Bridge health monitoring system calibration method and device based on sub-region loading test and storage medium

Abstract

The application provides a bridge health monitoring system calibration method and device based on a sub-region loading test and a storage medium, wherein the calibration method comprises the following steps: performing a sub-region loading test on a target real bridge based on an influence line, and obtaining monitoring data; calling and performing loading on a bridge benchmark state finite element model of the target real bridge at the same position and with the same load as the sub-region loading test on the real bridge, and obtaining corresponding structure response theoretical data; obtaining a first fitting relationship and a second fitting relationship based on a monitoring data sequence and a structure response theoretical data sequence; determining a target fitting relationship in the first fitting relationship and the second fitting relationship based on a first priority feature obtained from the monitoring data sequence and a preset standard threshold value, determining a fitting model corresponding to the target fitting relationship as a calibration model of the bridge health monitoring system, and calibrating the monitoring data by using the calibration model. The calibration method provided by the application is more convenient to implement and can improve the calibration effect of the monitoring data.

Description

Technical Field

[0001] This invention relates to the field of bridge monitoring technology, and in particular to a calibration method, device and storage medium for a bridge health monitoring system based on regional loading tests. Background Technology

[0002] Bridge health monitoring systems are an important technical means to monitor the structural condition of bridges in real time. Essentially, these systems aim to sense and acquire data on the bridge's structural response under external loads. By comparing and analyzing the monitored data with theoretically calculated data of the bridge's structural response under baseline conditions, the system can be used to assess the bridge's performance.

[0003] Among them, the bridge health monitoring system (also known as the bridge structure monitoring system or simply the bridge monitoring system) is an electronic information system that connects various sensors, data acquisition and transmission, data processing and management, data analysis and application hardware devices, software modules and supporting facilities distributed at the bridge site and monitoring center through network integration technology. It has the functions of continuous monitoring, automatic recording, data display and alarm assessment of bridge set parameters, and assists in bridge management and maintenance decision-making.

[0004] However, upon completion of the bridge structure monitoring system installation, the monitoring data may exhibit systematic biases, necessitating calibration of the initial states of the monitoring sensors for each structural response. While bridge structure monitoring sensors may be free of quality issues at the time of manufacture, their performance can be compromised after installation due to factors such as harsh environments, network anomalies, electromagnetic interference, and installation errors. This can result in initial errors in the monitoring data even upon sensor installation completion. For instance, the intended placement orientation of a resistance strain gauge should align with the principal strain direction on the measured component (assuming the measuring point is under one-dimensional stress). However, if the actual placement orientation deviates from the principal strain direction by an angle α, such as... Figure 1 As shown, there is an error between the measured strain value ε' and the true strain value ε, but it can be corrected using formula (1).

[0005]

[0006] In the formula, μ is the Poisson's coefficient of the measured component material, and a is the deviation angle.

[0007] Without calibration, subsequent monitoring values ​​from the sensor will remain erroneous. The sensor monitoring data displayed in the bridge health monitoring system (i.e., the obtained monitoring value) u can be expressed by formula (2). Data containing errors is difficult to reflect the true condition of the bridge structure and will also lead to inaccuracies in the results of various analysis and evaluation algorithms based on the monitoring data. Therefore, it is essential to calibrate the bridge monitoring data using theoretical calculation data of the structural response under the bridge's reference state.

[0008] u=u0+σ0+σ 其他 Formula (2)

[0009] In the formula, u0 represents the true value of the bridge structure response, σ0 represents the monitoring error caused by various factors when the sensor is installed, and σ 其他 This indicates the monitoring error caused by various factors during long-term use after the sensor is installed.

[0010] In the process of realizing this invention, the inventors discovered at least the following problems in the prior art:

[0011] There are two main categories of existing monitoring data calibration methods: those based on long-term monitoring data and those based on measured data.

[0012] (1) The main methods for repairing monitoring data calibration based on long-term monitoring data include data interpolation, data fitting and machine learning methods.

[0013] ① The numerical interpolation method uses existing normal data from the same bridge monitoring sensor at different time periods and employs interpolation algorithms (such as linear interpolation and nearest neighbor interpolation) to calibrate abnormal data (such as supplementing data for missing time periods).

[0014] ② The data fitting method is based on existing normal data from the same bridge monitoring sensor at different time periods. A data fitting algorithm (such as polynomial fitting) is used to establish a fitting curve for the sensor's monitoring data. This fitting curve is then used to calibrate abnormal data (e.g., using polynomial fitted values ​​to replace abnormal data values). Another literature, "Research on Bridge Health Monitoring System Based on Load Test Correction," proposes establishing a fitting curve between manual detection data from traditional load tests and monitoring data from the health monitoring system to calibrate the monitoring data.

[0015] ③ The machine learning method is to train a machine learning model (including deep neural networks and long short-term memory networks) on normal data from a bridge monitoring sensor at different time periods or on normal monitoring data from other sensors that are related to a certain sensor, and then use the machine learning model to calibrate abnormal data.

[0016] Existing calibration methods for bridge anomaly monitoring data, based on data interpolation, data fitting, and machine learning, all rely on sensor monitoring data from the bridge structure monitoring system itself. Bridge monitoring data itself may contain errors; for example, initial errors may exist when sensors are first installed. These existing methods essentially perform anomaly calibration based on monitoring data that may contain initial errors. In short, these methods struggle to obtain truly accurate monitoring data.

[0017] (2) Monitoring data calibration method based on measured data. The literature "Research on Bridge Health Monitoring System Based on Load Test Correction" proposes to establish a fitting curve (mainly a linear fitting function) between the manual detection data of traditional load test and the monitoring data of the health monitoring system to calibrate the monitoring data.

