Water conservancy dam deformation monitoring data analysis method

Abstract

The invention provides a water conservancy dam deformation monitoring data analysis method, and relates to the technical field of monitoring data analysis, and the method comprises the steps: carrying out the preprocessing of target deformation monitoring data of a target water conservancy dam dynamically monitored by a deformation monitoring equipment group, and obtaining target effective deformation monitoring data; obtaining a target deformation feature according to a predetermined feature extraction strategy; activating a deformation mode discriminator to obtain a target deformation mode; introducing a time sequence analysis mechanism to obtain a target time sequence analysis result; if a target deformation risk coefficient obtained by analyzing the target time sequence analysis result reaches a preset risk threshold value, a deformation early warning signal is sent out, and deformation risk early warning of the target deformation mode is carried out. The problems that the deformation mode of the dam is difficult to accurately identify and the deformation trend of the dam is difficult to predict in the prior art and effective decision support cannot be provided for safety management of the dam can be solved. The effects of accurately identifying the deformation mode of the dam and predicting the deformation trend of the dam are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of monitoring data analysis, and in particular to a method for analyzing deformation monitoring data of a water conservancy dam. Background Art

[0002] At present, deformation monitoring of hydropower dams mainly uses sensors, total stations, GPS and other equipment to collect data, which is then processed through manual analysis or simple data analysis software.

[0003] At present, the existing analysis efficiency of water conservancy dam data is low, which is difficult to meet the needs of large-scale data processing. At the same time, the data analysis accuracy is not high, making it difficult to accurately identify the deformation mode of the dam and predict the deformation trend of the dam. In addition, since the data analysis results are not intuitive, it is difficult to provide effective decision-making support for the safety management of the dam. Summary of the invention

[0004] The purpose of this application is to provide a method for analyzing water conservancy dam deformation monitoring data to solve the problem that the existing analysis efficiency of water conservancy dam data is low and it is difficult to meet the needs of large-scale data processing. At the same time, the data analysis accuracy is not high, and it is difficult to accurately identify the deformation mode of the dam and predict the deformation trend of the dam. In addition, since the data analysis results are not intuitive, it is difficult to provide effective decision-making support for the safety management of the dam.

[0005] In view of the above problems, the present application provides a method for analyzing deformation monitoring data of a water conservancy dam.

[0006] The present application provides a method for analyzing deformation monitoring data of a water conservancy dam, wherein the method comprises: preprocessing target deformation monitoring data of a target water conservancy dam dynamically monitored by a deformation monitoring equipment group to obtain target effective deformation monitoring data; reading a predetermined feature extraction strategy, and extracting features of the target effective deformation monitoring data according to the predetermined feature extraction strategy to obtain target deformation features; activating a deformation mode discriminator, and using the target deformation features as input of the deformation mode discriminator to obtain an output, wherein the output includes a target deformation mode; introducing a timing analysis mechanism to perform time series analysis on the target deformation features to obtain a target timing analysis result; if a target deformation risk coefficient obtained by analyzing the target timing analysis result reaches a predetermined risk threshold, issuing a deformation warning signal; and performing a deformation risk warning of the target deformation mode for the target water conservancy dam based on the deformation warning signal.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] The target deformation monitoring data of the target water conservancy dam dynamically monitored by the deformation monitoring equipment group is preprocessed to obtain the target effective deformation monitoring data; the predetermined feature extraction strategy is read, and the target effective deformation monitoring data is feature extracted according to the predetermined feature extraction strategy to obtain the target deformation feature; the deformation mode discriminator is activated, and the target deformation feature is used as the input of the deformation mode discriminator to obtain the output, wherein the output includes the target deformation mode; the time series analysis mechanism is introduced to perform time series analysis on the target deformation feature to obtain the target time series analysis result; if the target deformation risk coefficient obtained by analyzing the target time series analysis result reaches the predetermined risk threshold, a deformation warning signal is issued; based on the deformation warning signal, the target water conservancy dam is given a deformation risk warning of the target deformation mode. That is to say, the deformation monitoring data of the water conservancy dam is collected by using sensors, total stations, GPS and other equipment. Then the collected data is filtered, smoothed and normalized to remove noise and outliers and improve data quality. Then, the principal component analysis, independent component analysis and other methods are used to extract characteristic information such as displacement, tilt angle, strain, etc. from the monitoring data. Based on the extracted feature information, cluster analysis, support vector machine and other methods are used to identify the deformation type of the dam, such as settlement, tilt, cracks, etc. In addition, by performing trend analysis, periodicity analysis and randomness analysis on the monitoring data, the development trend and law of dam deformation are understood, and the deformation of the dam is predicted based on the data analysis results using neural network, time series prediction and other methods. Finally, based on the deformation prediction results, the safety of the dam is evaluated using hierarchical analysis method, fuzzy comprehensive evaluation and other methods to determine whether it is in a safe state and take corresponding measures in a timely manner. By adopting automated data collection and analysis methods, the needs of large-scale data processing can be met, and advanced data analysis methods such as principal component analysis, cluster analysis, support vector machine, neural network, etc. can be used to accurately identify the deformation mode of the dam and predict the deformation trend of the dam. In addition, the visual data analysis result display method is used to make the data analysis results more intuitive, which can provide effective decision support for the safety management of the dam.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0011] Figure 1 A schematic diagram of a flow chart of a method for analyzing deformation monitoring data of a water conservancy dam for this application;

[0012] Figure 2 A schematic diagram of a flow chart of obtaining a deformation pattern discriminator in a method for analyzing deformation monitoring data of a water conservancy dam in the present application. DETAILED DESCRIPTION

