; a memory; and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method described above, the method including: acquiring beam structure design data of a high-speed railway continuous rigid frame beam, a three-dimensional real-scene image of the rigid frame beam formwork, and construction load parameters of the cast-in-place section; based on the beam structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real-scene image of the rigid frame beam formwork, determining multiple key structural response points of the rigid frame beam formwork and the test pressure of each key structural response point; and applying the test pressure of each key structural response point to the corresponding key structure. The system identifies response points and acquires information on multiple full-field deformation points corresponding to each key structural response point. Based on this information, it determines multiple advanced test structural response points and the test pressure for each advanced test structural response point. The test pressure for each advanced test structural response point is then applied to the corresponding advanced test structural response point for pressure testing, and information on multiple full-field deformation points corresponding to each advanced test structural response point is acquired. Based on this information and the multiple full-field deformation points corresponding to each key structural response point, the system determines the construction access monitoring results for the continuous rigid frame beam formwork.

[0014] According to the fourth aspect, this embodiment provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements the aforementioned intelligent preloading control method for a continuous rigid frame beam of a high-speed railway. The method includes: acquiring beam structure design data of the continuous rigid frame beam of the high-speed railway, a three-dimensional real-scene image of the rigid frame beam formwork, and construction load parameters of the cast-in-place section; based on the beam structure design data of the continuous rigid frame beam of the high-speed railway and the three-dimensional real-scene image of the rigid frame beam formwork, determining multiple key structural response points of the rigid frame beam formwork and the test pressure of each key structural response point; applying the test pressure of each key structural response point to the corresponding key structural response point, and... Acquire information on multiple full-field deformation points corresponding to each key structural response point; determine multiple advanced test structural response points and test pressures for each advanced test structural response point based on the information on multiple full-field deformation points corresponding to each key structural response point; apply the test pressures for each advanced test structural response point to the corresponding advanced test structural response point for pressure testing, and acquire information on multiple full-field deformation points corresponding to each advanced test structural response point; determine the construction access monitoring results for the continuous rigid frame beam formwork based on the information on multiple full-field deformation points corresponding to each advanced test structural response point and the information on multiple full-field deformation points corresponding to each key structural response point.

[0015] This invention provides an intelligent preloading control method and system for a continuous rigid frame beam in high-speed railway. The method includes acquiring the beam structure design data, a three-dimensional real-world image of the formwork, and construction load parameters of the cast-in-place section of the continuous rigid frame beam; based on the beam structure design data and the three-dimensional real-world image of the formwork, determining multiple key structural response points of the formwork and the test pressure at each key structural response point; applying the test pressure at each key structural response point to the corresponding key structural response point, and acquiring multiple full-field deformation point information corresponding to each key structural response point; and based on the information at each key structural response point... Multiple full-field deformation point information determines multiple advanced test structure response points and the test pressure of each advanced test structure response point; the test pressure of each advanced test structure response point is applied to the corresponding advanced test structure response point for pressure testing, and multiple full-field deformation point information corresponding to each advanced test structure response point is obtained; based on the multiple full-field deformation point information corresponding to each advanced test structure response point and the multiple full-field deformation point information corresponding to each key structure response point, the construction access monitoring results of the continuous rigid frame beam formwork are determined. This method can accurately simulate the overall morphological evolution of the rigid frame beam formwork under complex loads and determine the construction access monitoring results. Attached Figure Description

[0016] Figure 1 A flowchart illustrating an intelligent preloading control method for a continuous rigid frame beam in high-speed railway, provided in an embodiment of the present invention;

[0017] Figure 2 A schematic diagram of a continuous rigid frame beam for high-speed railway provided in an embodiment of the present invention;

[0018] Figure 3 This is a schematic diagram of a cantilever casting formwork structure provided in an embodiment of the present invention;

[0019] Figure 4 This is a flowchart illustrating a method for determining multiple advanced test structure response points and the test pressure of each advanced test structure response point, as provided in an embodiment of the present invention.

[0020] Figure 5 This is a schematic diagram of a process for determining the construction access monitoring results of a continuous rigid frame beam hanging basket, provided by an embodiment of the present invention.

[0021] Figure 6 This is a schematic diagram of an intelligent prestressing control system for a continuous rigid frame beam in a high-speed railway, provided as an embodiment of the present invention. Detailed Implementation

[0022] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0023] In this embodiment of the invention, the following are provided: Figure 1 The above describes an intelligent preloading control method for a continuous rigid frame beam in a high-speed railway. The method includes steps S1 to S6:

[0024] Step S1: Obtain the beam structure design data of the high-speed railway continuous rigid frame beam, the three-dimensional real-scene image of the rigid frame beam hanging basket, and the construction load parameters of the cast-in-place section.

[0025] The beam structure design data for high-speed railway continuous rigid frame beams are design data describing the structure of the high-speed railway continuous rigid frame beams themselves. Figure 2 This is a schematic diagram of a continuous rigid frame beam for high-speed railway provided in an embodiment of the present invention.

[0026] The structural design data for continuous rigid frame beams in high-speed railways includes the beam's dimensions, material strength parameters, load-bearing design standards, and structural connection methods.

[0027] The three-dimensional real-scene image of the rigid frame beam hanging basket is generated by taking multi-angle photos of the hanging basket after on-site installation using laser scanning equipment and processing them with computer graphics technology. The three-dimensional real-scene image of the rigid frame beam hanging basket can reflect the actual assembly state of the hanging basket, the actual connection position of the components, and the surface texture. Figure 3 is a schematic diagram of a cantilever casting hanging basket structure provided by an embodiment of the present invention.

[0028] The construction load parameters for the cast-in-place section are the core construction technical parameters for the stage of concrete pouring. These parameters include the total amount of concrete to be poured, concrete unit weight, concrete pouring speed per hour, pouring sequence, and the movement trajectory of the pouring points.

[0029] Step S2: Based on the beam structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real-scene image of the rigid frame beam hanging basket, determine multiple key structural response points of the rigid frame beam hanging basket and the test pressure of each key structural response point.

