Water conservancy project construction anti-seepage detection method
By obtaining and analyzing observation data of water conservancy construction buildings in real time, screening related data and clustering error values, the accuracy and real-time problems of anti-seepage monitoring in the existing technology are solved, and the accuracy of seepage flow prediction and anti-seepage monitoring are improved. effect.
Patent Information
- Application Number
- CN202411698642.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-26
- Publication Date
- 2024-12-27
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing anti-seepage monitoring methods for water conservancy construction rely on manual inspection and regular inspection, making it difficult to achieve real-time monitoring and early warning. In addition, the automated monitoring system has poor accuracy in predicting seepage changes, and is greatly affected by human factors and multiple factors.
By obtaining observation data and seepage data of water conservancy construction buildings in real time, screening related data, performing error value clustering analysis, obtaining abnormal characteristic values and deviation indexes, predicting seepage data in a comprehensive degree of importance, and improving the accuracy of anti-seepage monitoring.
It improves the accuracy of seepage flow prediction, enhances the accuracy of anti-seepage monitoring results, reduces the influence of human factors, and realizes timely identification of leakage abnormalities.
Smart Images
Figure CN119202909A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of water conservancy project data processing, and in particular to a water conservancy project construction anti-seepage detection method. Background Art
[0002] In the field of water conservancy construction and building anti-seepage monitoring, ensuring the safety and stability of hydraulic structures is a vital task. Traditional anti-seepage monitoring methods mainly rely on manual inspections and regular testing, which are not only time-consuming and labor-intensive, but also often difficult to achieve real-time monitoring and early warning of seepage conditions due to the limitation of monitoring frequency. In addition, manual monitoring is easily affected by human factors, resulting in insufficient accuracy and reliability of monitoring data.
[0003] With the development of technology, although some automated monitoring methods have emerged, these technologies still have limitations. Existing automated monitoring systems usually monitor the changes in seepage in real time, combine the relationship between multiple factors and seepage changes, predict the change trend of seepage, and achieve monitoring. However, these factors that affect the change in seepage have different degrees of influence on the seepage. Using factors that have a smaller impact on the change in seepage to predict the change in seepage will lead to a large deviation in the analysis results, making the accuracy of anti-seepage monitoring poor. Summary of the invention
[0004] In order to solve the above technical problems, the present application provides a water conservancy project construction anti-seepage detection method to solve the existing problems.
[0005] A water conservancy project construction anti-seepage detection method of the present application adopts the following technical solution: An embodiment of the present application provides a method for detecting water conservancy project construction seepage, the method comprising the following steps: S1, real-time acquisition of observation values and seepage data of various observation data of water conservancy construction buildings at each monitoring point; S2, according to the correlation between each group of observation values of the observation data and the seepage data, select each item of related data from all items of observation data; S3, for each associated data, the error value is obtained according to the local difference of the change trend of each group of observations, and the error value is clustered to screen the abnormal clusters; the abnormal characteristic value of each group of observations is obtained according to the data discreteness of each group of observations, the overall change trend and the value of the error value in its abnormal cluster; S4, according to the difference in the change trend of each group of observations and all groups of observations and the difference in the values at the same time, the deviation index of each group of observations is obtained; after denoising each group of observations according to the frequency domain components of each group of observations, outliers are screened; according to the continuity of the distribution of outliers, the outlier index of each outlier is obtained; S5, comprehensively considering the deviation index, abnormal characteristic value and abnormal index of all abnormal values of each group of observation values, to obtain the importance of each group of observation values; predict the seepage data according to the importance of all groups of observation values of each associated data, and obtain the anti-seepage monitoring result by comparing the predicted value with the seepage data obtained in real time.
[0006] Furthermore, the method for screening each item of associated data includes: For each observation data, each group of observation values of the observation data is taken as the independent variable, and the seepage data is taken as the dependent variable, and a linear correlation diagram of each group of observation values and the seepage data is drawn; according to the linear distribution of all data points in the linear correlation diagram, the correlation between each group of observation values and the seepage data is obtained; The mean of the correlation between all group observation values of each observation data and the seepage data is taken as the influencing factor of each observation data; the observation data with an influencing factor less than the preset correlation threshold are eliminated, and the remaining observation data are recorded as the associated data.
