Steel box girder safety monitoring method and system
Through multi-source data fusion and deep learning technology, multi-dimensional safety monitoring and early warning of steel box girders is achieved, solving the problem of the inability to evaluate and predict safety hazards from multiple dimensions in the existing technology, and improving prediction accuracy and early warning ability.
Patent Information
- Application Number
- CN202411108637.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-13
- Publication Date
- 2024-12-06
AI Technical Summary
The existing steel box girder safety monitoring system cannot be comprehensively measured and evaluated from multiple dimensions, cannot predict safety hazards in advance, the early warning mechanism is incomplete, and there is a lack of flexible early warning plans.
Using multi-source data fusion and deep learning technology, data is collected through temperature, humidity, displacement, inclination and stress sensors, data normalization and weight assignment are carried out, LSTM model is used for prediction, and early warning judgment is made based on the preset safety warning level. , realize multi-dimensional safety monitoring and early warning of steel box girders.
It improves the accuracy of the prediction results of steel box girder safety monitoring, can eliminate hidden dangers before safety problems occur, and provides scientific and reliable construction and engineering structure health management methods.
Smart Images

Figure CN119089263A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of bridge safety monitoring, and relates to a steel box girder safety monitoring method and system. Background Art
[0002] Steel box girder is a common bridge structure. It plays an extremely important role in the construction of long-span bridges due to its many advantages such as convenient and fast construction and strong bearing capacity. Its safety is crucial to transportation. After the steel box girder is erected, how to monitor the safety status of the steel box girder is of great significance for safe construction.
[0003] In the existing steel box girder safety monitoring system, only the impact of single factors such as temperature on the deformation of the steel box girder is considered, while the joint effect of other factors is ignored, resulting in incomplete safety analysis and inability to comprehensively measure and evaluate the specific situation from multiple angles and dimensions.
[0004] In addition, the existing steel box girder safety monitoring system has the following problems: it only processes real-time data and analyzes the current status of the steel box girder. It is unable to predict the possibility of safety hazards in advance and unable to take measures in advance to prevent the occurrence of danger. At the same time, the early warning mechanism is imperfect and lacks flexible early warning plans. Summary of the invention
[0005] The purpose of the present invention is to solve the problem that in the prior art, when conducting safety monitoring of steel box girders, it is impossible to conduct comprehensive measurement and evaluation from multiple dimensions, and it is impossible to predict the occurrence of potential safety hazards in advance, resulting in the inability to effectively monitor the safety problems existing in the construction process of steel box girders, and to provide a steel box girder safety monitoring method and system.
[0006] In order to achieve the above object, the present invention adopts the following technical solutions:
[0007] A steel box girder safety monitoring method comprises the following steps:
[0008] Collect real-time monitoring data of steel box girders from different data sources;
[0009] The real-time monitoring data is subjected to data fusion to obtain fused data, and the fused data is input into a pre-trained prediction model, wherein the prediction model outputs the prediction value of the steel box girder at the next moment under different data source states; wherein the prediction model is obtained by training the fused values of the monitoring data under different data sources;
[0010] The predicted value output by the prediction model is compared with the preset safety warning level value, and the current warning level is obtained based on the comparison result.
[0011] A further improvement of the present invention is:
[0012] The acquisition of the fusion value includes the following steps:
[0013] Normalize the monitoring data under different data sources:
[0014]
[0015] Assign different weight values to the monitoring data under different data sources:
[0016]
[0017] In the formula, a i represents: the component value of the eigenvector corresponding to the maximum eigenvalue;
[0018] Fuse multi-source data through the centralized averaging method:
[0019]
[0020] In the formula, w i represents the component value of the normalized vector; y j i represents: the data obtained after preprocessing of various sensors at time j.
[0021] The prediction model is an LSTM model;
[0022] The Adam optimizer and the mean squared error (MSE) are used as the loss function of the model, and the MSE is obtained through the following formula:
[0023]
[0024] where, z i represents the true value of the i-th sample, represents the corresponding estimated value, and m represents the number of samples.
[0025] The preset safety warning level values include:
[0026] C1, first-level warning; C2, second-level warning and C3; third-level warning.