[0013] This application provides a method for analyzing water conservancy dam deformation monitoring data, which solves the problem that the existing analysis efficiency of water conservancy dam data is low, which is difficult to meet the needs of large-scale data processing. At the same time, the data analysis accuracy is not high, and it is difficult to accurately identify the deformation mode of the dam and predict the deformation trend of the dam. In addition, since the data analysis results are not intuitive, it is difficult to provide effective decision support for the safety management of the dam. By adopting automated data collection and analysis methods, it can meet the needs of large-scale data processing, and use advanced data analysis methods such as principal component analysis, cluster analysis, support vector machine, neural network, etc., it can accurately identify the deformation mode of the dam and predict the deformation trend of the dam. In addition, a visual data analysis result display method is used to make the data analysis results more intuitive, which can provide effective decision support for the safety management of the dam.

[0014] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0015] For example, please refer to the attached Figure 1 The present application provides a method for analyzing deformation monitoring data of a water conservancy dam, and the method for analyzing deformation monitoring data of a water conservancy dam specifically comprises the following steps:

[0016] Step P10: pre-processing the target deformation monitoring data of the target water conservancy dam dynamically monitored by the deformation monitoring equipment group to obtain target effective deformation monitoring data;

[0017] Specifically, by purifying and optimizing the original monitoring data, removing noise and outliers, the quality and availability of the data can be improved. This process is the basic link in the analysis of the deformation monitoring data of the entire hydropower dam. Its main technical goal is to provide accurate and reliable data support for subsequent deformation pattern recognition, trend analysis and risk assessment, and ensure the scientificity and effectiveness of subsequent analysis.

[0018] First, the deformation monitoring equipment group will collect a large amount of deformation monitoring data of the target hydropower dam during the dynamic monitoring process. These data contain information such as displacement, tilt, and strain that may occur during the operation of the dam. However, due to the complexity of the monitoring environment and the errors of the equipment itself, the original data is often mixed with noise and outliers. These interference factors will affect the accuracy of subsequent analysis. Therefore, it is necessary to preprocess these target deformation monitoring data. In the preprocessing process, the data will first be filtered to remove high-frequency noise and make the data smoother. Then, the data smoothing technology is used to further reduce the short-term fluctuations in the data and reveal its long-term trend. In addition, the data will be normalized to unify the data of different dimensions into the same interval for subsequent analysis and comparison. Among them, the normalization process is to map the data to the [0,1] interval through a specific mathematical formula, thereby eliminating the impact of dimensional differences. It should be noted that in the preprocessing process, the data will also be detected and eliminated for outliers to ensure the rationality of the data. Through this series of operations, the target effective deformation monitoring data is finally obtained, laying the foundation for the subsequent analysis process.

[0019] Through the above preprocessing steps, the deformation monitoring data of the target hydropower dam is effectively purified and optimized, which significantly improves the quality and reliability of the data. In this way, the subsequent deformation analysis can be based on more accurate data, thereby improving the accuracy of dam deformation pattern recognition and deformation trend prediction. Obviously, this preprocessing process is an indispensable and important link in the entire deformation monitoring data analysis, which is directly related to the scientificity and effectiveness of the final risk assessment, and provides a solid data foundation for ensuring the safe operation of hydropower dams.

[0020] Step P20: reading a predetermined feature extraction strategy, and performing feature extraction on the target effective deformation monitoring data according to the predetermined feature extraction strategy to obtain target deformation features;

[0021] Specifically, feature information that can reflect the deformation nature of the hydropower dam is extracted from the pre-processed high-quality data. Through this process, the complex and huge monitoring data can be transformed into more representative and interpretable features, providing a key basis for subsequent deformation pattern recognition and trend analysis, thereby improving the intelligence level and decision-making support capabilities of the entire monitoring system.

[0022] First, it is necessary to read the predetermined feature extraction strategy, which is based on the in-depth analysis and understanding of the deformation monitoring data of the hydropower dam, and includes the specific methods of which features to extract and how to extract them. Then, according to the predetermined feature extraction strategy, the target effective deformation monitoring data is processed. In this process, the system extracts features related to the dam deformation from the data, such as displacement, tilt angle, strain and other key indicators. Then, through mathematical methods such as principal component analysis (PCA) or independent component analysis (ICA), the effect of feature extraction is further optimized to ensure that the extracted features can retain the important information in the original data to the greatest extent. In addition, the feature extraction process also considers the dimension and correlation of the data, removes redundant information, and improves the independence and discrimination of the features. Finally, after this series of operations, the system obtains the target deformation features, which are the basis for subsequent analysis and decision-making.

[0023] The effective conversion from raw monitoring data to key features has been achieved, which significantly improves the data availability and analysis efficiency, not only reducing the complexity of the data, but also enhancing the perception of the dam's deformation state. Therefore, feature extraction provides accurate and efficient input for subsequent deformation pattern recognition and trend prediction, thereby ensuring the intelligence and reliability of the entire monitoring system and providing strong technical support for the safe operation of water conservancy dams.

[0024] Step P30: activating a deformation pattern discriminator, and using the target deformation feature as an input of the deformation pattern discriminator to obtain an output, wherein the output includes a target deformation pattern;

[0025] Specifically, the extracted target deformation features are analyzed using a pre-trained deformation pattern discriminator to identify the specific deformation pattern of the dam. This process is a key link in the dam deformation monitoring system. Its main technical goal is to quickly and accurately determine the deformation type of the dam through intelligent pattern recognition methods, provide a direct basis for subsequent risk assessment and safety management, and thus improve the automation and intelligence level of dam monitoring.