[0007] Furthermore, the method for obtaining the correlation between each group of observation values and the seepage data includes: A straight line is fitted to all data points in the linear correlation diagram to obtain a fitted straight line; the determination coefficient of the fitted straight line is calculated as the correlation between each group of observations and the seepage data.
[0008] Furthermore, the error values are obtained according to the local differences in the change trends of the observation values of each group, and cluster analysis is performed on the error values to screen abnormal clusters, including: For each group of observation values of each associated data, the difference between the observation value at each moment and the previous moment is calculated, and the ratio of the observation value difference to the observation value at each moment is used as the growth rate at each moment; a neighborhood interval centered at each moment is constructed; Calculate the average growth rate of all moments in the neighborhood interval of any moment; if the growth rate at any moment is greater than the average growth rate, the observed value at any moment is the error value; All error values are clustered according to the time differences between the error values to obtain clusters; the mean time differences of all pairwise combinations of error values in each cluster are calculated; and the cluster with the smallest mean time difference is recorded as an abnormal cluster.
[0009] Furthermore, the method for obtaining the abnormal characteristic value of each group of observation values includes: in, is the abnormal characteristic value of the observation value of group a; is the standard deviation of the observations in group a; is the absolute value of the slope of the fitted straight line of the a-th group of observations; is the growth rate of the i-th error value in the abnormal cluster of the a-th group of observations; I is the number of error values in the abnormal cluster of the a-th group of observations; is the maximum growth rate of all error values in the abnormal cluster of group a observations; is a linear normalization function.
[0010] Furthermore, the method for obtaining the deviation index of each group of observation values includes: recording the deviation index of the ath group of observation values as , ;in, is the absolute value of the slope of the fitted straight line of the a-th group of observations, is the mean of the absolute values of the slopes of the fitted lines of all groups of observations; is the tth observation in the ath group of observations, is the mean of the t-th observation in all groups of observations, and T is the number of observations in each group of observations.
[0011] Further, after denoising each group of observation values according to the frequency domain components of each group of observation values, outliers are screened, including: Perform fast Fourier transform on each group of observations to obtain high-frequency components and low-frequency components; perform inverse transform on all low-frequency components to obtain each group of denoised observations; according to the method for obtaining the error value in each group of observations, obtain the error value in each group of denoised observations and record it as an outlier.
[0012] Furthermore, the abnormality index of each abnormal value is obtained according to the continuity of the abnormal value distribution, including: The outliers that are continuous at the same time are grouped together; the inverse of the continuous duration of each group of outliers is calculated, and the difference between the value 1 and the inverse is used as the outlier index of each outlier in each group of outliers; the outlier index of each individual outlier is recorded as 0.
[0013] Furthermore, the method for obtaining the importance of each group of observation values includes: ; is the importance of the observations in group a, is the deviation index of the observations in group a, is the anomaly index of the xth outlier in the ath group of observations, is the abnormal characteristic value of the a-th group of observations, and X is the number of abnormal values of the a-th group of observations.
[0014] Furthermore, the seepage data is predicted according to the importance of all groups of observation values of each associated data, and the anti-seepage monitoring result is obtained by comparing the predicted value obtained by prediction with the seepage data obtained in real time, including: Obtain a group of observations with the greatest importance among all group observations of each associated data item as the characteristic data of each associated data item; The characteristic data of all items of associated data are taken as independent variables, and the seepage data are taken as dependent variables to construct a multiple linear regression model; wherein the importance of the characteristic data of each item of associated data is taken as the regression coefficient of each item of associated data in the multiple linear regression model; The mean of the observation values at all monitoring points of each associated data at the current moment is taken as the average observation value of each associated data; the average observation value of all associated data at the current moment is predicted using a multivariate linear regression model to obtain the predicted seepage flow at the current moment; if the seepage flow data obtained in real time is greater than the predicted seepage flow value, abnormal leakage occurs in the water conservancy construction building.
