Bridge safety monitoring and analysis system based on big data

Abstract

The present application relates to the technical field of bridge safety monitoring, comprising a bridge safety monitoring and analysis system based on big data, which comprises a data acquisition and processing module, a manifold feature dimension reduction module, an entropy change partition evaluation module, a health state dynamic analysis module and a bridge safety early warning module.In the present application, a bridge measurement point topology relationship matrix is established, the spatial distribution characteristics of the measurement points are analyzed, the spatial correlation of the bridge structure is reflected, the local curvature of the measurement points is calculated and a local curvature matrix is established, the neighborhood similarity matrix is combined, and low-dimensional projection transformation is performed, so that the data maintains local topological consistency during dimension reduction, an entropy change partition matrix is established, the health states of different structural parts of the bridge are differentiated based on mechanical properties, the entropy change trend and damage probability of the measurement points are calculated, and the overall safety state grade of the bridge is judged through a damage risk threshold, so that the safety evaluation can comprehensively consider the severity of local damage of the bridge, and the stability and adaptability of safety state determination are improved.

Description

Technical Field

[0001] This invention relates to the field of bridge safety monitoring technology, and in particular to a bridge safety monitoring and analysis system based on big data. Background Technology

[0002] The field of bridge safety monitoring technology encompasses technical methods and systems for real-time or periodic monitoring of bridge structural conditions. It involves collecting physical quantity data such as stress, vibration, temperature, and displacement of bridges using various sensing devices, and then employing data analysis to identify the structural health status of the bridges. The overall process covers data acquisition, transmission, storage, analysis, and early warning. Data acquisition typically utilizes fiber optic sensing, accelerometers, and displacement measurement; data transmission relies on wired or wireless communication protocols; data storage employs distributed or centralized storage models; and data analysis involves model calculations and feature extraction to determine and assess bridge structural anomalies. Developments in this field have enabled bridge maintenance to evolve from traditional periodic manual inspections to intelligent and automated monitoring, providing technical support for the full lifecycle management of bridges.

[0003] The bridge safety monitoring and analysis system based on big data refers to a system that comprehensively assesses the safety status of bridges using large-scale data analysis technology. This encompasses high-frequency acquisition and cleaning of bridge monitoring data, cross-temporal and spatial data correlation calculation, feature parameter extraction, and health status assessment based on multi-dimensional data. Specifically, high-frequency acquisition of bridge monitoring data relies on a distributed sensor network, acquiring data such as strain, vibration, and temperature of key bridge components through multi-node collaboration. Data cleaning employs statistical screening methods to remove outliers. Cross-temporal and spatial data correlation calculation is based on time-series matching of historical monitoring data and real-time acquired data, calculating the correlation degree through data fusion. Feature parameter extraction relies on calculation methods based on the physical characteristics of the monitored object to extract key indicators related to the structural health status. Health status assessment is based on multi-source data input, using parameter calculation methods to identify the bridge's safety level and analyzing potential structural hazards by combining dynamic trend analysis.

[0004] In current bridge safety monitoring processes, the data acquisition phase primarily relies on sensors to obtain physical quantity data, without establishing a topological relationship matrix for the measurement points. This results in a lack of effective expression of the spatial correlation between measurement points, making it difficult for the data to reflect the overall stress characteristics of the bridge during structural analysis. In the data processing phase, traditional dimensionality reduction methods are often used without incorporating local geometric features of the bridge. This leads to a lack of spatial structural information in the dimensionality-reduced data during damage identification, affecting the accuracy of identifying local damage. In the health status assessment phase, existing methods calculate health status based on global entropy changes, without partitioning the stress characteristics of different parts of the bridge. This causes local damage characteristics to be masked by global changes, affecting the sensitivity of damage identification. In the safety status analysis phase, current bridge safety monitoring and assessment rely heavily on static health status calculations, lacking time-series analysis of the health status change trends of measurement points. This makes it difficult to quantify the risk of long-term damage evolution in the bridge, resulting in insufficient responsiveness of safety assessments to potential structural hazards. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a bridge safety monitoring and analysis system based on big data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a bridge safety monitoring and analysis system based on big data includes:

[0007] The data acquisition and processing module collects bridge strain and vibration data through the bridge sensor network, records the measurement point data, determines the distribution of measurement points, establishes a measurement point topology matrix, calculates the measurement point data density, filters out abnormal data points, and performs normalization processing to obtain a normalized dataset.

[0008] The manifold feature dimensionality reduction module calculates the local curvature of the measurement points based on the normalized dataset, analyzes the spatial topological relationship of the measurement points and calculates the Euclidean distance, calculates the neighborhood similarity matrix in combination with the measurement point data density, performs low-dimensional projection transformation, and obtains the reconstructed feature prediction value.

[0009] The entropy change partitioning evaluation module calculates the entropy change data of each part of the bridge structure based on the predicted value of the reconstructed features, adjusts the entropy change calculation scale according to the data disturbance amplitude, and establishes an entropy change partitioning matrix.

[0010] Based on the entropy change partition matrix, the health status dynamic analysis module calculates the entropy change trend of the measuring points, extracts abnormal fluctuations of vibration modes, calculates the damage probability based on the characteristic change trend of the measuring points, and obtains the structural damage trend analysis results.

[0011] Based on the structural damage trend analysis results, the bridge safety early warning module sets a damage risk threshold, calculates the probability of damage exceeding limits in bridge structural parts, determines the bridge safety status level, and obtains bridge safety early warning indicators.

[0012] As a further aspect of the present invention, the normalized dataset includes normalized strain data, normalized vibration data, normalized measuring point spatial coordinates, normalized timestamps, and normalized sampling frequencies; the reconstructed feature prediction values ​​include projected strain features, projected vibration features, local curvature features, neighborhood similarity features, and low-dimensional embedding features; the entropy variation partitioning matrix includes strain entropy variation intervals, vibration entropy variation intervals, data disturbance rate partitioning results, entropy gradient distribution results, and local feature change rate partitioning results; the structural damage trend analysis results include measuring point entropy variation trend analysis results, vibration mode abnormal change rate, measuring point damage probability, health status change amplitude, and damage time series trend analysis results; the bridge safety early warning indicators include damage risk thresholds, damage exceeding probability, exceeding measuring point proportion, safety status level, and bridge structural component risk assessment records.