[0018] This method is primarily applicable to structural response monitoring data. Traditional load tests, used to assess bridge load-bearing capacity, require numerous vehicle loads, resulting in high costs. Furthermore, the load arrangement is relatively fixed, and structural responses at certain locations may be insignificant. Manually measured data may also contain errors. The established calibration function only involves a fitting algorithm, the accuracy of which may be limited. In conclusion, the applicability of this method needs further improvement.

[0019] Therefore, there is a need for a calibration method, device, and storage medium for a bridge health monitoring system based on regional loading tests, in order to at least partially solve the above-mentioned technical problems. Summary of the Invention

[0020] In view of this, embodiments of the present invention provide a calibration method, device and storage medium for a bridge health monitoring system based on regional loading tests, so as to at least solve one of the problems in the prior art.

[0021] In a first aspect, embodiments of the present invention provide a calibration method for a bridge health monitoring system based on a regional loading test, the calibration method comprising:

[0022] Based on the influence line corresponding to the pre-determined structural response to be calibrated, an actual loading test is conducted on the target bridge at the maximum longitudinal index of the influence line or its vicinity. Structural response monitoring data is obtained through the bridge health monitoring system, and a monitoring data sequence is constructed.

[0023] The finite element model of the target bridge's reference state is invoked and loaded with the same load at the same location as the actual load test of the target bridge. The corresponding structural response theoretical data is obtained from the finite element analysis results, and a structural response theoretical data sequence is constructed.

[0024] Based on the monitoring data sequence and the structural response theoretical data sequence, a first fitting model that meets the preset R-squared value requirement is used for fitting to obtain a first fitting relationship;

[0025] Based on the monitoring data sequence and the structural response theoretical data sequence, a second fitting model that meets the preset R-squared value requirement is used for fitting to obtain a second fitting relationship;

[0026] Based on the monitoring data sequence and the structural response theoretical data sequence, obtain the first priority feature representing the quantity of data;

[0027] Based on the first priority feature and the preset standard threshold, a target fitting relationship is determined between the first fitting relationship and the second fitting relationship; if the target fitting relationship is determined, the fitting model corresponding to the target fitting relationship is determined as the calibration model of the bridge health monitoring system.

[0028] The calibration model is used to calibrate the monitoring data of the bridge health monitoring system.

[0029] Secondly, embodiments of the present invention also provide a calibration device for a bridge health monitoring system based on a regional loading test, the calibration device comprising:

[0030] Memory is used to store executable instructions for a computer;

[0031] A processor, used to execute computer-executable instructions stored in the memory, implements the calibration method of the above-described technical solution.

[0032] Thirdly, embodiments of the present invention also provide a storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the calibration method of the above-described technical solution.

[0033] According to the calibration method of the present invention, this method is not based on monitoring data that may contain errors, which can improve the calibration effect of monitoring data. It also eliminates the need to manually measure the physical quantities of sensor positions during load tests, thus improving the convenience of the method. In addition, the load test is performed according to the influence line, requiring less vehicle load, which is simpler and more effective than load tests. Furthermore, it can selectively acquire sensor monitoring data and theoretical data calculated by finite element model for different areas of the bridge, enabling regional calibration of sensors.

[0034] Additional advantages, objects, and features of the invention will be set forth in part in the description which follows, and will also become apparent in part to those skilled in the art upon studying the description, or may be learned by practice of the invention. The objects and other advantages of the invention can be realized and obtained by means of the structures specifically pointed out in the description and drawings.

[0035] Those skilled in the art will understand that the objectives and advantages achievable with the present invention are not limited to those specifically described above, and that the above and other objectives achievable with the present invention will become clearer from the following detailed description. Attached Figure Description

[0036] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, are not intended to limit the scope of the invention. The components in the drawings are not drawn to scale but are merely illustrative of the principles of the invention. For ease of illustration and description of certain parts of the invention, corresponding portions in the drawings may be enlarged, i.e., may appear larger relative to other components in an exemplary device actually manufactured according to the invention. In the drawings:

[0037] Figure 1 A schematic diagram of the orientation deviation of resistance strain gauges in the prior art;

[0038] Figure 2 A flowchart of a bridge health monitoring system calibration method according to an embodiment of the present invention;

[0039] Figure 3 A flowchart of a bridge health monitoring system calibration method according to another embodiment of the present invention;

[0040] Figure 4 This is a schematic block diagram of a bridge health monitoring system calibration method according to an embodiment of the present invention;

[0041] Figure 5 This is a schematic diagram of the bending moment diagram and the bending moment influence line of a simply supported beam bridge in the bridge health monitoring system calibration method according to an embodiment of the present invention.

[0042] Figure 6 This is a schematic diagram of the bending moment diagram and bending moment influence line of another simply supported beam bridge in the calibration method of the bridge health monitoring system according to an embodiment of the present invention;

[0043] Figure 7 This is a schematic diagram of a calibration apparatus according to an embodiment of the present invention;

[0044] Figure 8 This is a schematic diagram of a calibration system according to an embodiment of the present invention. Detailed Implementation

[0045] 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 embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0046] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0047] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0048] It should also be noted that, unless otherwise specified, the term "connection" in this article can refer not only to a direct connection, but also to an indirect connection involving an intermediary.

[0049] In the following description, embodiments of the invention will be illustrated with reference to the accompanying drawings. In the drawings, the same reference numerals represent the same or similar parts, or the same or similar steps.