[0015] This application has at least the following beneficial effects: The present application analyzes the influence of different factors on the seepage volume change. In order to improve the prediction accuracy of the seepage volume, the application first analyzes the correlation between various observation data and the seepage volume data, screens out some influencing factors with less correlation, and retains only factors that have a significant impact on the seepage volume; further analyzes the fluctuation of various observation data, and obtains abnormal characteristic values to reflect the abnormal characteristics of leakage; further analyzes the similarity of the observation values of the same related data at different monitoring points at the same time, calculates the deviation index of each group of observation values, and amplifies the abnormality of those data that are significantly different from the whole through the difference performance in two dimensions of numerical value and change trend, thereby avoiding misjudgment caused by a single factor; finally, the importance of each group of observation values is analyzed based on the above factors, and the seepage volume data is predicted according to the importance of all groups of observation values of each related data. For factors with higher abnormality, higher weights are given in the seepage volume prediction process, thereby improving the prediction accuracy of the seepage volume, and then improving the accuracy of the anti-seepage monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0017] Figure 1 A flow chart of the steps of a water conservancy project construction anti-seepage detection method provided in this application; Figure 2 A block diagram for filtering various associated data provided for one embodiment of the present application. DETAILED DESCRIPTION
[0018] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0019] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
[0020] The specific scheme of the anti-seepage detection method for water conservancy project construction provided by the present application is described in detail below with reference to the accompanying drawings.
[0021] An embodiment of the present application provides a method for detecting water conservancy project construction seepage prevention. Specifically, the following method for detecting water conservancy project construction seepage prevention is provided. Figure 1 , the method comprises the following steps: S1, real-time acquisition of observation values and seepage data of various observation data of water conservancy construction buildings at each monitoring point.
[0022] Seepage observation is usually carried out by obtaining various observation data of water conservancy construction buildings that may have an impact on it, including: infiltration line position, seepage pressure, reservoir water level, etc., to monitor the seepage water volume of water conservancy construction buildings such as dams and reservoirs. Taking the dam as an example, each observation data obtains observation values at multiple monitoring points to obtain multiple groups of measurement data; in this embodiment, each observation data has a total of 20 groups of measurement data. Each group of measurement data is time series data with a period of one month and a frequency of once per hour. At the same time, the seepage of the dam is obtained in real time using an infiltration meter with the same frequency and period as the observation data to obtain a set of seepage data.
[0023] S2, according to the correlation degree between each group of observation values of the observation data and the seepage data, select each item of related data from all items of observation data.
[0024] The core goal of anti-seepage monitoring is to ensure the safe and stable operation of the dam. Seepage is a key parameter in anti-seepage monitoring, which directly reflects the anti-seepage effect of water conservancy construction. Changes in seepage can be caused by a variety of factors, including fluctuations in reservoir water levels, changes in the infiltration line, and seepage pressure on the dam body. Therefore, during the monitoring process, continuous observation of seepage can provide direct information about these internal and external changes, help engineers evaluate the effectiveness of existing anti-seepage measures, and guide necessary maintenance and reinforcement work.
[0025] By continuously tracking key performance indicators, potential risks can be quickly identified and responded to, and necessary preventive or remedial measures can be implemented. In the process of water conservancy construction and anti-seepage monitoring, a variety of monitoring indicators are involved. It is very important to rationally screen indicators with key factors to avoid unnecessary complex calculations in subsequent data analysis. At the same time, excluding irrelevant factors can improve the accuracy of permeability analysis.
[0026] Therefore, this embodiment first preliminarily screens the observation data that have a greater impact on the monitoring of seepage changes based on the correlation between the changes in each observation data and the seepage data. Specifically, for each observation data, each group of observation values of the observation data is used as an independent variable, and the seepage data is used as a dependent variable to draw a linear correlation diagram between each group of observation values and the seepage data; the process of drawing the linear correlation diagram is a well-known technique and will not be repeated. Use the least squares method to perform a straight line fit on all data points in the linear correlation diagram to obtain a fitted straight line; calculate the determination coefficient of the fitted straight line as the correlation between each group of observation values and the seepage data; the determination coefficient is one of the indicators for evaluating the goodness of fit of the straight line fit, and its calculation process is a well-known technique and will not be repeated. The range of the determination coefficient is [0,1]; the closer it is to 1, the better the regression model can explain the variation of the dependent variable, and the better the fitting effect; the closer it is to 0, the worse the fitting effect.