[0013] As a further embodiment of the present invention, the data acquisition and processing module includes a measurement point analysis submodule, a data filtering submodule, and a normalization calculation submodule;

[0014] The measuring point analysis submodule collects bridge strain and vibration data through sensors, records the spatial coordinates, timestamps, and sampling frequencies of the measuring points, determines the distribution of measuring points based on the bridge structural layout diagram, establishes a measuring point topology matrix, calculates the measuring point data density, and obtains the measuring point topology density matrix.

[0015] The data filtering submodule filters abnormal data points based on the measurement point topology density matrix, calculates the abnormal point threshold based on the measurement point data density, and removes data points that exceed the abnormal point threshold range to obtain abnormal data filtering results.

[0016] The normalization calculation submodule calculates the normalized value of the data based on the abnormal data filtering results, strain gradient, and vibration mode change rate, using the following formula:

[0017] ;

[0018] Calculate measuring points normalized value This yields a normalized dataset, where... Representative measuring point strain value, This represents the average strain value of the set of measuring points. The variance of strain at all measuring points Representative measuring point The rate of change of vibration modes, This represents the average rate of change of vibration modes of the set of measuring points.

[0019] As a further embodiment of the present invention, the manifold feature dimensionality reduction module includes a local curvature calculation submodule, a neighborhood similarity matrix construction submodule, and a low-dimensional projection transformation submodule;

[0020] The local curvature calculation submodule, based on the normalized dataset, statistically analyzes the strain gradient and vibration mode change rate data of the measuring point, calculates the local geometric features of the measuring point, establishes a local coordinate system for the local area of ​​the measuring point, calculates the spatial gradient difference of the measuring points around the measuring point, calculates the local curvature of the measuring point based on the spatial gradient change rate, and obtains the local curvature value of the measuring point.

[0021] The neighborhood similarity matrix construction submodule analyzes the spatial topological relationship of the measuring points of the bridge main beam, piers and towers based on the local curvature values ​​of the measuring points, calculates the Euclidean distance between each measuring point, obtains the spatial density value of the measuring point distribution, normalizes the relative density between measuring points, and calculates the similarity coefficient between neighboring measuring points by combining the Euclidean distance and density distribution of the neighborhood of the measuring points, and establishes the neighborhood similarity matrix.

[0022] The low-dimensional projection transformation submodule, based on the local curvature value of the measurement point and the similarity matrix of the neighborhood, and according to the distribution characteristics of the local curvature of the measurement point in the neighborhood, uses the following formula:

[0023] ;

[0024] Calculate measuring points Low-dimensional projection transformation value Obtain the reconstructed feature prediction values, where, Representative measuring point The local curvature, Representative measuring point The local curvature, Representative measuring point and measuring points Neighborhood similarity coefficient between them Representative measuring point and measuring points The Euclidean distance between them Representative measuring point The total Euclidean distance to all neighboring measurement points, This represents the total number of measurement points within the neighborhood of the measurement point.

[0025] As a further embodiment of the present invention, the entropy change partitioning evaluation module includes a feature data calculation submodule, an entropy change calculation submodule, and a data disturbance analysis submodule.

[0026] The feature data calculation submodule divides the bridge structure into beam ends, mid-span, supports, and towers based on the bridge structure topology, obtains strain and vibration data for each part, performs feature reconstruction on the data, extracts the predicted value of the reconstructed features for each part, adjusts the calculation scale according to the topological characteristics of the different parts, calculates the reconstructed feature gradient, and filters the local feature change rate to obtain the local feature change rate matrix.

[0027] The entropy change calculation submodule calls the local feature change rate matrix to calculate the entropy change value for strain and vibration data at different locations. It adjusts the calculation scale based on the data disturbance amplitude and calculates the entropy gradient for each location using the formula:

[0028] ;

[0029] Calculate the rate of change of entropy Obtain the entropy change rate matrix for each part, where, Representing the Each reconstructed feature prediction value, Represents time, Represents the entropy change adjustment coefficient. Represents the number of data points;

[0030] The data disturbance analysis submodule calculates the data disturbance rate based on the entropy change rate matrix of each part, analyzes the data disturbance trend based on the coupling relationship between the disturbance rate and the entropy change gradient, and establishes an entropy change partition matrix.

[0031] As a further aspect of the present invention, the health status dynamic analysis module includes an entropy change trend calculation submodule, a vibration mode fluctuation analysis submodule, and a damage probability calculation submodule.

[0032] The entropy change trend calculation submodule extracts the entropy change value of each measuring point based on the entropy change partition matrix, calculates the entropy change trend of the measuring point, analyzes the entropy change gradient according to the time series, filters the measuring points with abnormal change rates, and obtains the measuring point entropy change trend matrix.

[0033] The vibration mode fluctuation analysis submodule calculates the change in vibration signal at the measuring point based on the entropy change trend matrix, analyzes the abnormal fluctuation amplitude of vibration modes at different measuring points, identifies the vibration abrupt change intervals at the measuring points, and uses the following formula:

[0034] ;

[0035] Calculate the intensity of vibration anomalies Obtain abnormal fluctuation values ​​of vibration modes, where, Representing the Vibration signal values ​​at each measuring point Represents the mean value of the vibration signal. Represents the number of measurement points. This represents the change in vibration signal between adjacent time steps;

[0036] The damage probability calculation submodule calculates the characteristic change trend of the measuring points based on the abnormal fluctuation value of the vibration mode and the entropy change trend matrix of the measuring points, analyzes the damage probability of the measuring points, selects measuring points with high damage probability, and obtains the structural damage trend analysis results.

[0037] As a further embodiment of the present invention, the bridge safety early warning module includes a structural damage trend analysis submodule, a damage exceeding probability calculation submodule, and a safety status assessment submodule.

[0038] Based on the structural damage trend analysis results, the structural damage trend analysis submodule obtains damage monitoring data for each structural part of the bridge, calculates the damage change rate, cumulative damage degree and damage fluctuation range, analyzes the damage change trend, determines the damage development mode, calculates the damage trend index for each part, and obtains the structural damage trend index.