[0030] In some embodiments, a structural response point determination model can be used to determine multiple key structural response points of the rigid frame beam formwork and the test pressure at each key structural response point. The structural response point determination model is a convolutional neural network. The input to the structural response point determination model is the beam structure design data of the high-speed railway continuous rigid frame beam and a three-dimensional real-world image of the rigid frame beam formwork. The output of the structural response point determination model is the multiple key structural response points of the rigid frame beam formwork and the test pressure at each key structural response point.

[0031] A convolutional neural network (CNN) is a type of neural network that incorporates convolutional computations and has a deep structure. Each layer of a CNN consists of multiple feature maps, and each feature can be mapped to a plane, with all neurons on the plane sharing a single convolutional kernel. CNNs possess powerful spatial feature extraction and representation learning capabilities.

[0032] The key structural response points are determined by the structural response point determination model. They are the characteristic points on the hanging basket structure used to trigger full-field deformation feedback by applying pressure.

[0033] The test pressure at each critical structural response point is determined by the structural response point determination model, which simulates the pressure value of the actual load for each critical structural response point.

[0034] By applying corresponding test pressure to the key structural response points, the force transmission path of the hanging basket component can drive other parts of the structure to generate a linkage response, thereby making the hanging basket as a whole exhibit a distinctive deformation distribution pattern.

[0035] The structural design data of the continuous rigid frame beam of high-speed railway provides the theoretical mechanical support point position and strength limit value of the connection between the hanging basket and the main beam. The three-dimensional real scene image of the hanging basket of the rigid frame beam reflects the actual geometric posture of the hanging basket on site due to installation errors or deformation, so that the model can establish a mapping relationship between the theoretical strength distribution and the actual physical structure.

[0036] Convolutional neural networks (CNNs) can extract topological features from 3D real-world images of rigid frame beam formwork through convolution kernels and pooling operations, and establish global mechanical mapping relationships by combining these features with the structural design data of continuous rigid frame beams in high-speed railways. In the fully connected layers, the CNN uses nonlinear mapping functions to extrapolate stress distribution trends along each force transmission path and automatically identifies locations with significant curvature changes and high displacement gradients as multiple key structural response points for the rigid frame beam formwork. Subsequently, the CNN, combined with a preset construction load gradient, assigns a pressure value to each point through weight matrix calculations that can elicit a significant structural response but does not exceed the material limit. Finally, it outputs multiple key structural response points of the rigid frame beam formwork and the test pressure at each key structural response point.

[0037] In some embodiments, determining multiple key structural response points of the rigid frame beam formwork and the test pressure at each key structural response point based on the beam structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real-world image of the rigid frame beam formwork includes steps S21 to S23:

[0038] Step S21: Based on the beam structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real-scene image of the rigid frame beam hanging basket, determine the actual resistance moment sequence of the component section, the component connection stiffness distribution map, and the structural stress concentration area distribution map.

[0039] In some embodiments, convolutional neural networks can be used to determine the actual resistance moment sequence of component cross sections, the component connection stiffness distribution map, and the distribution map of structural stress concentration areas.

[0040] The actual moment sequence of the component sections is a numerical set of the bending and torsional resistance capabilities of each key component section of the hanging basket, output by a convolutional neural network. The actual moment sequence of the component sections includes the actual bending stiffness value, torsional stiffness value, and effective stress area data of the corresponding section, distributed along the axial direction of each component.

[0041] The component connection stiffness distribution diagram is output by a convolutional neural network, representing the distribution of physical properties that limit the deformation capacity of each connection part of the hanging basket under actual installation conditions. The component connection stiffness distribution diagram includes the deformation resistance values ​​of each connection part of the hanging basket in the direction of force.

[0042] The stress concentration area distribution map is a numerical diagram output by a convolutional neural network, showing the stress distribution at various locations on the formwork structure under design loads. The stress concentration area distribution map includes the equivalent stress values ​​at each location on the formwork structure.

[0043] The equivalent stress value is a uniaxial stress scalar that represents the overall stress intensity and yield tendency at a certain point inside the hanging basket structure.

[0044] Convolutional neural networks (CNNs) extract theoretical mechanical parameters from beam structure design data and geometric features from 3D real-world images. This allows them to deduce and output the actual moment sequence of component sections and the distribution map of component connection stiffness, reflecting the current physical state of the formwork. Subsequently, the CNN can combine this with the design load distribution to perform mechanical deductions, determining the equivalent stress values ​​at various locations across the entire field, thereby generating a distribution map of concentrated stress areas in the structure.

[0045] Step S22: Based on the actual resistance moment sequence of the component cross section, the component connection stiffness distribution diagram, and the structural stress concentration area distribution diagram, determine the stress concentration risk heat map and multiple key structural response points of the rigid frame beam hanging basket.

[0046] In some embodiments, convolutional neural networks can be used to determine the stress concentration risk heat map and multiple key structural response points of the rigid frame beam hanging basket.

[0047] The stress concentration risk heat map is a probability distribution image of the risk of physical damage or instability in various parts of the hanging basket structure under load conditions, output by a convolutional neural network.

[0048] Convolutional neural networks perform nonlinear correlation analysis on the actual resistance moment sequence of component sections, the distribution map of component connection stiffness, and the distribution map of structural stress concentration areas. At the same time, by comparing the equivalent stress value with the physical limit of the component's bearing capacity, a stress concentration risk heat map reflecting the distribution of structural danger can be derived. Based on this, multiple key structural response points of the rigid frame beam hanging basket can be determined in the geometric center area with extremely high risk coefficient.

[0049] By analyzing the actual moment sequence of component sections and the stiffness distribution map of component connections, convolutional neural networks (CNNs) can derive the theoretical bearing capacity limits for each spatial location of the formwork. Then, the CNN calculates the difference between the equivalent stress values ​​in the stress concentration distribution map and the corresponding theoretical bearing capacity limits. The model uses a nonlinear mapping function to identify regions where the equivalent stress values ​​approach the actual moment values ​​and where there are abrupt gradient changes in the connection stiffness values. It also determines the structural failure risk coefficient at each location by quantifying the redundancy between load strength and material strength, thereby generating a stress concentration risk heatmap. Finally, the CNN identifies the locally closed regions with the highest risk coefficients in the stress concentration risk heatmap and extracts the geometric center coordinates of these regions, thus identifying them as multiple key structural response points for the rigid frame beam formwork.