[0027] The correlation mean of all groups of observation values of each observation data and the seepage data is taken as the influencing factor of each observation data; the observation data with an influencing factor less than the preset correlation threshold of 0.4 are eliminated, and the remaining observation data are recorded as the associated data. Screening out some influencing factors with less correlation and retaining only factors that have a significant impact on the seepage will help simplify the seepage prediction model.
[0028] The selected correlation data has a similar change trend to the seepage volume, so the change trend of the correlation data has a guiding role in the prediction of the change of the seepage volume. In order to analyze the change of the correlation data, for each correlation data, under normal circumstances, the observation data of the dam will fluctuate within a specific range, which reflects the natural dynamic balance of groundwater flow. This fluctuation usually corresponds to slight natural hydrological cycle changes such as light rain or seasonal temperature changes.
[0029] The filter box diagram of each associated data, such as Figure 2 shown.
[0030] S3, for each associated data, the error value is obtained according to the local difference of the change trend of each group of observations, and the error value is clustered to screen the abnormal clusters; according to the data discreteness of each group of observations, the overall change trend and the value of the error value in its abnormal cluster, the abnormal characteristic value of each group of observations is obtained.
[0031] For each associated data, take the position of the seepage line, which usually has a greater impact on the seepage volume, as an example. If the water level of the seepage line continues to rise or fall abnormally quickly, this may indicate that the seepage conditions have changed. The reasons may include cracks inside the dam body or foundation, changes in the leakage channel, or damage to the anti-seepage system, which manifests as abnormal leakage. Therefore, this embodiment performs stability analysis on the data of each cycle.
[0032] First, for each group of observations of each associated data, calculate the difference between the observations at each moment and the previous moment, and take the ratio of the difference between the observations and the observations at each moment as the growth rate at each moment; construct a neighborhood interval centered at each moment; in this embodiment, the size of the neighborhood interval is the central moment and 7 moments before and after it, for a total of 15 moments; it should be noted that when the central moment is at the beginning and end of the cycle, the insufficient data in the neighborhood interval is supplemented by the mean filling method. Calculate the mean growth rate of all moments in the neighborhood interval of any moment; if the growth rate at any moment is greater than the mean growth rate, the observation value at any moment is the error value.
[0033] Furthermore, density clustering is performed according to the time interval of the error values. In this embodiment, the DBSCAN clustering algorithm is used to cluster all error values according to the time difference between the error values; wherein the parameters of the DBSCAN clustering algorithm are: the neighborhood radius is 0.5, the minimum number of neighboring points is 4, and each cluster is obtained.
[0034] Calculate the time difference mean of the error values of all pairwise combinations in each cluster; record the cluster with the smallest time difference mean as the abnormal cluster; the smaller the time difference mean, the faster the corresponding observation value grows in a short period of time and the longer the duration, the more likely it is that an abnormality exists.
[0035] in, is the abnormal characteristic value of the observation value of group a; is the standard deviation of the observations in group a; is the absolute value of the slope of the fitted straight line of the a-th group of observations; is the growth rate of the i-th error value in the abnormal cluster of the a-th group of observations; I is the number of error values in the abnormal cluster of the a-th group of observations; is the maximum growth rate of all error values in the abnormal cluster of group a observations; is a linear normalization function.
[0036] It should be noted that the standard deviation represents the fluctuation of the observed value, and smaller fluctuations usually indicate higher stability; the smaller the slope of the fitting line is and the closer it is to zero, indicating that the state is relatively stable; for the growth rate of each data point in the cycle, the smaller the value is and the more consistent the change is, the more worthy of reference the authenticity of the data point is, and the greater the difference between the error value and the maximum value is, indicating a persistent high water level, and the greater the degree of abnormality.
