Method and system for intelligent analysis and early warning of hydraulic engineering safety

Abstract

The invention relates to the technical field of water conservancy projects, and discloses a water conservancy project safety intelligent analysis and early warning method and system, and the method comprises the steps: respectively obtaining M groups of structure data of a plurality of detection points within a first preset time and corresponding structure health assessment data; preprocessing each group of structure data of each detection point location to obtain first effective structure data; extracting N types of feature data of each group of first effective structure data for feature fusion to obtain comprehensive structure feature vector data; training a health state evaluation model by using the comprehensive structure feature vector data and the structure health evaluation data; and obtaining to-be-detected structure data, inputting the to-be-detected structure data into the trained health state evaluation model to obtain structure health evaluation data, judging the safety state of the to-be-detected detection point according to the structure health evaluation data, and triggering corresponding danger early warning. According to the method, the dynamic performance and the health condition of the hydraulic engineering structure are comprehensively reflected, and accurate early warning is carried out.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy engineering, and in particular to a method and system for intelligent analysis and early warning of water conservancy engineering safety. Background Art

[0002] Water conservancy projects refer to projects that develop, utilize, manage and protect water resources through artificial means to meet the needs of human life and production. Water conservancy projects usually include dams, sluice gates, hydropower stations, reservoirs, irrigation systems, flood control systems, water supply systems, etc. Water conservancy projects are important infrastructure for national economic and social development, and play an important role in ensuring national food security, energy security, ecological security, flood control security and social stability.

[0003] The safety management of water conservancy projects is particularly important, as it is related to the safety of people's lives and property, national economic and social stability. Therefore, the water conservancy department has taken a series of measures, including strengthening the design, construction, operation and maintenance management of water conservancy projects, establishing a sound safety supervision system, and promoting the application of new technologies, new materials and new processes to improve the safety and reliability of water conservancy projects. At the same time, it has also actively carried out research and application of intelligent analysis and early warning systems for water conservancy project safety to further improve the safety management level of water conservancy projects.

[0004] However, with the continuous expansion of the scale and increase in complexity of water conservancy projects, the safety management of water conservancy projects faces huge challenges. Traditional water conservancy project safety monitoring mainly relies on manual inspections and simple sensor monitoring, which has problems such as low efficiency, poor real-time performance, and weak early warning capabilities.

[0005] In view of this, the present invention proposes a method and system for intelligent analysis and early warning of water conservancy project safety to solve the above problems. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a method and system for intelligent analysis and early warning of water conservancy project safety, which comprehensively reflects the dynamic performance and health status of the water conservancy project structure and provides accurate early warning, providing strong technical support for ensuring the safe operation of water conservancy projects.

[0007] In a first aspect, the present invention provides a method for intelligent analysis and early warning of water conservancy project safety, comprising the following steps: S1. Obtain M groups of structural data of several detection points of the water conservancy project within a first preset time and structural health assessment data corresponding to each detection point; M is a natural number not equal to 0; S2, pre-processing each group of structural data of each detection point respectively to obtain first valid structural data corresponding to each group of structural data of each detection point; S3, respectively extracting N types of feature data from each group of first valid structural data, and performing feature fusion on all feature data to obtain comprehensive structural feature vector data corresponding to each detection point; N is a natural number not equal to 0; S4, training the health status assessment model using the comprehensive structural feature vector data and the structural health assessment data to obtain a trained health status assessment model; S5. Obtain M groups of structural data to be tested at the test points of the water conservancy project, input the M groups of structural data to be tested into the trained health status assessment model to obtain structural health assessment data, judge the safety status of the test points to be tested according to the structural health assessment data, and trigger corresponding danger warnings.

[0008] Further, the method of preprocessing each group of structural data of each detection point to obtain first valid structural data corresponding to each group of structural data of each detection point includes the following steps: S21, constructing a data curve graph of M groups of structural data of each detection point changing with time; S22, obtaining missing data positions according to each of the data curve graphs and supplementing the data at the missing data positions to obtain complete data; S23, using the moving average method to process each group of complete data to obtain smoothed data; S24. Using the standard score method, remove abnormal data in each group of smoothed data to obtain first valid structure data.

[0009] Further, obtaining the missing data position according to each of the data curve graphs and supplementing the data at the missing data position to obtain complete data includes: Obtaining missing data positions according to each of the data curve graphs; Determine the previous bit data and the next bit data adjacent to the missing data position; Construct the missing spline function according to the previous point data and the next point data; The missing data are calculated according to the missing spline function, and the missing data at the missing data position are supplemented to obtain complete data.