[0050] Step S23: Determine the test pressure of each key structural response point based on the stress concentration risk heat map and the multiple key structural response points of the rigid frame beam hanging basket.

[0051] In some embodiments, a convolutional neural network can be used to determine the test stress at each critical structural response point.

[0052] The convolutional neural network can extract the structural failure risk coefficient corresponding to each key structural response point in the stress concentration risk heatmap, and combine it with the preset material elastic modulus and allowable stress threshold to calculate the maximum safe load range for each point without permanent plastic damage. Subsequently, within the maximum safe load range, the convolutional neural network can match the pressure value corresponding to the minimum displacement response that enables the key structural response point to reach the sensor recognition accuracy threshold, thereby accurately determining the test pressure for each key structural response point.

[0053] Step S3: Apply the test pressure to each key structural response point and obtain information on multiple full-field deformation points corresponding to each key structural response point.

[0054] The information on multiple full-field deformation points corresponding to each critical structural response point is obtained by applying test pressure to each critical structural response point, relying on the reaction frame to provide rigid force support and achieve precise pressure transmission, and then collecting displacement data of multiple deformation points on the hanging basket structure induced by the pressure through sensors. The information on each full-field deformation point includes the spatial coordinates of the deformation point, the direction of the displacement vector, the displacement amplitude, the strain rate, and the cumulative deviation relative to the initial state.

[0055] Step S4: Based on the information of multiple full-field deformation points corresponding to each key structural response point, determine multiple advanced test structural response points and the test pressure of each advanced test structural response point.

[0056] In some embodiments, Figure 4 This is a flowchart illustrating a method for determining multiple advanced test structure response points and the test pressure at each advanced test structure response point, as provided in an embodiment of the present invention. The determination of the multiple advanced test structure response points and the test pressure at each advanced test structure response point includes steps S41 to S43:

[0057] Step S41: Based on the information of multiple full-field deformation points corresponding to each key structural response point, clustering is performed to determine multiple deformation feature clusters corresponding to each key structural response point.

[0058] The clustering method used is K-means clustering, a commonly used unsupervised learning clustering algorithm. Its core idea is to divide the dataset into K predetermined clusters and continuously update the cluster centers through iterative calculations, so that each data point belongs to the cluster containing the nearest cluster center, ultimately achieving a clustering effect with high similarity among data points within clusters and low similarity among data points between clusters.

[0059] The multiple deformation feature clusters corresponding to each key structural response point are formed by dividing the information of multiple full-field deformation points corresponding to each key structural response point according to the similarity of deformation features using the K-means clustering algorithm. The full-field deformation points within each deformation feature cluster have similar characteristics such as deformation direction and deformation variation trend.

[0060] In some embodiments, the value of K can be determined using a preset relationship table between the value of K and the total number of deformation points of the hanging basket structure across the entire field. The greater the total number of deformation points of the hanging basket structure across the entire field, the larger the value of K. The preset relationship table between the value of K and the total number of deformation points of the hanging basket structure across the entire field is artificially constructed.

[0061] In some embodiments, the process of clustering multiple full-field deformation point information corresponding to each key structural response point using the K-means clustering algorithm is as follows: First, several cluster centers are initialized based on the spatial coordinates and displacement vectors of the multiple full-field deformation point information corresponding to each key structural response point. Simultaneously, the Euclidean distance from each deformation point to each cluster center is calculated, and each point is assigned to its corresponding category according to the minimum distance principle. Then, the positions of the cluster centers are continuously updated to minimize the displacement feature differences among deformation points within the same category, while maximizing the feature differences between different categories. During the iteration process, the algorithm considers the magnitude of the displacement values ​​and the synergy of the displacement directions, effectively separating regions with drastic responses, regions with smooth responses, and regions with abnormal response patterns on the hanging basket structure. When the cluster centers no longer move significantly or the preset number of iterations is reached, the algorithm completes the calculation and finally determines multiple deformation feature clusters corresponding to each key structural response point.

[0062] Clustering can effectively integrate complex information on all-field deformation points. Since the number of all-field deformation points corresponding to each key structural response point is large and their characteristics vary greatly, direct analysis of this information is difficult. However, clustering can group all-field deformation points with similar deformation characteristics into one class and form multiple deformation feature clusters, thus greatly simplifying the data structure and facilitating the extraction of valuable deformation feature information. By dividing the information on each all-field deformation point corresponding to each key structural response point into multiple deformation feature clusters, the different deformation modes and distributions of the hanging basket structure under specific pressure loading conditions can be intuitively displayed. Through the analysis of these deformation feature clusters, the main deformation characteristics and stress response of the hanging basket structure under that pressure condition can be quickly determined. Furthermore, by comparing multiple deformation feature clusters corresponding to different key structural response points, the influence of different pressure points on the deformation response of the hanging basket structure can be clearly discovered.

[0063] Step S42: Based on the multiple deformation feature clusters corresponding to each key structural response point, determine multiple high-level deformation regions and multiple low-level deformation regions corresponding to each key structural response point.

[0064] In some embodiments, a deformation-based analysis model can be used to determine multiple high-order deformation regions and multiple low-order deformation regions corresponding to each key structural response point. The deformation-based analysis model is a deep neural network. The input to the deformation-based analysis model is multiple deformation feature clusters corresponding to each key structural response point, and the output of the deformation-based analysis model is multiple high-order deformation regions and multiple low-order deformation regions corresponding to each key structural response point.

[0065] A deep neural network (DNN) is a multilayer perceptron model consisting of an input layer, multiple hidden layers, and an output layer. DNNs can learn complex nonlinear relationships through backpropagation and abstract data features layer by layer in a high-dimensional space. By connecting multiple layers of neurons and using nonlinear activation functions, DNNs can accurately identify subtle differences in input patterns.

