Bridge health detection method and device, electronic equipment and readable storage medium

Abstract

The application provides a bridge health detection method and device, electronic equipment and readable storage medium, and belongs to the technical field of automatic monitoring. The method comprises the following steps: acquiring an average value of a bridge structure parameter measured value in a current first period, a change rate of the bridge structure parameter measured value in the current first period, a prediction value of a bridge structure parameter threshold value in the current first period, and an aging state parameter of the bridge in the current first period; calculating an aging deviation based on the average value of the bridge structure parameter measured value and the prediction value of the bridge structure parameter threshold value; obtaining an anomaly score through a trained isolation forest model based on the aging deviation, the change rate of the bridge structure parameter measured value and the aging state parameter; and determining a health detection result of the bridge based on the relative size of the anomaly score and a preset score threshold value. The bridge health detection method and device, electronic equipment and readable storage medium provided by the application can improve the accuracy of bridge health detection.

Description

Technical Field

[0001] This application belongs to the field of automatic monitoring technology, and more specifically, relates to a bridge health detection method and device, electronic equipment, and readable storage medium. Background Technology

[0002] As a vital transportation infrastructure, the structural stability of bridges directly impacts the safe and efficient operation of transportation networks. However, during long-term use, bridges face multiple risks, including normal aging, load-bearing effects, environmental erosion, and sudden damage, which can create safety hazards. If these safety hazards are not identified and addressed in a timely manner, they will not only accelerate the evolution of bridge structural damage and lead to a significant increase in maintenance costs, but may also trigger serious accidents such as bridge collapse and traffic disruption under extreme conditions or sudden loads.

[0003] Bridge health monitoring, as a core technical means of preventative maintenance, can provide early warnings of risks and prevent accidents. However, most existing bridge health monitoring methods rely on fixed thresholds, which cannot match the dynamic service characteristics of bridge structures, easily leading to false alarms or missed alarms. Summary of the Invention

[0004] The purpose of this application is to provide a bridge health detection method and apparatus, electronic device, and readable storage medium to improve the accuracy of bridge health detection.

[0005] A first aspect of this application provides a bridge health detection method, comprising:

[0006] The average value of the measured bridge structural parameters in the current first time period, the rate of change of the measured bridge structural parameters in the current first time period, the predicted value of the bridge structural parameter threshold in the current first time period, and the aging status parameters of the bridge in the current first time period are obtained; wherein, the aging status parameters include the service life and the number of historical defects, and the predicted value of the bridge structural parameter threshold is obtained based on the trained time series prediction model.

[0007] The aging deviation is calculated based on the average value of the measured values ​​of the bridge structural parameters and the predicted value of the bridge structural parameter threshold. Based on the aging deviation, the rate of change of the measured values ​​of the bridge structural parameters, and the aging state parameters, anomaly scores are obtained through a trained isolated forest model. The trained isolated forest model is trained based on historical data under normal conditions.

[0008] The health detection result of the bridge is determined based on the relative magnitude of the abnormal score and the preset score threshold.

[0009] A second aspect of this application provides a bridge health detection device, comprising:

[0010] The data acquisition module is used to acquire the average value of the measured values ​​of the bridge structural parameters in the current first time period, the rate of change of the measured values ​​of the bridge structural parameters in the current first time period, the predicted value of the threshold of the bridge structural parameters in the current first time period, and the aging status parameters of the bridge in the current first time period; wherein, the aging status parameters include the usage time and the number of historical defects, and the predicted value of the bridge structural parameter threshold is obtained by prediction based on the trained time series prediction model.

[0011] The anomaly detection module is used to calculate the aging deviation based on the average value of the measured values ​​of the bridge structural parameters and the predicted value of the bridge structural parameter threshold. Based on the aging deviation, the rate of change of the measured values ​​of the bridge structural parameters, and the aging state parameters, an anomaly score is obtained through a trained isolated forest model. The trained isolated forest model is trained based on historical data under normal conditions.

[0012] The health detection module is used to determine the health detection result of the bridge based on the relative magnitude of the abnormal score and a preset score threshold.

[0013] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the bridge health detection method described above.

[0014] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the bridge health detection method described above.

[0015] The beneficial effects of the bridge health detection method and apparatus, electronic device, and readable storage medium provided in this application are as follows:

[0016] In this embodiment, considering the dynamic changes in structural parameters during bridge aging, a dynamic threshold is generated through a time-series prediction model. The dynamic threshold can be adjusted in real time as the bridge's service life increases and the number of historical defects accumulates, always conforming to the actual load-bearing capacity at different aging stages, thus reducing detection errors caused by threshold mismatch from the source. Based on this, aging deviation is calculated based on the dynamic threshold. Aging deviation can characterize the degree to which structural parameters deviate from the safety boundary, parameter change rate can characterize the degree of parameter mutation, and aging state parameters can quantify the risk accumulation caused by aging. Based on multi-dimensional features such as aging deviation, parameter change rate, and aging state parameters, the degree of deviation between the current state and the normal state is quantified through an isolated forest model to obtain a quantifiable anomaly score. This can accurately distinguish between fluctuations in the normal state and deviations in the abnormal state, effectively identifying minute anomalies and early latent anomalies that are difficult to detect by traditional methods, thereby achieving accurate identification of the bridge's health status. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic flowchart of a bridge health detection method provided in an embodiment of this application;

[0019] Figure 2 This is a structural block diagram of a bridge health monitoring device provided in one embodiment of this application;

[0020] Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0021] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0022] It is understood that in the embodiments of this application, data such as user information are involved. When the embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with relevant laws, regulations and standards.

[0023] It should be noted that the terms "first," "second," etc., used in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in sequences other than those illustrated or described herein.