[0010] Further, the method of using the moving average method to process each group of complete data to obtain smoothed data includes: Preset the window length for moving average; Select data in a corresponding window length range starting from the first data of the complete data, and calculate the average value; Keep the window length unchanged and move one data point backward, calculate the average value again, and repeat this step until the window length contains the last data point in each group of complete data points. The smoothed data corresponding to the complete data are obtained according to all average values ​​corresponding to each group of complete data.

[0011] Furthermore, the method of using the standard score method to remove abnormal data in each group of smoothed data to obtain the first effective structure data includes: Set different data threshold ranges for different structural data; The standard score method is used to calculate the discrete degree of each data in each group of smoothed data; If the discrete degree exceeds a set data threshold range, the corresponding data is discarded to obtain the first valid structure data.

[0012] Furthermore, in S3, N types of feature data are extracted from each group of first valid structure data, including: Dividing each group of the first valid structure data according to a second preset time to obtain second valid structure data; N types of feature data are extracted from each group of second effective structure data respectively.

[0013] Further, the feature fusion of all feature data includes: Standardizing the characteristic data of each group of second effective structural data to obtain standard characteristic data; Constructing a covariance matrix according to the standard feature data; the covariance matrix is ​​a matrix; Calculate the eigenvalues ​​and corresponding eigenvectors of the covariance matrix, sort the eigenvalues ​​from large to small, select the eigenvectors corresponding to the first S eigenvalues ​​for feature fusion, and obtain comprehensive structural feature vector data; wherein S is a natural number not equal to 0, and S is less than N.

[0014] Furthermore, the feature vectors corresponding to the first S feature values ​​in the selection order are subjected to feature fusion to obtain comprehensive structural feature vector data, including: The characteristic data of the characteristic vector corresponding to the first S characteristic values ​​are defined as original data, and several target points of each original data are calculated respectively; the target points are other original data whose distance from the original data is less than a preset value; According to the original data and the target point, a weight function is constructed to optimize the weights of all target points to obtain the optimal weight of each target point; Obtaining the minimum projection value of each original data according to the optimal weight calculation; A low-dimensional vector is constructed according to the minimum projection value, and the low-dimensional vector is used as comprehensive structural feature vector data.

[0015] Further, judging the safety status of the detection point to be tested according to the structural health assessment data and triggering a corresponding danger warning includes: Preset the danger interval corresponding to each group of structural data respectively; When the health assessment data to be tested is located in a dangerous interval corresponding to any set of structural data, the detection point to be tested corresponding to the health assessment data is determined to be a dangerous area, and a corresponding danger warning is issued.

[0016] In a second aspect, the present invention provides a system for intelligent analysis and early warning of water conservancy project safety, comprising: Data processing module: used to obtain M groups of structural data of several detection points of the water conservancy project within the first preset time and structural health assessment data corresponding to each detection point; M is a natural number not equal to 0; pre-process each group of structural data of each detection point to obtain the first valid structural data corresponding to each group of structural data of each detection point; extract N types of feature data of each group of first valid structural data respectively, perform feature fusion on all feature data, and obtain comprehensive structural feature vector data corresponding to each detection point; N is a natural number not equal to 0; Model building module: connected with the data processing module, used to train the health status assessment model using the comprehensive structural feature vector data and the structural health assessment data to obtain a trained health status assessment model; Output module: connected to the model building module, used to obtain M groups of structural data to be tested at the test points of the water conservancy project, input the M groups of structural data to be tested into the trained health status assessment model to obtain structural health assessment data, judge the safety status of the test points to be tested according to the structural health assessment data, and trigger corresponding danger warnings.