[0027] Comparing the predicted value output by the prediction model with the preset safety warning level value, and obtaining the current warning level based on the comparison result, includes the following steps:
[0028] If z t+1 < C3, it means that there is no safety risk in the safety monitoring and warning determination of the steel box girder;
[0029] If C2 < z t+1 ≤ C3, enter the first-level warning state;
[0030] If C1<z t+1 ≤C2, it enters the second-level warning state;
[0031] If t+1 ≥C1, it enters the third-level warning state;
[0032] Among them, z t+1 It represents the fusion value of the monitoring data from various data sources predicted at time t+1 for the steel box girder.
[0033] The monitoring data from different data sources include data collected by temperature sensors, humidity sensors, displacement sensors, inclination sensors and stress sensors.
[0034] It also includes the elimination of abnormal data and filling in missing data for the initially acquired data.
[0035] A steel box girder safety monitoring system, comprising:
[0036] Data acquisition module, used to collect real-time monitoring data of steel box girders from different data sources;
[0037] A prediction module is used to perform data fusion on the real-time monitoring data to obtain fused data, and input the fused data into a pre-trained prediction model, wherein the prediction model outputs the predicted value of the steel box girder at the next moment under different data source states; wherein the prediction model is obtained by training the fused values of the monitoring data under different data sources;
[0038] The warning module is used to compare the predicted value output by the prediction model with the pre-set safety warning level value, and obtain the current warning level based on the comparison result.
[0039] A terminal device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any method described in the present invention when executing the computer program.
[0040] A computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of any method described in the present invention.
[0041] Compared with the prior art, the present invention has the following beneficial effects:
[0042] The present invention discloses a steel box girder safety monitoring method, which collects monitoring data from different data sources and obtains multi-dimensional data, so as to facilitate the comprehensiveness of subsequent prediction results and improve the accuracy of the final prediction results. The data is predicted by a prediction model, and the current construction safety status of the steel box girder is judged according to the constructed early warning level. Hidden dangers can be eliminated before safety problems occur, potential safety risks can be predicted more accurately, and countermeasures can be taken, thereby providing scientific and reliable means and methods for safe construction and engineering structure health management.
[0043] Furthermore, in the present invention, different weights are assigned to relevant data from different data sources, and the acquired multi-source data are fused, so that the influence of different factors on data from different data sources can be more reasonably assigned. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.
[0045] Figure 1 is a flow chart of the monitoring method of the present invention;
[0046] Figure 2 This is a system distribution diagram of the present invention. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.
[0048] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0049] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0050] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. indicate an orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed when in use, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.
[0051] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but can be slightly tilted.
[0052] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0053] The present invention is further described in detail below in conjunction with the accompanying drawings:
[0054] See also Figure 1 to Figure 2 The present invention discloses a steel box girder safety monitoring method, which uses the Internet of Things, the Internet and other technologies to build an all-round, three-dimensional real-time monitoring and intelligent early warning platform. The system integrates multi-source data fusion and deep learning technology to efficiently monitor and predict the steel box girder structure, and realizes intelligent and digital safety management. By integrating multiple monitoring indicators and intelligent prediction, the steel box girder safety monitoring system can accurately evaluate the state of the steel box girder structure, and generate different levels of early warning information according to the prediction results, providing scientific and reliable means and methods for safe construction and engineering structure health management, and specifically includes the following steps:
[0055] The specific steps include:
[0056] Step 1: Sensor Data Preprocessing
[0057] In order to monitor the status of the steel box girder in a more comprehensive and multi-dimensional manner, the present invention deploys temperature sensors, humidity sensors, displacement sensors, inclination sensors and stress sensors to collect data.
[0058] Since sensors are affected by factors such as the environment and weather, the collected data usually contains abnormal data. Therefore, it is necessary to remove abnormal data and fill in missing data, including the following steps:
[0059] In step 1.1, bilateral Grubbs is used to remove abnormal values received by various sensors.
[0060] Get the time series data collected by various sensors, where the time series data collected by the temperature sensor is represented by X T ={x t1 ,x t2 ,…,x tn}, the time series data collected by the humidity sensor is represented by X W ={x w1 ,x w2 ,…,x wn}, the time series data collected by the displacement sensor is represented by X D ={x d1 ,x d2 ,…,x dn}, the time series data collected by the tilt sensor is represented by X A ={x a1 ,x a2 ,…,x an}, the time series data collected by the stress sensor is represented by X F ={x f1 ,x f2 ,…,x fn}, the time series data collected by various sensors have the same operations for data preprocessing and normalization steps. Taking the time series data collected by the temperature sensor as an example, it is recorded as X = {x1, x2, …, x n}, where n is the number of data points.