[0037] S4, according to the difference in the changing trend of each group of observations and all groups of observations and the difference in values at the same time, the deviation index of each group of observations is obtained; after denoising each group of observations according to the frequency domain components of each group of observations, outliers are screened; according to the continuity of the outlier distribution, the outlier index of each outlier is obtained.
[0038] In order to further evaluate the importance of each associated data, the same associated data collected at different monitoring points at the same time is compared in parallel to identify and consider the impact of local mutation factors. At the same time, since data fluctuations may be caused by random noise, which shows obvious randomness in multi-dimensional data, the stability of parallel data at the same time is compared to eliminate the impact of noise.
[0039] Based on the above analysis, the observed values of the same type of associated data at different monitoring points at the same time should be consistent. The anti-seepage monitoring target analyzed in this embodiment is a dam, with the same geological conditions and monitoring points located in the same geological unit, such as similar soil types, permeability and groundwater flow characteristics. Then the infiltration lines of these points should show high consistency. If the infiltration line of a certain monitoring point fluctuates separately, this may be due to errors caused by random factors, such as local minor geological differences, measurement errors or equipment failures. Under normal circumstances, the change trend of each monitoring point within a period should be consistent.
[0040] In order to distinguish whether each error value is caused by the noise of the acquisition instrument or the abnormal phenomenon, this embodiment analyzes that the fluctuation of the instrument noise is random, and the leakage abnormality may persist in time and the value is high. The deviation index of each group of observations is calculated, and the deviation index of the observation value of group a is recorded as ; , is the deviation index of the observations in group a, is the absolute value of the slope of the fitted straight line of the a-th group of observations, is the mean of the absolute values of the slopes of the fitted lines for all groups of observations. is the tth observation in the ath group of observations, is the mean of the t-th observation in all groups of observations, and T is the number of observations in each group of observations.
[0041] in, The smaller it is, the data of all monitoring points are basically consistent, and the difference with the value of a single monitoring point is not big. The data of the monitoring point is highly consistent with the overall trend, and the possibility of abnormality is small. The difference between the change of data at a single monitoring point and the overall change is measured. The smaller the difference, the more consistent the overall change trend of the associated data is, and the greater the possibility of normal performance. The final deviation index is expressed through the difference in two dimensions, namely value and slope, to amplify the abnormality of data that are significantly different from the overall data, thus avoiding misjudgment caused by a single factor. For example, the infiltration line value of a monitoring point may be abnormal due to random fluctuations, but its slope is consistent with the overall data. In this case, it is not judged as abnormal, reducing the possibility of misjudgment.
[0042] This embodiment uses Fourier transform to distinguish real signals from noise, performs fast Fourier transform on each set of observations, converts time series data into frequency domain data, and identifies frequency components for data in the frequency domain. Usually, noise is mostly manifested as high-frequency components, which are uniform background or multiple small peaks on the spectrum.
[0043] Based on the results of frequency analysis, only high-frequency components are converted back to the time domain through inverse Fourier transform to obtain each group of denoised observations; according to the method of obtaining the error value in each group of observations, the error value in each group of denoised observations is obtained and recorded as an outlier; the outliers with continuous moments are taken as a group; that is, each group of outliers contains at least two outliers, and the more continuous moments in each group, the greater the possibility of numerical abnormality. The reciprocal of the continuous duration of each group of outliers is calculated, and the difference between the value 1 and the reciprocal is used as the anomaly index of each outlier in each group of outliers. The anomaly index of the remaining discontinuous outliers is 0.
[0044] S5, comprehensively considering the deviation index, abnormal characteristic value and abnormal index of all abnormal values of each group of observation values, to obtain the importance of each group of observation values; predict the seepage data according to the importance of all groups of observation values of each associated data, and obtain the anti-seepage monitoring result by comparing the predicted value with the seepage data obtained in real time.
[0045] In order to analyze the abnormality of each group of observations, the importance of each associated data is analyzed according to the abnormality of the observations. This embodiment comprehensively analyzes the deviation index, abnormal characteristic value and abnormal index of the abnormal value of each group of observations to obtain the importance of each group of observations. The importance of the observations in group a is recorded as
[0046] ; is the importance of the observations in group a, is the deviation index of the observations in group a, is the anomaly index of the xth outlier in the ath group of observations, is the abnormal characteristic value of the a-th group of observations, and X is the number of abnormal values of the a-th group of observations.