[0050] First, refer to Figure 2 This application describes a calibration method 100 for a bridge health monitoring system based on a regional loading test, according to an embodiment of this application. For example... Figure 2 As shown, calibration method 100 may include steps S110 to S170, as detailed below:

[0051] In step S110, based on the influence line corresponding to the pre-determined structural response to be calibrated, a measured loading test is conducted on the target bridge at the maximum longitudinal index of the influence line or its vicinity. Structural response monitoring data is obtained through the bridge health monitoring system, and a monitoring data sequence is constructed.

[0052] In step S120, the finite element model of the bridge reference state of the target bridge is called and loaded with the same load at the same location as the actual load test of the target bridge. The corresponding structural response theoretical data is obtained from the finite element analysis results, and a structural response theoretical data sequence is constructed.

[0053] In step S130, based on the monitoring data sequence and the structural response theoretical data sequence, a first fitting model that meets the preset R-squared value requirement is used to fit the data and obtain the first fitting relationship.

[0054] In step S140, based on the monitoring data sequence and the structural response theoretical data sequence, a second fitting model that meets the preset R-squared value requirement is used to fit the data and obtain a second fitting relationship.

[0055] In step S150, a first priority feature representing the amount of data is obtained based on the monitoring data sequence and the structural response theoretical data sequence.

[0056] In step S160, based on the first priority feature and the preset standard threshold, a target fitting relationship is determined among the first fitting relationship and the second fitting relationship; if the target fitting relationship is determined, the fitting model corresponding to the target fitting relationship is determined as the calibration model of the bridge health monitoring system.

[0057] In step S170, the monitoring data of the bridge health monitoring system are calibrated using the calibration model.

[0058] In the embodiments of this application, firstly, a sustained loading test of vehicle loads is conducted on the actual bridge along the influence line. The bridge health monitoring system's sensors acquire monitoring data of the structural response. Simultaneously, based on the established finite element model of the bridge structure in its baseline state, the same load is applied at the corresponding position of the finite element model to calculate the theoretical calculation data of the structural response. Next, a fitting relationship (calibration function) between the monitoring data and the finite element theoretical calculation data is established using either a first fitting model or a second fitting model. Then, based on the first priority feature and the preset standard threshold, the first fitting relationship or the second fitting relationship is determined as the target fitting relationship, thereby determining the calibration model. Finally, the calibration model is used to calibrate the bridge monitoring data.

[0059] As can be seen from the above description, the calibration method 100 according to the embodiment of this application is not based on monitoring data that may contain errors, but on the theoretical calculation values ​​through the finite element model, which can improve the calibration effect of monitoring data. At the same time, the loading test is carried out according to the influence line, which requires less load. The bridge health monitoring system can obtain larger and more stable structural response monitoring data, which is simpler and more effective than the load test. In addition, depending on the situation of the monitoring data sequence and the theoretical structural response data sequence, the first fitting model or the second fitting model can be used for calibration on the basis of ensuring the calibration accuracy, which has high flexibility and applicability.

[0060] Among them, Figure 2 Steps S110 to S170 are shown to be performed sequentially, but this is only an example. It is understood that the order of steps S130, S140, and S160 is not restricted.

[0061] See Figure 3 The following will describe in detail the steps of the calibration method 200 according to the embodiments of this application.

[0062] In the embodiments of this application, in step S201, based on the influence line corresponding to the pre-determined structural response to be calibrated, a measured loading test is conducted on the target bridge at the position of the maximum longitudinal marker of the influence line or its vicinity. Structural response monitoring data is obtained through the bridge health monitoring system, and a monitoring data sequence is constructed.

[0063] It should be noted that the reference Figure 4 In this embodiment, the structural response to be calibrated may include structural displacement, strain, cable force, rotation angle, and vibration acceleration, such as the strain and vertical displacement of the main beam, and there are no specific limitations. Furthermore, the calibration method of this embodiment is applicable to the calibration and correction of various types of monitoring data for various bridge structures (not limited to structural response monitoring data, but also including environmental and load monitoring data), and has a wide range of applications, and can also be applied to various structural and architectural engineering projects.

[0064] Before proceeding to step S201, it is necessary to pre-determine the influence line corresponding to the structural response to be calibrated. Generally, for a specific structural response monitoring sensor that needs to be calibrated, the influence line of the internal forces at the sensor location (such as the bending moment influence line) can be established using a static or kinematic method. For the calibration of structural response monitoring data, this invention considers using a simple loading test based on the influence line, rather than necessarily a load test, to assist in the manual measurement of structural response data. The loading test is simple and efficient.

[0065] Specifically, on the target bridge, for example, vehicle loads or other loads can be applied at or near the point of maximum longitudinal elevation along the influence line, allowing the bridge health monitoring system to acquire large and stable structural response monitoring data through monitoring sensors corresponding to the structural response. This data can then be used to construct a monitoring data sequence.

[0066] For example, refer to Figure 5 ,for Figure 5 In the simply supported beam bridge, displacement sensors ①②③④⑤⑥ in the bridge health monitoring system are arranged in regions C and E, respectively, to monitor the deflection at different locations on the main beam. When a load (such as a vehicle load) is applied to section c, the main beam located at sensors ①②③ will undergo a significant displacement, and the bridge health monitoring system can acquire the monitoring data from sensors ①②③. Simultaneously, in the bridge finite element model, the displacement calculation values ​​at sensors ①②③ under the corresponding loading mode can be extracted. Similarly, when a load (such as a vehicle load) is applied to section e, the bridge health monitoring system can acquire the monitoring data from sensors ④⑤⑥, and the displacement calculation values ​​at sensors ④⑤⑥ can be extracted from the bridge finite element analysis results.