[0039] The damage exceedance probability calculation submodule sets a damage risk threshold based on the structural damage trend index, using the following formula:

[0040] ;

[0041] Calculate the probability of damage exceeding the limit for each structural component. ,in, The damage trend index represents the i-th structural part. Represents the damage risk threshold. Represents the total number of structural components;

[0042] The safety status assessment submodule calculates the proportion of over-limit measuring points based on the probability of damage exceeding limits for each structural component, assesses the overall safety status of the bridge based on the bridge safety level classification standard, and generates bridge safety early warning indicators.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0044] In this invention, a topological relationship matrix of bridge measuring points is established to analyze the spatial distribution characteristics of the measuring points, reflecting the spatial correlation of the bridge structure. The local curvature of the measuring points is calculated and a local curvature matrix is ​​established. Combined with the neighborhood similarity matrix, a low-dimensional projection transformation is performed to maintain local topological consistency of the data during dimensionality reduction. An entropy-change partitioning matrix is ​​established to perform differentiated analysis of the health status of different structural parts of the bridge based on mechanical characteristics. The entropy change trend and damage probability of the measuring points are calculated, and the overall safety status level of the bridge is determined by the damage risk threshold. This allows the safety assessment to comprehensively consider the severity of local damage to the bridge, improving the stability and adaptability of the safety status determination. Attached Figure Description

[0045] Figure 1 This is a system flowchart of the present invention;

[0046] Figure 2 This is a flowchart of the data acquisition and processing module of the present invention;

[0047] Figure 3 This is a flowchart of the manifold feature dimensionality reduction module of the present invention;

[0048] Figure 4 This is a flowchart of the entropy change partitioning evaluation module of the present invention;

[0049] Figure 5 This is a flowchart of the dynamic health status analysis module of the present invention;

[0050] Figure 6 This is a flowchart of the bridge safety early warning module of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the 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 merely illustrative and not intended to limit the invention.

[0052] Please see Figure 1 The bridge safety monitoring and analysis system based on big data includes:

[0053] The data acquisition and processing module collects bridge strain and vibration data through the bridge sensor network, records the spatial coordinates, timestamps, and sampling frequencies of the measuring points, determines the distribution of measuring points based on the bridge structure layout diagram, establishes a topological relationship matrix of measuring points, calculates the data density of measuring points, filters out abnormal data points, and performs normalization processing based on strain gradient and vibration mode change rate to obtain a normalized dataset.

[0054] The manifold feature dimensionality reduction module is based on a normalized dataset. It statistically analyzes strain gradient and vibration mode change rate data, calculates the local curvature of the measuring points, establishes a local curvature matrix, analyzes the spatial topological relationship of the measuring points based on the main beam, piers, and towers of the bridge, calculates the Euclidean distance, calculates the neighborhood similarity matrix in combination with the measuring point data density, and performs low-dimensional projection transformation based on the local curvature of the measuring points and the neighborhood similarity matrix to obtain the reconstructed feature prediction value.

[0055] The entropy change partitioning assessment module is based on the reconstructed feature prediction values. According to the bridge structure topology, it divides the beam ends, mid-span, supports, and towers, calculates the strain entropy change and vibration entropy change of each part, adjusts the entropy change calculation scale according to the data disturbance amplitude, calculates the entropy gradient, data disturbance rate, local feature change rate, and establishes the entropy change partitioning matrix.

[0056] The dynamic health status analysis module calculates the entropy change trend of the measuring points based on the entropy change partition matrix, extracts abnormal fluctuations of vibration modes, calculates the damage probability based on the characteristic change trend of the measuring points, and obtains the structural damage trend analysis results.

[0057] The bridge safety early warning module sets a damage risk threshold based on the structural damage trend analysis results, calculates the probability of damage exceeding the limit for each structural part of the bridge, calculates the overall safety status level based on the proportion of exceeding the limit measuring points, and obtains the bridge safety early warning index.

[0058] The normalized dataset includes normalized strain data, normalized vibration data, normalized measurement point spatial coordinates, normalized timestamps, and normalized sampling frequency; the reconstructed feature predictions include projected strain features, projected vibration features, local curvature features, neighborhood similarity features, and low-dimensional embedding features; the entropy variation partitioning matrix includes strain entropy variation intervals, vibration entropy variation intervals, data disturbance rate partitioning results, entropy gradient distribution results, and local feature change rate partitioning results; the structural damage trend analysis results include measurement point entropy variation trend analysis results, vibration mode abnormal change rate, measurement point damage probability, health status change amplitude, and damage time series trend analysis results; the bridge safety early warning indicators include damage risk threshold, damage exceeding probability, proportion of exceeding measurement points, safety status level, and risk assessment records of bridge structural parts.

[0059] Please see Figure 2 The data acquisition and processing module includes a measurement point analysis submodule, a data filtering submodule, and a normalization calculation submodule;

[0060] The measuring point analysis submodule collects bridge strain and vibration data through sensors, records the spatial coordinates, timestamps, and sampling frequencies of the measuring points, determines the distribution of measuring points based on the bridge structural layout diagram, establishes a measuring point topology matrix, calculates the measuring point data density, and obtains the measuring point topology density matrix.

[0061] The bridge monitoring system needs to collect bridge strain and vibration data in real time and record the spatial coordinates, timestamps, and sampling frequency of the measuring points. First, multiple high-precision strain and vibration sensors installed on the bridge synchronously collect data according to a set sampling frequency. For example, if a bridge has 100 strain sensors and 50 vibration sensors, and the sampling frequency is set to 100Hz, then 15,000 data points (100×100 and 50×100) can be obtained per second. Each data point includes the sensor number, timestamp, and strain or vibration data. Based on this, the distribution of measuring points is determined according to the bridge structural layout diagram, and the physical coordinate information of each measuring point is obtained. The specific calculation method can use a three-dimensional coordinate mapping method. For example, let the measuring points... Having coordinates Then its relative position can be calculated from the local coordinate system of the bridge and the coordinate system transformation relationship. Furthermore, a topological relationship matrix of the measuring points is established, and the elements of this matrix are... Reflecting the measuring point With measuring points Are they directly connected? If so, then... ,otherwise For example, the topological relationship matrix of five measurement points of a certain bridge is shown below:

[0062] Table 1. Topological relationship matrix of measurement points:

[0063]

[0064] As shown in Table 1, the connectivity between different measuring points reflects the topological characteristics of the bridge structure. Next, the data density of the measuring points is calculated, and the measuring points are defined. Data density The data density is the ratio of the number of its neighboring measurement points to the number of valid sampling points within the time window. For example, for measurement point 1, its neighboring measurement points are 2 and 5. Assuming the total number of sampling points within the time window is 1000, the data density is calculated as follows:

[0065] ;

[0066] The density of all measuring points is calculated in this way, and the topological density matrix of the measuring points is finally obtained.

[0067] The data filtering submodule filters out abnormal data points based on the topological density matrix of the measurement points, calculates the abnormal point threshold based on the measurement point data density, and removes data points that exceed the abnormal point threshold range to obtain the abnormal data filtering results.