[0026] First, the deformation pattern discriminator is activated. This is a model built based on machine learning or artificial intelligence algorithms. It is usually trained with a large amount of historical data and known deformation patterns, so that it has the ability to identify different deformation patterns. Then, the target deformation features obtained through feature extraction are input into the deformation pattern discriminator. These features contain key information about dam deformation, such as displacement, tilt angle, strain, etc., which are the basis for deformation pattern recognition. Next, the deformation pattern discriminator analyzes and processes the input features, and classifies and identifies the features through internal algorithm logic, such as cluster analysis, support vector machine (SVM) or deep learning algorithm. In addition, the discriminator matches the input deformation features with the known deformation patterns based on the pattern features learned during the training process, so as to determine the current deformation state of the dam. Finally, the deformation pattern discriminator outputs the target deformation pattern. This output result directly reflects the deformation type of the dam, such as settlement, tilt, cracks, etc., providing a clear basis for subsequent risk assessment and decision-making. By activating the deformation pattern discriminator and performing feature input and pattern recognition, an intelligent transition from feature extraction to deformation pattern determination is achieved, which significantly improves the efficiency and accuracy of dam deformation monitoring. It can not only quickly identify the deformation type of the dam, but also reduce the errors caused by manual intervention and subjective judgment.

[0027] Step P40: introducing a time series analysis mechanism to perform time series analysis on the target deformation characteristics to obtain a target time series analysis result;

[0028] Specifically, by conducting an in-depth analysis of the time series characteristics of the target deformation features, the laws and trends of dam deformation in the time dimension can be explored. This process can help the monitoring system better understand the dynamic behavior of dam deformation, thereby providing a scientific basis for predicting future deformation trends and early warning of potential risks, thereby enhancing the foresight and initiative of dam safety management.

[0029] First, the time series analysis mechanism is introduced. This is an analysis method specifically used to process time series data, which can reveal the temporal change law of data. Then, the target deformation characteristics are taken as the object of time series analysis. After the initial extraction and processing, these characteristics can accurately reflect the key information of dam deformation. Then, by establishing a time series model, the target deformation characteristics are subjected to trend analysis, periodicity analysis and randomness analysis. Among them, trend analysis is used to determine the long-term trend of dam deformation, periodicity analysis is used to identify the periodic law of deformation, and randomness analysis is used to evaluate the uncertainty of deformation. In addition, in order to ensure the accuracy of the analysis results, the system will select a suitable time series analysis model according to the data characteristics, such as the autoregressive moving average model (ARMA) or the autoregressive integrated moving average model (ARIMA). Finally, through these analysis steps, the target time series analysis results are obtained. These results are presented in the form of data charts or analysis reports, which intuitively show the time characteristics of dam deformation. By introducing the time series analysis mechanism to analyze the target deformation characteristics, the dynamic behavior of dam deformation is accurately described, which significantly improves the prediction ability and early warning accuracy of the monitoring system.

[0030] Step P50: If the target deformation risk coefficient obtained by analyzing the target time series analysis result reaches a predetermined risk threshold, a deformation warning signal is issued;

[0031] Specifically, by quantitatively analyzing the deformation risk coefficient in the target time series analysis results and comparing it with the predetermined risk threshold, real-time monitoring and early warning of the deformation risk of the hydropower dam can be achieved. By effectively identifying whether the dam deformation is within the safe range and issuing an early warning in time when the risk exceeds the safety threshold, relevant personnel can take measures to ensure the safe operation of the dam.

[0032] First, the deformation data in the target time series analysis results are further analyzed to extract features related to deformation risk, such as deformation rate, deformation acceleration, etc., and the target deformation risk coefficient is calculated based on these features. This risk coefficient is a quantitative indicator used to measure the potential danger of the current deformation state of the dam. Then, the calculated target deformation risk coefficient is compared with the pre-set risk threshold. If the risk coefficient does not reach the predetermined threshold, the deformation of the dam is considered to be within the safe range, and the system will continue to monitor and regularly update the risk assessment results. However, if the analysis results show that the target deformation risk coefficient reaches or exceeds the predetermined risk threshold, the system will trigger the early warning mechanism. Then, a deformation early warning signal is issued, which can be in the form of sound and light alarms, SMS notifications, or system alarms, to promptly notify relevant personnel to pay attention to the deformation of the dam. In addition, the issuance of the early warning signal not only provides a timely decision-making basis for dam safety management, but also prompts relevant personnel to take corresponding emergency measures, such as increasing the monitoring frequency, conducting on-site inspections, or implementing reinforcement projects. Finally, through this early warning mechanism, the system realizes dynamic monitoring and active early warning of dam deformation risks, effectively improving the timeliness and effectiveness of dam safety management. It has achieved real-time monitoring and early warning of dam deformation risks, significantly improved the initiative and reliability of dam safety management, and ensured the safety and stability of the dam in a complex operating environment.

[0033] Step P60: Based on the deformation warning signal, a deformation risk warning of the target deformation mode is issued to the target water conservancy dam.

[0034] Specifically, the deformation warning signals issued by the monitoring system are converted into specific risk warning information for the target deformation mode, and applied to the safety management of the target water conservancy dam. Through clear risk warnings, timely and accurate decision-making support is provided to dam managers so that targeted measures can be taken to deal with potential deformation risks, thereby ensuring the safe operation of water conservancy dams.