[0066] The multiple high-level deformation zones corresponding to each key structural response point are determined by the deformation quantitative analysis model. After each key structural response point is subjected to test pressure, the area on the hanging basket structure with large deformation that reaches the preset high-level standard is identified.

[0067] The multiple low-level deformation zones corresponding to each key structural response point are determined by the deformation quantitative analysis model. After each key structural response point is subjected to test pressure, the area on the hanging basket structure with small deformation is within the preset low-level range.

[0068] Multiple deformation feature clusters corresponding to each key structural response point can classify the full-field deformation point information with similar deformation features. Each deformation feature cluster contains specific deformation rules and magnitude features. These clustered feature clusters provide structured input data for deep neural networks, enabling the model to accurately determine the deformation level based on the deformation data within the cluster, thereby efficiently and accurately distinguishing between high-magnitude deformation regions and low-magnitude deformation regions.

[0069] Deep neural networks can extract features from multiple deformation feature clusters corresponding to each key structural response point. Then, through layers of hidden layers, they can uncover deeper features such as the magnitude and trend of deformation within each cluster. Based on these extracted features, the deep neural network can determine the magnitude of the region corresponding to each deformation feature cluster. Regions corresponding to clusters with deformation reaching a preset high magnitude standard are defined as high-magnitude deformation regions, while regions corresponding to clusters with deformation within a preset low magnitude range are defined as low-magnitude deformation regions.

[0070] Step S43: Based on the multiple high-level deformation zones corresponding to each key structural response point and the multiple low-level deformation zones, determine multiple advanced test structural response points and the test pressure of each advanced test structural response point.

[0071] In some embodiments, an advanced structural response point determination model can be used to determine multiple advanced test structural response points and the test stress of each advanced test structural response point. The advanced structural response point determination model is a deep neural network. The input to the advanced structural response point determination model consists of multiple high-order deformation regions and multiple low-order deformation regions corresponding to each key structural response point, and the output of the advanced structural response point determination model consists of multiple advanced test structural response points and the test stress of each advanced test structural response point.

[0072] Multiple advanced test structural response points are determined by the advanced structural response point determination model and are supplementary points used in the rigid frame beam hanging basket structure for further accurate testing of the stress performance of the hanging basket structure.

[0073] The test pressure at each advanced test structure response point is set specifically for each advanced test structure response point, and is used to further test its stress condition and structural bearing capacity.

[0074] Multiple high-order deformation zones corresponding to each critical structural response point reveal the possible direction of structural instability, while multiple low-order deformation zones corresponding to each critical structural response point provide a benchmark for stiffness support. By comparing the boundary characteristics and internal gradient changes of these two types of regions, the model can infer subtle response features that were not covered in the initial stress test, and can guide the model to find more representative supplementary test points at the edge of high-risk areas or in response blind zones, thereby improving the coverage accuracy of pre-stress monitoring.

[0075] Deep neural networks can utilize their deep architecture to perform topological analysis on the geometric contours of multiple high-order deformation zones and multiple low-order deformation zones corresponding to each key structural response point. By analyzing the transition zone between high-order and low-order deformation zones, deep neural networks can identify the gradient inflection points where stress transfer is most intense. By calculating the mechanical correlation strength between different regions, the model can deduce sensitive locations that may induce secondary deformation under specific pressures and identify them as multiple advanced test structural response points. Subsequently, based on the structural location and material limits of the identified points, and combined with the nonlinear regression results of historical pre-stress data, the deep neural network matches a corresponding test pressure value to each newly identified point. If the point is located in the core of a high-order deformation zone, the model will allocate a smaller pressure step size to prevent structural damage. If the point is located in the transition zone of a low-order deformation zone, the model will allocate an appropriate pressure value to stimulate its implicit structural characteristics.

[0076] Step S5: Apply the test pressure of each advanced test structure response point to the corresponding advanced test structure response point for stress testing, and obtain information on multiple full-field deformation points corresponding to each advanced test structure response point.

[0077] The information on multiple full-field deformation points corresponding to each advanced test structure response point is obtained by applying test pressure to each advanced test structure response point, relying on the reaction frame to provide rigid force support and achieve precise pressure transmission, and then collecting displacement data of multiple deformation points on the hanging basket structure induced by the pressure through sensors. The information on each full-field deformation point includes the spatial coordinates of the deformation point, the direction of the displacement vector, the displacement amplitude, the strain rate, and the cumulative deviation relative to the initial state.

[0078] Step S6: Based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point, determine the construction access monitoring results of the continuous rigid frame beam formwork.

[0079] In some embodiments, Figure 5This is a flowchart illustrating the process for determining the construction access monitoring results of a continuous rigid frame beam formwork according to an embodiment of the present invention. The determination of the construction access monitoring results of the continuous rigid frame beam formwork includes steps S61 to S66:

[0080] Step S61: Based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point, generate multiple full-field deformation point information corresponding to multiple simulated structure response points.

[0081] In some embodiments, a response point simulation model can be used to generate multiple full-field deformation point information corresponding to multiple simulated structural response points. The response point simulation model is a variational autoencoder. The inputs to the response point simulation model are the multiple full-field deformation point information corresponding to each advanced test structural response point and the multiple full-field deformation point information corresponding to each key structural response point; the output of the response point simulation model is the multiple full-field deformation point information corresponding to multiple simulated structural response points.

[0082] A variational autoencoder (VAE) is a generative model consisting of an encoder and a decoder. By learning the latent probability distribution of the input data, a VAE can generate new samples with similar features. The VAE introduces randomness and normal distribution constraints during the encoding process to ensure the continuity and completeness of the generated latent space, enabling it to be used for high-quality data interpolation and augmentation.

[0083] Multiple simulated structural response points are virtual response locations derived from the measured deformation characteristics by the response point simulation model, used to supplement the overall response data of the hanging basket structure.

[0084] The information of multiple full-field deformation points corresponding to each simulated structural response point is generated by the response point simulation model. After applying simulation test pressure to each simulated structural response point, the displacement data of multiple deformation points on the hanging basket structure induced by the pressure are obtained.