[0068] To filter out outlier data points, the topology density matrix of the measurement points is invoked. First, the average data density of each measurement point is calculated. And set a threshold for filtering outlier data points. Take the range of twice the standard deviation of the mean, that is:

[0069] ;

[0070] in, Let be the standard deviation of the data density. For example, for 10 measurement points, the data densities are as follows: Then calculate its mean:

[0071] ;

[0072] Calculate the standard deviation:

[0073] ;

[0074] Calculated The filtering threshold range is:

[0075] ;

[0076] Data points that exceed the threshold range are removed. For example, the data densities of measurement points 3 and 8 are 0.001 and 0.007 respectively, both exceeding the threshold range. They are identified as abnormal data points and removed from subsequent calculations, thus obtaining the final abnormal data filtering results.

[0077] The normalization calculation submodule calculates the normalized value of the data based on the abnormal data filtering results, strain gradient, and vibration mode change rate, using the following formula:

[0078] ;

[0079] Calculate measuring points normalized value This yields a normalized dataset, where... Representative measuring point strain value, This represents the average strain value of the set of measuring points. The variance of strain at all measuring points Representative measuring point The rate of change of vibration modes, The average rate of change of vibration modes of the set of measuring points;

[0080] The abnormal data filtering results were retrieved, and the sample data is as follows:

[0081] Table 2 Strain and vibration data at measurement points:

[0082]

[0083] Calculate the mean based on Table 2:

[0084] ;

[0085] ;

[0086] Calculate the variance:

[0087] ;

[0088] Calculate the normalized value of measuring point 1:

[0089] ;

[0090] ;

[0091] The normalized values ​​for all measurement points are calculated as above, and the normalized dataset is finally obtained.

[0092] Please see Figure 3 The manifold feature dimensionality reduction module includes a local curvature calculation submodule, a neighborhood similarity matrix construction submodule, and a low-dimensional projection transformation submodule;

[0093] The local curvature calculation submodule is based on a normalized dataset. It statistically analyzes the strain gradient and vibration mode change rate data of the measuring point, calculates the local geometric features of the measuring point, establishes a local coordinate system for the local area of ​​the measuring point, calculates the spatial gradient difference of the measuring points around the measuring point, calculates the local curvature of the measuring point based on the spatial gradient change rate, and obtains the local curvature value of the measuring point.

[0094] In bridge health monitoring, the accuracy of strain gradient and vibration mode change rate data at measuring points is crucial for calculating the local curvature of these points. First, data from multiple measuring points within the bridge structure are acquired, including strain gradient data from key components such as the bridge deck, main beam, piers, and towers. This data can be collected in real-time using strain gauges or fiber optic sensors. For example, a bridge might have 100 measuring points on its deck, each with an initial strain gradient between 0.002 and 0.008, and an initial vibration mode change rate range of 0.1% to 0.5%. Second, a local coordinate system is calculated for the area where the measuring point is located, using the direction of the main beam as the X-axis, the vertical direction as the Y-axis, and the direction of the tower as the Z-axis, thus constructing a three-dimensional coordinate system. The finite difference method is used to calculate the strain gradient change at neighboring points around the measuring point. For example, measuring point P1 (0.002) has strain gradients of 0.004, 0.005, and 0.0035 with its neighboring measuring points P2, P3, and P4, respectively. Then, the curvature is calculated using the following formula:

[0095] ;

[0096] Calculate the local curvature value of measuring point P1 and extend it to all measuring points to finally obtain the local curvature value of the measuring points.

[0097] Table 3 shows the results of local curvature calculations at some measuring points on a certain bridge.

[0098] Table 3: Calculation of Local Curvature at Measuring Points

[0099]

[0100] As shown in Table 3, the local curvature values ​​of each measuring point reflect its morphological changes in the overall structure. Higher curvature values ​​indicate that the strain changes at that measuring point are more significant.

[0101] The neighborhood similarity matrix construction submodule analyzes the spatial topological relationship of the measuring points of the bridge main beam, piers, and towers based on the local curvature values ​​of the measuring points, calculates the Euclidean distance between each measuring point, obtains the spatial density value of the measuring point distribution, normalizes the relative density between measuring points, and calculates the similarity coefficient between neighboring measuring points by combining the Euclidean distance and density distribution of the neighborhood of the measuring points, and establishes the neighborhood similarity matrix.

[0102] After calling the local curvature values ​​of the measuring points, the spatial topological relationship of the measuring points on the bridge is analyzed. First, the Euclidean distance between the measuring points is calculated. Assuming that measuring points A, B, C, and D are set up on the main beam of the bridge with coordinates (0,0,0), (3,0,0), (6,0,0), and (9,0,0) respectively, the Euclidean distance from measuring point A to the other measuring points is calculated as follows:

[0103] ;

[0104] ;

[0105] ;

[0106] Secondly, the spatial density value of the measuring point is calculated. The density value is based on the number of measuring points within a fixed radius R around a unit measuring point. Assuming the radius R is 5m, the spatial density of measuring point A is:

[0107] ;

[0108] The density values ​​of other measuring points are calculated sequentially, then normalized to map them to the range [0, 1] to obtain standardized density values. The neighborhood similarity coefficient is then calculated by combining the Euclidean distance and density distribution of the measuring point's neighborhood.

[0109] ;

[0110] Calculate the neighborhood similarity between measurement point A and other measurement points. Assuming the normalized density values ​​of A, B, C, and D are 0.1, 0.15, 0.2, and 0.25, respectively, then:

[0111] ;

[0112] ;

[0113] ;

[0114] Finally, a neighborhood similarity matrix is ​​constructed for subsequent calculations.

[0115] The low-dimensional projection transformation submodule is based on the local curvature value of the measurement point and the similarity matrix of the neighborhood. According to the distribution characteristics of the local curvature of the measurement point in the neighborhood, it uses the following formula:

[0116] ;

[0117] Calculate measuring points Low-dimensional projection transformation value Obtain the reconstructed feature prediction values, where, Representative measuring point The local curvature, Representative measuring point The local curvature, Representative measuring point and measuring points Neighborhood similarity coefficient between them Representative measuring point and measuring points The Euclidean distance between them Representative measuring point The total Euclidean distance to all neighboring measurement points, The total number of measurement points within the neighborhood of the measurement point;

[0118] Based on the local curvature value of the measuring point and the similarity matrix of the neighborhood, the low-dimensional projected coordinates of the measuring point are calculated. First, the projection transformation value of the measuring point is calculated based on the distribution characteristics of the local curvature value of the measuring point in the neighborhood.