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0025] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a bridge health detection method according to an embodiment of this application. The bridge health detection method provided in this embodiment can be executed by an electronic device, and the method may include:

[0026] S101: Obtain the average value of the measured bridge structural parameters in the current first time period, the rate of change of the measured bridge structural parameters in the current first time period, the predicted value of the bridge structural parameter threshold in the current first time period, and the aging status parameters of the bridge in the current first time period; wherein, the aging status parameters include the service life and the number of historical defects, and the predicted value of the bridge structural parameter threshold is obtained by prediction based on the trained time series prediction model.

[0027] In this embodiment, the bridge structural parameters may include parameters such as strain, displacement, and crack width. The measured values ​​of the above parameters can be obtained by corresponding sensors. For example, the measured value of the strain parameter can be obtained by a strain sensor, the measured value of the displacement parameter can be obtained by a displacement sensor, and the measured value of the crack parameter can be obtained by a crack monitoring instrument.

[0028] Based on this, by acquiring measured values ​​of bridge structural parameters at multiple detection times within the current first time period, multiple measured values ​​of bridge structural parameters can be obtained. The average value of these multiple measured values ​​is then calculated to obtain the average value of the bridge structural parameters for the current first time period. Calculating the average value can eliminate detection errors caused by short-term fluctuations. Simultaneously, a maximum and a minimum value can be selected from the multiple measured values ​​of the bridge structural parameters. The difference between the maximum and minimum values ​​is calculated, and this difference is divided by the duration of the first time period (e.g., 10 minutes) to obtain the rate of change of the measured values ​​of the bridge structural parameters for the current first time period.

[0029] Furthermore, considering that the actual load-bearing capacity of a bridge will continuously decrease with the increase of service time and the accumulation of historical defects, and that structural parameters such as strain and displacement will continuously increase, this embodiment uses a time-series prediction model to dynamically generate predicted values ​​of the bridge structural parameter thresholds for the current first time period, i.e., dynamic thresholds. These dynamic thresholds are used to determine whether the structural parameters exceed the safety boundary. The service time and the number of historical defects can be obtained from the bridge's maintenance records. These two parameters can quantify the normal aging level of the bridge; therefore, they are used as aging state parameters.

[0030] S102: Calculate the aging deviation based on the average value of the measured values ​​of the bridge structural parameters and the predicted values ​​of the bridge structural parameter thresholds. Based on the aging deviation, the rate of change of the measured values ​​of the bridge structural parameters, and the aging state parameters, obtain the anomaly score through the trained isolated forest model. The trained isolated forest model is trained based on historical data under normal conditions.

[0031] In this embodiment, the difference between the average value of the measured bridge structural parameters and the predicted value of the bridge structural parameter threshold can be calculated. This difference is used as the aging deviation, which characterizes the degree of deviation between the actual aging state of the bridge and the normal aging trend. Specifically, a positive aging deviation indicates that the actual aging degree is higher than the normal aging trend, while a negative aging deviation indicates that the actual aging degree is lower than the normal trend.

[0032] Meanwhile, the rate of change of the measured values ​​of bridge structural parameters can characterize the degree of abrupt change in the bridge structural parameters. During normal aging, the measured values ​​of bridge structural parameters change slowly, while when the structure is damaged, the measured values ​​of bridge structural parameters may change rapidly. Therefore, the greater the rate of change of the measured values ​​of bridge structural parameters, the more drastic the change is, and the more likely it is to be a sudden injury.

[0033] In this embodiment, aging deviation, the rate of change of measured bridge structural parameters, and aging state parameters are used to characterize bridge anomalies from three dimensions: trend deviation, dynamic mutation, and risk accumulation caused by aging. Therefore, data on these three dimensions under normal bridge conditions can be collected in advance as sample data to train the Isolation Forest model. Based on a large amount of normal state data, the Isolation Forest model constructs multiple decision trees by randomly partitioning the feature space, learning the characteristic distribution patterns of aging deviation, parameter change rate, and aging state parameters under normal conditions. Normal samples will be assigned to deeper decision tree nodes (difficult to isolate).

[0034] Based on this, using real-time acquired aging deviations, the rate of change of measured bridge structural parameters, and aging state parameters, the isolation degree (path length) of features in the decision tree is calculated using the isolated forest model to output anomaly scores. Abnormal samples are quickly isolated (short path length) because their feature distribution deviates from the normal range, and their corresponding scores are close to 1; normal samples are difficult to isolate (long path length) because they conform to the feature distribution pattern, and their corresponding scores are close to 0.

[0035] S103: Determine the health inspection results of the bridge based on the relative magnitude of the abnormal scores and the preset score thresholds.

[0036] In this embodiment, those skilled in the art can pre-calculate a score threshold based on the bridge's design safety level, operation and maintenance experience, and historical anomaly data; for example, 0.65 for extra-large bridges and 0.75 for medium-sized bridges. Based on this, when the anomaly score is below the score threshold, it indicates that the current feature distribution is within the normal distribution range, and the bridge structural parameters have not deviated from the safety boundary corresponding to the current bearing capacity, thus being judged as normal. When the anomaly score is above the score threshold, it indicates that at least one type of feature exceeds the normal distribution range, and the bridge has a structural anomaly risk, thus being judged as abnormal.

[0037] As can be seen from the above, this embodiment takes into account the dynamic changes of structural parameters during the aging process of bridges. A dynamic threshold is generated through a time-series prediction model. The dynamic threshold can be adjusted in real time as the bridge's service life increases and the number of historical defects accumulates, always conforming to the actual load-bearing capacity at different aging stages, thus reducing detection errors caused by threshold mismatch from the source. On this basis, aging deviation is calculated based on the dynamic threshold. Aging deviation can characterize the degree to which structural parameters deviate from the safety boundary, parameter change rate can characterize the degree of parameter mutation, and aging state parameters can quantify the risk accumulation caused by aging. Based on multi-dimensional features such as aging deviation, parameter change rate, and aging state parameters, the degree of deviation between the current state and the normal state is quantified through the isolated forest model to obtain a quantifiable anomaly score. This can accurately distinguish between fluctuations in the normal state and deviations in the abnormal state, effectively identifying minute anomalies and early latent anomalies that are difficult to detect by traditional methods, thereby achieving accurate identification of the bridge's health status.

