Safety assessment method, system and equipment for intelligently building bridge

Abstract

The invention provides a safety assessment method, system and equipment for intelligently constructing a bridge. The method comprises the following steps: acquiring associated data in a bridge construction process; extracting vibration mode parameters of the bridge structure based on the associated data; carrying out similar theory modeling on geometric dimensions and material attributes in the bridge design drawing, constructing a reduced scale proportion physical test model, simulating mechanical response of the bridge under static loading and dynamic loading, and obtaining model test data; acquiring field measured data of the bridge, performing dimensionless processing on the field measured data and the model test data, and calculating to obtain a similar criterion deviation degree; performing clustering analysis on historical bridge disease case data, establishing a disease feature classification system based on Euclidean distance, matching current detection data of a bridge with the disease feature classification system, and identifying a disease type; and comprehensively evaluating the identified disease type, the similarity criterion deviation degree and the vibration mode parameter, and determining the bridge safety risk level. According to the invention, the evaluation result is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a safety assessment method, system and equipment for intelligently building bridges. Background Art

[0002] In modern bridge construction, traditional bridge safety assessment methods rely primarily on single data sources or simple empirical judgments, making it difficult to fully and accurately reflect the actual safety status of bridges. For example, early assessment methods often relied on manual inspections to observe visible defects such as cracks and deformations on the bridge surface. This method is not only inefficient but also fails to effectively detect internal structural damage and hidden defects, posing a significant safety hazard.

[0003] With technological advancements, some assessment methods have begun to incorporate sensor technology to collect data such as bridge stress and displacement. However, existing technologies still have limitations in data processing and analysis. On the one hand, conventional methods for collecting dynamic data such as vibration and stress often use simple time-domain analysis, failing to fully exploit the structural characteristics contained in the data. This results in the inability to accurately identify key indicators such as the bridge's vibration mode parameters, which affects the assessment of the bridge's dynamic performance. On the other hand, in terms of structural performance assessment, existing analysis methods based on finite element models often ignore factors such as fluctuations in material properties and geometric dimension deviations during actual construction. This leads to significant discrepancies between the model and the actual structure, making it difficult to accurately simulate the mechanical response of the bridge.

[0004] In addition, in the process of disease identification and risk assessment, existing technologies usually use a single indicator threshold to determine the type of disease, lack comprehensive analysis of multi-source data, and cannot consider the mutual influence between different diseases and the synergistic effect of disease development on the overall safety of the structure.

[0005] Therefore, the assessment results of current bridge safety assessment methods lack accuracy. Summary of the Invention

[0006] The main purpose of the present invention is to provide a safety assessment method, system and equipment for intelligently constructed bridges, aiming to overcome the defect that the assessment results of current bridge safety assessment methods lack accuracy.

[0007] To achieve the above object, the present invention provides a safety assessment method for intelligently constructed bridges, comprising the following steps:

[0008] Collect relevant data during the bridge construction process;

[0009] Performing Fourier transform on the correlation data to obtain frequency domain signal data to extract mode shape parameters of the bridge structure;

[0010] Conduct similarity theory modeling on the geometric dimensions and material properties in the bridge design drawings, construct a scaled physical test model, and simulate the mechanical response of the bridge under static and dynamic loads to obtain model test data;

[0011] Obtain the field measured data of the bridge, perform dimensionless processing on the field measured data and the model test data, and calculate the deviation of the similarity criterion;

[0012] Cluster analysis was performed on historical bridge damage case data to establish a damage feature classification system based on Euclidean distance. The current bridge inspection data was matched with the system to identify the damage type.

[0013] A comprehensive assessment is conducted on the identified damage types, similarity criterion deviations, and vibration mode parameters to determine the bridge safety risk level.

[0014] Furthermore, relevant data during the bridge construction process is collected, including:

[0015] The stress data, displacement data, temperature and humidity data during the bridge construction process are collected based on multi-source sensors; abnormal data points are eliminated based on the median filtering algorithm, and the related data are obtained.

[0016] Furthermore, the field measured data of the bridge is obtained, and the field measured data and the model test data are dimensionlessly processed to calculate the similarity criterion deviation, including:

[0017] The raw data collected by sensors at key parts of the bridge are segmented and filtered to eliminate invalid data segments with missing data to obtain valid on-site measured data;

[0018] Based on similarity theory, geometric dimension ratio, stress distribution ratio, and time response ratio are selected as basic similarity criteria. The ratios of field measured data and model test data under each similarity criterion are calculated respectively, and each ratio is compared with the theoretical benchmark value to obtain the degree of deviation of each similarity criterion.

[0019] According to the influence of various physical quantities on structural safety in the bridge design specifications, different weights are assigned to the deviation degrees of various similarity criteria, and the deviation degrees are weighted and summarized to obtain the similarity criterion deviation degree.

[0020] Furthermore, the raw data is collected by acceleration sensors, displacement sensors and strain sensors installed in the middle of the span of the bridge main beam, the top of the pier, and the connection of the support.

[0021] Furthermore, a comprehensive evaluation is conducted on the identified damage types, similarity criterion deviations, and mode parameters to determine the bridge safety risk level, including:

[0022] The identified damage type, similarity criterion deviation, and mode parameters are used as indicators and network nodes, respectively. Based on the bridge structure mechanical conduction path and historical accident data, the risk propagation correlation between network nodes is established.

[0023] Using the tolerance band analysis method, a safety threshold interval is set for each indicator. When any indicator exceeds the safety threshold interval, the coupling effect calculation is triggered. The coupling effect calculation includes: calculating the impact coefficient of the abnormal indicator on other indicators based on the risk propagation correlation relationship;

[0024] A three-dimensional assessment coordinate system was established, with the severity of the disease type as the X-axis, the deviation of the similarity criterion as the Y-axis, and the change in the vibration parameter as the Z-axis, and 27 assessment areas were divided; each area corresponds to a different risk level standard, and the bridge safety risk level is determined through spatial coordinate positioning.

[0025] Furthermore, a comprehensive evaluation is conducted on the identified damage types, similarity criterion deviations, and mode parameters to determine the bridge safety risk level, including:

[0026] The identified damage types, similarity criterion deviations, and mode parameters are tensor-encoded, and a feature correlation tensor is constructed in combination with the bridge structure topology information to obtain a risk feature tensor.

[0027] Mapping the risk feature tensor to a low-dimensional non-Euclidean space through manifold learning to obtain a risk manifold with geometric features;

[0028] A butterfly effect simulation is performed on the risk manifold, and the initial disturbance threshold of each indicator is set. A Lagrangian mechanics model is used to simulate the propagation trajectory of data changes on the risk manifold. The risk amplification effect is quantified by calculating the Lyapunov exponent to obtain the risk sensitivity assessment result.

[0029] Performing topological clustering on the risk sensitivity assessment results, analyzing the topological structure of the risk manifold, identifying holes and connected components and performing clustering to obtain risk evolution pattern classification results;

[0030] The risk evolution pattern classification results are subjected to probabilistic reasoning based on the Bayesian decision tree, and the decision weight is dynamically adjusted in combination with the real-time working conditions of the bridge to ultimately determine the safety risk level of the bridge.

[0031] Furthermore, a comprehensive evaluation is conducted on the identified damage types, similarity criterion deviations, and mode parameters to determine the bridge safety risk level, including:

[0032] The identified disease types, similarity criterion deviations, and mode shape parameters are subjected to biological lattice encoding processing. The disease types are mapped to lattice defect types, the similarity criterion deviations are quantified as the degree of lattice distortion, and the mode shape parameters are converted into lattice vibration frequencies. A three-dimensional data lattice with biomimetic mechanical properties is constructed to obtain a risk characteristic lattice model.