[0119] Assuming the local curvature value of measurement point A is 0.00075, and the local curvature values ​​of measurement points B, C, and D are 0.0012, 0.0015, and 0.0011 respectively, substituting these values ​​into the previously calculated neighborhood similarity matrix:

[0120] ;

[0121] The final calculated low-dimensional projection transformation value of measurement point A is 0.0012. The projection transformation values ​​of other measurement points were calculated sequentially to obtain the predicted values ​​of the reconstructed features. The results show that the projection transformation value of measurement point A is close to a certain proportion of the original curvature value, so that the integrity of the topological relationship can still be maintained after dimensionality reduction.

[0122] Please see Figure 4 The entropy change partitioning evaluation module includes a feature data calculation submodule, an entropy change calculation submodule, and a data perturbation analysis submodule.

[0123] The feature data calculation submodule divides the bridge structure into beam ends, mid-span, supports, and towers based on the bridge structure topology, obtains strain and vibration data for each part, performs feature reconstruction on the data, extracts the reconstructed feature prediction values ​​for each part, adjusts the calculation scale according to the topological characteristics of different parts, calculates the reconstructed feature gradient, and filters the local feature change rate to obtain the local feature change rate matrix.

[0124] Based on the bridge's structural topology, the structure is divided into four regions: beam ends, mid-span, supports, and towers. A coordinate system is established with the towers as the center, and the spatial position of each region within the coordinate system is determined according to its geometric characteristics. Distributed sensors are used to collect strain and vibration data for each region, with a sampling frequency of 1000Hz and a recording period of 10 minutes, which is then stored as time-series data. Data interpolation is used to complete missing data, wavelet transform is used to decompose the signal, and characteristic frequency bands are extracted. Short-time Fourier transform is calculated based on a sliding window strategy to obtain frequency domain features, and Kurtosis is combined to calculate time domain features, resulting in strain and vibration characteristic values ​​for each region. The characteristic threshold range is set to 10%-9%. The 0% quantile interval is used to filter out outliers. This threshold is set based on the statistical distribution of historical strain monitoring data of the bridge to ensure that extreme disturbances are eliminated while retaining the actual structural response information. On this basis, a reconstructed feature matrix is ​​constructed, and the features are standardized according to the topological characteristics to make their mean 0 and standard deviation 1. The feature gradient matrix is ​​then calculated, and the gradient change rate between each feature point is calculated using Euclidean distance. In the gradient calculation process, the benchmark value is set as the overall gradient mean. This benchmark value is obtained through the distribution of sample data to ensure that areas with large feature changes can be captured. The top 20% of data points with a change rate higher than the mean are selected, and their local feature change rates are calculated to establish a local feature change rate matrix.

[0125] The entropy change calculation submodule calls the local feature change rate matrix to calculate the entropy change value for strain and vibration data at different locations. It adjusts the calculation scale based on the data disturbance amplitude and calculates the entropy gradient for each location using the formula:

[0126] ;

[0127] Calculate the rate of change of entropy Obtain the entropy change rate matrix for each part, where, Representing the Each reconstructed feature prediction value, Represents time, Represents the entropy change adjustment coefficient. Represents the number of data points;

[0128] The local feature change rate matrix is ​​invoked to calculate the entropy change values ​​of strain and vibration data for beam ends, mid-span, supports, and towers, respectively. The time step is set to 0.1s, and the entropy value of each time step is calculated. The standard deviation is used to calculate the data disturbance amplitude. If the data disturbance amplitude exceeds 1.5 times the median absolute deviation (MAD), the calculation scale is adjusted. This threshold is set according to the steady-state distribution characteristics of the data. 1.5 times MAD is used to capture sudden changes and avoid misjudging normal fluctuations. The calculation is compared according to local time windows (5s, 10s, 15s), and the entropy gradient under each window is calculated using a formula.

[0129] This represents the entropy change adjustment coefficient, which controls the influence of the second derivative on the final entropy change calculation. Its setting references the fluctuation range of bridge structure vibration data, typically set to 0.05 to ensure appropriate suppression of drastically changing data. The entropy change rate is calculated within a 10-second window to obtain time series data of the entropy change rate of each part, and an entropy change rate matrix of each part is constructed.

[0130] Predicted values ​​of reconstructed features in a certain region Time step .

[0131] Calculate the first derivative:

[0132] ;

[0133] Calculate the second derivative:

[0134] ;

[0135] Substitute into the formula:

[0136] ;

[0137] ;

[0138] ;

[0139] The calculation results show that the entropy change rate of this region is 0.665, which represents the degree of change in the structural state of this region. This value can be compared with the set benchmark entropy change value (usually the long-term statistical average value under the healthy state of the structure, such as 0.5). If it is higher than this value, it indicates that there is local stress concentration or abnormal vibration in the structure.

[0140] The data disturbance analysis submodule calculates the data disturbance rate based on the entropy change rate matrix of each part, analyzes the data disturbance trend based on the coupling relationship between the disturbance rate and the entropy change gradient, and establishes an entropy change partition matrix.

[0141] By calling the entropy change rate matrix for each part, the perturbation rate is calculated for the time series data. The calculation formula is as follows:

[0142] ;

[0143] in, The rate of change of entropy. Its mean, Standard deviation, The number of data points is specified. A data disturbance threshold of 1.5 standard deviations is set, based on long-term structural health monitoring data. 1.5 standard deviations can be used to determine whether data fluctuations exceed the normal range. The disturbance rate of each part is calculated. If the disturbance rate exceeds the threshold, there is abnormal disturbance in that area. The spatial distribution of the disturbance area is statistically analyzed, and the disturbance intensity index is calculated to form a disturbance trend matrix and establish an entropy change partition matrix.

[0144] The entropy change rate data for a certain region is as follows:

[0145] ;

[0146] Calculate the mean:

[0147] ;

[0148] Calculate the standard deviation:

[0149] ;

[0150] ;

[0151] Calculate the disturbance rate:

[0152] ;

[0153] ;

[0154] Set disturbance threshold This value is set based on bridge monitoring data and is typically taken as 1.5 times the standard deviation. ,because The system determined that there was an anomaly in the region, and further analyzed its spatiotemporal distribution, ultimately establishing an entropy-variable partitioning matrix.

[0155] Please see Figure 5 The dynamic health status analysis module includes an entropy change trend calculation submodule, a vibration mode fluctuation analysis submodule, and a damage probability calculation submodule.