[0061] Further, determine the significance level α for outlier detection. The specific value of the significance level depends on the user's needs and tolerance for error types. Perform cyclic detection and delete outliers, repeating the following steps until no outliers are found:
[0062] For each data point x in X i Perform a two-sided Grubbs test and if the test statistic exceeds the critical value g, set x i The data points marked as outliers are deleted and the data set X is updated. The above steps are repeated until the stop condition is met, that is, the number of cycles reaches n.
[0063] The calculation formula of the statistic G of the two-sided Grubbs test is shown in (1):
[0064]
[0065] in, is the sample mean, s is the sample standard deviation, x i is the data point being tested. If G exceeds the critical value g, then x i is considered an outlier and x is removed i .
[0066] Step 1.2: The data sequence after outlier processing in step 1.1 above is recorded as X' = {x1, x2, ..., x n}, in order to solve the missing problems in the collected data, Newton interpolation method is used to perform interpolation estimation to supplement the missing values.
[0067] Construct an n-1th order Newton interpolation polynomial f(t), where t represents the time point of data collection and f(t) represents the data collected at time t, which is given by the following formula (2):
[0068]
[0069] Among them, f(t i ) represents the function value of the ith known data point, called the basic node value; f[t i ,t i+1 ] indicates that (t i ,t i+1 ) is the first-order difference quotient of the node, called the first-order difference quotient; f[t i ,t i+1 ,…,t j ] indicates that (t i ,t i+1 ,…,t j ) is the high-order difference quotient of the node.
[0070] Calculate f(t) at t i The function value at i ) as an estimate of the missing values.
[0071] In this embodiment, the preprocessing process of deleting abnormal data and filling missing data for various sensor data is to ensure the quality of the data and reduce errors and uncertainties in the subsequent analysis process. This processing can improve the reliability and effectiveness of the data and provide a reliable basis for the subsequent data analysis and modeling of the system;
[0072] Step 2: Data filtering
[0073] Used to improve data quality by removing error points, considering the complex environment during the construction of steel box girders and the influence of data noise during the sensor acquisition process, which will cause errors in the data collected by various sensors. By using a moving average filter, the purpose of effectively removing the influence of noise is achieved, the error of sensor data acquisition is reduced, and the estimated value closest to the true value is obtained, which enhances the continuity and stability of the collected data.
[0074] The specific steps include:
[0075] Step 2.1, calculation of moving window N.
[0076] First, we need to select the filter window size N, that is, the number of data points used to calculate the moving average. The size of the moving window is calculated using formula (3):
[0077] N=φ×f S ×μ (3)
[0078] Where N is the filter window size. S is the sampling frequency of the sensor device. μ is the time interval between two consecutive sampling points. φ is a coefficient with a value range of (0,1).
[0079] Step 2.2, calculate the moving average
[0080] Starting from the starting point x1 of the time series data, sliding with the window size N, calculate the average value of the data points in each window, and use this average value as the new data point, representing the moving average value of the original data. The calculation of the moving average is shown in formula (4):
[0081]
[0082] Among them, SMA i is the ith moving average, x i are the raw data points in the time series.
[0083] In step 2.3, the root mean square error is used to evaluate the denoising effect.
[0084] Further collect and process the outlier data X'={x1,x2,…,x n} and the denoised data Y = {y1,y2,…,y n}. Then calculate the residual at the corresponding time step, the residual formula (5) is:
[0085] e i =x i -y i ,i=1,2,…,n (5)
[0086] Further calculation of square error: square each residual value to get the square error e i 2 Then the root mean square error is calculated through this square error, the average of all square error values is calculated, and the square root of the result is taken to get the root mean square error. The root mean square error formula (6) is:
[0087]
[0088] The denoising effect is judged by the calculated root mean square error, and the size of the parameter φ is adjusted. The smaller the root mean square error, the better the denoising effect.
[0089] Step 3: Multi-source monitoring data fusion processing:
[0090] It aims to comprehensively consider the joint effects of temperature, humidity, displacement, inclination and stress on steel box girders. If only a single data affecting the steel box girder is considered, the analysis will be incomplete. Comprehensive analysis of different data sources can monitor the safety of steel box girders from multiple perspectives and dimensions.