[0033] Performing multi-body dynamic modeling on the risk characteristic lattice model, treating the nodes of the lattice as rigid bodies, establishing elastic connection relationships between the nodes, simulating the dynamic response of the bridge structure under load, and solving to obtain a dynamically evolving multi-body risk model;

[0034] The multi-body risk model is subjected to network deconstruction, where the nodes of the lattice are abstracted into network nodes. The connection strength between nodes corresponds to the degree of dynamic coupling. The network topology characteristics are analyzed using betweenness centrality and clustering coefficient in graph theory to identify the key nodes and core paths of risk propagation, thus obtaining a network topology map of risk propagation.

[0035] Perform cascading failure simulation on the network topology of the risk propagation, set a node failure threshold, and trigger a cascading failure when the node risk value exceeds the failure threshold; simulate the avalanche-like spread of risk in the network, simulate and calculate the failure probability distribution under different working conditions, and obtain the probability assessment result of risk spread;

[0036] A multi-party game model involving bridge management, maintenance teams, and transportation departments is constructed. The probability assessment results are used as profit function parameters, and game decisions are made on the probability assessment results to determine the bridge safety risk level.

[0037] The present invention also provides a safety assessment system for intelligently constructed bridges, comprising:

[0038] The acquisition module is used to collect relevant data during the bridge construction process;

[0039] An extraction module, configured to perform Fourier transform on the correlation data to obtain frequency domain signal data to extract mode shape parameters of the bridge structure;

[0040] The simulation module is used to perform similarity theoretical modeling on the geometric dimensions and material properties in the bridge design drawings, construct a scaled physical test model, and simulate the mechanical response of the bridge under static and dynamic loads to obtain model test data;

[0041] The calculation module is used to obtain the field measured data of the bridge, perform dimensionless processing on the field measured data and the model test data, and calculate the deviation of the similarity criterion;

[0042] The recognition module is used to perform cluster analysis on historical bridge damage case data, establish a damage feature classification system based on Euclidean distance, match the current bridge inspection data with it, and identify the damage type;

[0043] The evaluation module is used to comprehensively evaluate the identified damage types, similarity criterion deviations and vibration mode parameters to determine the bridge safety risk level.

[0044] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0045] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0046] The present invention provides a safety assessment method, system, and device for intelligently constructed bridges, comprising: collecting relevant data from the bridge construction process; performing a Fourier transform on the relevant data to obtain frequency domain signal data to extract the modal parameters of the bridge structure; performing similarity theory modeling on the geometric dimensions and material properties in the bridge design drawings, constructing a scaled physical test model, and simulating the mechanical response of the bridge under static and dynamic loads to obtain model test data; obtaining field-measured data of the bridge, dimensionlessly processing the field-measured data and the model test data, and calculating the similarity criterion deviation; performing cluster analysis on historical bridge defect case data, establishing a defect feature classification system based on Euclidean distance, matching the current bridge inspection data with the system, and identifying the defect type; and comprehensively evaluating the identified defect type, similarity criterion deviation, and modal parameters to determine the bridge safety risk level. In the present invention, by identifying the defect type, similarity criterion deviation, and modal parameters during the bridge construction process and then performing a comprehensive assessment, the final result takes into account the influence of dimensionality, resulting in a more accurate assessment result. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a schematic diagram of the steps of a safety assessment method for intelligently building a bridge in one embodiment of the present invention;

[0048] Figure 2 This is a structural block diagram of a safety assessment system for intelligently constructed bridges according to an embodiment of the present invention;

[0049] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0050] The implementation, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0052] Reference Figure 1 In one embodiment of the present invention, a safety assessment method for intelligently constructed bridges is provided, comprising the following steps:

[0053] Step S1, collecting relevant data during the bridge construction process;

[0054] Step S2, performing Fourier transform on the correlation data to obtain frequency domain signal data to extract mode shape parameters of the bridge structure;

[0055] Step S3: Perform similarity theory modeling on the geometric dimensions and material properties in the bridge design drawings, construct a scaled physical test model, and simulate the mechanical response of the bridge under static and dynamic loads to obtain model test data;

[0056] Step S4, obtaining the field measured data of the bridge, performing dimensionless processing on the field measured data and the model test data, and calculating the similarity criterion deviation;

[0057] Step S5: Perform cluster analysis on historical bridge defect case data, establish a defect feature classification system based on Euclidean distance, match the current bridge inspection data with it, and identify the defect type;

[0058] Step S6: Comprehensively evaluate the identified damage types, similarity criterion deviations, and mode parameters to determine the bridge safety risk level.

[0059] In this embodiment, as described in step S1 above, the collection of relevant data during the bridge construction process involves deploying accelerometers, strain sensors, and displacement sensors at key locations throughout the bridge construction process (e.g., pier tops, girder midspans, and support connections) to collect dynamic data such as vibration acceleration, stress and strain, and displacement deformation generated by the structure under construction loads and environmental excitations. Static data such as material strength test reports and component machining error records from the construction log are also collected simultaneously. This multi-source, heterogeneous data constitutes a complete relevant dataset, providing foundational information for subsequent analysis.

[0060] As described in step S2 above, the associated data is Fourier transformed to obtain frequency domain signal data to extract the modal parameters of the bridge structure. The time domain dynamic data collected in step S1 is input into a Fourier transform algorithm, which decomposes the time domain signal into a superposition of different frequency components and converts it into frequency domain signal data. Based on this, frequency domain peak recognition technology is used to extract the natural frequency of the bridge structure. Combined with phase difference analysis, the relationship between the vibration amplitude and phase at each measuring point at a specific frequency is calculated, and the modal vectors of the bridge structure, i.e., the modal parameters, are derived. As core indicators reflecting the dynamic characteristics of a bridge, modal parameters can effectively characterize the structural stiffness distribution and modal characteristics.

[0061] As described in step S3 above, similarity theory modeling is performed on the geometric dimensions and material properties in the bridge design drawings, a scaled physical test model is constructed, and the mechanical response of the bridge under static and dynamic loads is simulated. Based on similarity theory, basic similarity criteria such as geometric similarity constants, stress similarity constants, and time similarity constants are selected. The three-dimensional geometric dimensions and material elastic modulus parameters in the design drawings are scaled to produce a physical test model. A hydraulic loading system is used to apply graded static loads to the model. A vibrator is used to simulate dynamic excitations such as vehicle and wind loads. High-precision sensors are simultaneously used to collect stress, strain, and displacement response data of the model under different load conditions, forming a model test dataset containing the structural mechanical properties.

[0062] As described in step S4 above, the field measured data of the bridge is obtained, and the field measured data and the model test data are dimensionlessly processed to calculate the similarity criterion deviation. The field measured data is obtained (which can be the data in step S1 or the real-time data of additional measurements), and the normalization algorithm is used to map the numerical range of physical quantities such as stress, displacement, and vibration frequency to the interval [0,1] to eliminate the dimension difference; the model test data is subjected to the same dimensionless processing. Based on the similarity theory π theorem, the numerical ratios of the field measurements and the model tests under the criteria of geometric similarity, stress similarity, and time similarity are calculated respectively. By comparing the theoretical similarity constant with the actual calculated ratio, the degree of deviation of each similarity criterion is obtained; further, the weights are set according to the bridge design specifications, and the weighted sum is used to obtain the comprehensive similarity criterion deviation, which quantifies the difference between the actual structure and the theoretical model.