[0156] The entropy change trend calculation submodule extracts the entropy change value of each measuring point based on the entropy change partition matrix, calculates the entropy change trend of the measuring point, analyzes the entropy change gradient based on the time series, filters the measuring points with abnormal change rate, and obtains the entropy change trend matrix of the measuring point.

[0157] Based on the entropy change partitioning matrix, the data source of the measuring points is first selected. These measuring points are usually located at key parts of the bridge structure, such as beam ends, mid-span, supports, and towers. The time series data recorded by each measuring point includes strain values ​​and vibration signals. The basic fluctuation range of the measuring point data is determined based on historical data, and the entropy change values ​​of the measuring points at different time steps are extracted. In specific calculations, the entropy change values ​​of each measuring point at different time steps are considered. Entropy value at time The entropy gradient is obtained by calculating the difference between the two time points. For example, at a certain measuring point The entropy value at time t is 0.52. The entropy at time t is 0.48, then the gradient is calculated as follows: The complete gradient matrix is ​​obtained by traversing all measurement points. The gradient matrix is ​​then filtered, and a threshold for judging abnormal gradient changes is set. The threshold is set based on the historical fluctuation range of the entropy gradient at the measuring point. The upper limit of the 90% confidence interval is selected as the threshold for abnormal changes. For example, the mean and standard deviation of the entropy gradient at the measuring point of a certain bridge are calculated as follows:

[0158] ;

[0159] ;

[0160] ;

[0161] Therefore, an entropy gradient threshold is set. If the absolute value of the entropy change gradient of a certain measuring point at a certain time step is greater than 0.045, then the entropy change trend of that measuring point is marked as having an abnormal change, and the entropy change trend matrix of the measuring point is finally obtained.

[0162] The vibration mode fluctuation analysis submodule calculates the change in vibration signal at the measuring point based on the entropy change trend matrix, analyzes the abnormal fluctuation amplitude of vibration modes at different measuring points, and identifies the vibration abrupt change intervals at the measuring points using the following formula:

[0163] ;

[0164] Calculate the intensity of vibration anomalies Obtain abnormal fluctuation values ​​of vibration modes, where, Representing the Vibration signal values ​​at each measuring point Represents the mean value of the vibration signal. Represents the number of measurement points. This represents the change in vibration signal between adjacent time steps;

[0165] The entropy change trend matrix of the measuring points is called to obtain the vibration signal changes of all measuring points. The vibration data of each measuring point comes from the accelerometer and is recorded once every 1 second to form time series data. The difference in vibration change of the measuring point at adjacent time steps is calculated. For example, at a certain measuring point The vibration value at time t is 0.35. The vibration value at time 0.41 is calculated as follows: The data from all time steps are summed to obtain the total change, and the average value of the data is used as the basis for the calculation. And calculate the root mean square rate of change.

[0166] mean of vibration data The calculation method is as follows:

[0167] ;

[0168] Assuming the vibration data recorded at a certain measuring point is [0.35, 0.41, 0.38, 0.44, 0.39], the mean value is calculated as follows:

[0169] ;

[0170] The root mean square rate of change is calculated as follows:

[0171] ;

[0172] Cumulative vibration variation difference:

[0173] ;

[0174] The final vibration anomaly intensity is obtained as follows:

[0175] ;

[0176] Vibration anomaly detection threshold The root mean square error range is calculated based on historical vibration data from measurement points, and an upper limit is set as the anomaly judgment criterion. Historical average intensity of vibration anomalies at measuring points Standard deviation Then calculate:

[0177] ;

[0178] If the vibration intensity is abnormal If the value exceeds 0.20, the measuring point is marked as an anomaly, and the abnormal fluctuation value of the vibration mode is finally obtained.

[0179] The damage probability calculation submodule calculates the characteristic change trend of the measuring points based on the abnormal fluctuation value of the vibration mode and the entropy change trend matrix of the measuring points, analyzes the damage probability of the measuring points, screens the measuring points with high damage probability, and obtains the structural damage trend analysis results.

[0180] By calling the abnormal fluctuation values ​​of the vibration mode and combining them with the entropy change trend matrix of the measuring point, the characteristic change trend of the measuring point is calculated. For the identified abnormal measuring points, the abnormal vibration amplitude is used as the basis for further analysis. and entropy gradient Calculate the probability of damage The formula used is:

[0181] ;

[0182] Vibration anomaly intensity at a certain measuring point Entropy gradient Then calculate the damage probability:

[0183] ;

[0184] Damage probability determination threshold The setting is based on the bridge fatigue life assessment standard, and is usually set at 0.2, that is, if If the value exceeds 0.2, the measuring point has a high risk of damage. The average damage probability is calculated from historical data of the measuring point's lifespan assessment. Standard deviation Then adopt Ultimately, the structural damage trend analysis results are obtained.

[0185] Table 4. Example of vibration measurement point data:

[0186]

[0187] Table 4 shows the vibration value changes of some vibration measurement points over a continuous time step. These data are used to calculate the abnormal fluctuation values ​​of vibration modes and are further used for damage probability assessment.

[0188] Please see Figure 6 The bridge safety early warning module includes a structural damage trend analysis submodule, a damage exceedance probability calculation submodule, and a safety status assessment submodule.

[0189] Based on the structural damage trend analysis results, the structural damage trend analysis submodule obtains damage monitoring data of various structural parts of the bridge, calculates the damage change rate, cumulative damage degree and damage fluctuation range, analyzes the damage change trend, judges the damage development mode, calculates the damage trend index of each part, and obtains the structural damage trend index.

[0190] Based on the structural damage trend analysis results, damage monitoring data for various structural components of the bridge are obtained. Data is collected from multiple sensors deployed on the bridge, including strain sensors and displacement measurement devices. Each sensor periodically collects data and uploads it to the monitoring center, recording the trend of damage changes over time. For example, a strain sensor is installed on a critical structural component of a bridge, with its initial strain value set to zero. When the stress exceeds a set range (e.g., a strain change exceeding 5‰), the system records the change value and calculates the cumulative damage degree. Taking a certain monitoring point as an example, if the initially monitored strain value is 0.0002, and it is measured to be 0.0015 after six months, then the cumulative damage degree is 0. .0013, and simultaneously calculate the historical damage fluctuation range. Calculate the damage change trend index using multi-period historical data, setting the calculation period to 30 days. Calculate the damage increment for each period. Assuming the damage change values ​​measured within a period are 0.0002, 0.0003, and 0.00025 respectively, the damage change rate for the current period is 0.00025. Calculate the damage trend index for each part. The trend index is calculated as the average of the damage rates for all historical periods. Finally, obtain the structural damage trend index. If the damage trend index for a certain bridge monitoring point is 0.002, it indicates that its damage growth rate is relatively fast, and it is necessary to proceed to the next step of calculating the probability of damage exceeding the limit.