[0017] The embodiments of the present invention have the following technical effects: This application effectively reduces random data changes by preprocessing structural data, improves data stability and reliability, and ensures data integrity; Among them, by extracting and fusing comprehensive features of structural data, the dynamic performance and health status of the structure can be fully reflected, thus improving the accuracy and reliability of the assessment; This application uses feature fusion and principal component analysis to perform feature dimensionality reduction, effectively reducing feature dimensions, reducing redundant information, improving the computational efficiency and generalization ability of the model, and enhancing the accuracy of the evaluation model; This application can effectively prevent structural failure and accidents by performing model health assessment and early warning on the collected real-time structural data, ensuring the safe operation of water conservancy projects, and providing strong technical support for ensuring the safe operation of water conservancy projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a flow chart of a method for intelligent analysis and early warning of water conservancy project safety provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0021] Example 1 Figure 1 Flow chart of a method for intelligent analysis and early warning of water conservancy project safety provided by an embodiment of the present invention. Figure 1 , including: S1. Obtain M groups of structural data of a plurality of detection points of a water conservancy project within a first preset time and structural health assessment data corresponding to each detection point; M is a natural number not equal to 0.

[0022] Exemplarily, the structural data in the present application may be strain data, displacement data, and vibration data. In this case, M is 3. This case is used as an example for the following description.

[0023] Strain data is the degree of deformation of the structure after being subjected to force, which is obtained through strain sensors; displacement data is the change in the position of the structure over time after being subjected to force, which is obtained through displacement sensors; vibration data is the data measured and recorded by vibration sensors when the structure vibrates under the action of external force.

[0024] The sensors in this application are all existing sensors in existing water conservancy projects, which reduces additional equipment and reduces costs. It should be noted that in this application, three data types, namely strain, displacement and vibration, are selected for description, which can better reflect the safety performance of water conservancy projects; in actual use, more structural data can be selected according to other situations, or these three data can be replaced to meet actual requirements, which is not limited in this application.

[0025] Structural health assessment data is data used to evaluate the status of water conservancy projects through monitoring and collection. Safety judgments are made based on this data. This data needs to be evaluated based on the collected structural data.

[0026] Regarding the acquisition of samples, in view of the particularity of water conservancy projects, samples in different time periods will show great differences. Therefore, this application collects data within a preset time period to avoid excessive calculations or reduced data value caused by too long or too short a time. For example, the structural data of different detection points within one month can be collected. Different time periods can also be preset for different detection points. This application uniformly presets the time for explanation. In addition, as time changes, the structure of water conservancy projects will undergo different changes. Therefore, it is very necessary to maintain uniformity within a certain time period, and the sample data of different detection points should be as much as possible to ensure the accuracy of the subsequent model.

[0027] S2. Preprocessing each group of structural data of each detection point respectively to obtain first valid structural data corresponding to each group of structural data of each detection point.

[0028] Specifically, preprocessing includes the following processes: S21, constructing a data curve graph of M groups of structural data of each detection point changing with time; For the above three types of structural data, data visualization is performed for each detection point and the corresponding curve graph is constructed.

[0029] S22, obtaining missing data positions according to each of the data curve graphs and supplementing the data at the missing data positions to obtain complete data; The blanks and abnormal areas in each data are identified as missing data through the curve graph, and the location of the missing data is recorded. The missing data may reduce the data quality, generate errors, and limit the training effect of the model, so it needs to be supplemented.

[0030] Obtaining missing data positions according to each of the data curve graphs; Determine the previous bit data and the next bit data adjacent to the missing data position; the before and after here are adjacent data defined from the perspective of time, the adjacent data collected before the missing data position is the previous bit data, and the adjacent data collected after it is the next bit data.

[0031] The missing spline function is constructed based on the previous point data and the next point data; in this application, a cubic missing spline function is used, as shown in the following formula: , in, , , and are the spline coefficients, is the coordinate horizontal axis value of the next point data. is the function variable, Represents the function value when the variable is x, and i represents different types of data structures.

[0032] The spline coefficients are determined by the boundary conditions, which are: , and , in, It is the coordinate horizontal axis value of the previous point data. is the coordinate horizontal axis value of the next point data. It is the vertical axis value of the coordinate of the previous point data. is the coordinate vertical axis value of the next point data. Finally, the missing data position can be supplemented with the missing spline function to obtain the complete data. represents the value of the second-order derivative when the variable is x0, The representative variable is x n The value of the second derivative when .

[0033] By performing the above steps on the strain data curve graph, the displacement data curve graph and the vibration data curve graph respectively, complete strain data, complete displacement data and complete vibration data can be obtained.

[0034] S23, using the moving average method to process each group of complete data to obtain smoothed data; The supplemented data are smoothed to improve data quality and obtain stable data, which specifically includes the following steps: Preset the window length for moving average; Select data in a corresponding window length range starting from the first data of the complete data, and calculate the average value; Keep the window length unchanged and move one data backward, calculate the average value again, and repeat this step until the last data in each group of complete data is included in the window length range, and obtain the smoothed data corresponding to the complete data according to all average values ​​corresponding to each group of complete data.