[0038] In one embodiment of this application, the method for determining the predicted value of the bridge structural parameter threshold for the current first time period includes:

[0039] The threshold update step is performed according to the preset update cycle, and the predicted value of the bridge structural parameter threshold obtained from the most recent threshold update step is used as the predicted value of the bridge structural parameter threshold for the current first time period.

[0040] The threshold update steps include:

[0041] Using the time point of the threshold update step as the reference time point, the measured values ​​of bridge structural parameters and the aging status parameters of the bridge during the second period are obtained at multiple detection times before the reference time point.

[0042] Based on the measured values ​​of bridge structural parameters at multiple detection times during the second time period and the aging state parameters of the bridge during the second time period, the predicted values ​​of the bridge structural parameter thresholds for the first time period after the reference time point are obtained through the trained time series prediction model. The trained time series prediction model is trained based on multiple sets of normal historical sample data. Each set of historical sample data includes the measured values ​​of bridge structural parameters at multiple detection times during the historical second time period under normal conditions, the aging state parameters of the bridge during the historical second time period, and the true values ​​of the bridge structural parameter thresholds during the historical first time period.

[0043] In this embodiment, multiple sets of normal historical sample data can be collected in advance. Each set of historical sample data includes the measured values ​​of bridge structural parameters at multiple detection times within a second historical time period (e.g., 1 hour) under normal conditions, the aging state parameters of the bridge within the second historical time period, and the true values ​​of the bridge structural parameter thresholds within the first historical time period. The measured values ​​of bridge structural parameters at multiple detection times within the second historical time period and the aging state parameters of the bridge within the second historical time period are used as input data, and the true values ​​of the bridge structural parameter thresholds within the first historical time period are used as output data. This trains a time-series prediction model, enabling the model to learn the intrinsic relationship between aging state, parameter fluctuations, and safety thresholds from normal historical samples. This ensures that the dynamic threshold always closely matches the actual safety boundary under normal bridge aging, while eliminating the influence of abnormal situations on parameter fluctuations, thus improving the accuracy and stability of threshold prediction. Specifically, the time-series prediction model can be implemented using existing models such as Long Short-Term Memory (LSTM) networks or Gated Recurrent Units (GRUs).

[0044] Based on this, the predicted values ​​of bridge structural parameter thresholds can be updated periodically according to a preset update cycle. Specifically, each time a threshold update is performed, the time point of the threshold update step is used as the reference time point. The measured values ​​of bridge structural parameters at multiple detection times within the second time period before the reference time point, as well as the aging state parameters of the bridge within the second time period, are obtained. The correlation between structural parameter fluctuations and aging state parameters is then mined using a time-series prediction model to obtain the predicted values ​​of bridge structural parameter thresholds for the first time period after the reference time point. The predicted values ​​of bridge structural parameter thresholds for the first time period are primarily used for real-time bridge health monitoring. The first time period is relatively short (e.g., 10 minutes), allowing for timely identification of abnormal fluctuations within a short timeframe, meeting the needs of real-time health monitoring. The predicted values ​​of bridge structural parameter thresholds for the first time period are obtained by predicting the measured values ​​of bridge structural parameters at multiple detection times within the second time period, as well as the aging state parameters of the bridge within the second time period. The second time period is longer than the first time period, allowing for the accumulation of sufficient measured parameter values ​​and aging state data to support feature learning in the time-series prediction model and ensure the effectiveness of the model's predictions.

[0045] Furthermore, considering that the performance degradation of bridge structures (i.e., changes in bridge structural parameters) is a long-term and slow process, the update cycle of the predicted values ​​of bridge structural parameter thresholds can be longer. Specifically, it can be set according to the load conditions of the bridge, for example, 1 month for high-load bridges and 3 months for ordinary bridges.

[0046] As can be seen from the above, this embodiment uses normal historical sample data to train the time-series prediction model, which can accurately learn the inherent correlation between structural parameter fluctuations, aging state, and safety thresholds during the normal aging process of bridges. This effectively eliminates the interference of abnormal situations on parameter patterns, ensuring that the dynamic thresholds output by the model always closely match the actual safety boundaries under normal aging, avoiding threshold deviations caused by abnormal data. Furthermore, by periodically updating the predicted values ​​of bridge structural parameter thresholds according to a preset update cycle, it is possible to further adapt to changes in load-bearing capacity and reduce misjudgments and omissions.

[0047] In one embodiment of this application, the training process of the time series prediction model includes:

[0048] For each set of historical sample data, the aging stage assessment value of the set of historical sample data is determined based on the aging state parameters in the set of historical sample data.

[0049] Based on the aging stage assessment values ​​corresponding to each set of historical sample data, multiple sets of historical sample data are divided into sample data of multiple stages; the sample data of multiple stages include healthy period sample data, early aging sample data, mid-aging sample data, and late aging sample data.

[0050] Train the corresponding time series prediction sub-model for each stage based on the sample data of each stage;

[0051] The time series prediction sub-models trained in multiple stages are used as the completed time series prediction model.

[0052] Specifically, based on the measured values ​​of bridge structural parameters at multiple detection times within the second time period and the aging state parameters of the bridge within the second time period, the predicted values ​​of the bridge structural parameter thresholds for the first time period after the reference time point are obtained through a trained time-series prediction model, including:

[0053] The aging stage assessment value of the bridge in the second time period is determined based on the aging status parameters of the bridge in the second time period.