[0091] First, the purpose of eliminating dimension is achieved by normalizing the data;
[0092] Secondly, the analytic hierarchy process is used to give different weights to different data sources;
[0093] Finally, weighted average is used to obtain the fused value.
[0094] The specific process is:
[0095] Step 3.1: Normalization using the deviation standardization method
[0096] Furthermore, the data collected by the temperature sensor are sorted in chronological order to obtain Y = {y1, y2, ..., y n}, the maximum value Max and the minimum value Min are obtained according to the sorted data, and then normalized according to the following formula (7):
[0097]
[0098] y' i This is the normalized result.
[0099] Step 3.2: Use the analytic hierarchy process to assign different weights to the data of different categories of sensors.
[0100] Temperature, humidity, displacement, inclination and stress are five factors that have different impacts on the safety of steel box girders. In order to distinguish their influence, different weights are given to the data collected by different types of sensors.
[0101] Furthermore, in order to determine the specific weight assignments of different categories of sensors, the hierarchical analysis method is adopted. This method regards the weight assignment of different categories of sensors as a decision problem and decomposes the decision problem. The goal of this problem is the safety of the steel box girder, and the criteria are temperature sensors, humidity sensors, displacement sensors, inclination sensors and stress sensors.
[0102] Furthermore, the factors in the criteria are compared in pairs, and the relative importance between the two is obtained based on experience. The standard of [1 / 9,9] is used for evaluation, where 1 represents the same importance, that is, the former is equally important as the latter, 9 and 1 / 9 represent extreme importance, 9 represents that the former is extremely important compared with the latter, and 1 / 9 represents that the former is extremely unimportant compared with the latter. Through the above comparison, the judgment matrix A is constructed.
[0103]
[0104] In particular, the value of the elements on the main diagonal of the matrix is 1, and the product of two elements at symmetric positions about the main diagonal is 1.
[0105] Furthermore, the eigenvalue and eigenvector are solved according to the judgment matrix A using the following formula (8).
[0106] Αβ=λβ (8)
[0107] Get the eigenvalues of the matrix λ1,λ2,…,λ k , select its largest eigenvalue λ max , and normalize the corresponding feature vector β = [a1, a2, a3, a4, a5]. The components in the normalized vector w = [w1, w2, w3, w4, w5] are the weights of the corresponding category sensors as shown in formula (9).
[0108]
[0109] Step 3.3: Apply weighted average method to fuse multi-source data
[0110] Furthermore, the weights of various sensors are multiplied by the corresponding data and then added, as shown in formula (10), to obtain the fused data Z = {z1, z2, …, z n}.
[0111]
[0112] Step 4: Sensor sequence data training and prediction:
[0113] The aim is to use the fused time series data to input into the Long Short-Term Memory (LSTM) network model. Traditional statistics-based analysis does not have predictive capabilities.
[0114] First, through learning and pattern recognition of historical data, we can establish an understanding of the inherent laws of the data.
[0115] The trained model is then used to predict data for future time steps to provide estimates of expected events or changes. Through the training and prediction of the LSTM model, the system can more accurately predict potential safety risks and provide the necessary warnings for taking timely and effective measures.
[0116] Step 4.1, prepare data.
[0117] According to the data fusion process in step 3, a set of fused data sequences Z = {z1, z2, …, z n}.
[0118] Step 4.2, divide the data set.
[0119] The fused data is divided into training set and test set, represented by train_data and test_data respectively.
[0120] Step 4.3, prepare training data.
[0121] Convert the time series data into the input features z_train and target values u_train required by the model. Among them, z_train is a three-dimensional array with a shape of (number of samples, time_steps, number of features), where the number of samples refers to the number of samples in the training set, that is, how many time series are there in the data set. time_steps represents the time step of each time series, that is, how many consecutive time points each time series contains. The number of features represents the number of features at each time point, that is, the number of attributes describing each time point. That is, the three-dimensional array with a shape of (number of samples, time_steps, number of features) can be regarded as a data structure consisting of the number of samples, time steps, and number of features. u_train is a one-dimensional array including the number of samples. Through a loop, each sample (or time point) in the data is traversed, starting from the first sample until the last time_steps+1 sample. For each sample, the data of the first time_steps time steps are used as input features, and the data of the next time step of the sample is used as the target value.
[0122] Furthermore, each iteration of the loop will perform the following two steps:
[0123] Divide the data of the current time step and the data of the previous time_steps time steps in the data and add them to z_train.