[0035] First, after receiving the deformation warning signal, the system will determine the specific deformation risk type according to the target deformation mode information contained in the warning signal. For example, if the warning signal indicates that the dam has a risk of settlement deformation, the system will further analyze the scope, rate and possible influencing factors of the settlement. Then, the system will combine the characteristics of the target deformation mode to evaluate its impact on the overall structural safety of the dam. This evaluation process usually refers to historical data, deformation rate and dam design parameters to ensure the scientificity and accuracy of the evaluation results. In addition, the system will generate a detailed deformation risk warning report, which not only contains the type and risk level of the deformation mode, but also provides corresponding recommended measures, such as whether reinforcement measures need to be taken immediately and whether downstream residents need to be evacuated. Finally, the system transmits the deformation risk warning information to the dam management personnel or relevant emergency departments so that they can take effective countermeasures in time according to the warning information. The transformation from deformation warning signal to specific risk warning information is realized, and it is applied to the safety management of the target water conservancy dam, which provides an important guarantee for the safe operation of water conservancy dams, effectively improves the scientificity and efficiency of dam safety management, and also provides a reliable reference for risk warning and emergency management of similar projects.

[0036] Furthermore, the deformation monitoring equipment group includes sensors, a total station and a GPS.

[0037] Further, the target deformation monitoring data of the target water conservancy dam dynamically monitored by the deformation monitoring equipment group is preprocessed to obtain target effective deformation monitoring data, including:

[0038] Performing filtering processing on the target deformation monitoring data to obtain target filtered data;

[0039] Smoothing the target filtered data to obtain target smoothed data;

[0040] The target smoothing data is normalized to obtain the target effective deformation monitoring data.

[0041] Specifically, sensors are installed at key locations on the dam, including but not limited to: monitoring the horizontal and vertical displacements of the top of the dam to detect possible cracks or settlements; monitoring the pressure distribution inside the dam body to prevent internal voids or uneven settlements; monitoring the stability of the dam foundation to prevent deformation caused by geological changes. The specific arrangements include: Displacement sensor: a displacement sensor is set at a certain distance (for example, every 50 meters) to comprehensively monitor the displacement of the dam surface. Tilt sensor: installed on both sides of the dam to monitor the tilt angle of the dam. Strain sensor: arranged inside the dam body to monitor the strain of the dam body material to evaluate its strength and stability. In addition, it is also possible to consider installing temperature and humidity sensors at specific locations to monitor the impact of environmental factors on the dam structure.

[0042] In addition, the frequency of data collection depends on the importance of the dam and actual needs. For important dams, a higher collection frequency can be set, such as once an hour. High-frequency collection helps to detect potential problems in a timely manner. For ordinary dams, the collection frequency can be appropriately reduced, such as once a day. This can ensure the effectiveness of monitoring and save resources.

[0043] In addition, while collecting relevant data, the quality of the data also needs to be ensured. Sensors need to be calibrated regularly to ensure the accuracy and reliability of the data, and sensors suitable for the environmental conditions of the dam should be selected, such as corrosion resistance and electromagnetic interference resistance. Redundant sensors should be installed at key locations so that valid data can still be obtained when a sensor fails. In addition, data consistency and rationality can be ensured by cross-validating data from different sensors.

[0044] Finally, after data collection is completed, it needs to be stored and transmitted to the data center, using highly reliable storage devices and regularly backing up data to prevent loss. Data is transmitted to the data center via wired or wireless networks to ensure the security and timeliness of data transmission.

[0045] Furthermore, in dam monitoring, the quality of data processing is crucial to accuracy. First, we use the median filter algorithm to purify the data in response to the noise interference collected. For a data sequence of length n }, take the middle value as the output after filtering. If n is an odd number, the middle value is the (n+1) / 2th data; if n is an even number, the middle value is the average of the n / 2th data and the (n / 2+1)th data. The advantage of median filtering is that it is insensitive to extreme values ​​and effectively reduces the impact of noise on the data.

[0046] For the filtered data, the moving average method is used for smoothing. The formula of the moving average method is:

[0047]

[0048] in, is the smoothed data, is the window size of the moving average, where k=m, m+1,...,n. This process helps to eliminate short-term fluctuations and reveal long-term trends and cyclical changes in the data.

[0049] In order to facilitate the analysis and comparison of data of different dimensions, the present invention adopts the maximum and minimum normalization method to map the data to the interval [0,1]. The formula is:

[0050]

[0051] in is the normalized data, is the original data, is the minimum value in the data sequence, is the maximum value in the data sequence.

[0052] Further, a predetermined feature extraction strategy is read, and feature extraction is performed on the target effective deformation monitoring data according to the predetermined feature extraction strategy to obtain target deformation features, including:

[0053] Extracting a parallel extraction scheme in the predetermined feature extraction strategy, wherein the parallel extraction scheme refers to a scheme for collaboratively performing feature extraction based on principal component analysis and independent component analysis;

[0054] Extracting features from the target effective deformation monitoring data according to the principal component analysis to obtain a first deformation feature;

[0055] Extracting features from the target effective deformation monitoring data according to the independent component analysis to obtain a second deformation feature;

[0056] The target deformation feature is constructed based on the first deformation feature and the second deformation feature.