[0054] The bridge's stage is determined based on the aging stage assessment values ​​of the bridge during the second time period.

[0055] Select the corresponding time-series prediction sub-model based on the stage in which the bridge is located;

[0056] The predicted values ​​of bridge structural parameter thresholds for the first time period after the reference time point are obtained through the time-series prediction sub-model of the corresponding stage.

[0057] In this embodiment, considering that bridge aging is a gradual process, the correlation between structural parameter fluctuations, aging status, and safety thresholds differs at different stages. For example, parameters fluctuate gently and thresholds are stable during the healthy period, while parameters fluctuate drastically and threshold decays rapidly during the later aging period. Therefore, this embodiment first determines the corresponding aging stage assessment value based on the aging status parameters in each set of historical sample data. According to the magnitude of the corresponding aging stage assessment value, multiple sets of historical sample data are divided into sample data for multiple stages. For example, multiple sets of historical sample data can be divided into healthy period sample data, early aging sample data, mid-term aging sample data, and late aging sample data.

[0058] Based on this, a time-series prediction sub-model can be trained for each stage using sample data from each stage. Each sub-model is trained solely on sample data from its corresponding stage, focusing on learning the unique correlation patterns specific to that stage. Furthermore, considering the differences in sample data across stages, different hyperparameters can be set for different time-series prediction sub-models. For example, in the healthy phase, the structural parameters of the sample data fluctuate smoothly, and the aging state is stable. The correlation between structural parameter fluctuations and thresholds is simple and highly linear, making fitting easy. Therefore, a larger learning rate can be used to accelerate parameter updates and quickly converge to the optimal solution, capturing simple patterns without excessive training time. In the early aging phase, the fluctuations in sample data begin to increase, and the complexity of the correlation patterns slightly increases. Therefore, the learning rate can be appropriately reduced to avoid parameter update oscillations caused by increased feature fluctuations, ensuring stable convergence. Qualitatively, the mid-term aging sample data fluctuates dramatically, the cumulative aging effect is significant, the nonlinear correlation between structural parameter fluctuations, aging state and threshold is enhanced, and the complexity of feature patterns is greatly increased. The learning rate can be further reduced to make the parameter update step size more refined, which can gradually approach the optimal mapping relationship of complex features and avoid skipping the optimal solution due to excessive step size. The late-term aging sample data fluctuates irregularly, the mapping relationship between aging state and threshold is highly nonlinear, and the samples may contain a small amount of implicit noise. Further reducing the learning rate can ensure more robust parameter updates, gradually correct complex errors, avoid model oscillations caused by noise interference, and achieve accurate convergence.

[0059] Therefore, as the aging stage changes—from the healthy period to the early aging stage, the middle aging stage, and the late aging stage—the learning rate of the time-series prediction sub-model decreases sequentially, while the corresponding number of iterations increases sequentially. For example, for healthy period sample data, the learning rate of the time-series prediction sub-model can be set to 0.001 and the number of iterations to 500; for early aging sample data, the learning rate can be set to 0.0008 and the number of iterations to 600; for middle aging sample data, the learning rate can be set to 0.0005 and the number of iterations to 800; and for late aging sample data, the learning rate can be set to 0.0003 and the number of iterations to 1000.

[0060] Therefore, when performing the threshold update step, the aging stage assessment value of the bridge in the second time period can be determined first based on the aging state parameters of the bridge in the second time period before the reference time point. Then, the stage of the bridge is determined according to the aging stage assessment value, and the time series prediction sub-model of the corresponding stage is selected to predict the threshold of the bridge structural parameters, so as to obtain the predicted value of the threshold of the bridge structural parameters in the first time period after the reference time point.

[0061] As can be seen from the above, this embodiment divides multiple sets of sample data based on aging state parameters and trains the corresponding time-series prediction sub-model for each stage to adapt to the fluctuation pattern of structural parameters in each aging stage. In practical applications, the corresponding time-series prediction sub-model can be called according to the actual aging stage to make the threshold prediction fit the current bearing capacity of the bridge.

[0062] In one embodiment of this application, the step of determining the aging stage assessment value based on aging state parameters includes:

[0063] The first assessment value of aging status is determined based on the usage duration and the preset service life in the aging status parameters.

[0064] The second assessment value of aging status is determined based on the number of historical defects in the aging status parameters;

[0065] The aging stage assessment value is determined based on the first assessment value and the second assessment value.

[0066] In this embodiment, usage duration can characterize the gradual process of normal aging of a bridge. Therefore, a first assessment value of the aging state can be determined based on the usage duration and the preset service life in the aging state parameters. The closer the usage duration is to the preset service life, the larger the first assessment value, and the deeper the aging degree. Simultaneously, the number of historical defects can characterize the damage accumulation process of the bridge. The more damage accumulated, the deeper the aging degree. Therefore, a second assessment value of the aging state can be determined based on the number of historical defects in the aging state parameters. Based on this, the aging stage assessment value of the bridge is determined by comprehensively considering the first and second assessment values.

[0067] Specifically, the aging stage assessment value can be determined using the following formula:

[0068] ;

[0069] in, Indicates the aging stage assessment value. Indicates usage duration. This indicates the preset service life (e.g., 50 years). Indicates the number of historical diseases / diseases. This represents a preset scaling factor, which can be set by those skilled in the art according to actual needs. The specific values, for example .