[0063] As described in step S5 above, cluster analysis is performed on historical bridge defect case data, and a defect feature classification system based on Euclidean distance is established. The current inspection data of the bridge is matched with it to identify the type of defect. The historical bridge defect database is collected and organized, and characteristic parameters such as crack width, corrosion area, and deformation angle are extracted. The feature similarity between each case is calculated using the Euclidean distance measurement method. The case data is classified using the K-means clustering algorithm to construct a classification system containing typical defect patterns. The defect feature data obtained from the current bridge inspection (such as internal defect data obtained by ultrasonic testing and surface crack data identified by visual recognition) are calculated with the Euclidean distance of each cluster center in the classification system, and matched to the cluster category with the closest distance, thereby determining the type of defect currently existing on the bridge.

[0064] As described in step S6 above, a comprehensive assessment is conducted on the identified damage types, similarity criterion deviations, and modal parameters to determine the bridge safety risk level and establish a hierarchical analysis model. From the three dimensions of structural safety, durability, and applicability, an evaluation system is constructed that includes indicators such as damage severity, similarity criterion deviations, and modal parameter variation. The weights of each indicator are determined by combining expert scoring with historical data statistics. A fuzzy comprehensive evaluation method is used to map the hazard level of the damage type, the numerical range of the similarity criterion deviations, and the modal parameter change rate into fuzzy membership. A weighted synthesis calculation is then performed to determine the bridge's membership values ​​for the four levels of safety, concern, warning, and danger. Based on the maximum membership principle, the level with the highest membership is selected as the final bridge safety risk level, and corresponding treatment recommendations are generated.

[0065] In one embodiment, collecting relevant data during the bridge construction process includes:

[0066] The stress data, displacement data, temperature and humidity data during the bridge construction process are collected based on multi-source sensors; abnormal data points are eliminated based on the median filtering algorithm, and the related data are obtained.

[0067] In this embodiment, during the construction of a bridge, the mechanical properties and environmental conditions of the structure are in dynamic change. In order to accurately assess the safety of the bridge, it is necessary to fully obtain key data reflecting its status. By deploying various types of sensors at key locations on the bridge (such as the top of the pier, the mid-span of the main beam, and the connection between supports), real-time monitoring of different physical quantities can be achieved. Among them, stress sensors can capture the changes in internal forces borne by bridge components under the action of construction loads, deadweight, etc., and reflect whether the structural stress is within a safe range; displacement sensors are used to measure the position movement of various parts of the bridge under stress or environmental influences to determine whether the structure has abnormal deformation; temperature and humidity sensors can monitor changes in ambient temperature and humidity, because temperature changes may cause thermal expansion and contraction of materials, and humidity changes may affect the durability of materials. Both of these will indirectly affect the performance of the bridge structure.

[0068] However, during sensor data collection, factors such as equipment errors and electromagnetic interference may generate abnormal data points that deviate from the true value. If these data are not processed, they will interfere with subsequent analysis and reduce the accuracy of the assessment. Therefore, a median filtering algorithm is used to eliminate abnormal data points. The principle of this algorithm is to sort the data sequence by numerical value and take the value in the middle as the filtered output value of the sequence. When processing bridge monitoring data, the data within a certain time window before and after each time point is sorted, and the median value is used to replace the current data point. This filter eliminates abnormal values ​​that deviate from the normal range, retains valid data that can truly reflect the bridge status, and ultimately obtains correlated data that can be used for subsequent analysis. Through this multi-source acquisition and data purification method, a reliable data foundation is provided for bridge safety assessment.

[0069] In one embodiment, field measured data of a bridge is obtained, the field measured data and model test data are dimensionlessly processed, and the similarity criterion deviation is calculated, including:

[0070] The raw data collected by sensors at key parts of the bridge are segmented and filtered to eliminate invalid data segments with missing data to obtain valid on-site measured data;

[0071] Based on similarity theory, geometric dimension ratio, stress distribution ratio, and time response ratio are selected as basic similarity criteria. The ratios of field measured data and model test data under each similarity criterion are calculated respectively, and each ratio is compared with the theoretical benchmark value to obtain the degree of deviation of each similarity criterion.

[0072] According to the influence of various physical quantities on structural safety in the bridge design specifications, different weights are assigned to the deviation degrees of various similarity criteria, and the deviation degrees are weighted and summarized to obtain the similarity criterion deviation degree.

[0073] In this embodiment, during the construction or operation of the bridge, stress, displacement, and vibration sensors deployed at key locations such as the top of the pier, the middle of the main beam, and the support connection will continuously collect raw data. Due to equipment failure, signal transmission interruption, and other reasons, the raw data may be missing in some time periods. To ensure the reliability of the analysis results, the continuously collected raw data needs to be divided into independent data segments at fixed time intervals (such as every 30 minutes), and each data segment needs to be checked for integrity. Data segments with a data missing ratio exceeding a set threshold (such as 5%) are eliminated, and data segments with complete data that can truly reflect the state of the bridge structure are retained. Finally, effective on-site measured data are integrated. This step eliminates interference information through data screening to provide an accurate data basis for subsequent analysis.

[0074] Similarity theory, a scientific theory that studies the similarities between physical phenomena, can be used to correlate the mechanical properties of actual structures with those from model tests. In this step, three dimensions that significantly influence bridge structural performance are first selected as basic similarity criteria based on similarity theory: geometric dimension ratio reflects the degree of similarity in shape and size between the actual bridge and the model; stress distribution ratio reflects the consistency of the internal force transmission patterns under load; and time response ratio measures the similarity of the structure's response speed and period under dynamic loads (such as vehicle traffic and wind). For each similarity criterion, the ratio of the field-measured data to the model test data is calculated (for example, the stress value of a measured bridge component is divided by the stress value of the corresponding model component to obtain the calculated stress distribution ratio). This ratio is then compared with a theoretical benchmark value pre-determined based on design requirements. The absolute value of the difference between the two represents the degree of deviation from the similarity criterion; a larger value indicates a more significant difference between the actual bridge and the model in that aspect.

[0075] Different similarity criteria have varying degrees of influence on the overall safety performance of bridges. To more accurately reflect the overall differences between actual bridges and models, the importance of each similarity criterion must be assessed based on bridge design specifications and engineering experience. For example, stress distribution is directly related to the bearing capacity of bridge structures, and its impact on safety is greater than minor deviations in geometric dimensions. Therefore, a higher weight (e.g., 0.5) can be assigned to the stress distribution ratio; while the geometric dimension ratio and time response ratio can be assigned weights of 0.3 and 0.2, respectively. The degree of deviation for each similarity criterion calculated in step 2 is multiplied by the corresponding weight coefficient and summed. The resulting value is the similarity criterion deviation. This metric, through a weighted aggregation approach, comprehensively quantifies the overall differences between actual bridges and model tests across multiple key performance dimensions, providing a quantitative basis for bridge safety risk assessment.

[0076] In one embodiment, the raw data is collected by acceleration sensors, displacement sensors, and strain sensors installed at the middle of the span of the bridge main beam, the top of the pier, and the connection of the support.

[0077] In this embodiment, acceleration sensors, displacement sensors and strain sensors are installed at the middle of the span of the bridge main beam, the top of the pier, and the support connection to collect corresponding raw data.

[0078] In one embodiment, a comprehensive evaluation is performed on the identified damage type, similarity criterion deviation, and mode shape parameters to determine the bridge safety risk level, including:

[0079] The identified damage type, similarity criterion deviation, and mode parameters are used as indicators and network nodes, respectively. Based on the bridge structure mechanical conduction path and historical accident data, the risk propagation correlation between network nodes is established.