[0057] Specifically, feature extraction is the process of converting raw monitoring data into meaningful information that can help identify the pattern of dam deformation. Through principal component analysis and independent component analysis, useful features can be extracted from the raw data to provide a basis for subsequent deformation pattern recognition. These features may include displacement, tilt angle, strain, etc., which are important indicators of the health of the dam. The specific steps are as follows:

[0058] Displacement, tilt angle and strain data from different locations of the dam are collected to form a data matrix X, which contains n rows (samples) and m columns (features). Next, the covariance matrix C of the data matrix X is calculated. The calculation formula of the covariance matrix C is

[0059]

[0060] Perform eigenvalue decomposition on the covariance matrix C and obtain the eigenvalue and the corresponding eigenvector .

[0061] Select the eigenvectors corresponding to the first k largest eigenvalues ​​to form a projection matrix The features selected should preserve the variability of the original data to the greatest extent possible.

[0062] Use the selected eigenvector matrix P to project the original data X to obtain a new feature matrix Y, which contains the information that best represents the changes in the original data. The formula is:

[0063]

[0064] matrix It contains the information that best represents the changes in the original data, which is helpful for subsequent analysis and modeling.

[0065] Further, as attached Figure 2 As shown, a deformation pattern discriminator is activated, and the target deformation feature is used as an input of the deformation pattern discriminator to obtain an output, wherein the output includes a target deformation pattern, including:

[0066] Acquire a deformation database, wherein the deformation database includes first historical deformation data of a first deformation mode;

[0067] Extracting features from the first historical deformation data according to the predetermined feature extraction strategy to obtain first historical deformation features;

[0068] forming a first training data set based on the first historical deformation feature and the first deformation mode;

[0069] Based on the support vector machine principle, supervised training is performed on the first training data group to obtain a deformation pattern support vector machine;

[0070] The deformation pattern support vector machine is used as the deformation pattern discriminator.

[0071] Specifically, a model that can accurately identify deformation patterns is trained through the analysis of historical deformation data and the support vector machine (SVM) principle of machine learning methods. This solution extracts features from historical data, constructs a training data set, and conducts supervised training to ultimately obtain an efficient deformation pattern discriminator, thereby providing an intelligent pattern recognition tool for water conservancy dam deformation monitoring and improving the accuracy and reliability of deformation monitoring.

[0072] First, the system obtains a deformation database containing the first historical deformation data of the first deformation mode. These historical deformation data are data related to dam deformation accumulated during past monitoring processes, covering specific manifestations under different deformation modes. Then, the first historical deformation data is feature extracted according to a predetermined feature extraction strategy to obtain the first historical deformation feature. This process extracts key features that can reflect the deformation mode from the historical data through specific algorithms, such as principal component analysis or independent component analysis, to provide a basis for subsequent model training. Then, based on the extracted first historical deformation features and the corresponding first deformation mode, a first training data group is formed. This data group is the basis for support vector machine training. It associates features with known deformation modes and provides a clear input and output relationship for model learning. Next, supervised training is performed on the first training data group based on the support vector machine principle. In this process, the support vector machine constructs a classifier that can distinguish different deformation modes by learning the relationship between features and deformation modes in the training data group. Finally, the trained deformation mode support vector machine is used as a deformation mode discriminator for subsequent recognition and classification of target deformation features.

[0073] Through the above process, the technical solution realizes the construction from historical deformation data to an efficient deformation pattern discriminator, significantly improving the accuracy and efficiency of deformation pattern recognition.

[0074] Further, a deformation database is obtained, wherein the deformation database includes first historical deformation data of a first deformation mode, including:

[0075] Clustering the deformation database to obtain a deformation clustering result, wherein the deformation clustering result includes a plurality of clustering clusters;

[0076] Extracting a first cluster from the plurality of clusters, wherein the first cluster refers to a first database corresponding to the first deformation mode;

[0077] Any set of data in the first database is randomly extracted and recorded as the first historical deformation data.

[0078] Specifically, a large amount of deformation data is classified and sorted through the clustering method in data mining technology, so as to efficiently obtain historical data related to a specific deformation mode. First, cluster analysis is performed on the deformation database, which is an unsupervised learning method used to group similar data points into different clusters. Through clustering algorithms such as K-Means or hierarchical clustering, the deformation data is divided into multiple clusters according to the intrinsic characteristics and similarities of the data. Each cluster represents a specific deformation mode or a set of deformation features, thereby achieving preliminary classification and sorting of the deformation data. Then, the first cluster corresponding to the first deformation mode is extracted from the multiple clusters. This process is based on the analysis and marking of the clustering results, identifying the clusters associated with the specific deformation mode, thereby providing a clear target for subsequent data extraction. Among them, the first cluster refers to the database containing the data corresponding to the first deformation mode, which is the basis for subsequent analysis. Finally, any set of data is randomly extracted from the first database and recorded as the first historical deformation data. This random extraction process ensures the diversity and representativeness of the data, and provides high-quality sample data for subsequent feature extraction and model training.

[0079] Through the above process, the goal of efficiently extracting historical data related to specific deformation patterns from a large amount of deformation data was achieved, significantly improving the efficiency and accuracy of data processing. Data management was optimized through cluster analysis, enabling the deformation monitoring system to process and utilize data more efficiently, providing a more scientific and efficient data basis for dam safety management.

[0080] Furthermore, a time series analysis mechanism is introduced to perform time series analysis on the target deformation characteristics to obtain target time series analysis results, including:

[0081] Establishing a time series model, and arranging the target deformation features in time sequence through the time series model to form a target feature time series;

[0082] Performing trend analysis on the target feature time series to obtain target trend features;

[0083] Performing periodic analysis on the target feature time series to obtain target periodic features;

[0084] Performing randomness analysis on the target feature time series to obtain target randomness features;

[0085] The target trend characteristics, the target periodic characteristics and the target random characteristics constitute the target time series analysis result.