[0067] In the embodiments of this application, in step S203, the finite element model of the target bridge structure is called and loaded with the same load at the same location as the actual load test of the target bridge. The corresponding structural response theoretical data is obtained from the finite element analysis results, and a structural response theoretical data sequence is constructed.

[0068] Of course, before proceeding to step S203, it is necessary to first construct a finite element model of the target bridge's reference state.

[0069] Regarding the selection of the bridge reference state, when calculating the theoretical value of the structural response, the bridge reference state is best chosen to be the state before the bridge is opened to traffic. After long-term operation, bridges may suffer various damages, such as concrete cracking and steel corrosion. If the bridge state during long-term operation is selected as the reference state for finite element modeling calculations, it is difficult to realistically simulate various damages and obtain a true structural response. For newly built bridges, the bridge performance condition is intact, and the various performance parameters of the bridge are basically consistent with the design state. For a small number of performance parameters that are inconsistent with the design due to construction, they can be obtained through actual measurements, thereby establishing a finite element model consistent with the actual performance condition of the bridge. Generally, monitoring systems are gradually implemented during bridge construction, and the monitoring systems also need to be handed over and accepted during bridge handover and acceptance. In addition, for extra-large and complex bridges, load tests are also conducted during handover and acceptance to verify the bridge's load-bearing capacity. Therefore, when calibrating the bridge health monitoring system, a finite element model of the bridge baseline state is established based on the bridge's handover and acceptance as the baseline state to calculate the theoretical value of the structural response. That is, the bridge finite element model is established based on the bridge's state before it is opened to traffic. This finite element model can more accurately reflect the bridge's true performance status compared to the bridge finite element model during the operation period.

[0070] Specifically, after determining the finite element model of the target bridge's reference state, the same load was applied to the finite element model at the same location as the actual bridge loading test. The corresponding structural response theoretical data was then extracted from the finite element analysis results. For details, please refer to [link to relevant documentation]. Figure 5 And the relevant descriptions above.

[0071] In the embodiments of this application, in step S205, based on the monitoring data sequence and the structural response theoretical data sequence, a first fitting model that meets the preset R-squared value requirement is used to fit the data and obtain a first fitting relationship.

[0072] The first fitting model uses a regression model (such as a linear regression model, a multinomial regression model, etc.).

[0073] Specifically, after the sensors of the bridge health monitoring system obtain the monitoring data sequence X = (x1, x2, ..., xn) and the finite element theory calculation value sequence Y = (y1, y2, ..., yn) (i.e., the structural response theory data sequence), a regression model (such as a linear regression model, a multinomial regression model, etc.) is used to determine the functional relationship f between X and Y, such as y = f(x).

[0074] Using mathematical analysis software, multiple fitting functions (known functions), including polynomial functions, linear functions, and logarithmic functions, are used to fit the acquired structural response monitoring data and structural response theoretical data one by one, and the R-squared value used to measure the degree of fit of the model to the data is calculated.

[0075] When the R-squared value reaches a preset value, such as 0.95, the fitting function at this time is determined as the first fitting model that meets the requirements, and the fitting function at this time is the first fitting relationship.

[0076] The R-squared value, which measures the goodness of fit of the model to the data, is calculated using formula (3).

[0077]

[0078] In the formula, y i Let p represent the theoretical data of the i-th structural response, m represent the average value of the n theoretical structural response data, and p represent the theoretical data of the i-th structural response. i This represents the fitted value corresponding to the i-th structural response monitoring data calculated according to a certain fitting function.

[0079] For example, refer to Figure 6 Assume a simply supported beam bridge is equipped with a bridge health monitoring system. Sensors are installed at three locations (sections a, b, and c) to monitor the vertical displacement of the main beam. The sensor arrangement is as follows: Figure 6 As shown, Figure 6The influence lines of bending moments at the three cross-section locations calculated using the static method are also presented. From the Ma bending moment influence lines, it can be seen that when load F is applied at location ③ and its vicinity, a large vertical displacement will occur at the main beam cross-section a, while when load F is applied at location ⑨, the vertical displacement at the main beam cross-section a is smaller. Therefore, to calibrate the displacement sensor at cross-section a, load F can be applied at locations ① to ⑥ respectively to obtain six large and stable vertical displacement monitoring values. Simultaneously, in the finite element model, loads of the same weight were applied to six locations identical to the actual bridge loading locations (each loading location represents a working condition), and theoretical calculation values ​​of the vertical displacement at cross-section a under the six working conditions were extracted from the finite element analysis results. Regression analysis was performed on these six theoretical calculation values ​​and six monitoring data (see Table 1). When using a polynomial function to fit the relationship between the theoretical calculation data and the monitoring data, the R-squared value was 0.9806, and the corresponding polynomial function is as follows:

[0080] y = 0.6719x 2 -3.5672x + 7.8275 Formula (4)

[0081] Therefore, the function represented by formula (4) can be used as the calibration function (i.e., the first fitting relationship) for calibrating the displacement monitoring data at section a of the main beam. Then, using this calibration function, the displacement monitoring data at section a can be corrected. For example, if the displacement monitoring value at section a is 4.1, it can be corrected to 4.5 (with one significant decimal place) using the correction function.

[0082] Table 1. Theoretical calculation data and monitoring data of vertical displacement at section a (unit: mm)

[0083] Theoretical calculation data Y 3.4 3.5 4.2 3.9 3.8 3.2 Monitoring data X 3.3 3.4 3.9 3.8 3.7 3.1

[0084] It should be noted that when the load is applied at positions ⑦, ⑧, and ⑨, the displacement generated at section a is relatively small. Therefore, the vertical displacement value at section a calculated by the finite element model when the load is applied at positions ⑦, ⑧, and ⑨ was not used to calibrate and correct the displacement sensor monitoring data at section a. However, the vertical displacement value at section c calculated by the finite element model when the load is applied at positions ⑦, ⑧, and ⑨ can be used to calibrate and correct the displacement sensor monitoring data at section c.