[0124] Add the data of the next time step of the current time step in the data to u_train to form the target value.
[0125] Finally, we get z_train and u_train.
[0126] Step 4.4, build the LSTM model.
[0127] Create a model object and add an LSTM layer and a fully connected layer. The LSTM layer has 64 LSTM units and uses ReLU as the activation function. The input shape is (time_steps, 1), which means that each input sequence has time_steps time steps and each time step has 1 feature.
[0128] Step 4.5, compile and train the model.
[0129] Use the compile function to configure the optimizer and loss function of the model. Here we choose the Adam optimizer and mean square error (MSE) as the loss function. The MSE calculation formula (11) is as follows.
[0130]
[0131] Among them, z i represents the true value of the i-th sample, represents the corresponding estimated value, and m represents the number of samples.
[0132] Furthermore, the training set z_train and u_train are passed to the model. The model will be trained based on this data and the model parameters will be adjusted to minimize the loss function.
[0133] Step 4.6, prepare input data and prediction.
[0134] Convert the test set data into the form required by the model. Then input the data into the trained model for prediction, and predict the fusion data value z of various sensors at the steel box girder at time t+1 t+1 .
[0135] Step 5: Set safety thresholds and issue warning information
[0136] The purpose of this step is to define different warning levels. If the prediction results exceed the set threshold range, the steel box girder safety monitoring system will generate warning information of the corresponding level. Timely release of warning information can remind on-site staff or relevant managers of impending safety issues, so that they can take appropriate measures in time to mitigate or prevent potential risks. By setting different levels of thresholds, the steel box girder safety monitoring system can generate warning signals of corresponding levels according to the changes in the prediction result values, thereby improving the ability to perceive and respond to potential safety hazards.
[0137] Step 5.1, monitoring and early warning
[0138] The value z predicted by the above steps t+1 , and compare it with the safety warning value C = {C1, C2, C3}, where C1 represents the first-level warning: emergency level, C2 represents the second-level warning: serious warning level, and C3 represents the third-level warning: warning level. Determine whether to issue a safety warning based on the comparison result. If z t+1 <C3, it means that the steel box girder safety monitoring and early warning judgment does not exist safety risks, and the staff can continue construction; if C2<z t+1 ≤C3, it enters the first-level warning state, indicating that the steel box girder safety monitoring and warning judgment has a slight safety risk and needs attention; if C1<z t+1 ≤C2, it enters the second-level warning state, indicating that the steel box girder safety monitoring and early warning judgment has a large safety risk and requires immediate attention and measures, such as shutdown, maintenance, load adjustment, etc.; if z t+1 ≥C1, it enters the third-level warning state, indicating that the steel box girder safety monitoring and early warning judgment has serious safety risks and emergency measures need to be taken immediately.
[0139] Step 5.2, warning log recording
[0140] Early warning log recording is essential for performance analysis and improvement of steel box girder safety monitoring system, including:
[0141] (1) Detailed description of the event: Record the detailed information of each warning event of the safety status of the steel box girder, including the duration and solution. The duration should describe the entire time process from the triggering of the warning signal to the resolution. The solution should clearly state what measures were taken to deal with the warning event.
[0142] (2) Multi-level classification: Steel box girder safety status warning events can be divided into different levels according to the degree of urgency, which helps to better analyze and manage the safety risks of steel box girders. The levels can be divided into three categories: Level 1 warning represents the emergency level, Level 2 warning represents the serious warning level, and Level 3 warning represents the warning level. According to these levels, construction personnel can respond quickly and take corresponding measures to reduce or eliminate potential safety risks.
[0143] The monitoring method disclosed in this embodiment can give early warnings for safety issues of steel box girders, eliminate hidden dangers before safety issues occur, use long-term and short-term memory networks to train historical data of sensors monitoring steel box girders, predict relevant data of steel box girders at future moments, compare with set thresholds, and give early warnings according to levels. Data fusion is performed on multiple data affecting the deformation of steel box girders measured by sensors, and different weights are assigned to different categories of sensors using the hierarchical analysis method. The joint effects of multiple factors are comprehensively considered to scientifically judge the safety issues existing in the construction process of steel box girders.