[0086] Specifically, in the time series analysis phase, the displacement values ​​of the dam monitored over a period of time are analyzed using statistical and machine learning methods to analyze the changes in the monitoring data over time. The trends, periodicity, and randomness of these data are analyzed.

[0087] First, establish a time series model: set up time series data ,in Representing time point Monitoring data on.

[0088] Second, perform trend analysis and fit the linear regression model: Assume that the trend line is

[0089]

[0090] Where t is the time variable, a is the intercept, and b is the slope of the trend. The least squares method is used to estimate the parameters a and b so that the residual sum of squares is minimized:

[0091]

[0092] Evaluate whether the data has a clear upward or downward trend. Take the partial derivatives of S with respect to a and b and set them to 0, and you get the equation system:

[0093]

[0094]

[0095] Solving this system of equations gives us estimated values ​​for parameters a and b. By analyzing the sign and size of b, we can determine whether the data show a clear upward or downward trend.

[0096] Third, the Discrete Fourier Transform (DFT) is used to calculate the periodicity analysis: In order to detect the periodic components in the data, the time series data is converted from the time domain to the frequency domain using the Discrete Fourier Transform:

[0097]

[0098] Where j is the imaginary unit and N is the length of the sequence. By analyzing the spectrum after Fourier transformation, we can determine whether there is a periodic component in the time series and the size of the period. If there is an obvious peak, it means that the data has a periodic component.

[0099] Fourth, randomness analysis is performed by calculating the autocorrelation function (ACF) and partial autocorrelation function (PACF) to evaluate the random fluctuation properties of the data:

[0100] Calculate the autocorrelation function (ACF) and partial autocorrelation function (PACF) of the time series. Analyze the ACF and PACF graphs to see whether they tend to 0 when the lag order is large. If the autocorrelation function and partial autocorrelation function quickly approach 0 when the lag order is large, it means that the time series is random. Autocorrelation function The calculation formula is

[0101]

[0102] in is the mean of the time series. The autocorrelation function describes the correlation between a data point and its past data point at a certain time lag.

[0103] Further, after introducing a time series analysis mechanism to perform time series analysis on the target deformation feature and obtaining the target time series analysis result, the method includes:

[0104] activating a neural network prediction model, and using the target time series analysis result as input information of the neural network prediction model to obtain a first deformation risk coefficient;

[0105] activating a time series prediction model, and using the target time series analysis result as input information of the time series prediction model to obtain a second deformation risk coefficient;

[0106] The first deformation risk coefficient and the second deformation risk coefficient are compared and verified to obtain the target deformation risk coefficient.

[0107] Specifically, deformation prediction is based on the previously extracted feature information and the data patterns obtained through time series analysis to predict the future deformation of the dam. In the monitoring process, neural networks and time series prediction models are used to verify each other. The time series prediction model is suitable for short-term situations, and the neural network is suitable for medium- and long-term situations.

[0108] Neural networks are a powerful tool that can be used to predict nonlinear relationships. Construct a multi-layer feedforward neural network for dam deformation prediction, including input layer, hidden layer and output layer. Assume that features have been extracted from historical monitoring data and these features have been used to train a neural network model. Now you need to use this model to predict deformation:

[0109] The input layer receives feature vectors, which can be extracted information such as displacement, tilt angle, strain, etc. The network performs calculations through the hidden layer to predict the deformation value of the dam, such as the change in displacement or tilt angle. When training the neural network, historical data is used and the network weights are adjusted through the back propagation algorithm to minimize the prediction error. The back propagation algorithm continuously adjusts the weights and biases of the network to ensure that the error between the predicted value and the actual value is minimized.

[0110] In order to evaluate the prediction effect of the model, the mean square error (MSE) can be used as the loss function, the formula is:

[0111]

[0112] in is the actual value, is the predicted value.

[0113] Finally, the trained neural network can make predictions on the latest monitoring data to obtain deformation prediction values ​​for a period of time in the future.

[0114] Time series forecasting using the Autoregressive Moving Average (ARMA) model:

[0115] First, the model order needs to be determined. The appropriate autoregressive order p and moving average order q need to be determined by analyzing the time series data. The model formula is

[0116]

[0117] in is the value of the time series at time t, c is a constant term, is the autoregressive coefficient, is the moving average coefficient, is a white noise sequence, p and q are the orders of autoregression and moving average, respectively.

[0118] The parameters of the ARMA model are estimated using the least squares method. By analyzing the time series data, the order p and q of the model are determined, and then the parameters of the model are estimated using the least squares method.

[0119] Finally, the estimated model is used to predict the deformation in the future, thus providing decision support for the safety management of the dam.

[0120] Through these steps, historical monitoring data can be used to predict future dam deformation and provide decision support for dam safety management.

[0121] Furthermore, it also includes:

[0122] Read the predetermined evaluation plan;

[0123] Evaluate and analyze the target effective deformation monitoring data according to the predetermined evaluation scheme to obtain a fuzzy deformation risk level;

[0124] The target deformation risk coefficient is verified using the fuzzy deformation risk level.

[0125] Further, the target effective deformation monitoring data is evaluated and analyzed according to the predetermined evaluation scheme to obtain a fuzzy deformation risk level, including:

[0126] Extracting a predetermined evaluation factor set and a predetermined evaluation grade set in the predetermined evaluation scheme;

[0127] Constructing a membership matrix according to the predetermined evaluation factor set and the predetermined evaluation grade set;

[0128] A fuzzy relationship matrix is ​​obtained according to the membership matrix, and the fuzzy deformation risk level is calculated.