[0085] Each full-field deformation point information includes the location of the simulated deformation point, the simulated test pressure value of the corresponding simulated structural response point, the direction of the simulated displacement vector of the full-field deformation point, the simulated displacement amplitude, the simulated strain rate of change, and the simulated cumulative deviation relative to the initial state.

[0086] By generating multiple full-field deformation point information corresponding to each simulated structural response point, the coverage gaps in the measured deformation data of the hanging basket structure can be effectively supplemented. Since the number of measured points is limited and it is difficult to completely characterize the stress response state of the entire structure, the simulated deformation data and the measured data follow the same physical evolution law, which can fill the gaps in the response distribution between measured points, and thus construct a continuous, high-density full-domain structural deformation response dataset.

[0087] The information on multiple full-field deformation points corresponding to each key structural response point records the basic force feedback of the hanging basket, while the information on multiple full-field deformation points corresponding to each advanced test structural response point provides more detailed local deformation corrections. These two sets of data can serve as real observation samples, thus providing a complete data basis for the response point simulation model to learn the global response law of the structure, and enabling the simulated deformation point information generated by the model to have both physical logic and high realism.

[0088] The variational autoencoder (VAE) first extracts and encodes features from multiple full-field deformation points corresponding to each advanced test structural response point and multiple full-field deformation points corresponding to each key structural response point. The encoder maps the features of both types of deformation data to a latent probability distribution space and calculates the mean and variance of the latent variables. By randomly sampling the latent variables, the encoder obtains latent vectors that conform to this probability distribution, containing the core distribution features of both types of deformation data. The decoder receives these latent vectors and, through inverse mapping using a multi-layer neural network, transforms them into simulated data with a structure consistent with the original deformation data. The model can minimize the reconstruction error between the generated data and the original data and continuously optimize the parameters of the encoder and decoder to ensure that the generated simulated data accurately reflects the patterns of the original deformation data. Ultimately, the VAE generates multiple full-field deformation point information corresponding to multiple simulated structural response points. These simulated data have similar distribution characteristics and force response patterns to the deformation data obtained from real tests.

[0089] Step S62: Construct a structural response map. The structural response map includes multiple key structural response point nodes, multiple advanced test structural response point nodes, and multiple simulated structural response point nodes. The node features of each key structural response point node are multiple full-field deformation point information corresponding to each key structural response point. The node features of each advanced test structural response point node are multiple full-field deformation point information corresponding to each advanced test structural response point. The node features of each simulated structural response point node are multiple full-field deformation point information corresponding to each simulated structural response point.

[0090] A structural response map is a graph data structure composed of multiple response point nodes and the relationships between nodes, used to represent the association status of different response points and their deformation data.

[0091] The key structural response point node represents the basic test location of the hanging basket structure. Each key structural response point node has node characteristics, which include information on multiple full-field deformation points corresponding to that key structural response point.

[0092] The advanced test structure response point node represents the refined test position of the hanging basket structure. Each advanced test structure response point node has node characteristics, which include information on multiple full-field deformation points corresponding to that advanced test structure response point.

[0093] The simulated structural response point node represents the deduced supplementary position of the hanging basket structure. Each simulated structural response point node has node characteristics, which include information on multiple full-field deformation points corresponding to that simulated structural response point.

[0094] Edges are used to connect adjacent response point nodes. Each edge is characterized by the positional relationship between adjacent response point nodes.

[0095] Positional relationships include the relative orientation and geometric distance between adjacent points in three-dimensional space.

[0096] Step S63: Process the structural response map based on graph neural network to determine the overall morphological evolution sequence information of the rigid frame beam hanging basket under load gradient.

[0097] Graph Neural Networks (GNNs) are a type of neural network that can directly process graph data. By transmitting and aggregating messages between nodes, GNNs can capture topological features and inter-node dependencies in the graph. GNNs are suitable for analyzing structural systems with complex connections and can extract global evolutionary patterns from local interactions. The input to the GNN is the structural response graph, and the output is the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient.

[0098] The overall morphological evolution sequence information of the rigid frame beam formwork under load gradient is predicted and output by a graph neural network. The data information shows that the overall structural morphology of the formwork continuously evolves with the load as the load level increases.

[0099] Load gradient refers to a pre-defined sequence of load magnitudes arranged in specific increments. For example, a graded load sequence based on the maximum construction load of a continuous rigid frame beam, divided into proportions of 20%, 40%, 60%, 80%, 100%, and 120%.

[0100] The overall morphological evolution sequence information of the rigid frame beam formwork under load gradient includes the load values ​​corresponding to each load gradient, the three-dimensional displacement vector of the formwork structure, the rotation angle values ​​of each key structural section, and the cumulative deflection and lateral displacement relative to the initial state.

[0101] In the structural response map, each node represents a response point on the formwork structure, and the node features include the full-field deformation information of the corresponding point. These features directly reflect the deformation state of each response point under load. The edges between nodes represent the positional relationship between different response points, including relative orientation and geometric distance. This positional relationship information reflects the mutual influence between the deformations of different parts of the formwork, such as the synergistic relationship between stress transmission and deformation. By processing the map data through a graph neural network, the overall morphological evolution sequence information of the rigid frame beam formwork under load gradients can be determined. This helps to fully characterize the continuous deformation process of the formwork from low load to high load and accurately capture the evolution law of structural morphology.

[0102] Graph neural networks (GNNs), through graph convolution operations, can dynamically calculate the influence weights of adjacent nodes based on the distance and physical connectivity attributes between nodes in the structural response map. Through a message-passing mechanism, GNNs can aggregate the node features of each key structural response point, each advanced test structural response point, and each simulated structural response point in multiple rounds. During aggregation, the model corrects the transmission of displacement information based on the truss stiffness or sling tension represented by the edges. Using multi-layer nonlinear activation functions, the model can identify the cooperative deformation patterns of the formwork structure under stress. As the depth of graph convolution increases, node features are gradually integrated into global topological dependencies to form an embedded representation reflecting the overall deformation. Next, the GNN can sequentially model the node states under different load steps and analyze the rate of displacement change and acceleration with increasing load gradient. Finally, the model transforms these evolutionary features into a sequence of overall morphological evolution information of the rigid frame formwork under load gradients, accurately depicting the entire process of the formwork's morphological change from unloaded to fully loaded.