[0047] In the above formula, the abnormal index of all abnormal values is combined, and the abnormal degree of the group of observations is judged by the continuity of the abnormal values. The larger the abnormal index, the greater the continuity of the abnormal values, and the more likely the observation abnormality will occur; the deviation index and the abnormal characteristic value are combined to comprehensively evaluate the abnormal degree of the group of observations; the greater the abnormal degree, the greater the degree of attention to the data. This makes the observations show a higher importance in weight, which helps to identify leakage anomalies and reduce the possibility of misjudgment. For example, the infiltration value of a monitoring point may be abnormally high due to random fluctuations, but its abnormality and periodic weight are consistent with the overall, in which case it may not be judged as important.
[0048] A group of observation values with the greatest importance among all group observation values of each associated data item is obtained as the feature data of each associated data item.
[0049] The characteristic data of all items of associated data were used as independent variables and the seepage data as dependent variables to construct a multiple linear regression model; the importance of the characteristic data of each item of associated data was used as the regression coefficient of each item of associated data in the multiple linear regression model.
[0050] The error term of the model is obtained by calculating the residual after fitting the multiple linear regression model, and the final corrected model is obtained for future data monitoring. The construction process of the multiple linear regression model is a well-known technology and will not be described in detail.
[0051] The mean of the observation values at all monitoring points of each associated data at the current moment is taken as the average observation value of each associated data; the average observation value of all associated data at the current moment is predicted using a multivariate linear regression model to obtain the predicted seepage flow at the current moment; if the seepage flow data obtained in real time is greater than the predicted seepage flow value, abnormal leakage occurs in the water conservancy construction building.
[0052] Through the above description of the implementation method in combination with the accompanying drawings, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.
[0053] The above contents are only specific implementation methods of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be covered by the protection scope of the present application.
Claims
1. A method for detecting anti-seepage in water conservancy project construction, characterized in that: The method comprises the following steps: S1, real-time acquisition of observation values and seepage data of various observation data of water conservancy construction buildings at each monitoring point; S2, according to the correlation between each group of observation values of the observation data and the seepage data, select each item of related data from all items of observation data; S3, for each associated data, the error value is obtained according to the local difference of the change trend of each group of observations, and the error value is clustered to screen the abnormal clusters; the abnormal characteristic value of each group of observations is obtained according to the data discreteness of each group of observations, the overall change trend and the value of the error value in its abnormal cluster; S4, according to the difference in the change trend of each group of observations and all groups of observations and the difference in the values at the same time, the deviation index of each group of observations is obtained; after denoising each group of observations according to the frequency domain components of each group of observations, outliers are screened; according to the continuity of the distribution of outliers, the outlier index of each outlier is obtained; S5, comprehensively considering the deviation index, abnormal characteristic value and abnormal index of all abnormal values of each group of observation values, to obtain the importance of each group of observation values; predict the seepage data according to the importance of all groups of observation values of each associated data, and obtain the anti-seepage monitoring result by comparing the predicted value with the seepage data obtained in real time.
2. A water conservancy project construction anti-seepage detection method as claimed in claim 1, characterized in that: The method for screening each item of associated data includes: For each observation data, each group of observation values of the observation data is taken as the independent variable, and the seepage data is taken as the dependent variable, and a linear correlation diagram of each group of observation values and the seepage data is drawn; according to the linear distribution of all data points in the linear correlation diagram, the correlation between each group of observation values and the seepage data is obtained; The mean of the correlation between all group observation values of each observation data and the seepage data is taken as the influencing factor of each observation data; the observation data with an influencing factor less than the preset correlation threshold are eliminated, and the remaining observation data are recorded as the associated data.
3. A water conservancy project construction anti-seepage detection method as claimed in claim 2, characterized in that: The method for obtaining the correlation between each group of observation values and the seepage data includes: A straight line is fitted to all data points in the linear correlation diagram to obtain a fitted straight line; the determination coefficient of the fitted straight line is calculated as the correlation between each group of observations and the seepage data.