[0035] S24. Using the standard score method, remove abnormal data in each group of smoothed data to obtain first valid structure data.

[0036] The specific steps include: Setting different data threshold ranges for different structural data; such as strain data threshold range, displacement data threshold range and vibration data threshold range; The standard score method is used to calculate the discrete degree of each data in each group of smoothed data; If the discrete degree exceeds a set data threshold range, the corresponding data is discarded to obtain the first valid structure data.

[0037] Specifically, the standard score method, namely the Z-score method, is used to remove outliers from the smoothed data. The calculation formula is as follows: ; in, Indicates the first The degree of dispersion of the data, Indicates the first data, represents the average, represents the standard deviation, according to Compare with the preset data threshold range and Eliminate data that is larger than a preset data threshold range to obtain first valid structure data; It should be noted that the strain data threshold range, the displacement data threshold range and the vibration data threshold range can all be set according to actual conditions.

[0038] In summary, the preprocessing of the collected structural data as samples has been completed, which effectively reduces the random variation of the data, improves the stability and reliability of the data, and ensures the integrity of the data.

[0039] S3. Extract N types of feature data from each group of first effective structural data respectively, perform feature fusion on all feature data, and obtain comprehensive structural feature vector data corresponding to each detection point; N is a natural number not equal to 0.

[0040] (1) Feature extraction Dividing each group of the first valid structure data according to a second preset time to obtain second valid structure data; N types of feature data are extracted from each group of second effective structure data respectively.

[0041] Exemplarily, the second preset time may be 24 hours, that is, the feature acquisition of the second effective structure data is performed in units of one day, thereby improving the accuracy and quality of the feature acquisition data.

[0042] When the above three types of structural data are subjected to feature extraction, they are divided according to the second preset time to become second effective strain structural data, second effective displacement structural data, and second effective vibration structural data, respectively. The following contents can be extracted respectively: The strain feature data of each group of second effective strain structure data are extracted respectively. The strain feature data include the first time domain feature and the change rate feature. The first time domain feature includes the maximum value, minimum value, mean value and standard deviation of the effective strain data; the change rate feature is the slope of the line connecting adjacent data points in the effective strain data; therefore, there are 5 types of features in total.

[0043] The displacement feature data of each group of second effective displacement structure data are extracted respectively, and the displacement feature data include second time domain features and cumulative features. The second time domain features include the maximum displacement, minimum displacement, mean and standard deviation of the effective displacement data, and the cumulative features are the cumulative sum of the effective displacement data over time; therefore, there are 5 types of features in total.

[0044] Each group of second effective vibration structure data is transformed into frequency domain to obtain frequency domain components, and vibration feature data is extracted according to the frequency domain components. The vibration feature data includes peak frequency features, frequency features and frequency band energy features. Therefore, there are only 3 types of features. In order to unify with the previous ones, the remaining two types of features need to be padded with 0 to become 5 types.

[0045] Among them, the frequency domain processing formula is: , in, represents the transformed frequency domain data, is the corresponding effective vibration data, L is the data volume, is a plural unit, is the frequency index, t is the time index, is the frequency domain parameter, K is the frequency domain bias, analyze the spectrum after frequency domain processing, extract the peak frequency characteristics, frequency characteristics and band energy characteristics, Indicates Amount of data.

[0046] It should be noted that the various features extracted above are all conventionally acquired features, and this application will not explain them in detail. Other structural data can be feature extracted according to their actual conditions.

[0047] (2) Feature Fusion Standardizing the characteristic data of each group of second effective structural data to obtain standard characteristic data; A covariance matrix is ​​constructed based on the standard feature data; the covariance matrix is Matrix of Calculate the eigenvalues ​​and corresponding eigenvectors of the covariance matrix, sort the eigenvalues ​​from large to small, select the eigenvectors corresponding to the first S eigenvalues ​​for feature fusion, and obtain comprehensive structural feature vector data; wherein S is a natural number not equal to 0, and S is less than N.

[0048] Specifically, based on the setting of the aforementioned second preset time, the mean and standard deviation of the displacement feature data are standardized to 0 and 1, so as to facilitate subsequent fusion.