[0085] In the embodiments of this application, in step S207, a second fitting model that meets the preset R-squared value requirement is used to fit the data based on the monitoring data sequence and the structural response theoretical data sequence to obtain a second fitting relationship.

[0086] The second fitting model uses a neural network model (such as a deep neural network, a long short-term memory network, etc.).

[0087] Similarly, after the sensors of the bridge health monitoring system obtain the monitoring data sequence X = (x1, x2, ..., xn) and the finite element theory calculation value sequence Y = (y1, y2, ..., yn) (i.e., the structural response theory data sequence), a neural network model (such as a deep neural network, a long short-term memory network, etc.) is used to determine the functional relationship f between X and Y.

[0088] Specifically, first, it is necessary to obtain a neural network model that meets the preset requirement for the R-squared value. The specific process is as follows:

[0089] The acquired monitoring data sequences and structural response theory data sequences are combined into a dataset, which is then divided into mutually exclusive training and validation sets in a certain ratio, for example, an 8:2 ratio. The monitoring data sequence X serves as the input layer, and the corresponding structural response theory data sequence Y serves as the output layer.

[0090] Then, a common and well-suited neural network model is selected, and the number of network layers, neurons, learning rate, and loss function parameters of the second fitting model are preset. The model is trained using the training set, and the R-squared value is calculated as an indicator to evaluate the performance of the model. The calculation method of the R-squared value can refer to formula (3) above, and will not be repeated here.

[0091] If the R-squared value reaches the preset value, then the second fitting model at this time is determined as the second fitting model that meets the requirements. The relationship between the monitoring data sequence and the structural response theoretical data sequence determined by the second fitting model at this time is the second fitting relationship.

[0092] If the R-squared value does not reach the preset value, continue to adjust the model parameters and iterate the above steps until the R-squared value reaches the preset value.

[0093] It is understandable that if the initially selected neural network model fails to meet the requirements, it can be replaced with another known neural network model that is more suitable, and the same steps described above can be used to process it until the R-squared value reaches the preset value.

[0094] For example, for Figure 6A simply supported beam bridge equipped with a health monitoring system was used. Assuming a vehicle slowly moves and loads the bridge at and near section a, 500 theoretical and monitoring data points of vertical displacement at section a were obtained. This dataset was divided into mutually exclusive training and validation sets in an 8:2 ratio. A deep neural network with two hidden layers and seven neurons per layer, using the sigmoid activation function, was trained on the training set and its hyperparameters were fine-tuned on the validation set, resulting in a deep learning model with an R-squared value of 0.98. This deep learning model can then be used to correct the displacement monitoring data at section a. For example, if the displacement monitoring value at section a is 4.2, its corrected value is 4.6 (rounded to one decimal place).

[0095] In an embodiment of this application, step S209 involves obtaining a first priority feature representing the quantity of data based on the monitoring data sequence and the structural response theoretical data sequence.

[0096] Specifically, the first priority feature refers to the total number of data points in the monitoring data sequence and the structural response theoretical data sequence. For example, if both the monitoring data sequence and the structural response theoretical data sequence contain 500 data points, then the first priority feature is 1000.

[0097] In the embodiments of this application, step S211 compares the first priority feature with a preset standard threshold to determine whether the first priority feature has reached the preset standard threshold. For example, if there are 800 first priority features and the preset standard threshold is set to 1000, then the first priority feature has not reached the preset standard threshold.

[0098] In the embodiments of this application, in step S213, if the first priority feature bit does not reach the preset standard threshold, the first fitting relationship is determined as the target fitting relationship; if the target fitting relationship is determined, the first fitting model corresponding to the first fitting relationship is determined as the calibration model of the bridge health monitoring system.

[0099] In the embodiments of this application, in step S215, when the monitoring data obtained by the bridge health monitoring system is calibrated using the first fitting model, a second priority feature representing the fitting error rate corresponding to the fitting result is obtained.

[0100] Specifically, once the first fitting relationship is determined as the target fitting relationship and the first fitting model corresponding to the first fitting relationship is determined as the calibration model, the monitoring data acquired by the bridge health monitoring system is calibrated using the first fitting model. For example, by obtaining the fitting result after calibrating the monitoring data acquired by the bridge health monitoring system using the first fitting model within a set time period, the corresponding second priority feature characterizing the fitting error rate can be obtained based on the fitting result, and the fitting error value within that time period can be obtained based on the second priority feature.

[0101] In the embodiments of this application, step S217 compares the fitting error value obtained based on the second priority feature with a preset error threshold to determine whether the fitting error value obtained based on the second priority feature reaches the preset error threshold. For example, if the second priority feature (fitting error rate) obtained within a certain time period is 15%, then the fitting error value obtained based on the second priority feature is 0.15, while the preset error threshold is set to 0.05. Therefore, the fitting error value obtained based on the second priority feature is greater than the preset error threshold. If the fitting error value obtained based on the second priority feature is less than the preset error threshold, then the first fitting model is still determined as the calibration model of the bridge health monitoring system.

[0102] In the embodiments of this application, in step S219, when the first priority feature bit reaches a preset standard threshold, the second fitting relationship is determined as the target fitting relationship; and when the fitting error value obtained based on the second priority feature is greater than the preset error threshold, the second fitting relationship is determined as the target fitting relationship. When the target fitting relationship is determined, the second fitting model corresponding to the second fitting relationship is determined as the calibration model of the bridge health monitoring system.