[0080] Using the tolerance band analysis method, a safety threshold interval is set for each indicator. When any indicator exceeds the safety threshold interval, the coupling effect calculation is triggered. The coupling effect calculation includes: calculating the impact coefficient of the abnormal indicator on other indicators based on the risk propagation correlation relationship;

[0081] A three-dimensional assessment coordinate system was established, with the severity of the disease type as the X-axis, the deviation of the similarity criterion as the Y-axis, and the change in the vibration parameter as the Z-axis, and 27 assessment areas were divided; each area corresponds to a different risk level standard, and the bridge safety risk level is determined through spatial coordinate positioning.

[0082] In this embodiment, within the bridge safety assessment system, damage type, similarity criterion deviation, and modal parameters are core indicators reflecting the structural health status. To further explore the interplay between these indicators, these three indicators are first abstracted into network nodes. Damage types encompass various defect types, such as cracks, corrosion, and deformation. Similarity criterion deviation characterizes the degree of discrepancy between the actual structure and the theoretical model, and modal parameters reflect the structural dynamics. Subsequently, based on the mechanical transmission pathways of bridge structures—the physical process by which loads are transferred from the bridge deck to the piers and foundation—and incorporating the patterns of chain reactions triggered by abnormalities in various indicators from historical accident data, the connections and weights between the nodes are determined. For example, if historical data indicates that pier cracks can significantly alter the structural modal parameters, a higher weight is assigned to the connection between the "Damage Type" node and the "Modal Parameter" node in the network. This creates a network model that reflects the patterns of risk transmission.

[0083] To promptly identify abnormal bridge structural conditions, it is necessary to set reasonable safety thresholds for each assessment indicator. Using a tolerance band analysis method, based on bridge design specifications, material performance standards, and historical monitoring data, safety threshold intervals are defined for each damage type, similarity criterion deviation, and modal parameter. For example, the safety interval for similarity criterion deviation is set between 0 and 10%. When the actual monitored value exceeds this range, the indicator is deemed abnormal and the coupling effect calculation mechanism is triggered. This coupling effect calculation relies on the risk propagation relationships established in the previous step, quantitatively determining the impact of the abnormal indicator on other indicators through analysis. Specifically, based on the connection weights between network nodes, the impact coefficient of the abnormal indicator on other indicators through the risk propagation path is calculated. For example, if the "Disease Type" node is abnormal, the degree of change in the modal parameters that the damage may cause is calculated based on its connection weight with the "Modal Parameter" node, thereby assessing the risk diffusion effect across different indicators.

[0084] To achieve intuitive and quantitative assessment of bridge safety risks, a three-dimensional assessment coordinate system was constructed based on the dimensions of defect severity, similarity criterion deviation, and modal parameter variation. Defect severity was categorized into low, medium, and high levels based on factors such as defect size and development speed. Similarity criterion deviation and modal parameter variation were similarly categorized into three levels based on their numerical ranges. By combining these three dimensional levels, a total of 27 assessment regions, arranged in a 3×3×3 configuration, were formed in the coordinate system. Each region was assigned a predefined risk level. For example, if the defect type was "high severity," the similarity criterion deviation was "high deviation," and the modal parameter variation was "high variation," the region would be assigned a "high risk level." By mapping the values ​​of the three indicators obtained from actual monitoring into the three-dimensional coordinate system and determining their spatial coordinates, the current safety risk level of the bridge could be quickly and accurately determined based on the risk level of the region.

[0085] In one embodiment, a comprehensive evaluation is performed on the identified damage type, similarity criterion deviation, and mode shape parameters to determine the bridge safety risk level, including:

[0086] The identified damage types, similarity criterion deviations, and mode parameters are tensor-encoded, and a feature correlation tensor is constructed in combination with the bridge structure topology information to obtain a risk feature tensor.

[0087] Mapping the risk feature tensor to a low-dimensional non-Euclidean space through manifold learning to obtain a risk manifold with geometric features;

[0088] A butterfly effect simulation is performed on the risk manifold, and the initial disturbance threshold of each indicator is set. A Lagrangian mechanics model is used to simulate the propagation trajectory of data changes on the risk manifold. The risk amplification effect is quantified by calculating the Lyapunov exponent to obtain the risk sensitivity assessment result.

[0089] Performing topological clustering on the risk sensitivity assessment results, analyzing the topological structure of the risk manifold, identifying holes and connected components and performing clustering to obtain risk evolution pattern classification results;

[0090] The risk evolution pattern classification results are subjected to probabilistic reasoning based on the Bayesian decision tree, and the decision weight is dynamically adjusted in combination with the real-time working conditions of the bridge to ultimately determine the safety risk level of the bridge.

[0091] In this embodiment, in the bridge safety assessment, the type of disease, similarity criterion deviation and mode parameters are used as core evaluation indicators, and their data has the characteristics of multi-dimensionality and heterogeneity. In order to achieve efficient processing and feature extraction of these complex data, tensor encoding technology is used to convert the disease type (discrete categories such as cracks and rust) into a vector form through one-hot encoding, and the numerical data such as similarity criterion deviation and mode parameters are kept in the original format, and the three are integrated into a multidimensional array. At the same time, the bridge structure topology information is introduced, that is, the connection relationship and mechanical conduction path of each bridge component, which is converted into an adjacency matrix through graph theory methods and embedded in the tensor structure. In this way, a feature association tensor that can simultaneously express data features and structural associations is constructed, and finally a risk feature tensor containing bridge risk information is formed, providing a structured data foundation for subsequent analysis.

[0092] Because the high-dimensional nature of the risk signature tensor leads to data processing complexity and the curse of dimensionality, a manifold learning algorithm is used to reduce its dimensionality. Based on the assumption that data has an inherent geometric structure in low-dimensional space, manifold learning calculates the local similarities and global topological relationships between data points in the tensor. While maintaining the essential characteristics of the data, it maps the risk signature tensor from high-dimensional Euclidean space to low-dimensional non-Euclidean space. In non-Euclidean space, the distribution of data points reflects the geometric characteristics of the bridge's risk status, forming a risk manifold with complex geometric structures such as bends and folds. This mapping not only reduces the data dimension but also reveals potential nonlinear relationships between data, facilitating subsequent analysis of risk evolution patterns.

[0093] Considering that even small initial changes in bridge structures can trigger chain reactions, a butterfly effect simulation method was used to explore the dynamic evolution of risk. First, initial disturbance thresholds were set for the type of damage, similarity criterion deviation, and mode parameters. These thresholds, determined based on bridge design specifications and historical monitoring data, were used to determine whether data changes were sufficient to trigger risk evolution. Then, using a Lagrangian mechanics model, data points on the risk manifold were treated as energy-carrying particles. The propagation trajectory of data changes on the manifold was simulated by solving the Lagrangian equations, analyzing how risk diffuses over time and structural relationships. During this process, the Lyapunov exponent was calculated to quantify the system's sensitivity to initial conditions, i.e., the risk amplification effect. A larger Lyapunov exponent indicates that small changes in the data can lead to significant differences in the risk evolution outcomes. This results in a risk sensitivity assessment that reflects the sensitivity of bridge risk to initial disturbances.

[0094] To further reveal the inherent laws of bridge risk evolution, a topological cluster analysis was performed on the risk sensitivity assessment results. Based on the theory of topological data analysis (TDA), topological features such as holes and connected components in the risk manifold are identified by calculating the topological invariants of the risk manifold, such as the Betti number. These topological features correspond to different risk evolution patterns. For example, holes in the manifold represent unstable states during the risk evolution process, while connected components represent data sets with similar risk evolution trends. By clustering these topological features, the risk manifold is divided into different categories, each corresponding to a specific risk evolution pattern, thereby obtaining a risk evolution pattern classification result that can reflect the laws of bridge risk evolution.