[0103] Step S64: Based on the construction load parameters of the cast-in-place section and the three-dimensional real-scene image of the rigid frame beam hanging basket, determine multiple construction stress points of the rigid frame beam hanging basket and the load intensity change sequence corresponding to each construction stress point.

[0104] In some embodiments, a construction stress point analysis model can be used to determine multiple construction stress points of the rigid frame beam formwork and the load intensity variation sequence corresponding to each construction stress point. The construction stress point analysis model is a deep neural network. The inputs to the construction stress point analysis model are the construction load parameters of the cast-in-place section and a three-dimensional real-world image of the rigid frame beam formwork. The outputs of the construction stress point analysis model are the multiple construction stress points of the rigid frame beam formwork and the load intensity variation sequence corresponding to each construction stress point.

[0105] The multiple stress points of the rigid frame beam hanging basket are output from the construction stress point analysis model, and are the key locations that need to bear the construction load during the construction of the cast-in-place section to be poured.

[0106] The load intensity change sequence corresponding to each construction stress point is output by the construction stress point analysis model. It is a time series data sequence of the load intensity borne by each construction stress point during the construction of the cast-in-place section to be poured, which varies with time.

[0107] The construction load parameters of the cast-in-place section clearly define key information such as the load magnitude, loading speed, and load distribution during the pouring process. The three-dimensional real-scene image of the rigid frame beam hanging basket clearly presents the structural layout, load-bearing component positions, support methods, and other structural features of the hanging basket. Through these load data and structural features, the model can accurately identify the stress points during the construction process and analyze the load intensity change sequence corresponding to each construction stress point.

[0108] Deep neural networks (DNNs) perform mesh-based analysis of the geometric surfaces in the 3D real-world image of the rigid frame beam formwork using multiple hidden layers. The DNN can transform the pouring speed and movement trajectory in the construction load parameters of the cast-in-place section into spatial load vectors, and utilize the model's feature extraction layer to identify high-frequency regions where the formwork surface contacts the poured material. During calculations, the model can consider the normal pressure and frictional forces generated by the concrete's unit weight on formwork with different slopes. By learning the rheological properties of concrete, the DNN can simulate the material accumulation process on the formwork's bottom plate and automatically identify key load-bearing points, thus defining them as multiple construction stress points of the rigid frame beam formwork. Subsequently, by combining the construction timeline, the model can dynamically simulate the mass accumulation at each stress point, and then calculate the stress amplitude at each point at each time point through a time-series processing unit. Finally, the model can generate a load intensity change sequence corresponding to each construction stress point.

[0109] Step S65: Based on the multiple construction stress points of the rigid frame beam formwork, the load intensity change sequence corresponding to each construction stress point, and the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient, a dynamic morphological simulation video of the continuous rigid frame beam formwork during the pouring process is generated.

[0110] In some embodiments, a casting simulation model can be used to generate a dynamic morphological simulation video of the continuous rigid frame beam formwork during the casting process. The casting simulation model is a variational autoencoder. The inputs to the casting simulation model are multiple construction stress points of the rigid frame beam formwork, the load intensity change sequence corresponding to each construction stress point, and the overall morphological evolution sequence information of the rigid frame beam formwork under load gradients. The output of the casting simulation model is a dynamic morphological simulation video of the continuous rigid frame beam formwork during the casting process.

[0111] The dynamic morphology simulation video of the continuous rigid frame beam formwork during the pouring process is generated by the pouring simulation model. It is a simulation video data of the overall morphology of the rigid frame beam formwork dynamically changing with the pouring progress during the construction of the in-situ section to be poured.

[0112] The dynamic morphology simulation video of the continuous rigid frame beam formwork during the pouring process can simulate the structural deformation state of the formwork at various construction time points, the deformation details of key parts, and the dynamic evolution of the overall shape, as well as other visual information.

[0113] The multiple stress points of the rigid frame beam formwork and the corresponding load intensity change sequence at each stress point constitute the real external driving force during the pouring process. Meanwhile, the overall morphological evolution sequence of the rigid frame beam formwork under load gradients reflects the formwork's own response resistance and deformation modes. Based on this information, the pouring simulation model can accurately predict the pouring process that has not yet occurred, relying on real mechanical feedback laws.

[0114] Variational autoencoders (VAEs) possess powerful cross-modal data generation capabilities. The encoder transforms the overall morphological evolution sequence of a rigid frame beam formwork under load gradients into a structural stiffness map in the implicit space. Simultaneously, the encoder encodes multiple construction stress points of the rigid frame beam formwork and the corresponding load intensity change sequence for each stress point into dynamic excitation vectors. Within the implicit space of the VAE, a physical correlation between external loads and structural response can be established. The decoder utilizes these fused features, combined with the geometric topological constraints of the formwork, to generate a deformation mesh of the formwork frame-by-frame at different pouring times. The model employs nonlinear smoothing techniques to ensure displacement continuity between video frames and simulates the dynamic fluctuation effects caused by load movement, ensuring that each generated frame accurately reflects the stress state and deformation response under the current construction progress. Ultimately, the VAE can output a dynamic morphological simulation video of a continuous rigid frame beam formwork during the pouring process, enabling perception of the morphological evolution throughout the entire construction process.

[0115] Step S66: Based on the dynamic morphology simulation video of the continuous rigid frame beam formwork during the pouring process, determine the construction access monitoring results of the continuous rigid frame beam formwork.

[0116] In some embodiments, a formwork analysis model can be used to determine the construction access monitoring results of the formwork for continuous rigid frame beams. The formwork analysis model is a Transformer model. The input to the formwork analysis model is a dynamic morphological simulation video of the formwork during the pouring process, and the output of the formwork analysis model is the construction access monitoring results of the formwork for continuous rigid frame beams.

[0117] The Transformer model is a deep learning model based on a self-attention mechanism. Through multi-head self-attention, the Transformer model can capture long-range dependencies in the input sequence and perform global weighted analysis on each element of the sequence. In video processing tasks, the Transformer can efficiently identify key anomalous features in the temporal dimension.