[0049] Within the first preset time, the average value of all data of various types of standard characteristic data is calculated. For example, based on the above settings, the average value of all minimum displacement data in the displacement characteristic data within the first preset time is calculated, and then the variance is calculated based on the average value; similarly, the mean is calculated, and the covariance is obtained based on the mean calculation.

[0050] The covariance matrix is ​​constructed based on the variance and covariance, where the diagonal elements are the variances and the off-diagonal elements are the corresponding covariance values.

[0051] Perform eigenvalue decomposition on the feature covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. Define a new coordinate system for these eigenvectors, where each coordinate axis corresponds to a principal component in the data. Sorting is performed in reverse order according to the size of the eigenvalues, and the eigenvectors corresponding to the first S eigenvalues ​​are selected for feature fusion.

[0052] Among them, feature fusion is dimensionality reduction, which can be achieved by the following methods: The feature data of the feature vectors corresponding to the first S eigenvalues, that is, the feature data before standardization processing, are obtained and defined as the original data. Several target points of each original data are calculated respectively; the target point is other original data whose distance from the original data is less than a preset value; the Euclidean distance formula can be used to calculate several target points, where the target point is the neighboring point, and the preset value is set according to actual needs; it should be noted that S needs to be selected according to the dimension after dimensionality reduction, and usually the order of dimensionality reduction is small, only 1-3 orders.

[0053] Construct a weight function based on the original data and the target point to optimize the weights of all target points and obtain the optimal weight corresponding to each target point; By minimizing the weight function of the local weight, the optimal weight of each target point is calculated. The weight function is as follows: , in, represents the weight, Representative The original data, Representative The original data target points, is the total number of original data, and the optimal weight meets the conditions: .

[0054] The weight function can reflect the "importance" or "correlation" between a data point and its target point.

[0055] The projection of the original data in the low-dimensional space is calculated according to the optimal weight. The minimum projection function of the projection of the original data in the low-dimensional space is minimized to obtain the minimum projection value of each original data in the low-dimensional space. The minimum projection function is: , in, Representative The projection of the original data in the low-dimensional space, Representative The first target points, is the magnitude of the selected eigenvector.

[0056] The goal of optimization is to find the optimal low-dimensional space projection, which can best preserve the local structure of the original data in the high-dimensional space. The low-dimensional vector obtained by the dimensionality reduction projection is composed of the minimum projection value of each original data. This vector is the comprehensive structural feature vector data.

[0057] In summary, feature fusion and principal component analysis are used to reduce feature dimensionality, effectively reducing feature dimensions and redundant information, and obtaining the best training data set, namely, comprehensive structural feature vector data. Using this data to train the model can improve the computational efficiency and generalization ability of the model and enhance the accuracy of the evaluation model.

[0058] S4. Train the health status assessment model using the comprehensive structural feature vector data and the structural health assessment data to obtain a trained health status assessment model.

[0059] Exemplarily, the comprehensive structural feature vector data and the structural health assessment data can be divided into a test set and a training set, the comprehensive structural feature vector is used as the input of the health status assessment model, the structural health assessment data is used as the output, the health status assessment model is constructed based on the training set, and the health status assessment model is evaluated through the test set according to the root mean square error formula, and the optimal health status assessment model is obtained by taking the smaller the root mean square error, the better the model is as the evaluation standard; The health status assessment model can be any one of the support vector machine model, random forest model or gradient boosting decision tree model, and the root mean square error formula is: , in, is the root mean square error, is the sample size, It is Structural health assessment data, It is Structural health assessment data calculated by the model.

[0060] In summary, a health status assessment model is constructed, and the input structural data can output structural health assessment data.

[0061] S5. Obtain M groups of structural data to be tested at the test points of the water conservancy project, input the M groups of structural data to be tested into the trained health status assessment model to obtain structural health assessment data, judge the safety status of the test points to be tested according to the structural health assessment data, and trigger corresponding danger warnings.

[0062] The danger intervals corresponding to the groups of structural data are preset respectively; when the health assessment data to be tested is located in the danger interval corresponding to any group of structural data, the detection point to be tested corresponding to the health assessment data is determined to be a danger zone, and a corresponding danger warning is issued.

[0063] Exemplarily, a strain danger interval, a displacement danger interval and a vibration danger interval are preset respectively; When the health assessment data to be tested is in the strain danger zone, the detection point is determined to be a strain danger zone and a strain danger warning is issued; When the health assessment data to be tested is in the displacement danger zone, the monitoring point is determined to be a displacement danger zone and a displacement danger warning is issued; When the health assessment data to be tested is in the vibration danger range, the detection point is determined to be a vibration danger area and a vibration danger warning is issued.