[0144] The embodiment of the present invention further discloses a steel box girder safety monitoring system, comprising:
[0145] Data acquisition module, used to collect real-time monitoring data of steel box girders from different data sources;
[0146] A prediction module is used to perform data fusion on the real-time monitoring data to obtain fused data, and input the fused data into a pre-trained prediction model, wherein the prediction model outputs the predicted value of the steel box girder at the next moment under different data source states; wherein the prediction model is obtained by training the fused values of the monitoring data under different data sources;
[0147] The warning module is used to compare the predicted value output by the prediction model with the pre-set safety warning level value, and obtain the current warning level based on the comparison result.
[0148] A schematic diagram of a terminal device provided in an embodiment of the present invention. The terminal device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor executes the computer program, the functions of the modules / units in the above-mentioned device embodiments are implemented.
[0149] The computer program may be divided into one or more modules / units, and the one or more modules / units are stored in the memory and executed by the processor to accomplish the present invention.
[0150] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.
[0151] The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
[0152] The memory may be used to store the computer programs and / or modules, and the processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory.
[0153] If the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.
[0154] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for safety monitoring of a steel box girder, characterized in that: The following steps are involved: Collect real-time monitoring data of steel box girders from different data sources; The real-time monitoring data is subjected to data fusion to obtain fused data, and the fused data is input into a pre-trained prediction model, wherein the prediction model outputs the prediction value of the steel box girder at the next moment under different data source states; wherein the prediction model is obtained by training the fused values of the monitoring data under different data sources; The predicted value output by the prediction model is compared with the preset safety warning level value, and the current warning level is obtained based on the comparison result.
2. A steel box girder safety monitoring method according to claim 1, characterized in that: The acquisition of the fusion value comprises the following steps: Normalize the monitoring data from different data sources: Assign different weights to monitoring data from different data sources: In the formula, a i Indicates: the maximum eigenvalue corresponds to the component value of its eigenvector; Fusion of multi-source data through weighted average method: In the formula, w i Represents the component value of the normalized vector; y j i Represents: the data obtained by various sensors after preprocessing at time j.
3. A steel box girder safety monitoring method according to claim 1, characterized in that: The prediction model is an LSTM model; The Adam optimizer and mean square error (MSE) are selected as the loss function of the model. The MSE is obtained by the following formula: Among them, z i represents the true value of the i-th sample, represents the corresponding estimated value, and m represents the number of samples.
4. A steel box girder safety monitoring method according to claim 1, characterized in that: The preset safety warning level values include: C1, level one warning; C2, level two warning and C3; level three warning.
5. A steel box girder safety monitoring method according to claim 1, characterized in that: The method of comparing the predicted value output by the prediction model with the preset safety warning level value and obtaining the current warning level based on the comparison result includes the following steps: If z t+1 <C3, it indicates that there is no safety risk in the safety monitoring and early warning judgment of the steel box girder; If C2<z t+1 ≤C3, it enters the first-level warning state; If C1<z t+1 ≤C2, it enters the second-level warning state; If t+1 ≥C1, it enters the third-level warning state; Among them, z t+1 It represents the fusion value of the monitoring data from various data sources predicted at time t+1 for the steel box girder.
6. A steel box girder safety monitoring method according to claim 1, characterized in that: The monitoring data from different data sources include data collected by temperature sensors, humidity sensors, displacement sensors, inclination sensors and stress sensors.
7. A steel box girder safety monitoring method according to claim 1, characterized in that: It also includes the elimination of abnormal data and filling in missing data for the initially acquired data.
8. A steel box girder safety monitoring system, characterized in that: include: Data acquisition module, used to collect real-time monitoring data of steel box girders from different data sources; A prediction module is used to perform data fusion on the real-time monitoring data to obtain fused data, and input the fused data into a pre-trained prediction model, wherein the prediction model outputs the predicted value of the steel box girder at the next moment under different data source states; wherein the prediction model is obtained by training the fused values of the monitoring data under different data sources; The warning module is used to compare the predicted value output by the prediction model with the pre-set safety warning level value, and obtain the current warning level based on the comparison result.
9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Bridge structure safety monitoring data prediction method
CN104268658A
Full-bridge structure safety state monitoring method and system based on multi-source data
CN111310273A
Bridge health degree Beidou monitoring method and system based on long and short time memory learning
CN117537996A
Urban bridge real-time monitoring and early warning system
CN117671921A
Cited By
Intelligent dynamic compaction system and method based on Beidou precise positioning
CN121115078A