[0191] The threshold for the rate of damage change is determined by the material properties, design life, and service environment of the bridge structure. It is typically set to 10%-15% of the strain rate corresponding to the material's yield strength. For example, if the yield strength strain rate of a reinforced concrete bridge is 0.01, then the threshold for the rate of damage change can be set as follows:

[0192] ;

[0193] That is, if the damage rate at a certain monitoring point is higher than 0.001, its trend index will be marked as abnormal and further analysis is required.

[0194] The damage exceedance probability calculation submodule, based on the structural damage trend index, sets a damage risk threshold and uses the following formula:

[0195] ;

[0196] Calculate the probability of damage exceeding the limit for each structural component. ,in, The damage trend index represents the i-th structural part. Represents the damage risk threshold. Represents the total number of structural components;

[0197] Based on the structural damage trend index, a damage risk threshold is set. This threshold is determined based on bridge design standards and historical damage data, typically taking 20% ​​to 30% of the design limit value as a reference. Assuming the design limit damage value for a critical component of the bridge is 0.01, the damage risk threshold can be set as follows:

[0198] ;

[0199] Assuming the damage trend indices for a certain structural component of the bridge are 0.0021, 0.0028, and 0.0030, and the risk threshold is set to 0.0025, the probability of damage exceeding the limit for each structural component is calculated as follows:

[0200] ;

[0201] The probability of damage exceeding the limit for each structural component is calculated. If the value exceeds 50%, it indicates that the structural component has a high risk of damage and needs to enter the safety status assessment stage.

[0202] The damage risk threshold is set based on the fatigue resistance of bridge materials, traffic load standards, and a safety reserve factor. The safety reserve factor is typically taken as 1.2-1.5. If the bridge's fatigue damage limit is 0.01, and a safety reserve factor of 1.4 is used, then:

[0203] ;

[0204] This value reflects the damage warning threshold under normal use conditions; once the limit is exceeded, further evaluation will be required.

[0205] The safety status assessment submodule calculates the proportion of over-limit measuring points based on the probability of damage exceeding limits for each structural part, assesses the overall safety status of the bridge based on the bridge safety level classification standard, and generates bridge safety early warning indicators.

[0206] The probability of damage exceeding limits for each structural component is retrieved, and the proportion of exceeding measurement points is calculated. The calculation method for the overall proportion of exceeding measurement points for the bridge is as follows:

[0207] ;

[0208] in, Given the total number of monitoring points, assuming there are 100 monitoring points on the bridge, and 40 of these points have a damage exceedance probability exceeding 50%, then calculate:

[0209] ;

[0210] According to the bridge safety level classification standard, if the proportion of monitoring points exceeding the limit is less than 20%, it is rated as a safe state; 20% to 50% is a warning state; and more than 50% is a dangerous state. The proportion of monitoring points exceeding the limit in this bridge is 40%, so it is rated as a warning state. A bridge safety early warning indicator is generated. This result indicates that there is a certain risk of damage to the overall structure of the bridge. Monitoring should be further strengthened and corresponding maintenance measures should be formulated.

[0211] The benchmark values ​​for safety level assessment are set based on domestic and international standards for bridge damage level assessment, and are typically set as follows:

[0212] Safe status ( <20%): The structure is in good condition and can be used normally;

[0213] Alert status (20%≤) ≤50%): Monitoring needs to be strengthened; repairs may be required.

[0214] Dangerous conditions ( >50%: Immediate repair or traffic restriction measures are required.

[0215] Suppose a city's bridge monitoring system requires... If more than 30% of the bridges are included in the key monitoring list, then this bridge (40%) needs to be included in the key monitoring scope, and corresponding safety warnings and maintenance measures should be taken.

[0216] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A bridge safety monitoring and analysis system based on big data, characterized in that, The system includes: The data acquisition and processing module collects bridge strain and vibration data through the bridge sensor network, records the measurement point data, determines the distribution of measurement points, establishes a measurement point topology matrix, calculates the measurement point data density, filters out abnormal data points, and performs normalization processing to obtain a normalized dataset. The data density of the measurement point specifically refers to the ratio of the number of adjacent measurement points around the corresponding measurement point to the number of effective sampling points within a time period; The manifold feature dimensionality reduction module calculates the local curvature of the measurement points based on the normalized dataset, analyzes the spatial topological relationship of the measurement points and calculates the Euclidean distance, calculates the neighborhood similarity matrix in combination with the measurement point data density, performs low-dimensional projection transformation, and obtains the reconstructed feature prediction value. The entropy change partitioning evaluation module calculates the entropy change data of each part of the bridge structure based on the predicted value of the reconstructed features, adjusts the entropy change calculation scale according to the data disturbance amplitude, and establishes an entropy change partitioning matrix. The health status dynamic analysis module calculates the entropy change trend of the measuring points based on the entropy change partition matrix, extracts abnormal fluctuations of vibration modes, calculates the damage probability based on the characteristic change trend of the measuring points, and obtains the structural damage trend analysis results. Based on the structural damage trend analysis results, the bridge safety early warning module sets a damage risk threshold, calculates the probability of damage exceeding the limit in the bridge structural parts, determines the bridge safety status level, and obtains the bridge safety early warning index. The manifold feature dimensionality reduction module includes a local curvature calculation submodule, a neighborhood similarity matrix construction submodule, and a low-dimensional projection transformation submodule; The local curvature calculation submodule, based on the normalized dataset, statistically analyzes the strain gradient and vibration mode change rate data of the measuring point, calculates the local geometric features of the measuring point, establishes a local coordinate system for the local area of ​​the measuring point, calculates the spatial gradient difference of the measuring points around the measuring point, calculates the local curvature of the measuring point based on the spatial gradient change rate, and obtains the local curvature value of the measuring point. The neighborhood similarity matrix construction submodule analyzes the spatial topological relationship of the measuring points of the bridge main beam, piers and towers based on the local curvature values ​​of the measuring points, calculates the Euclidean distance between each measuring point, obtains the spatial density value of the measuring point distribution, normalizes the relative density between measuring points, and calculates the similarity coefficient between neighboring measuring points by combining the Euclidean distance and density distribution of the neighborhood of the measuring points, and establishes the neighborhood similarity matrix. The low-dimensional projection transformation submodule, based on the local curvature value of the measurement point and the similarity matrix of the neighborhood, and according to the distribution characteristics of the local curvature of the measurement point in the neighborhood, uses the following formula: ； Calculate measuring points Low-dimensional projection transformation value Obtain the predicted values ​​of the reconstructed features, where, Representative measuring point The local curvature, Representative measuring point The local curvature, Representative measuring point and measuring points Neighborhood similarity coefficient between them Representative measuring point and measuring points The Euclidean distance between them Representative measuring point The total Euclidean distance to all neighboring measurement points, This represents the total number of measurement points within the neighborhood of the measurement point.