[0095] To accurately determine the safety risk level of bridges, a Bayesian decision tree algorithm is used to perform probabilistic reasoning on the classification results of risk evolution patterns. The Bayesian decision tree combines prior knowledge and real-time monitoring data to evaluate the likelihood of a bridge being at each risk level by calculating the posterior probability of different risk levels. At the same time, taking into account the differences in working conditions of bridges at different construction stages or operating environments, real-time working condition data of bridges (such as traffic loads, ambient temperature and humidity, etc.) is introduced to dynamically adjust the weights of each node in the decision tree according to changes in working conditions. In this way, the risk evolution model, real-time working condition information and probabilistic reasoning are combined to ultimately determine the risk level that can reflect the current actual safety status of the bridge, providing a scientific decision-making basis for bridge maintenance and management.

[0096] In one embodiment, a comprehensive evaluation is performed on the identified damage type, similarity criterion deviation, and mode shape parameters to determine the bridge safety risk level, including:

[0097] The identified disease types, similarity criterion deviations, and mode shape parameters are subjected to biological lattice encoding processing. The disease types are mapped to lattice defect types, the similarity criterion deviations are quantified as the degree of lattice distortion, and the mode shape parameters are converted into lattice vibration frequencies. A three-dimensional data lattice with biomimetic mechanical properties is constructed to obtain a risk characteristic lattice model.

[0098] Performing multi-body dynamic modeling on the risk characteristic lattice model, treating the nodes of the lattice as rigid bodies, establishing elastic connection relationships between the nodes, simulating the dynamic response of the bridge structure under load, and solving to obtain a dynamically evolving multi-body risk model;

[0099] The multi-body risk model is subjected to network deconstruction, where the nodes of the lattice are abstracted into network nodes. The connection strength between nodes corresponds to the degree of dynamic coupling. The network topology characteristics are analyzed using betweenness centrality and clustering coefficient in graph theory to identify the key nodes and core paths of risk propagation, thus obtaining a network topology map of risk propagation.

[0100] Perform cascading failure simulation on the network topology of the risk propagation, set a node failure threshold, and trigger a cascading failure when the node risk value exceeds the failure threshold; simulate the avalanche-like spread of risk in the network, simulate and calculate the failure probability distribution under different working conditions, and obtain the probability assessment result of risk spread;

[0101] A multi-party game model involving bridge management, maintenance teams, and transportation departments is constructed. The probability assessment results are used as profit function parameters, and game decisions are made on the probability assessment results to determine the bridge safety risk level.

[0102] In this embodiment, in the bridge safety assessment, the type of disease, the deviation of similarity criteria and the mode parameters are the core indicators reflecting the health status of the structure. In order to more intuitively show the relationship between these indicators and their impact on the structure, a biological lattice encoding processing method is adopted. In nature, biological lattice structures (such as honeycombs and bones) have excellent mechanical stability. Drawing on this characteristic, the types of diseases (such as cracks, rust, and deformation) are mapped to the types of defects in the lattice structure. For example, cracks can be compared to fracture defects in the lattice; the deviation of the similarity criteria is quantified into the degree of distortion of the lattice unit through numerical conversion. The greater the deviation, the more obvious the lattice distortion; the mode parameters are converted into lattice vibration frequency to reflect the dynamic response characteristics of the structure. Through this mapping relationship, the originally abstract multi-source data is constructed into a three-dimensional data lattice. Each lattice unit carries the corresponding indicator information, and finally forms a risk characteristic lattice model with bionic mechanical characteristics. This model can display the risk status of the bridge and the interaction between various indicators in an intuitive geometric structure.

[0103] After obtaining the risk characteristic lattice model, the model was further refined based on multibody dynamics theory to further explore the dynamic behavior of bridge structures under actual loads. Each node in the lattice model was abstracted as a rigid body to simulate the mechanical properties of bridge components. Based on the actual connection methods and mechanical transmission paths of the various parts of the bridge structure, elastic connection relationships were established between the nodes to characterize the interactions between the components. By introducing external excitation conditions such as construction loads, traffic loads, and wind loads, and applying mechanical principles such as Newton's laws of motion and Hooke's law, dynamic equations describing the motion state of the lattice nodes were established. This set of equations was solved using numerical calculation methods to simulate the changes in displacement, velocity, and acceleration of the lattice nodes under load, thereby obtaining a multibody risk model that can reflect the dynamic response process of the bridge structure. This model dynamically presents the evolution of risk with time and load changes.

[0104] To more clearly analyze the propagation patterns of risks in bridge structures, the multi-body risk model was converted into a network structure for analysis. The nodes in the lattice model were abstracted into network nodes, with each node representing a key part of the bridge structure. The connection strength between nodes is determined by the degree of dynamic coupling between nodes; the higher the degree of coupling, the greater the connection strength. Using the betweenness centrality metric from graph theory, the importance of each node in risk transmission in the entire network was calculated. Nodes with higher betweenness centrality are more critical in the risk propagation process. The clustering coefficient was used to analyze the degree of clustering between nodes and identify closely related groups of nodes. Combining the analysis results of betweenness centrality and clustering coefficient, key nodes (such as vulnerable components) and core paths (the main risk transmission channels) in the risk propagation process were precisely located. Ultimately, a complete risk propagation network topology map was drawn, which intuitively displays the paths and key links of risk propagation in bridge structures.

[0105] Based on the network topology of risk propagation, cascading failure simulations were conducted to predict the failure modes of bridge structures under extreme conditions. A failure threshold was pre-set for each network node, determined based on the design load-bearing capacity and material properties of the bridge components. When the risk value (e.g., stress, deformation) carried by a node exceeded the failure threshold, the node was deemed failed, triggering a chain reaction among its connected nodes. By simulating the risk diffusion process in the network and employing methods such as Monte Carlo simulation, the risk propagation paths and node failures under different operating conditions (e.g., overload, earthquake, and strong wind) were repeatedly simulated. The number of failures of each node in a large number of simulations was counted, and the failure probability of each node under different operating conditions was calculated. These failure probabilities were summarized and analyzed to form a complete risk diffusion probability assessment result, which quantified the possibility of bridge structure failure under different operating conditions and the range of risk diffusion.

[0106] Considering that bridge safety management involves multiple stakeholders, a multi-party game model was constructed involving bridge management, maintenance teams, and transportation departments. Bridge management focuses on management costs and social impact, maintenance teams prioritize maintenance expenses and workload, and transportation departments prioritize traffic efficiency and safety. These stakeholders engage in a competitive game of interests regarding bridge risk response strategies. The probabilistic assessment of risk diffusion serves as the payoff function parameter of the game model. For example, when the probability of bridge failure corresponding to a certain risk level is high, the benefits (e.g., avoiding major accident losses) and costs (capital investment) of emergency reinforcement measures implemented by management, the workload of maintenance teams, and the costs of traffic restrictions imposed by transportation departments all change accordingly. By solving the Nash equilibrium of the game model, the optimal decision-making strategies of each party under different risk scenarios are determined. Based on a comprehensive analysis of these decision-making strategies, the bridge's safety risk level is ultimately determined.

[0107] In one embodiment, after determining the bridge safety risk level, the following steps are included:

[0108] Based on the bridge safety risk level, a corresponding undirected graph template is matched; wherein different risk levels correspond to different row and column structure attributes of the undirected graph;

[0109] Obtain bridge design information and convert it into a character sequence; sequentially add the characters in the character sequence to each node of the undirected graph template to form an undirected graph of design information;

[0110] Mapping the disease type, similarity criterion deviation, and mode shape parameter into a node position offset and an edge length variation coefficient of a design information undirected graph; deforming the design information undirected graph based on the node position offset and the edge length variation coefficient to obtain a deformed undirected graph;

[0111] Based on the deformed undirected graph, a coding table is generated; based on the coding table, the associated data is encoded and then sent to an associated management terminal.