[0118] The construction access monitoring results for continuous rigid frame beam formwork are evaluation indicators used to determine whether the formwork meets construction conditions and safety standards through a formwork analysis model. The construction access monitoring results for continuous rigid frame beam formwork include a conclusion of either "qualified" or "unqualified."

[0119] Each frame of the dynamic morphology simulation video of the continuous rigid frame beam formwork during the casting process contains displacement information of various parts of the formwork, and the correlation between frames reflects the rate of deformation development. This highly integrated visual physical information provides the model with an intuitive and comprehensive analysis object. By identifying abnormal deformation trends in the video, the model can comprehensively judge the safety of the formwork under extreme casting conditions, thereby determining the admission conclusion.

[0120] The Transformer model utilizes its self-attention mechanism to perform multi-scale scanning of simulated videos of the dynamic morphology of the formwork during the pouring process of a continuous rigid frame beam. First, the Transformer model converts the simulated video frame sequence into feature embedding vectors and introduces positional encoding to preserve the temporal sequence information of the construction. The multi-head self-attention layer of the Transformer model focuses on key parts of the formwork in each frame, thereby identifying sensitive areas where displacement exceeds safety thresholds or deformation rates abruptly change. By calculating the correlation between features at different time points, the Transformer model can analyze whether the deformation has divergence risks or signs of instability. The model compares the identified deformation features with preset safety evaluation standards for continuous rigid frame beams in high-speed railways. If the model finds that the deflection or nodal displacement difference at a certain moment in the video sequence exceeds the design allowable range, the output layer will generate a high-probability risk signal. The model then comprehensively scores the stability of the entire process and calculates the minimum safety factor for the formwork during the entire pouring process. Finally, the Transformer model determines the construction access monitoring results of the continuous rigid frame beam formwork based on the scoring results.

[0121] Based on the same inventive concept Figure 6 This invention provides a schematic diagram of an intelligent preloading control system for a continuous rigid frame beam in high-speed railway. The intelligent preloading control system for the continuous rigid frame beam in high-speed railway includes:

[0122] Data acquisition module 71 is used to acquire beam structure design data of continuous rigid frame beams of high-speed railway, three-dimensional real-scene images of rigid frame beam hanging baskets, and construction load parameters of cast-in-place sections.

[0123] The key point determination module 72 is used to determine multiple key structural response points of the rigid frame beam hanging basket and the test pressure of each key structural response point based on the beam structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real-scene image of the rigid frame beam hanging basket.

[0124] The key point testing module 73 is used to apply the test pressure of each key structural response point to the corresponding key structural response point, and to obtain information on multiple full-field deformation points corresponding to each key structural response point.

[0125] The advanced point determination module 74 is used to determine multiple advanced test structure response points and the test pressure of each advanced test structure response point based on the information of multiple full-field deformation points corresponding to each key structure response point.

[0126] The advanced test point module 75 is used to apply the test pressure of each advanced test structure response point to the corresponding advanced test structure response point for stress testing, and to obtain multiple full-field deformation point information corresponding to each advanced test structure response point.

[0127] The monitoring result determination module 76 is used to determine the construction access monitoring results of the continuous rigid frame beam hanging basket based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point.

[0128] It should be noted that, in order to simplify the descriptions disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments of this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.

[0129] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A method for intelligent preloading control of continuous rigid frame beams for high-speed railways, characterized in that, include: Obtain structural design data of continuous rigid frame beams for high-speed railways, 3D real-scene images of the rigid frame beam hanging basket, and construction load parameters of the cast-in-place section; Based on the beam structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real-scene image of the rigid frame beam hanging basket, the key structural response points of the rigid frame beam hanging basket and the test pressure of each key structural response point are determined. The test pressure of each key structural response point is applied to the corresponding key structural response point, and the information of multiple full-field deformation points corresponding to each key structural response point is obtained. Based on the information of multiple full-field deformation points corresponding to each key structural response point, multiple advanced test structural response points and the test pressure of each advanced test structural response point are determined. The determination of multiple advanced test structural response points and the test pressure of each advanced test structural response point based on the information of multiple full-field deformation points corresponding to each key structural response point includes: Based on the information of multiple full-field deformation points corresponding to each key structural response point, clustering is performed to determine multiple deformation feature clusters corresponding to each key structural response point. Based on the multiple deformation feature clusters corresponding to each key structural response point, multiple high-level deformation regions and multiple low-level deformation regions are determined for each key structural response point. Based on the multiple high-level deformation regions and multiple low-level deformation regions corresponding to each key structural response point, multiple advanced test structural response points are determined, along with the test pressure for each advanced test structural response point. The determination of these points, and the test pressure for each advanced test structural response point, includes: using an advanced structural response point determination model to determine the multiple advanced test structural response points and the test pressure for each advanced test structural response point. The advanced structural response point determination model is a deep neural network, and its input is... The model defines multiple high-order deformation zones and multiple low-order deformation zones corresponding to each key structural response point. The output of the advanced structural response point determination model is multiple advanced test structural response points and the test pressure of each advanced test structural response point. By calculating the mechanical correlation strength between different regions, the advanced structural response point determination model can deduce sensitive locations that may induce secondary deformation under specific pressure and identify them as multiple advanced test structural response points. Subsequently, based on the structural location and material limits of the identified points and combined with the nonlinear regression results of historical pre-stress data, the advanced structural response point determination model matches the corresponding test pressure value for each newly determined point. The test pressure of each advanced test structure response point is applied to the corresponding advanced test structure response point for stress testing, and the information of multiple full-field deformation points corresponding to each advanced test structure response point is obtained. Based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point, the construction access monitoring results of the continuous rigid frame beam formwork are determined.