[0070] Based on this, the aging stage of a bridge can be determined by the magnitude of the aging stage assessment value. For example, when the aging stage assessment value is less than 0.2, the bridge is determined to be in the healthy period; when the aging stage assessment value is between [0.2, 0.5), the bridge is determined to be in the early aging stage; when the aging stage assessment value is between [0.5, 0.8), the bridge is determined to be in the middle aging stage; and when the aging stage assessment value is greater than 0.8, the bridge is determined to be in the late aging stage.

[0071] In one embodiment of this application, for each stage, a time-series prediction sub-model corresponding to that stage is trained based on the sample data of that stage, including:

[0072] Based on the sample data of this stage, perform multiple iterative training until the stopping condition is met to obtain the trained time series prediction sub-model;

[0073] Each iteration of training includes:

[0074] Based on the measured values ​​of bridge structural parameters at multiple detection times during the second historical period, as well as the aging status parameters of the bridge during the second historical period, the predicted values ​​of the current bridge structural parameter thresholds are obtained through the current time-series prediction sub-model.

[0075] The prediction loss is calculated based on the difference between the predicted value of the current bridge structural parameter threshold and the actual value of the bridge structural parameter threshold.

[0076] The weight parameters of the current time series prediction sub-model are adjusted based on the relative magnitude of the prediction loss and the preset loss threshold.

[0077] During the first iteration of training, the initialized time series prediction sub-model is used as the current time series prediction sub-model;

[0078] In other iterations of training, the current time series prediction sub-model obtained after the previous iteration is used as the current time series prediction sub-model;

[0079] The stopping condition is: the predicted loss is less than the preset loss threshold, or the number of iterations reaches the preset number of iterations.

[0080] In this embodiment, for each stage, multiple iterations of training can be performed based on the sample data corresponding to that stage to obtain the time series prediction sub-model corresponding to that stage.

[0081] Specifically, during the first iteration of training, the measured values ​​of bridge structural parameters at multiple detection times within the historical second time period, as well as the aging state parameters of the bridge within the historical second time period, can be input into the initialized time-series prediction sub-model to obtain the predicted value of the current bridge structural parameter threshold. The mean squared error is used as the loss function to calculate the difference between the predicted value of the current bridge structural parameter threshold and the true value of the bridge structural parameter threshold, which is used as the prediction loss. With the goal of reducing the prediction loss, the weight parameters of the current time-series prediction sub-model are adjusted through the backpropagation algorithm.

[0082] During the second iteration of training, the measured values ​​of bridge structural parameters at multiple detection times within the historical second time period, as well as the aging state parameters of the bridge within the historical second time period, can be input into the current time-series prediction sub-model to obtain the predicted values ​​of the current bridge structural parameter thresholds. The mean squared error is used as the loss function to calculate the difference between the predicted value and the true value of the current bridge structural parameter thresholds, which is used as the prediction loss. With the goal of reducing the prediction loss, the weight parameters of the current time-series prediction sub-model are adjusted through the backpropagation algorithm.

[0083] This process is repeated until the predicted loss is less than or equal to a preset loss threshold (e.g., 0.001), or the number of iterations reaches a preset number of iterations, such as 500.

[0084] In one embodiment of this application, the method for determining the true value of the bridge structural parameter threshold within the first historical time period includes:

[0085] Obtain the measured values ​​of bridge structural parameters at multiple detection times within the first historical period under normal conditions;

[0086] Among the measured values ​​of bridge structural parameters at multiple detection times, the measured value of the bridge structural parameter at a specified quantile is taken as the true value of the bridge structural parameter threshold in the first historical time period.

[0087] In this embodiment, a sequence of measured bridge structural parameters without alarms or abnormal maintenance records within the first historical time period can be extracted from the historical monitoring database. This sequence is then sorted from largest to smallest, and the 95th or 99th percentile is selected as the true value for the bridge structural parameter threshold. Taking bridge structural strain parameters as an example, the 95th percentile is taken from the measured strain values ​​at multiple detection times within the first historical time period. This value covers 95% of normal operating condition fluctuations and can serve as a safety boundary for normal samples.

[0088] As can be seen from the above, this embodiment takes into account the normal fluctuation upper limit of the measured values ​​of bridge structural parameters, and takes the measured values ​​of bridge structural parameters at a specified quantile as the true values ​​of the bridge structural parameter thresholds in the first historical time period. This can cover the safety thresholds of most normal working conditions and ensure that the true values ​​match the actual operating status of the bridge.

[0089] In one embodiment of this application, the bridge health detection method further includes:

[0090] For each set of historical sample data, if the set of historical sample data belongs to the healthy period sample data, the early aging sample data, or the mid-aging sample data, the true value of the bridge structural parameter threshold in the first historical period of the set of historical sample data is determined based on the first quantile; if the set of historical sample data belongs to the late aging sample data, the true value of the bridge structural parameter threshold in the first historical period of the set of historical sample data is determined based on the second quantile.

[0091] The first quantile is smaller than the second quantile.

[0092] In this embodiment, considering that the bridge structure performance is stable during the healthy period, early aging stage, and mid-aging stage, and the normal fluctuation dispersion of parameters is low, a lower first quantile (such as the 95th quantile) can be used to accurately anchor its normal safety boundary, which covers most normal operating conditions and will not mask early minor anomalies due to an excessively high threshold. In the later aging stage, the bridge structure performance deteriorates rapidly, the parameter fluctuation dispersion increases, and hidden defects are prone to occur. A higher second quantile (such as the 99th quantile) can be used to improve safety redundancy, filter more extreme fluctuation noise, and ensure that the threshold can be adapted to the high-risk operating conditions of this stage.

[0093] As can be seen from the above, this embodiment fine-tunes the statistical results of the bridge structural parameters in the first historical period based on the stage in which the bridge is located. This fine-tunes the results and uses them as the true values ​​of the bridge structural parameter thresholds for the corresponding stage. This allows the true values ​​of the thresholds to accurately match the structural performance of the corresponding stage, thereby further improving the accuracy of bridge health detection.