[0103] In the embodiments of this application, step S221 uses the calibration model to calibrate the monitoring data of the bridge health monitoring system. That is, after determining the first fitting model or the second fitting model in steps S217 and S219 as the calibration model for the bridge health monitoring system, the subsequent monitoring data of the bridge health monitoring system are calibrated using the first fitting model or the second fitting model.

[0104] For example, based on a predetermined calibration function, the monitoring data x0 of one (or more) sensors to be calibrated is obtained, and the calibration function y = f(x) is used to calculate y0 = f(x0), which is then used to replace x0. If the calibration function is established using a neural network model, the x0 is input into the trained neural network model, and the neural network model can output the calibration value y0 of the monitoring value x0.

[0105] Based on the above description, the calibration method according to the embodiments of this application establishes a calibration function relationship between the theoretically calculated data and the monitoring data based on the finite element analysis theoretical calculation data from the theoretical loading test and the monitoring data from the bridge health monitoring system. This calibration method uses the theoretically calculated values ​​from the finite element analysis rather than the monitoring data itself. Compared to calibrating the monitoring data based on monitoring data that may contain initial errors, this method can more accurately correct the monitoring data. Furthermore, the embodiments of this invention select the bridge's state before it is opened to traffic as the baseline state to establish the bridge finite element model. This finite element model can more accurately reflect the bridge's true performance state compared to the bridge finite element model during the operation period. Furthermore, the loading test is conducted along the influence line, requiring less load. The bridge health monitoring system can acquire larger and more stable structural response monitoring data, which is simpler and more effective than the load test. It also eliminates the need for manual measurement of sensor positions during the load test, improving the ease of implementation. Additionally, it allows for targeted acquisition of sensor monitoring data and theoretical data calculated by the finite element model for different regions, enabling regional sensor calibration. Moreover, based on the monitoring data sequence and the theoretical structural response data sequence, and while ensuring calibration accuracy, it allows for calibration using either the first or second fitting model, offering greater flexibility and applicability.

[0106] Furthermore, the calibration method in this embodiment of the invention differs from that in "Research on Bridge Health Monitoring System Based on Load Test Correction" in that:

[0107] ① Regarding the loading method, this invention directly loads a vehicle load of a fixed weight G1 onto bridge location c along the influence line, and continues loading for a period of time to allow the structural response monitoring sensors deployed near location c to collect large and stable monitoring values. Then, the vehicle load is slowly moved to bridge location e, and loading is continued for a period of time to allow the structural response monitoring sensors deployed near location e to collect large and stable monitoring values. The method in "Research on Bridge Health Monitoring System Based on Load Test Correction" is based on a traditional load test implemented to evaluate the load-bearing capacity of a bridge. This load test involves progressively loading a vehicle load to a fixed weight G2 (G2>G1) at a fixed location on the bridge. The embodiment of this invention requires a lower vehicle load and has a lower implementation cost.

[0108] ② Regarding calibration data, this invention uses structural response data calculated by finite element analysis of the bridge structure during loading tests to calibrate the monitoring data. The method in "Research on Bridge Health Monitoring System Based on Load Test Correction" uses manually measured structural response data from load tests to calibrate the monitoring data.

[0109] ③ Regarding the calibration function, this invention proposes using statistical regression and neural networks to establish a correction function between the monitoring data and the finite element theory calculation data. The method in "Research on Bridge Health Monitoring System Based on Load Test Correction" only uses regression fitting to establish the correction function. Regression fitting methods sometimes fail to achieve satisfactory results. The neural network model used in this invention can theoretically fit any functional relationship; therefore, this invention has a wider range of applications and stronger applicability.

[0110] refer to Figure 7 A calibration apparatus 300 for implementing the calibration method according to an embodiment of this application includes a processor 310 and a memory 320. The calibration apparatus 300 may include one or more processors 310 and one or more memories 320. The memory 320 stores an executable program that is run by the processor 310. When the executable program is run by the processor 310, it causes the processor 310 to execute the calibration method 200 described above according to an embodiment of this application.

[0111] The processor 310 may be a central processing unit (CPU) or other processing units with data processing capabilities and / or instruction execution capabilities.

[0112] The memory 320 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 310 may execute the program instructions to implement the client functions (implemented by the processor) in the embodiments of this application described herein, and / or other desired functions. Various applications and various data may also be stored in the computer-readable storage medium, such as various data used and / or generated by the applications.

[0113] The calibration device 300 may also include input and output devices, which are interconnected via a bus system and / or other forms of connection mechanisms. It should be noted that... Figure 7 The components and structure of the calibration device 300 shown are merely exemplary and not limiting; the calibration device 300 may also have other components and structures as needed.

[0114] The input device can be a device used by a user to input commands, and can include one or more of a keyboard, mouse, microphone, and touchscreen. Furthermore, the input device can also be any interface for receiving information.

[0115] The output device can output various information (e.g., images or sounds) to the outside (e.g., a user), and may include one or more of a display, speaker, etc. Furthermore, the output device can also be any other device with output functionality.

[0116] For example, the example calibration device 300 for implementing the calibration method 200 according to the embodiments of this application can be applied to terminal devices (such as mobile phones), tablet computers, laptop computers, ultra-mobile personal computers (UMPCs), handheld computers, netbooks, personal digital assistants (PDAs), wearable devices (such as smartwatches, smart glasses, or smart helmets), augmented reality (AR) devices, virtual reality (VR) devices, smart home devices, in-vehicle computers, and other electronic devices. The embodiments of this application do not impose any limitations on this.