[0129] Specifically, the fuzzy comprehensive evaluation method is used to quantitatively evaluate the deformation risk of the dam. This process can not only comprehensively consider the impact of multiple evaluation factors on the deformation risk, but also process uncertain information through fuzzy logic, thereby providing a more scientific and comprehensive risk assessment result for dam safety management.

[0130] First, the system reads the predetermined evaluation scheme, which is based on expert experience and historical data and contains rules and methods for dam deformation risk assessment. According to the predetermined evaluation scheme, the system evaluates and analyzes the target effective deformation monitoring data. This process involves comprehensive consideration of multiple evaluation factors in the data, such as deformation rate, displacement, tilt angle, etc., and converts these factors into fuzzy deformation risk levels through fuzzy mathematics. Specifically, the system first extracts the predetermined evaluation factor set and the predetermined evaluation level set in the predetermined evaluation scheme. These factor sets and level sets are the basis for constructing a fuzzy comprehensive evaluation model. Then, a membership matrix is ​​constructed based on the extracted evaluation factor set and evaluation level set, which reflects the degree of membership of each evaluation factor under different risk levels. Then, the fuzzy relationship matrix is ​​obtained based on the membership matrix, and the fuzzy deformation risk level is calculated through fuzzy operation. This process handles the uncertainty and fuzziness between the evaluation factors through fuzzy logic, making the risk assessment results more in line with the actual situation.

[0131] In the risk assessment stage, the safety of the dam will be evaluated based on the data analysis results in the previous steps, and corresponding risk management decisions will be made accordingly. Fuzzy comprehensive evaluation is a decision-making method that takes uncertainty and fuzziness into account and is suitable for dealing with subjective judgments and fuzzy information.

[0132] In the process of dam risk assessment, we first need to determine the evaluation factor set and level set:

[0133] Evaluation factor set , such as crack width, settlement, maintenance frequency, etc.

[0134] Rating Set , such as safety, basic safety, attention, danger, etc.

[0135] Next, determine the membership matrix R: For each evaluation factor , determine its evaluation level for each Membership , thus obtaining the fuzzy relationship matrix

[0136]

[0137] The weights of the evaluation factors are determined through pairwise comparison, and the factors such as structural integrity, geological stability, and maintenance status are compared pairwise to establish a judgment matrix. For factors at the same level, their relative importance is determined through pairwise comparison to establish a judgment matrix.

[0138]

[0139] in, Indicates the importance of factor i relative to factor j, satisfying , , This matrix allows decision makers to evaluate the relative importance of each factor and ultimately calculate a weight vector W such that the sum of all weights is 1.

[0140] Finally, the fuzzy comprehensive evaluation method is used to obtain the comprehensive evaluation results.

[0141]

[0142] It is an n-dimensional vector, which indicates the degree of membership of the dam to each evaluation level. The safety level of the dam is determined according to the principle of maximum membership.

[0143] Furthermore, the system uses the fuzzy deformation risk level as the target deformation risk coefficient for verification. This verification process aims to verify and correct the previously calculated target deformation risk coefficient through the results of fuzzy comprehensive evaluation, so as to improve the accuracy and reliability of risk assessment. The risk level obtained by fuzzy comprehensive evaluation not only considers the influence of a single factor, but also integrates the interaction of multiple factors, so it can more comprehensively reflect the actual deformation risk status of the dam. Obviously, this process provides a more scientific and systematic method for dam deformation risk assessment, which can effectively make up for the shortcomings of traditional risk assessment methods. The conversion from target effective deformation monitoring data to fuzzy deformation risk level is realized, and the accuracy and reliability of risk assessment are further improved by verifying the target deformation risk coefficient.

[0144] In summary, the method for analyzing water conservancy dam deformation monitoring data provided by this application has the following technical effects:

[0145] The target deformation monitoring data of the target water conservancy dam dynamically monitored by the deformation monitoring equipment group is preprocessed to obtain the target effective deformation monitoring data; the predetermined feature extraction strategy is read, and the target effective deformation monitoring data is feature extracted according to the predetermined feature extraction strategy to obtain the target deformation feature; the deformation mode discriminator is activated, and the target deformation feature is used as the input of the deformation mode discriminator to obtain the output, wherein the output includes the target deformation mode; the time series analysis mechanism is introduced to perform time series analysis on the target deformation feature to obtain the target time series analysis result; if the target deformation risk coefficient obtained by analyzing the target time series analysis result reaches the predetermined risk threshold, a deformation warning signal is issued; based on the deformation warning signal, the target water conservancy dam is given a deformation risk warning of the target deformation mode. That is to say, the deformation monitoring data of the water conservancy dam is collected by using sensors, total stations, GPS and other equipment. Then the collected data is filtered, smoothed and normalized to remove noise and outliers and improve data quality. Then, the principal component analysis, independent component analysis and other methods are used to extract characteristic information such as displacement, tilt angle, strain, etc. from the monitoring data. Based on the extracted feature information, cluster analysis, support vector machine and other methods are used to identify the deformation type of the dam, such as settlement, tilt, cracks, etc. In addition, by performing trend analysis, periodicity analysis and randomness analysis on the monitoring data, the development trend and law of dam deformation are understood, and the deformation of the dam is predicted based on the data analysis results using neural network, time series prediction and other methods. Finally, based on the deformation prediction results, the safety of the dam is evaluated using hierarchical analysis method, fuzzy comprehensive evaluation and other methods to determine whether it is in a safe state and take corresponding measures in a timely manner. By adopting automated data collection and analysis methods, the needs of large-scale data processing can be met, and advanced data analysis methods such as principal component analysis, cluster analysis, support vector machine, neural network, etc. can be used to accurately identify the deformation mode of the dam and predict the deformation trend of the dam. In addition, the visual data analysis result display method is used to make the data analysis results more intuitive, which can provide effective decision support for the safety management of the dam.