[0064] It should be noted that the strain danger zone, displacement danger zone and vibration danger zone can be set according to actual conditions to meet the needs.

[0065] Based on the above technical solution, this application comprehensively reflects the dynamic performance and health status of the water conservancy project structure, and provides accurate early warning, providing strong technical support for ensuring the safe operation of the water conservancy project.

[0066] Example 2 Based on the content of Example 1, the present application provides a system for intelligent analysis and early warning of water conservancy project safety, including: Data processing module: used to obtain M groups of structural data of several detection points of the water conservancy project within the first preset time and structural health assessment data corresponding to each detection point; M is a natural number not equal to 0; pre-process each group of structural data of each detection point to obtain the first valid structural data corresponding to each group of structural data of each detection point; extract N types of feature data of each group of first valid structural data respectively, perform feature fusion on all feature data, and obtain comprehensive structural feature vector data corresponding to each detection point; N is a natural number not equal to 0; Model building module: connected with the data processing module, used to train the health status assessment model using the comprehensive structural feature vector data and the structural health assessment data to obtain a trained health status assessment model; Output module: connected to the model building module, used to obtain M groups of structural data to be tested at the test points of the water conservancy project, input the M groups of structural data to be tested into the trained health status assessment model to obtain structural health assessment data, judge the safety status of the test points to be tested according to the structural health assessment data, and trigger corresponding danger warnings.

[0067] For other contents, please refer to Example 1 and will not be repeated here.

[0068] Example 3 The present application provides an electronic device, the electronic device includes a processor and a memory for storing instructions executable by the processor; when the processor is configured to execute the instructions, the electronic device implements the method in Embodiment 1. The electronic device may be a tablet computer, a handheld computer, a personal computer PC, or other electronic device. The electronic device may directly display the calibration result, which is convenient for the operator to directly view.

[0069] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When the use is implemented in whole or in part in the form of a computer program product, the computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL) or wireless (e.g., infrared, wireless, microwave, etc.) mode) to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk Solid State Disk (SSD)), etc.

[0070] It should be noted that the terms used in the present invention are only for describing specific embodiments, rather than limiting the scope of the present application. As shown in the present specification, unless the context clearly indicates an exception, the words "one", "a", "a kind of" and / or "the" do not specifically refer to the singular, but may also include the plural. The terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of more restrictions, the elements defined by the sentence "include one..." do not exclude the presence of other identical elements in the process, method or device including the elements.

[0071] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying 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. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. 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 an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent analysis and early warning of water conservancy project safety, characterized in that: The steps include: S1. Obtain M groups of structural data of several detection points of the water conservancy project within a first preset time and structural health assessment data corresponding to each detection point; M is a natural number not equal to 0; S2, pre-processing each group of structural data of each detection point respectively to obtain first valid structural data corresponding to each group of structural data of each detection point; S3, respectively extracting N types of feature data of each group of first effective structural data, performing feature fusion on all feature data, and obtaining comprehensive structural feature vector data corresponding to each detection point; N is a natural number not equal to 0; S4, training the health status assessment model using the comprehensive structural feature vector data and the structural health assessment data to obtain a trained health status assessment model; S5. Obtain M groups of structural data to be tested at the test points of the water conservancy project, input the M groups of structural data to be tested into the trained health status assessment model to obtain structural health assessment data, judge the safety status of the test points to be tested according to the structural health assessment data, and trigger corresponding danger warnings.

2. The method for intelligent analysis and early warning of water conservancy project safety according to claim 1 is characterized in that: The method of preprocessing each group of structural data of each detection point to obtain first valid structural data corresponding to each group of structural data of each detection point comprises the following steps: S21, constructing a data curve graph of M groups of structural data of each detection point changing with time; S22, obtaining missing data positions according to each of the data curve graphs and supplementing the data at the missing data positions to obtain complete data; S23, using the moving average method to process each group of complete data to obtain smoothed data; S24. Using the standard score method, remove abnormal data in each group of smoothed data to obtain first valid structure data.