[0117] Those skilled in the art can understand the specific operation of the calibration device 300 for implementing the calibration method 200 according to the embodiments of this application in conjunction with the content described above. For the sake of brevity, the specific details will not be repeated here, but only some main operations of the processor 310 will be described.

[0118] In one embodiment of this application, when the executable program is run by the processor 310, the processor 310 performs the following steps: Based on a predetermined influence line corresponding to the structural response to be calibrated, a measured loading test is conducted on the target bridge at the location of the maximum longitudinal axis of the influence line or its vicinity; structural response monitoring data is acquired through a bridge health monitoring system, and a monitoring data sequence is constructed; the bridge structure finite element model of the target bridge is called and loaded with the same position and load as the measured loading test of the target bridge; the corresponding structural response theoretical data is obtained from the finite element analysis results, and a structural response theoretical data sequence is constructed; based on the monitoring data sequence and the structural response theoretical data sequence, a first fitting model that meets a preset R-squared value requirement is used for fitting to obtain a first fitting relationship; based on the monitoring data sequence and the structural response theoretical data sequence, a second fitting model that meets a preset R-squared value requirement is used for fitting to obtain a second fitting relationship; based on the monitoring data sequence... A first priority feature representing the amount of data is obtained from the structural response theory data sequence; the first priority feature is compared with a preset standard threshold to determine whether the first priority feature reaches the preset standard threshold; if the first priority feature does not reach the preset standard threshold, the first fitting relationship is determined as the target fitting relationship, and the first fitting model is determined as the calibration model; when calibrating the monitoring data obtained by the bridge health monitoring system using the first fitting model, a second priority feature representing the fitting error rate corresponding to the fitting result is obtained; the fitting error value obtained based on the second priority feature is compared with a preset error threshold to determine whether the error value obtained based on the second priority feature reaches the preset error threshold; if the first priority feature reaches the preset standard threshold, and the fitting error value obtained based on the second priority feature is greater than the preset error threshold, the second fitting relationship is determined as the target fitting relationship, and the second fitting model is determined as the calibration model; the calibration model is used to calibrate the subsequent monitoring data of the bridge health monitoring system.

[0119] The calibration method 200 according to an embodiment of this application has been illustrated above. The following is in conjunction with... Figure 8 This application describes a calibration system 400 provided in another aspect of an embodiment.

[0120] Reference Figure 8 This document describes an example calibration system 400 for implementing the calibration method of the embodiments of this application. The calibration system 400 may include a first construction module 402, a second construction module 404, a first fitting relationship acquisition module 406, a second fitting relationship acquisition module 408, a first priority feature acquisition module 410, a first judgment module 412, a first determination module 414, a second priority feature acquisition module 416, a second judgment module 418, a second determination module 420, and a calibration module 422. Wherein:

[0121] The first construction module 402 is used to: conduct a measured loading test on the target bridge at the maximum longitudinal index of the influence line or its vicinity based on the influence line corresponding to the pre-determined structural response to be calibrated; acquire structural response monitoring data through the bridge health monitoring system; and construct a monitoring data sequence.

[0122] The second construction module 404 is used to: call and load the target bridge's bridge structure finite element model with the same position and load as the target bridge's measured loading test, obtain the corresponding structural response theoretical data from the finite element analysis results, and construct a structural response theoretical data sequence.

[0123] The first fitting relationship acquisition module 406 is used to: fit the monitoring data sequence and the structural response theoretical data sequence using a first fitting model that meets the preset R-squared value requirement, and obtain the first fitting relationship.

[0124] The second fitting relationship acquisition module 408 is used to: fit the monitoring data sequence and the structural response theoretical data sequence using a second fitting model that meets the preset R-squared value requirement, and obtain the second fitting relationship.

[0125] The first priority feature acquisition module 410 is used to: acquire a first priority feature representing the amount of data based on the monitoring data sequence and the structural response theoretical data sequence.

[0126] The first judgment module 412 is used to: compare the first priority feature with a preset standard threshold and determine whether the first priority feature reaches the preset standard threshold.

[0127] The first determining module 414 is used to: determine the first fitting relationship as the target fitting relationship and the first fitting model as the calibration model when the first priority feature does not reach the preset standard threshold.

[0128] The second priority feature acquisition module 416 is used to: acquire the second priority feature corresponding to the fitting result, representing the fitting error rate, when calibrating the monitoring data acquired by the bridge health monitoring system using the first fitting model.

[0129] The second judgment module 418 is used to: compare the fitting error value obtained based on the second priority feature with a preset error threshold, and determine whether the error value obtained based on the second priority feature reaches the preset error threshold.

[0130] The second determining module 420 is used to: determine the second fitting relationship as the target fitting relationship and the second fitting model as the calibration model when the first priority feature reaches a preset standard threshold and the error value obtained based on the second priority feature is greater than the preset error threshold.

[0131] The calibration module 422 is used to calibrate the monitoring data of the bridge health monitoring system using the calibration model.

[0132] The calibration system 400 proposed in this embodiment of the invention uses finite element analysis to calculate values ​​rather than the monitoring data itself. Compared with calibrating monitoring data based on monitoring data that may contain initial errors, it can more accurately correct the monitoring data. At the same time, the loading test is performed according to the influence line, requiring less load. The bridge health monitoring system can obtain larger and more stable structural response monitoring data, which is simpler and more effective than the load test. In addition, based on the monitoring data sequence and the structural response theoretical data sequence, it can use either the first fitting model or the second fitting model for calibration while ensuring the calibration accuracy. This provides high flexibility and applicability.