2. The bridge safety monitoring and analysis system based on big data according to claim 1, characterized in that, The normalized dataset includes normalized strain data, normalized vibration data, normalized measurement point spatial coordinates, normalized timestamps, and normalized sampling frequency; the reconstructed feature predictions include projected strain features, projected vibration features, local curvature features, neighborhood similarity features, and low-dimensional embedding features. The entropy change partitioning matrix includes strain entropy change interval, vibration entropy change interval, data disturbance rate partitioning results, entropy gradient distribution results, and local feature change rate partitioning results; the structural damage trend analysis results include measurement point entropy change trend analysis results, vibration mode abnormal change rate, measurement point damage probability, health status change amplitude, and damage time series trend analysis results. The bridge safety early warning indicators include damage risk threshold, probability of damage exceeding limits, proportion of exceeding limit measuring points, safety status level, and risk assessment records of bridge structural parts.

3. The bridge safety monitoring and analysis system based on big data according to claim 1, characterized in that, The data acquisition and processing module includes a measurement point analysis submodule, a data filtering submodule, and a normalization calculation submodule. The measuring point analysis submodule collects bridge strain and vibration data through sensors, records the spatial coordinates, timestamps, and sampling frequencies of the measuring points, determines the distribution of measuring points based on the bridge structural layout diagram, establishes a measuring point topology matrix, calculates the measuring point data density, and obtains the measuring point topology density matrix. The data filtering submodule filters abnormal data points based on the measurement point topology density matrix, calculates the abnormal point threshold based on the measurement point data density, and removes data points that exceed the abnormal point threshold range to obtain abnormal data filtering results. The normalization calculation submodule calculates the normalized value of the data based on the abnormal data filtering results, strain gradient, and vibration mode change rate, using the following formula: ； Calculate measuring points normalized value This yields a normalized dataset, where... Representative measuring point strain value, This represents the average strain value of the set of measuring points. The variance of strain at all measuring points Representative measuring point The rate of change of vibration modes, This represents the average rate of change of vibration modes of the set of measuring points.

4. The bridge safety monitoring and analysis system based on big data according to claim 1, characterized in that, The entropy change partitioning evaluation module includes a feature data calculation submodule, an entropy change calculation submodule, and a data perturbation analysis submodule. The feature data calculation submodule divides the bridge structure into beam ends, mid-span, supports, and towers based on the bridge structure topology, obtains strain and vibration data for each part, performs feature reconstruction on the data, extracts the predicted value of the reconstructed features for each part, adjusts the calculation scale according to the topological characteristics of the different parts, calculates the reconstructed feature gradient, and filters the local feature change rate to obtain the local feature change rate matrix. The entropy change calculation submodule calls the local feature change rate matrix to calculate the entropy change value for strain and vibration data at different locations. It adjusts the calculation scale based on the data disturbance amplitude and calculates the entropy gradient for each location using the formula: ； Calculate the rate of change of entropy Obtain the entropy change rate matrix for each part, where, Representing the Each reconstructed feature prediction value, Represents time, Represents the entropy change adjustment coefficient. Represents the number of data points; The data disturbance analysis submodule calculates the data disturbance rate based on the entropy change rate matrix of each part, analyzes the data disturbance trend based on the coupling relationship between the disturbance rate and the entropy change gradient, and establishes an entropy change partition matrix.

5. The bridge safety monitoring and analysis system based on big data according to claim 1, characterized in that, The health status dynamic analysis module includes an entropy change trend calculation submodule, a vibration mode fluctuation analysis submodule, and a damage probability calculation submodule. The entropy change trend calculation submodule extracts the entropy change value of each measuring point based on the entropy change partition matrix, calculates the entropy change trend of the measuring point, analyzes the entropy change gradient according to the time series, filters the measuring points with abnormal change rates, and obtains the measuring point entropy change trend matrix. The vibration mode fluctuation analysis submodule calculates the change in vibration signal at the measuring point based on the entropy change trend matrix, analyzes the abnormal fluctuation amplitude of vibration modes at different measuring points, identifies the vibration abrupt change intervals at the measuring points, and uses the following formula: ； Calculate the intensity of vibration anomalies Obtain abnormal fluctuation values ​​of vibration modes, where, Representing the Vibration signal values ​​at each measuring point Represents the mean value of the vibration signal. Represents the number of measurement points. This represents the change in vibration signal between adjacent time steps; The damage probability calculation submodule calculates the characteristic change trend of the measuring points based on the abnormal fluctuation value of the vibration mode and the entropy change trend matrix of the measuring points, analyzes the damage probability of the measuring points, selects measuring points with high damage probability, and obtains the structural damage trend analysis results.

6. The bridge safety monitoring and analysis system based on big data according to claim 1, characterized in that, The bridge safety early warning module includes a structural damage trend analysis submodule, a damage exceeding probability calculation submodule, and a safety status assessment submodule. Based on the structural damage trend analysis results, the structural damage trend analysis submodule obtains damage monitoring data for each structural part of the bridge, calculates the damage change rate, cumulative damage degree and damage fluctuation range, analyzes the damage change trend, determines the damage development mode, calculates the damage trend index for each part, and obtains the structural damage trend index. The damage exceedance probability calculation submodule sets a damage risk threshold based on the structural damage trend index, using the following formula: ； Calculate the probability of damage exceeding the limit for each structural component. ,in, The damage trend index represents the i-th structural part. Represents the damage risk threshold. Represents the total number of structural components; The safety status assessment submodule calculates the proportion of over-limit measuring points based on the probability of damage exceeding limits for each structural component, assesses the overall safety status of the bridge based on the bridge safety level classification standard, and generates bridge safety early warning indicators.

Images