[0112] In this embodiment, after the bridge safety assessment is completed and the risk level is determined, in order to achieve visualization and structured expression of risk information, the abstract risk level needs to be mapped into a concrete undirected graph template. In the pre-built template library, low-risk levels correspond to undirected graphs with fewer rows and columns and sparse node connections. Such templates, through their concise structure, imply a simple risk transmission path and a limited impact range. High-risk levels are matched with templates with more rows and columns and dense node connections. Their complex topological structure intuitively reflects the intertwined risk transmission paths and wide-ranging impact. Based on the bridge safety risk level currently assessed, the corresponding undirected graph template is retrieved from the template library. This template serves as the basic framework for subsequent data carrying and processing, providing a visual carrier for the integration and analysis of risk information.

[0113] Bridge design information includes key parameters such as geometric dimensions, material properties, and structural forms, and is an important basis for evaluating bridge performance. By reading design drawings and technical documents, the above information is extracted and converted into a character sequence according to specific encoding rules (such as ASCII or Unicode), so that the design information has a digital expression. Subsequently, each character in the character sequence is assigned to a node of the undirected graph template in turn as an attribute label of the node. For example, the bridge span dimension information is assigned to the node representing the key part of the main beam, and the material strength parameters are assigned to the corresponding component nodes. In this way, each node of the undirected graph carries specific design information, and the originally discrete design parameters are integrated into structured graphic data to form an undirected graph containing complete design information.

[0114] Defect type, similarity criterion deviation, and modal parameters are core indicators of a bridge's current health. To visualize the impact of these indicators on the bridge structure, a mapping relationship between these indicators and the geometric properties of an undirected graph is established: the severity of the defect type is quantified and mapped to the spatial offset of the corresponding node; the more severe the defect, the greater the node offset; the similarity criterion deviation is converted into the length variation coefficient of the edge connecting the nodes; the higher the deviation, the greater the adjustment of the edge length; the degree of change in the modal parameters determines the elastic coefficient of the edge, affecting its deformation tendency under load. Based on this mapping relationship, a geometric deformation operation is performed on the undirected graph containing design information, allowing the graph structure to dynamically adjust according to real-time monitoring data. For example, if a severe crack is detected in a certain area, the corresponding node in the graph will be significantly offset, and the edges connecting the node will also change in length and shape accordingly, ultimately generating a deformed undirected graph.

[0115] Deformable undirected graphs integrate bridge design information and real-time monitoring data through topological and geometric changes. To achieve efficient data transmission and storage, the graph information must be converted into a computer-readable encoding format. First, node numbering rules, edge connectivity representation methods, and geometric attribute quantification standards are defined. Mapping rules from undirected graph elements to coded characters or numbers are established, generating a dedicated encoding table. For example, node position coordinates are converted into numerical codes in a specific format, and the edge connectivity sequence is represented as a binary sequence. Then, based on the encoding table, associated data from the bridge construction process (such as stress monitoring data and construction progress information) is encoded and converted into strings or numerical sequences that conform to the encoding rules. Finally, the encoded data is transmitted to associated management terminals, such as bridge management departments and construction units, via a secure communication protocol. The receiving terminal can decode the data using the same encoding table to restore bridge safety risk information.

[0116] In one embodiment, after determining the bridge safety risk level, the following steps are included:

[0117] Based on the bridge safety risk level, a corresponding undirected graph template is matched; wherein different risk levels correspond to different row and column structure attributes of the undirected graph;

[0118] Obtain bridge design information and convert it into a fixed-length character sequence using a hash coding algorithm; embed the characters in the character sequence into nodes with even degrees in an undirected graph template to form a character undirected graph;

[0119] The mode parameters are mapped to the gravitational coefficients between nodes, the similarity criterion deviation is mapped to the elastic modulus of the edge, and the disease type is mapped to the topological dimension change parameter of the node. Based on the finite element mesh deformation algorithm, the gravitational coefficients between nodes, the elastic modulus, and the topological dimension change parameter are used to perform overall topological optimization on the character undirected graph to generate a deformable undirected graph.

[0120] Through the eigenvalue decomposition of the adjacency matrix in graph theory, the deformed undirected graph information is converted into a eigenvector; combined with the preset coding mapping table, the eigenvector is converted into a hexadecimal coding unit containing a 128-bit check code, and a unique coding table is formed to perform multi-dimensional encoding on the collected related data and send it to the associated management terminal.

[0121] In this embodiment, after completing the bridge safety risk level assessment, in order to achieve a structured expression of risk information, the abstract risk level needs to be mapped into a concrete undirected graph template. In the pre-built template library, different risk levels form a specific mapping relationship with the row and column structure attributes of the undirected graph.

[0122] Bridge design information covers core parameters such as geometric dimensions, material properties, and structural forms, and is an important basis for evaluating bridge performance. After obtaining the above information by reading design drawings and technical documents, a hash coding algorithm (such as SHA-256) is used to process it, converting the original design information into a fixed-length character sequence. This sequence is collision-resistant and irreversible, ensuring information integrity and security. Subsequently, the nodes with even degrees (i.e., the number of edges connecting the nodes) in the undirected graph template are screened, and the characters in the character sequence are embedded in the attribute fields of these nodes in sequence. For example, the hash code characters of the bridge span size are assigned to the even-degree nodes representing the key parts of the main beam, and the hash values ​​of the material strength parameters are embedded in the corresponding component nodes. In this way, specific nodes of the undirected graph carry bridge design information, forming a character undirected graph that integrates design data.

[0123] Modal parameters, similarity criterion deviation, and damage type are key indicators of a bridge's real-time health. To intuitively quantify their impact on the bridge structure, a mapping relationship between these indicators and the geometric and physical properties of an undirected graph is established: Modal parameters are converted into internode gravitational coefficients. Changes in these parameters adjust the gravitational coefficients accordingly, simulating the effect of structural vibration on the relative positions of nodes. Similarity criterion deviations are mapped to edge elastic moduli; higher deviations indicate more significant changes in the edge's ability to resist deformation. Damage types correspond to topological dimension variation parameters of nodes; different damage types trigger node splitting or merging operations in high-dimensional space. Based on a finite element mesh deformation algorithm, these parameters are input as control variables. By solving mechanical equilibrium equations and geometric constraints, the overall topology of the undirected graph is optimized. For example, when a severe crack is detected in a certain area, the topological dimension of the corresponding node increases and undergoes a morphological change. The edges connecting the nodes change their elastic moduli due to the increased similarity criterion deviation. Ultimately, a deformable undirected graph is generated that dynamically reflects the actual condition of the bridge.

[0124] Deformable undirected graphs integrate bridge design information and real-time monitoring data through changes in topology and geometric properties. To achieve efficient data transmission and storage, the graph information must be converted into a computer-readable encoding format. First, based on graph theory principles, the deformable undirected graph is converted into an adjacency matrix. Eigenvalue decomposition is then used to extract the eigenvectors of this matrix, which condense core information such as the graph's topology, node connectivity, and geometric properties. Then, using a pre-defined encoding mapping table, each component of the eigenvector is converted into a hexadecimal encoding unit. A 128-bit checksum is embedded in the encoding unit to facilitate error detection and correction during data transmission. Finally, all encoding units are combined according to specific rules to generate a unique encoding table. Based on this encoding table, multi-dimensional encoding is performed on relevant data from the bridge construction process (such as stress monitoring data and construction progress information). This data is converted into strings or numeric sequences that conform to the encoding rules and transmitted via a secure communication protocol to relevant management terminals, such as bridge management departments and construction units. The receiving terminal decodes the data using the same encoding table to recover bridge safety risk information, providing accurate data support for management decision-making and maintenance planning.