3. The method for intelligent analysis and early warning of water conservancy project safety according to claim 2 is characterized in that: The step of obtaining the missing data position according to each of the data curve graphs and supplementing the data at the missing data position to obtain complete data includes: Obtaining missing data positions according to each of the data curve graphs; Determine the previous bit data and the next bit data adjacent to the missing data position; Construct the missing spline function according to the previous point data and the next point data; The missing data are calculated according to the missing spline function, and the missing data at the missing data position are supplemented to obtain complete data.

4. The method for intelligent analysis and early warning of water conservancy project safety according to claim 3 is characterized in that: The method of using the moving average method to process each group of complete data to obtain smoothed data includes: Preset the window length for moving average; Select data in a corresponding window length range starting from the first data of the complete data, and calculate the average value; Keep the window length unchanged and move one data point backward, calculate the average value again, and repeat this step until the window length contains the last data point in each group of complete data points. The smoothed data corresponding to the complete data are obtained according to all average values ​​corresponding to each group of complete data.

5. The method for intelligent analysis and early warning of water conservancy project safety according to claim 2 is characterized in that: The method of using the standard score method to remove abnormal data in each group of smoothed data to obtain first valid structure data includes: Set different data threshold ranges for different structural data; The standard score method is used to calculate the discrete degree of each data in each group of smoothed data; If the discrete degree exceeds a set data threshold range, the corresponding data is discarded to obtain the first valid structure data.

6. The method for intelligent analysis and early warning of water conservancy project safety according to claim 1 is characterized in that: In S3, N types of feature data of each group of first valid structure data are extracted respectively, including: Dividing each group of the first valid structure data according to a second preset time to obtain second valid structure data; N types of feature data are extracted from each group of second effective structure data.

7. A method for intelligent analysis and early warning of water conservancy project safety according to claim 6, characterized in that: The feature fusion of all feature data includes: Standardizing the characteristic data of each group of second effective structural data to obtain standard characteristic data; A covariance matrix is ​​constructed based on the standard feature data; the covariance matrix is Matrix of Calculate the eigenvalues ​​and corresponding eigenvectors of the covariance matrix, sort the eigenvalues ​​from large to small, select the eigenvectors corresponding to the first S eigenvalues ​​for feature fusion, and obtain comprehensive structural feature vector data; wherein S is a natural number not equal to 0, and S is less than N.

8. The method for intelligent analysis and early warning of water conservancy project safety according to claim 7 is characterized in that: The feature vectors corresponding to the first S feature values ​​in the selection order are subjected to feature fusion to obtain comprehensive structural feature vector data, including: The characteristic data of the characteristic vector corresponding to the first S characteristic values ​​are defined as original data, and several target points of each original data are calculated respectively; the target points are other original data whose distance from the original data is less than a preset value; Construct a weight function based on the original data and the target point to optimize the weights of all target points and obtain the optimal weight of each target point; Obtaining the minimum projection value of each original data according to the optimal weight calculation; A low-dimensional vector is constructed according to the minimum projection value, and the low-dimensional vector is used as comprehensive structural feature vector data.

9. The method for intelligent analysis and early warning of water conservancy project safety according to claim 1 is characterized in that: The method of judging the safety status of the test points to be tested based on the structural health assessment data and triggering corresponding danger warnings includes: Preset the danger interval corresponding to each group of structural data respectively; When the health assessment data to be tested is located in a dangerous interval corresponding to any set of structural data, the detection point to be tested corresponding to the health assessment data is determined to be a dangerous area, and a corresponding danger warning is issued.

10. A system for intelligent analysis and early warning of water conservancy project safety, characterized in that: include: Data processing module: used to obtain M groups of structural data of several detection points of the water conservancy project within the first preset time and structural health assessment data corresponding to each detection point; M is a natural number not equal to 0; pre-process each group of structural data of each detection point respectively to obtain first valid structural data corresponding to each group of structural data of each detection point; Extract N types of feature data of each group of first effective structural data respectively, perform feature fusion on all feature data, and obtain comprehensive structural feature vector data corresponding to each detection point; N is a natural number not equal to 0; Model building module: connected with the data processing module, used to train the health status assessment model using the comprehensive structural feature vector data and the structural health assessment data to obtain a trained health status assessment model; Output module: connected to the model building module, used to obtain M groups of structural data to be tested at the test points of the water conservancy project, input the M groups of structural data to be tested into the trained health status assessment model to obtain structural health assessment data, judge the safety status of the test points to be tested according to the structural health assessment data, and trigger corresponding danger warnings.

Images