4. A water conservancy project construction anti-seepage detection method as claimed in claim 1, characterized in that: The error values are obtained according to the local differences in the change trends of the observation values of each group, and cluster analysis is performed on the error values to screen abnormal clusters, including: For each group of observation values of each associated data, the difference between the observation value at each moment and the previous moment is calculated, and the ratio of the observation value difference to the observation value at each moment is used as the growth rate at each moment; a neighborhood interval centered at each moment is constructed; Calculate the average growth rate of all moments in the neighborhood interval of any moment; if the growth rate at any moment is greater than the average growth rate, the observed value at any moment is the error value; All error values are clustered according to the time differences between the error values to obtain clusters; the mean time differences of all pairwise combinations of error values in each cluster are calculated; and the cluster with the smallest mean time difference is recorded as an abnormal cluster.
5. A water conservancy project construction anti-seepage detection method as claimed in claim 4, characterized in that: The method for obtaining the abnormal characteristic value of each group of observation values comprises: in, is the abnormal characteristic value of the observation value of group a; is the standard deviation of the observations in group a; is the absolute value of the slope of the fitted straight line of the a-th group of observations; is the growth rate of the i-th error value in the abnormal cluster of the a-th group of observations; I is the number of error values in the abnormal cluster of the a-th group of observations; is the maximum growth rate of all error values in the abnormal cluster of group a observations; is a linear normalization function.
6. A water conservancy project construction anti-seepage detection method as claimed in claim 1, characterized in that: The method for obtaining the deviation index of each group of observation values includes: recording the deviation index of the a-th group of observation values as , ;in, is the absolute value of the slope of the fitted straight line of the a-th group of observations, is the mean of the absolute values of the slopes of the fitted lines of all groups of observations; is the tth observation in the ath group of observations, is the mean of the t-th observation in all groups of observations, and T is the number of observations in each group of observations.
7. A water conservancy project construction anti-seepage detection method as claimed in claim 4, characterized in that: After denoising each group of observation values according to the frequency domain components of each group of observation values, screening outliers includes: Perform fast Fourier transform on each group of observations to obtain high-frequency components and low-frequency components; perform inverse transform on all low-frequency components to obtain each group of denoised observations; according to the method for obtaining the error value in each group of observations, obtain the error value in each group of denoised observations and record it as an outlier.
8. A water conservancy project construction anti-seepage detection method as claimed in claim 1, characterized in that: According to the continuity of the outlier distribution, the outlier index of each outlier is obtained, including: The outliers that are continuous at the same time are grouped together; the inverse of the continuous duration of each group of outliers is calculated, and the difference between the value 1 and the inverse is used as the outlier index of each outlier in each group of outliers; the outlier index of each individual outlier is recorded as 0.
9. A water conservancy project construction anti-seepage detection method as claimed in claim 1, characterized in that: The method for obtaining the importance of each group of observation values includes: ; is the importance of the observations in group a, is the deviation index of the observations in group a, is the anomaly index of the xth outlier in the ath group of observations, is the abnormal characteristic value of the a-th group of observations, and X is the number of abnormal values of the a-th group of observations.
10. A water conservancy project construction anti-seepage detection method as claimed in claim 1, characterized in that: The seepage data is predicted according to the importance of all groups of observation values of each associated data, and the anti-seepage monitoring result is obtained by comparing the predicted value with the seepage data obtained in real time, including: Obtain a group of observations with the greatest importance among all group observations of each associated data item as the characteristic data of each associated data item; The characteristic data of all items of associated data are taken as independent variables, and the seepage data are taken as dependent variables to construct a multiple linear regression model; wherein the importance of the characteristic data of each item of associated data is taken as the regression coefficient of each item of associated data in the multiple linear regression model; The mean of the observation values at all monitoring points of each associated data at the current moment is taken as the average observation value of each associated data; the average observation value of all associated data at the current moment is predicted using a multivariate linear regression model to obtain the predicted seepage flow at the current moment; if the seepage flow data obtained in real time is greater than the predicted seepage flow value, abnormal leakage occurs in the water conservancy construction building.