[0146] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0147] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A method for analyzing deformation monitoring data of a hydraulic dam, characterized in that: include: Preprocessing the target deformation monitoring data of the target water conservancy dam dynamically monitored by the deformation monitoring equipment group to obtain target effective deformation monitoring data; Reading a predetermined feature extraction strategy, and performing feature extraction on the target effective deformation monitoring data according to the predetermined feature extraction strategy to obtain target deformation features; activating a deformation pattern discriminator and using the target deformation feature as an input of the deformation pattern discriminator to obtain an output, wherein the output includes a target deformation pattern; Introducing a time series analysis mechanism to perform time series analysis on the target deformation characteristics to obtain a target time series analysis result; If the target deformation risk coefficient obtained by analyzing the target time series analysis result reaches a predetermined risk threshold, a deformation warning signal is issued; Based on the deformation warning signal, a deformation risk warning of the target deformation mode is performed on the target water conservancy dam.

2. The method according to claim 1, characterized in that: The deformation monitoring equipment group includes sensors, a total station and a GPS.

3. The method according to claim 2, characterized in that The target deformation monitoring data of the target water conservancy dam dynamically monitored by the deformation monitoring equipment group is preprocessed to obtain the target effective deformation monitoring data, including: Performing filtering processing on the target deformation monitoring data to obtain target filtered data; Smoothing the target filtered data to obtain target smoothed data; The target smoothing data is normalized to obtain the target effective deformation monitoring data.

4. The method according to claim 3, characterized in that Reading a predetermined feature extraction strategy, and performing feature extraction on the target effective deformation monitoring data according to the predetermined feature extraction strategy to obtain target deformation features, including: Extracting a parallel extraction scheme in the predetermined feature extraction strategy, wherein the parallel extraction scheme refers to a scheme for collaboratively performing feature extraction based on principal component analysis and independent component analysis; Extracting features from the target effective deformation monitoring data according to the principal component analysis to obtain a first deformation feature; Extracting features from the target effective deformation monitoring data according to the independent component analysis to obtain a second deformation feature; The target deformation feature is constructed based on the first deformation feature and the second deformation feature.

5. The method according to claim 1, characterized in that activating a deformation pattern discriminator and using the target deformation feature as an input of the deformation pattern discriminator to obtain an output, wherein the output includes a target deformation pattern, including: Acquire a deformation database, wherein the deformation database includes first historical deformation data of a first deformation mode; Extracting features from the first historical deformation data according to the predetermined feature extraction strategy to obtain first historical deformation features; forming a first training data set based on the first historical deformation feature and the first deformation mode; Based on the support vector machine principle, supervised training is performed on the first training data group to obtain a deformation pattern support vector machine; The deformation pattern support vector machine is used as the deformation pattern discriminator.

6. The method according to claim 5, characterized in that Acquire a deformation database, wherein the deformation database includes first historical deformation data of a first deformation mode, including: Clustering the deformation database to obtain a deformation clustering result, wherein the deformation clustering result includes a plurality of clustering clusters; Extracting a first cluster from the plurality of clusters, wherein the first cluster refers to a first database corresponding to the first deformation mode; Any set of data in the first database is randomly extracted and recorded as the first historical deformation data.

7. The method according to claim 1, characterized in that A time series analysis mechanism is introduced to perform time series analysis on the target deformation characteristics to obtain target time series analysis results, including: Establishing a time series model, and arranging the target deformation features in time sequence through the time series model to form a target feature time series; Performing trend analysis on the target feature time series to obtain target trend features; Performing periodic analysis on the target feature time series to obtain target periodic features; Performing randomness analysis on the target feature time series to obtain target randomness features; The target trend characteristics, the target periodic characteristics and the target random characteristics constitute the target time series analysis result.

8. The method according to claim 7, characterized in that After introducing a time series analysis mechanism to perform time series analysis on the target deformation feature and obtaining the target time series analysis result, the method includes: activating a neural network prediction model, and using the target time series analysis result as input information of the neural network prediction model to obtain a first deformation risk coefficient; activating a time series prediction model, and using the target time series analysis result as input information of the time series prediction model to obtain a second deformation risk coefficient; The first deformation risk coefficient and the second deformation risk coefficient are compared and verified to obtain the target deformation risk coefficient.

9. The method according to claim 1, characterized in that: Also includes: Read the predetermined evaluation plan; Evaluate and analyze the target effective deformation monitoring data according to the predetermined evaluation scheme to obtain a fuzzy deformation risk level; The target deformation risk coefficient is verified using the fuzzy deformation risk level.

10. The method according to claim 9, characterized in that The target effective deformation monitoring data is evaluated and analyzed according to the predetermined evaluation scheme to obtain a fuzzy deformation risk level, including: Extracting a predetermined evaluation factor set and a predetermined evaluation grade set in the predetermined evaluation scheme; Constructing a membership matrix according to the predetermined evaluation factor set and the predetermined evaluation grade set; A fuzzy relationship matrix is ​​obtained according to the membership matrix, and the fuzzy deformation risk level is calculated.

Images