[0133] Furthermore, according to embodiments of this application, this application also provides a storage medium on which a computer program is stored. When the computer program is run by a processor, it is used to execute corresponding steps of the calibration method 200 of this application. The storage medium may, for example, include a memory card of a smartphone, a storage component of a tablet computer, a hard disk of a personal computer, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable compact disc read-only memory (CD-ROM), a USB memory, or any combination of the above storage media. The computer-readable storage medium may be any combination of one or more computer-readable storage media.

[0134] Furthermore, according to embodiments of this application, this application also provides a computer program product, including computer instructions that, when executed by a processor, implement the steps of the calibration method 200 of embodiments of this application.

[0135] Although exemplary embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above exemplary embodiments are merely illustrative and are not intended to limit the scope of this application. Various changes and modifications can be made therein by those skilled in the art without departing from the scope and spirit of this application. All such changes and modifications are intended to be included within the scope of this application as claimed in the appended claims.

[0136] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. 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 implementation should not be considered beyond the scope of this application.

[0137] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed.

[0138] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features but not others included in other embodiments, combinations of features from different embodiments are intended to be within the scope of this application and form different embodiments. For example, in the claims, any one of the claimed embodiments can be used in any combination.

[0139] It should be noted that the above embodiments are illustrative of this application and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

[0140] The above description is merely a specific embodiment or illustration of the embodiments of this application. The scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. The scope of protection of this application shall be determined by the scope of the claims.

Claims

1. A calibration method for a bridge health monitoring system based on regional loading tests, characterized in that, The calibration method includes: Based on the influence line corresponding to the pre-determined structural response to be calibrated, an actual loading test is conducted on the target bridge at the maximum longitudinal index of the influence line or its vicinity. Structural response monitoring data is obtained through the bridge health monitoring system, and a monitoring data sequence is constructed. The finite element model of the target bridge's reference state is invoked and loaded with the same load at the same location as the actual load test of the target bridge. The corresponding structural response theoretical data is obtained from the finite element analysis results, and a structural response theoretical data sequence is constructed. Based on the monitoring data sequence and the structural response theoretical data sequence, a first fitting model that meets the preset R-squared value requirement is used for fitting to obtain a first fitting relationship; Based on the monitoring data sequence and the structural response theoretical data sequence, a second fitting model that meets the preset R-squared value requirement is used for fitting to obtain a second fitting relationship; Based on the monitoring data sequence and the structural response theoretical data sequence, obtain the first priority feature representing the quantity of data; Based on the first priority feature and the preset standard threshold, a target fitting relationship is determined between the first fitting relationship and the second fitting relationship; if the target fitting relationship is determined, the fitting model corresponding to the target fitting relationship is determined as the calibration model of the bridge health monitoring system. The calibration model is used to calibrate the monitoring data of the bridge health monitoring system.

2. The calibration method according to claim 1, characterized in that, The calibration method further includes: When calibrating the monitoring data acquired by the bridge health monitoring system using the first fitting model, a second priority feature representing the fitting error rate corresponding to the fitting result is obtained. If the second priority feature is greater than the preset error threshold, the second fitting model is determined as the calibration model.

3. The calibration method according to claim 1, characterized in that, The first fitting model is used for fitting to obtain the first fitting relationship, which specifically includes: Using mathematical analysis software, multiple fitting functions, including polynomial functions, linear functions, and logarithmic functions, were employed to fit the acquired structural response monitoring data and structural response theoretical data one by one, and the R-squared value, which is used to measure the degree of fit of the model to the data, was calculated. When the R-squared value reaches the preset value, the fitting function at this time is determined as the first fitting model that meets the requirements, and the fitting function at this time is the first fitting relationship.

4. The calibration method according to claim 3, characterized in that, The calculation used to measure the R-squared value, which measures the goodness of fit of the model to the data, specifically refers to: In the formula, y i Let p represent the theoretical data of the i-th structural response, m represent the average value of the n theoretical data of the structural response, and p represent the theoretical data of the n-th structural response. i This represents the fitted value corresponding to the i-th structural response monitoring data calculated according to a certain fitting function.

5. The calibration method according to claim 1, characterized in that, The second fitting model is used for fitting to obtain the second fitting relationship, which specifically includes: The acquired monitoring data sequence and structural response theoretical data sequence are divided into mutually exclusive training and validation sets according to a certain ratio, with the monitoring data sequence serving as the input layer and the corresponding structural response theoretical data sequence serving as the output layer. The number of network layers, number of neurons, learning rate, and loss function parameters of the second fitting model are preset. The model is trained using the training set, and the R-squared value is calculated as an indicator to evaluate the performance of the model. If the R-squared value reaches the preset value, then the second fitting model at this time is determined as the second fitting model that meets the requirements. The relationship between the monitoring data sequence and the structural response theoretical data sequence determined by the second fitting model at this time is the second fitting relationship. If the R-squared value does not reach the preset value, continue to adjust the model parameters and iterate the above steps until the R-squared value reaches the preset value.

6. The calibration method according to claim 1, characterized in that, The first fitting model is a regression model.

7. The calibration method according to claim 1, characterized in that, The second fitting model is a neural network model.

8. The calibration method according to claim 1, characterized in that, The first priority feature refers to the total number of data points in the monitoring data sequence and the structural response theoretical data sequence; The requirement to meet the preset R-squared value means that the R-squared value is at least 0.

95.

9. A calibration device for a bridge health monitoring system based on regional loading tests, characterized in that, The calibration device includes: Memory is used to store executable instructions for a computer; A processor, when executing computer-executable instructions stored in the memory, implements the calibration method according to any one of claims 1 to 8.

10. A storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the calibration method according to any one of claims 1 to 8.

Images