[0125] In one embodiment, after determining the bridge safety risk level, the following steps are included:

[0126] Obtain bridge monitoring equipment numbers, construction stage identification information, construction log data, and monitoring sensor layout data, extract the complete construction monitoring path associated with the risk, and arrange the data collection node numbers in the complete construction monitoring path in chronological order to form an ordered number sequence;

[0127] Mapping each number in the ordered number sequence to four dimensions of a four-dimensional space according to its corresponding acquisition timestamp, sensor installation location coordinates, monitoring data type code, and construction process stage code, determining multiple coordinate points in the four-dimensional space, and connecting each coordinate point to construct a four-dimensional hypercube;

[0128] Calculating the length difference of each edge of the four-dimensional hypercube, combining the weight coefficients of different monitoring indicators in the bridge design specification, and using the weighted sum of the length difference and the weight coefficient as a parameter to perform a topological distortion transformation on the four-dimensional hypercube to generate a distorted hypercube;

[0129] The geometric elements of the distorted hypercube are converted into nodes, edges, and subtree structures of a ternary tree according to a preset mapping rule to form an initial ternary tree; according to the severity coefficient corresponding to the risk warning, the initial ternary tree is subjected to node splitting or merging operations to obtain a variant ternary tree;

[0130] The variant ternary tree is subjected to character screening based on a preset graph, and the characters are combined into a character string as an encryption password to encrypt the associated data and send it to the management terminal.

[0131] In this embodiment, after determining the bridge safety risk level, in order to trace the entire process of risk generation, it is necessary to integrate multi-source data to build a complete construction monitoring path. Through the bridge information management system, the unique number of the monitoring equipment is obtained to locate the source of data collection, and the construction stage identification information (such as pile foundation construction, main beam erection, etc.), the process records in the construction log, and the installation location data in the sensor layout drawings are extracted. Based on the temporal and spatial correlation of the risk occurrence, the data collection nodes related to the current risk are screened out, and the numbers of these nodes are arranged in ascending order according to the data collection timestamp to form an ordered number sequence that can reflect the risk evolution process. This sequence not only connects the key monitoring nodes in the construction process in series, but also provides a logical main line of the time dimension for subsequent data structured processing.

[0132] In order to achieve visualization and structured expression of construction monitoring data, each node number in the ordered number sequence is converted into a coordinate point in four-dimensional space through multi-dimensional mapping. Among them, the acquisition timestamp is mapped to the time dimension coordinate to accurately identify the moment of data acquisition; the latitude, longitude and elevation information of the sensor installation location correspond to the three-dimensional spatial coordinates to determine the physical location of data acquisition; the monitoring data type (such as stress, displacement, temperature and humidity) and the construction process stage (such as concrete pouring, prestressing tensioning) are respectively encoded as discrete numerical values ​​as the coordinate values ​​of the other two dimensions. In four-dimensional space, each number corresponds to a unique coordinate point, and these coordinate points are connected in sequence to form a four-dimensional hypercube containing spatiotemporal and business attribute information. The above four-dimensional hypercube intuitively displays the spatiotemporal distribution and business association of monitoring data during the construction process in a geometric form.

[0133] In order to highlight the impact of key risk factors on bridge structures, it is necessary to perform topological optimization on the four-dimensional hypercube. First, the length difference values ​​of each edge of the hypercube in different dimensions are calculated. This difference value reflects the degree of change in the temporal, spatial or business attributes of adjacent monitoring nodes. For example, the edge length in the time dimension represents the monitoring time interval. A large length difference indicates a significant change in the data collection frequency. Then, based on the weight coefficients set for different monitoring indicators in the bridge design specifications (such as stress monitoring is more important than temperature and humidity monitoring), the length difference value and the weight coefficient are weighted and summed to obtain the comprehensive impact parameter of each edge. Based on this parameter, the four-dimensional hypercube is topologically distorted, so that the areas most affected by the risk are significantly deformed in the hypercube, generating a distorted hypercube that can intuitively reflect the risk propagation path and the degree of impact.

[0134] To further simplify the data structure and enhance its algorithmic processing efficiency, the geometric elements (vertices, edges, and faces) of the twisted hypercube are converted into a ternary tree structure using preset mapping rules. Specifically, the vertices of the hypercube correspond to nodes in the ternary tree, the edges are converted into connecting edges in the tree, and the topological relationships of the faces are constructed as subtrees. This conversion forms an initial ternary tree that can express the hierarchical relationships of the data and the associations between risks. Subsequently, the initial ternary tree is dynamically adjusted based on the severity coefficient of the risk warning (e.g., a low risk coefficient of 0.1 and a high risk coefficient of 0.9): When the coefficient is above a threshold, key nodes are split to refine the risk analysis hierarchy; when the coefficient is low, secondary nodes are merged to simplify the structure. Ultimately, a variant ternary tree that can adaptively reflect the severity of the risk is obtained.

[0135] To generate a password for data encryption, first construct a preset graph containing a specific character set (such as a matrix consisting of letters, numbers, and special symbols). This graph corresponds to the node attributes of the variant ternary tree. Traverse the nodes of the variant ternary tree and filter the corresponding characters from the preset graph based on the node's hierarchy, position, and attribute information. For example, the root node of the tree corresponds to the character in the upper left corner of the graph, and the leaf node corresponds to the character on the edge of the graph. The filtered characters are combined in sequence according to the tree traversal order (such as pre-order traversal) to form a unique encrypted string. This string is used as an encryption password, and the collected related data (such as raw monitoring data and risk assessment reports) are encrypted using a symmetric encryption or asymmetric encryption algorithm and sent to the bridge management terminal through a secure communication protocol. The receiving terminal uses the same password and decryption algorithm to restore the data to ensure the confidentiality and integrity of the information during transmission.

[0136] Reference Figure 2 In one embodiment of the present invention, a safety assessment system for intelligently constructed bridges is provided, comprising:

[0137] The acquisition module is used to collect relevant data during the bridge construction process;

[0138] An extraction module, configured to perform Fourier transform on the correlation data to obtain frequency domain signal data to extract mode shape parameters of the bridge structure;

[0139] The simulation module is used to perform similarity theoretical modeling on the geometric dimensions and material properties in the bridge design drawings, construct a scaled physical test model, and simulate the mechanical response of the bridge under static and dynamic loads to obtain model test data;

[0140] The calculation module is used to obtain the field measured data of the bridge, perform dimensionless processing on the field measured data and the model test data, and calculate the deviation of the similarity criterion;

[0141] The recognition module is used to perform cluster analysis on historical bridge damage case data, establish a damage feature classification system based on Euclidean distance, match the current bridge inspection data with it, and identify the damage type;

[0142] The evaluation module is used to comprehensively evaluate the identified damage types, similarity criterion deviations and vibration mode parameters to determine the bridge safety risk level.

[0143] In this embodiment, for the specific implementation of each module in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0144] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, memory, display screen, input device, network interface and database connected via a system bus. The processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the above method is implemented.

[0145] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0146] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-described method when executed by a processor. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0147] In summary, the safety assessment method, system, and device for intelligently constructed bridges provided in embodiments of the present invention include: collecting relevant data during the bridge construction process; performing Fourier transform on the relevant data to obtain frequency domain signal data to extract the modal parameters of the bridge structure; performing similarity theory modeling on the geometric dimensions and material properties in the bridge design drawings, constructing a scaled physical test model, and simulating the mechanical response of the bridge under static and dynamic loads to obtain model test data; obtaining field measured data of the bridge, dimensionlessly processing the field measured data and the model test data, and calculating the similarity criterion deviation; performing cluster analysis on historical bridge defect case data, establishing a defect feature classification system based on Euclidean distance, matching the current bridge inspection data with the system, and identifying the defect type; and comprehensively evaluating the identified defect type, similarity criterion deviation, and modal parameters to determine the bridge safety risk level. In the present invention, by identifying the defect type, similarity criterion deviation, and modal parameters during the bridge construction process and then performing a comprehensive evaluation, the final result takes into account the influence of dimensionality, making the evaluation result more accurate.