[0094] In one embodiment of this application, the bridge health detection method further includes:

[0095] If the rate of change of the predicted value of the bridge structural parameter threshold obtained by two consecutive threshold updates is greater than the preset rate of change threshold, then the predicted value of the bridge structural parameter threshold obtained by the next threshold update is corrected based on the predicted value of the bridge structural parameter threshold obtained by the previous threshold update and the rate of change threshold.

[0096] In this embodiment, considering the gradual nature of bridge aging, the predicted values ​​of the bridge structural parameter thresholds obtained by performing the threshold update step twice consecutively will not change abruptly. Therefore, the predicted values ​​of the bridge structural parameter thresholds obtained by performing the threshold update step each time can be verified.

[0097] Specifically, the absolute value of the difference between the two predicted values ​​obtained from two consecutive threshold updates can be calculated, and then divided by the predicted value obtained from the previous threshold update to obtain the rate of change of the predicted values ​​of the bridge structural parameter thresholds obtained from the two consecutive threshold updates. If the rate of change of the predicted value is greater than a preset rate of change threshold (e.g., 10%), the predicted value of the bridge structural parameter threshold obtained from the previous threshold update is recalculated based on the predicted value of the bridge structural parameter threshold obtained from the previous threshold update and the rate of change threshold. For example, the predicted value of the bridge structural parameter threshold obtained from the subsequent threshold update can be recalculated using the following formula:

[0098] ;

[0099] in, This represents the predicted value of the bridge structural parameter threshold obtained from the last threshold update. This represents the predicted value of the bridge structural parameter threshold obtained from the previous threshold update. This represents the threshold for the rate of change.

[0100] If the rate of change of the predicted value is less than or equal to the preset rate of change threshold, then the predicted value of the bridge structural parameter threshold obtained from the last threshold update will remain unchanged.

[0101] By using the above method, the rate of change of the predicted values ​​of bridge structural parameters obtained by two consecutive threshold updates can be limited to within a preset rate of change threshold. This avoids sudden data fluctuations that could cause abrupt changes in the predicted values ​​of bridge structural parameters, ensuring that the threshold always conforms to the gradual aging pattern of the bridge and reducing the risk of misjudgment or missed judgment caused by sudden threshold changes.

[0102] Based on the same inventive concept, this application also provides a bridge health detection device for implementing the bridge health detection method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more embodiments of the bridge health detection device provided below can be found in the limitations of the bridge health detection method described above, and will not be repeated here.

[0103] This application provides a bridge health detection device, such as... Figure 2 As shown, the bridge health detection device 20 includes: a data acquisition module 21, an anomaly judgment module 22, and a health detection module 23.

[0104] The data acquisition module 21 is used to acquire the average value of the measured values ​​of the bridge structural parameters in the current first time period, the rate of change of the measured values ​​of the bridge structural parameters in the current first time period, the predicted value of the threshold of the bridge structural parameters in the current first time period, and the aging status parameters of the bridge in the current first time period. The aging status parameters include the usage time and the number of historical defects. The predicted value of the threshold of the bridge structural parameters is obtained by prediction based on the trained time series prediction model.

[0105] The anomaly detection module 22 is used to calculate the aging deviation based on the average value of the measured values ​​of the bridge structural parameters and the predicted value of the bridge structural parameter threshold. Based on the aging deviation, the rate of change of the measured values ​​of the bridge structural parameters and the aging state parameters, the anomaly score is obtained through the trained isolated forest model. The trained isolated forest model is trained based on historical data under normal conditions.

[0106] The health detection module 23 is used to determine the health detection result of the bridge based on the relative magnitude of the abnormal score and a preset score threshold.

[0107] In one embodiment of this application, the data acquisition module 21 is specifically used for:

[0108] The threshold update step is performed according to the preset update cycle, and the predicted value of the bridge structural parameter threshold obtained from the most recent threshold update step is used as the predicted value of the bridge structural parameter threshold for the current first time period.

[0109] The threshold update steps include:

[0110] Using the time point of the threshold update step as the reference time point, the measured values ​​of bridge structural parameters and the aging status parameters of the bridge during the second period are obtained at multiple detection times before the reference time point.

[0111] Based on the measured values ​​of bridge structural parameters at multiple detection times during the second time period and the aging state parameters of the bridge during the second time period, the predicted values ​​of the bridge structural parameter thresholds for the first time period after the reference time point are obtained through the trained time series prediction model. The trained time series prediction model is trained based on multiple sets of normal historical sample data. Each set of historical sample data includes the measured values ​​of bridge structural parameters at multiple detection times during the historical second time period under normal conditions, the aging state parameters of the bridge during the historical second time period, and the true values ​​of the bridge structural parameter thresholds during the historical first time period.

[0112] In one embodiment of this application, the data acquisition module 21 is further configured to:

[0113] For each set of historical sample data, the aging stage assessment value of the set of historical sample data is determined based on the aging state parameters in the set of historical sample data.

[0114] Based on the aging stage assessment values ​​corresponding to each set of historical sample data, multiple sets of historical sample data are divided into sample data of multiple stages; the sample data of multiple stages include healthy period sample data, early aging sample data, mid-aging sample data, and late aging sample data.

[0115] Train the corresponding time series prediction sub-model for each stage based on the sample data of each stage;

[0116] The time series prediction sub-models trained in multiple stages are used as the completed time series prediction model.

[0117] Specifically, based on the measured values ​​of bridge structural parameters at multiple detection times within the second time period and the aging state parameters of the bridge within the second time period, the predicted values ​​of the bridge structural parameter thresholds for the first time period after the reference time point are obtained through a trained time-series prediction model, including:

[0118] The aging stage assessment value of the bridge in the second time period is determined based on the aging status parameters of the bridge in the second time period.