2. The intelligent preloading control method for high-speed railway continuous rigid frame beams as described in claim 1, characterized in that, The construction access monitoring results for the continuous rigid frame beam formwork, determined based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point, include: Based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point, multiple full-field deformation point information corresponding to multiple simulated structure response points is generated. A structural response map is constructed, which includes multiple key structural response point nodes, multiple advanced test structural response point nodes, and multiple simulated structural response point nodes. The node features of each key structural response point node are multiple full-field deformation point information corresponding to each key structural response point. The node features of each advanced test structural response point node are multiple full-field deformation point information corresponding to each advanced test structural response point. The node features of each simulated structural response point node are multiple full-field deformation point information corresponding to each simulated structural response point. The structural response spectrum is processed based on graph neural network to determine the overall morphological evolution sequence information of the rigid frame beam hanging basket under load gradient. Based on the construction load parameters of the cast-in-place section and the three-dimensional real-scene image of the rigid frame beam hanging basket, multiple construction stress points of the rigid frame beam hanging basket and the load intensity change sequence corresponding to each construction stress point are determined. Based on the multiple construction stress points of the rigid frame beam formwork, the load intensity change sequence corresponding to each construction stress point, and the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient, a dynamic morphological simulation video of the continuous rigid frame beam formwork during the pouring process is generated. Based on the dynamic morphology simulation video of the continuous rigid frame beam formwork during the pouring process, the construction access monitoring results of the continuous rigid frame beam formwork were determined.

3. The intelligent preloading control method for high-speed railway continuous rigid frame beams as described in claim 2, characterized in that, The input of the graph neural network is the structural response map, and the output of the graph neural network is the overall morphological evolution sequence information of the rigid beam hanging basket under the load gradient.

4. An intelligent preloading control system for a continuous rigid frame beam in high-speed railway, characterized in that, include: The data acquisition module is used to acquire beam structure design data of continuous rigid frame beams for high-speed railways, three-dimensional real-scene images of the rigid frame beam hanging basket, and construction load parameters of the cast-in-place section. The key point determination module is used to determine multiple key structural response points of the rigid frame beam hanging basket and the test pressure of each key structural response point based on the beam structure design data of the high-speed railway continuous rigid frame beam and the three-dimensional real-scene image of the rigid frame beam hanging basket. The key point testing module is used to apply the test pressure to each key structural response point and obtain information on multiple full-field deformation points corresponding to each key structural response point. The advanced point determination module is used to determine multiple advanced test structure response points and the test pressure of each advanced test structure response point based on multiple full-field deformation point information corresponding to each key structural response point. The advanced point determination module is also used for: Based on the information of multiple full-field deformation points corresponding to each key structural response point, clustering is performed to determine multiple deformation feature clusters corresponding to each key structural response point. Based on the multiple deformation feature clusters corresponding to each key structural response point, multiple high-level deformation regions and multiple low-level deformation regions are determined for each key structural response point. Based on the multiple high-level deformation regions and multiple low-level deformation regions corresponding to each key structural response point, multiple advanced test structural response points are determined, along with the test pressure for each advanced test structural response point. The determination of these points, and the test pressure for each advanced test structural response point, includes: using an advanced structural response point determination model to determine the multiple advanced test structural response points and the test pressure for each advanced test structural response point. The advanced structural response point determination model is a deep neural network, and its input is... The model defines multiple high-order deformation zones and multiple low-order deformation zones corresponding to each key structural response point. The output of the advanced structural response point determination model is multiple advanced test structural response points and the test pressure of each advanced test structural response point. By calculating the mechanical correlation strength between different regions, the advanced structural response point determination model can deduce sensitive locations that may induce secondary deformation under specific pressure and identify them as multiple advanced test structural response points. Subsequently, based on the structural location and material limits of the identified points and combined with the nonlinear regression results of historical pre-stress data, the advanced structural response point determination model matches the corresponding test pressure value for each newly determined point. The advanced test point module is used to apply the test pressure of each advanced test structure response point to the corresponding advanced test structure response point for stress testing, and to obtain information on multiple full-field deformation points corresponding to each advanced test structure response point. The monitoring result determination module is used to determine the construction access monitoring results of the continuous rigid frame beam hanging basket based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point.

5. The intelligent preloading control system for high-speed railway continuous rigid frame beams as described in claim 4, characterized in that, The monitoring result determination module is also used for: Based on the information of multiple full-field deformation points corresponding to each advanced test structure response point and the information of multiple full-field deformation points corresponding to each key structure response point, multiple full-field deformation point information corresponding to multiple simulated structure response points is generated. A structural response map is constructed, which includes multiple key structural response point nodes, multiple advanced test structural response point nodes, and multiple simulated structural response point nodes. The node features of each key structural response point node are multiple full-field deformation point information corresponding to each key structural response point. The node features of each advanced test structural response point node are multiple full-field deformation point information corresponding to each advanced test structural response point. The node features of each simulated structural response point node are multiple full-field deformation point information corresponding to each simulated structural response point. The structural response spectrum is processed based on graph neural network to determine the overall morphological evolution sequence information of the rigid frame beam hanging basket under load gradient. Based on the construction load parameters of the cast-in-place section and the three-dimensional real-scene image of the rigid frame beam hanging basket, multiple construction stress points of the rigid frame beam hanging basket and the load intensity change sequence corresponding to each construction stress point are determined. Based on the multiple construction stress points of the rigid frame beam formwork, the load intensity change sequence corresponding to each construction stress point, and the overall morphological evolution sequence information of the rigid frame beam formwork under load gradient, a dynamic morphological simulation video of the continuous rigid frame beam formwork during the pouring process is generated. Based on the dynamic morphology simulation video of the continuous rigid frame beam formwork during the pouring process, the construction access monitoring results of the continuous rigid frame beam formwork were determined.

6. The intelligent preloading control system for high-speed railway continuous rigid frame beams as described in claim 5, characterized in that, The input of the graph neural network is the structural response map, and the output of the graph neural network is the overall morphological evolution sequence information of the rigid beam hanging basket under the load gradient.

7. An electronic device, characterized in that, include: processor; Memory; And a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the intelligent preloading control method for high-speed railway continuous rigid frame beams as described in any one of claims 1 to 3.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the intelligent preloading control method for continuous rigid frame beams of high-speed railways as described in any one of claims 1 to 3.

Images