[0148] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described method embodiments. Among them, any reference to memory, storage, database, or other media provided by the present invention and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM.

[0149] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0150] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A safety assessment method for intelligently constructed bridges, characterized in that: The following steps are involved: Collect relevant data during the bridge construction process; Performing Fourier transform on the correlation data to obtain frequency domain signal data to extract mode shape parameters of the bridge structure; Conduct similarity theory modeling on the geometric dimensions and material properties in the bridge design drawings, construct a scaled physical test model, and simulate the mechanical response of the bridge under static and dynamic loads to obtain model test data; Obtain the field measured data of the bridge, perform dimensionless processing on the field measured data and the model test data, and calculate the deviation of the similarity criterion; Cluster analysis was performed on historical bridge damage case data to establish a damage feature classification system based on Euclidean distance. The current bridge inspection data was matched with the system to identify the damage type. A comprehensive assessment is conducted on the identified damage types, similarity criterion deviations, and vibration mode parameters to determine the bridge safety risk level.

2. The safety assessment method for intelligently constructed bridges according to claim 1, characterized in that: Collect relevant data during the bridge construction process, including: The stress data, displacement data, temperature and humidity data during the bridge construction process are collected based on multi-source sensors; abnormal data points are eliminated based on the median filtering algorithm, and the related data are obtained.

3. The safety assessment method for intelligently constructed bridges according to claim 1, characterized in that: Obtain the field measured data of the bridge, perform dimensionless processing on the field measured data and the model test data, and calculate the deviation of the similarity criterion, including: The raw data collected by sensors at key parts of the bridge are segmented and filtered to eliminate invalid data segments with missing data to obtain valid on-site measured data; Based on similarity theory, geometric dimension ratio, stress distribution ratio, and time response ratio are selected as basic similarity criteria. The ratios of field measured data and model test data under each similarity criterion are calculated respectively, and each ratio is compared with the theoretical benchmark value to obtain the degree of deviation of each similarity criterion. According to the influence of various physical quantities on structural safety in the bridge design specifications, different weights are assigned to the deviation degrees of various similarity criteria, and the deviation degrees are weighted and summarized to obtain the similarity criterion deviation degree.

4. The safety assessment method for intelligently constructed bridges according to claim 3, characterized in that: The raw data are collected by acceleration sensors, displacement sensors and strain sensors installed in the middle of the span of the bridge main beam, the top of the pier and the connection of the support.

5. The safety assessment method for intelligently constructed bridges according to claim 1, characterized in that: Comprehensively evaluate the identified damage types, similarity criterion deviations, and mode parameters to determine the bridge safety risk level, including: The identified damage type, similarity criterion deviation, and mode parameters are used as indicators and network nodes, respectively. Based on the bridge structure mechanical conduction path and historical accident data, the risk propagation correlation between network nodes is established. Using the tolerance band analysis method, a safety threshold interval is set for each indicator. When any indicator exceeds the safety threshold interval, the coupling effect calculation is triggered. The coupling effect calculation includes: calculating the impact coefficient of the abnormal indicator on other indicators based on the risk propagation correlation relationship; A three-dimensional assessment coordinate system was established, with the severity of the disease type as the X-axis, the deviation of the similarity criterion as the Y-axis, and the change in the vibration parameter as the Z-axis, and 27 assessment areas were divided; each area corresponds to a different risk level standard, and the bridge safety risk level is determined through spatial coordinate positioning.

6. The safety assessment method for intelligently constructed bridges according to claim 1, characterized in that: Comprehensively evaluate the identified damage types, similarity criterion deviations, and mode parameters to determine the bridge safety risk level, including: The identified damage types, similarity criterion deviations, and mode parameters are tensor-encoded, and a feature correlation tensor is constructed in combination with the bridge structure topology information to obtain a risk feature tensor. Mapping the risk feature tensor to a low-dimensional non-Euclidean space through manifold learning to obtain a risk manifold with geometric features; A butterfly effect simulation is performed on the risk manifold, and the initial disturbance threshold of each indicator is set. A Lagrangian mechanics model is used to simulate the propagation trajectory of data changes on the risk manifold. The risk amplification effect is quantified by calculating the Lyapunov exponent to obtain the risk sensitivity assessment result. Performing topological clustering on the risk sensitivity assessment results, analyzing the topological structure of the risk manifold, identifying holes and connected components and performing clustering to obtain risk evolution pattern classification results; The risk evolution pattern classification results are subjected to probabilistic reasoning based on the Bayesian decision tree, and the decision weight is dynamically adjusted in combination with the real-time working conditions of the bridge to ultimately determine the safety risk level of the bridge.

7. The safety assessment method for intelligently constructed bridges according to claim 1, characterized in that: Comprehensively evaluate the identified damage types, similarity criterion deviations, and mode parameters to determine the bridge safety risk level, including: The identified disease types, similarity criterion deviations, and mode shape parameters are subjected to biological lattice encoding processing. The disease types are mapped to lattice defect types, the similarity criterion deviations are quantified as the degree of lattice distortion, and the mode shape parameters are converted into lattice vibration frequencies. A three-dimensional data lattice with biomimetic mechanical properties is constructed to obtain a risk characteristic lattice model. Performing multi-body dynamic modeling on the risk characteristic lattice model, treating the nodes of the lattice as rigid bodies, establishing elastic connection relationships between the nodes, simulating the dynamic response of the bridge structure under load, and solving to obtain a dynamically evolving multi-body risk model; The multi-body risk model is subjected to network deconstruction, where the nodes of the lattice are abstracted into network nodes. The connection strength between nodes corresponds to the degree of dynamic coupling. The network topology characteristics are analyzed using betweenness centrality and clustering coefficient in graph theory to identify the key nodes and core paths of risk propagation, thus obtaining a network topology map of risk propagation. Perform cascading failure simulation on the network topology of the risk propagation, set a node failure threshold, and trigger a cascading failure when the node risk value exceeds the failure threshold; simulate the avalanche-like spread of risk in the network, simulate and calculate the failure probability distribution under different working conditions, and obtain the probability assessment result of risk spread; A multi-party game model involving bridge management, maintenance teams, and transportation departments is constructed. The probability assessment results are used as profit function parameters, and game decisions are made on the probability assessment results to determine the bridge safety risk level.

8. A safety assessment system for intelligently constructed bridges, characterized in that: include: The acquisition module is used to collect relevant data during the bridge construction process; An extraction module, configured to perform Fourier transform on the correlation data to obtain frequency domain signal data to extract mode shape parameters of the bridge structure; The simulation module is used to perform similarity theoretical modeling on the geometric dimensions and material properties in the bridge design drawings, construct a scaled physical test model, and simulate the mechanical response of the bridge under static and dynamic loads to obtain model test data; The calculation module is used to obtain the field measured data of the bridge, perform dimensionless processing on the field measured data and the model test data, and calculate the deviation of the similarity criterion; The recognition module is used to perform cluster analysis on historical bridge damage case data, establish a damage feature classification system based on Euclidean distance, match the current bridge inspection data with it, and identify the damage type; The evaluation module is used to comprehensively evaluate the identified damage types, similarity criterion deviations and vibration mode parameters to determine the bridge safety risk level.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images