[0119] The bridge's stage is determined based on the aging stage assessment values ​​of the bridge during the second time period.

[0120] Select the corresponding time-series prediction sub-model based on the stage in which the bridge is located;

[0121] The predicted values ​​of bridge structural parameter thresholds for the first time period after the reference time point are obtained through the time-series prediction sub-model of the corresponding stage.

[0122] In one embodiment of this application, the data acquisition module 21 is further configured to:

[0123] The first assessment value of aging status is determined based on the usage duration and the preset service life in the aging status parameters.

[0124] The second assessment value of aging status is determined based on the number of historical defects in the aging status parameters;

[0125] The aging stage assessment value is determined based on the first assessment value and the second assessment value.

[0126] In one embodiment of this application, for each stage, the data acquisition module 21 is further configured to:

[0127] Based on the sample data of this stage, perform multiple iterative training until the stopping condition is met to obtain the trained time series prediction sub-model;

[0128] Each iteration of training includes:

[0129] Based on the measured values ​​of bridge structural parameters at multiple detection times during the second historical period, as well as the aging status parameters of the bridge during the second historical period, the predicted values ​​of the current bridge structural parameter thresholds are obtained through the current time-series prediction sub-model.

[0130] The prediction loss is calculated based on the difference between the predicted value of the current bridge structural parameter threshold and the actual value of the bridge structural parameter threshold.

[0131] The weight parameters of the current time series prediction sub-model are adjusted based on the relative magnitude of the prediction loss and the preset loss threshold.

[0132] During the first iteration of training, the initialized time series prediction sub-model is used as the current time series prediction sub-model;

[0133] In other iterations of training, the current time series prediction sub-model obtained after the previous iteration is used as the current time series prediction sub-model;

[0134] The stopping condition is: the predicted loss is less than the preset loss threshold, or the number of iterations reaches the preset number of iterations.

[0135] In one embodiment of this application, the data acquisition module 21 is further configured to:

[0136] Obtain the measured values ​​of bridge structural parameters at multiple detection times within the first historical period under normal conditions;

[0137] Among the measured values ​​of bridge structural parameters at multiple detection times, the measured value of the bridge structural parameter at a specified quantile is taken as the true value of the bridge structural parameter threshold in the first historical time period.

[0138] In one embodiment of this application, the data acquisition module 21 is further configured to:

[0139] If the rate of change of the predicted value of the bridge structural parameter threshold obtained by two consecutive threshold updates is greater than the preset rate of change threshold, then the predicted value of the bridge structural parameter threshold obtained by the next threshold update is corrected based on the predicted value of the bridge structural parameter threshold obtained by the previous threshold update and the rate of change threshold.

[0140] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of each module / unit in the above-described device embodiments, for example... Figure 2 The functions of the data acquisition module 21, the anomaly detection module 22, and the health detection module 23 are shown.

[0141] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0142] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0143] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store preset constants such as preset score thresholds and preset loss thresholds.

[0144] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation method described in the bridge health detection method provided in the embodiments of this application, or they can execute the implementation method of the electronic device described in the embodiments of this application, which will not be repeated here.

[0145] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to complete the process. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or device capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0146] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., provided on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0147] Those skilled in the art will recognize that the modules / units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0148] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0149] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules, units, or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces or modules / units, or it may be an electrical, mechanical, or other form of connection.

[0150] The modules / units described as separate components may or may not be physically separate. Similarly, the components shown as modules / units may or may not be physical modules / units; they may be located in one place or distributed across multiple network modules / units. Some or all of the modules / units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0151] Furthermore, the functional modules / units in the various embodiments of this application can be integrated into one processing module / unit, or each module / unit can exist physically separately, or two or more modules / units can be integrated into one module / unit. The integrated modules / units described above can be implemented in hardware or in the form of software functional modules / units.

[0152] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for bridge health detection, characterized in that, include: The average value of the measured bridge structural parameters in the current first time period, the rate of change of the measured bridge structural parameters in the current first time period, the predicted value of the bridge structural parameter threshold in the current first time period, and the aging status parameters of the bridge in the current first time period are obtained; wherein, the aging status parameters include the service life and the number of historical defects, and the predicted value of the bridge structural parameter threshold is obtained based on the trained time series prediction model. The aging deviation is calculated based on the average value of the measured values ​​of the bridge structural parameters and the predicted value of the bridge structural parameter threshold. Based on the aging deviation, the rate of change of the measured values ​​of the bridge structural parameters, and the aging state parameters, anomaly scores are obtained through a trained isolated forest model. The trained isolated forest model is trained based on historical data under normal conditions. The health detection result of the bridge is determined based on the relative magnitude of the abnormal score and the preset score threshold. The method for determining the predicted values ​​of the bridge structural parameter thresholds for the current first time period includes: The threshold update step is performed according to the preset update cycle, and the predicted value of the bridge structural parameter threshold obtained from the most recent threshold update step is used as the predicted value of the bridge structural parameter threshold for the current first time period. The threshold update step includes: Using the time point at which the threshold update step is executed as a reference time point, the measured values ​​of bridge structural parameters at multiple detection times within the second time period before the reference time point, as well as the aging state parameters of the bridge within the second time period, are obtained. Based on the measured values ​​of bridge structural parameters at multiple detection times during the second time period and the aging state parameters of the bridge during the second time period, the predicted values ​​of the bridge structural parameter thresholds for the first time period after the reference time point are obtained through a trained time-series prediction model. The trained time-series prediction model is trained based on multiple sets of normal historical sample data. Each set of historical sample data includes the measured values ​​of bridge structural parameters at multiple detection times during the historical second time period under normal conditions, the aging state parameters of the bridge during the historical second time period, and the true values ​​of the bridge structural parameter thresholds during the historical first time period.

2. The bridge health detection method as described in claim 1, characterized in that, The training process of the time-series prediction model includes: The aging stage assessment value for each set of historical sample data is determined based on the aging state parameters in each set of historical sample data. Based on the aging stage assessment value corresponding to each set of historical sample data, the multiple sets of historical sample data are divided into sample data of multiple stages; the sample data of multiple stages include healthy period sample data, early aging sample data, mid-aging sample data, and late aging sample data. Train the corresponding time series prediction sub-model for each stage based on the sample data of each stage; The time series prediction sub-models trained in multiple stages are used as the completed time series prediction model. Specifically, based on the measured values ​​of bridge structural parameters at multiple detection times within the second time period and the aging state parameters of the bridge within the second time period, the predicted values ​​of the bridge structural parameter thresholds for the first time period after the reference time point are obtained through a trained time-series prediction model, including: The aging stage assessment value of the bridge in the second time period is determined based on the aging status parameters of the bridge in the second time period. The bridge's stage is determined based on the bridge's aging stage assessment value during the second time period. Select the corresponding time-series prediction sub-model based on the stage in which the bridge is located; The predicted values ​​of bridge structural parameter thresholds for the first time period after the reference time point are obtained through the time-series prediction sub-model of the corresponding stage.

3. The bridge health detection method as described in claim 2, characterized in that, The step of determining the aging stage assessment value based on the aging state parameters includes: The first assessment value of aging status is determined based on the usage duration and the preset service life in the aging status parameters. The second assessment value of aging status is determined based on the number of historical defects in the aging status parameters; The aging stage assessment value is determined based on the first assessment value and the second assessment value.

4. The bridge health detection method as described in claim 2, characterized in that, For each stage, a time-series prediction sub-model corresponding to that stage is trained based on the sample data of that stage, including: Based on the sample data of this stage, perform multiple iterative training until the stopping condition is met to obtain the trained time series prediction sub-model; Each iteration of training includes: Based on the measured values ​​of bridge structural parameters at multiple detection times during the second historical period and the aging status parameters of the bridge during the second historical period, the predicted values ​​of the current bridge structural parameter thresholds are obtained through the current time-series prediction sub-model. The prediction loss is calculated based on the difference between the predicted value of the current bridge structural parameter threshold and the actual value of the bridge structural parameter threshold. Based on the relative magnitude of the predicted loss and the preset loss threshold, the weight parameters of the current time series prediction sub-model are adjusted. During the first iteration of training, the initialized temporal prediction sub-model is used as the current temporal prediction sub-model; In other iterations of training, the current time-series prediction sub-model obtained after the previous iteration is used as the current time-series prediction sub-model; The stopping condition is: the predicted loss is less than a preset loss threshold, or the number of iterations reaches a preset number of iterations.

5. The bridge health detection method as described in claim 1, characterized in that, The method for determining the true values ​​of the bridge structural parameter thresholds within the first historical time period includes: Obtain the measured values ​​of bridge structural parameters at multiple detection times within the first historical period under normal conditions; Among the measured values ​​of bridge structural parameters at the multiple detection times, the measured value of the bridge structural parameters at a specified quantile is taken as the true value of the bridge structural parameter threshold in the first historical time period.

6. The bridge health detection method as described in claim 1, characterized in that, Also includes: If the rate of change of the predicted value of the bridge structural parameter threshold obtained by two consecutive threshold updates is greater than the preset rate of change threshold, then the predicted value of the bridge structural parameter threshold obtained by the previous threshold update and the rate of change threshold are used to correct the predicted value of the bridge structural parameter threshold obtained by the next threshold update.

7. A bridge health monitoring device, characterized in that, include: The data acquisition module is used to acquire the average value of the measured values ​​of the bridge structural parameters in the current first time period, the rate of change of the measured values ​​of the bridge structural parameters in the current first time period, the predicted value of the threshold of the bridge structural parameters in the current first time period, and the aging status parameters of the bridge in the current first time period; wherein, the aging status parameters include the usage time and the number of historical defects, and the predicted value of the bridge structural parameter threshold is obtained by prediction based on the trained time series prediction model. The anomaly detection module is used to calculate the aging deviation based on the average value of the measured values ​​of the bridge structural parameters and the predicted value of the bridge structural parameter threshold. Based on the aging deviation, the rate of change of the measured values ​​of the bridge structural parameters, and the aging state parameters, an anomaly score is obtained through a trained isolated forest model. The trained isolated forest model is trained based on historical data under normal conditions. The health detection module is used to determine the health detection result of the bridge based on the relative magnitude of the abnormal score and a preset score threshold. The data acquisition module is specifically used for: The threshold update step is performed according to the preset update cycle, and the predicted value of the bridge structural parameter threshold obtained from the most recent threshold update step is used as the predicted value of the bridge structural parameter threshold for the current first time period. The threshold update step includes: Using the time point at which the threshold update step is executed as a reference time point, the measured values ​​of bridge structural parameters at multiple detection times within the second time period before the reference time point, as well as the aging state parameters of the bridge within the second time period, are obtained. Based on the measured values ​​of bridge structural parameters at multiple detection times during the second time period and the aging state parameters of the bridge during the second time period, the predicted values ​​of the bridge structural parameter thresholds for the first time period after the reference time point are obtained through a trained time-series prediction model. The trained time-series prediction model is trained based on multiple sets of normal historical sample data. Each set of historical sample data includes the measured values ​​of bridge structural parameters at multiple detection times during the historical second time period under normal conditions, the aging state parameters of the bridge during the historical second time period, and the true values ​​of the bridge structural parameter thresholds during the historical first time period.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 6.

Images