A reservoir seepage pressure data intelligent analysis and early warning method

Abstract

The application discloses a reservoir seepage pressure data intelligent analysis and early warning method, extracts the seepage pressure value corresponding to each water level value at each moment from the preliminary screening database; combines the seepage pressure measuring point pipe mouth and the sensor installation elevation data to preliminarily judge and eliminate logical error data; uses the 3Sigma principle to eliminate abnormal seepage pressure data under each water level; adopts feature engineering and Z-score method to generate features and normalize them; maps the relationship between the reservoir water level and the piezometer water level through support vector regression fitting, calculates the model residual and generates the early warning range; finally, compares and analyzes the seepage pressure data and the early warning range, and starts the corresponding early warning mechanism according to the difference result. Based on the characteristics of the reservoir dam seepage pressure monitoring data, the machine learning technology and various statistical techniques are integrated, the accuracy and reliability of data analysis are significantly improved, and the application is especially suitable for the field of reservoir dam seepage pressure data analysis and early warning which needs high precision.

Description

TECHNICAL FIELD

[0001] The present application relates to a kind of reservoir seepage pressure data intelligent analysis early warning method, specifically for the small reservoir seepage pressure monitoring data intelligent analysis and early warning technology under complex conditions. BACKGROUND

[0002] It is known that reservoir dam, as the core of water conservancy infrastructure, not only plays a vital role in flood control, irrigation and power generation, but also its safe operation is directly related to the social and economic stability of the downstream area. Seepage pressure monitoring data is an important indicator to evaluate the operation condition of dam. However, the existing technology mainly relies on finite element analysis and finite difference method to calculate the threshold value of seepage pressure data, but this method has limitations in practical application. First, finite element analysis usually relies on idealized models and assumptions, which cannot fully reflect the complexity of dam materials in actual environment. Second, the finite difference method has low accuracy for complex geometry and nonlinear materials, and the grid division is fixed, which is difficult to adapt to complex boundary conditions and local precision requirements. In addition, the dam materials of many reservoirs are sourced from the local area, and the permeability coefficient is difficult to accurately determine. After years of operation and multiple reinforcement and repair, the uniformity of dam structure may have changed. These factors lead to significant differences between the theoretically calculated seepage pressure threshold and the actual monitoring data, affecting the accuracy of seepage pressure data interpretation and causing potential risks to the safety assessment of dam.

[0003] In view of the above challenges, it is urgent to develop a new seepage pressure data processing and diagnosis method to more accurately process and analyze seepage pressure data, thereby improving the quality of monitoring data and the reliability of analysis results. Therefore, the present application adopts logic error discrimination, 3Sigma principle, feature engineering, Z-score normalization and support vector regression technology to provide a more efficient and accurate seepage pressure data processing scheme to better adapt to the needs of reservoir dam safety monitoring and ensure the safety and stability of dam operation. SUMMARY

[0004] The present application aims to provide a reservoir seepage pressure data intelligent analysis early warning method. This technology can automatically identify logical errors and exclude abnormal data according to data characteristics, and optimize the feature expression of data through feature engineering and Z-score normalization processing. In addition, this technology also uses support vector regression technology for accurate fitting and prediction, and calculates the prediction residual to generate a reliable early warning range, effectively improving the accuracy and reliability of seepage pressure data analysis, especially suitable for data processing and diagnosis in reservoir dam safety monitoring.

[0005] The technical solution adopted by the present application to solve its technical problems is:

[0006] A reservoir seepage pressure data intelligent analysis and early warning method comprises the following steps:

[0007] Step S1: extracting the seepage pressure value corresponding to each water level value at each moment from the long-term continuous water level monitoring data of the reservoir;

[0008] Step S2: Using the sensor calibration data and other information, the extracted osmotic pressure values ​​are preliminarily evaluated, and data with logical errors are excluded. Outliers are processed using the 3Sigma principle.

[0009] Step S3: Generate basic features and derived features through feature engineering, and normalize these features using the Z-score standardization method;

[0010] Step S4: Based on cross-validation, select the SVR kernel function and corresponding hyperparameter combination that performs best on the current dataset;

[0011] Step S5: Solve the optimization problem of the SVR model and its dual problem, and construct a regression function for prediction;

[0012] Step S6: Through MSE ( M ean- S square E rror,MSE) and R 2 The score evaluates the model performance;

[0013] Step S7: Generate a warning range by calculating the residual of the SVR model and combining the standard deviation of the residual;

[0014] Step S8: Compare and analyze the actual seepage pressure data with the warning range, and activate the warning mechanism based on the difference results.

[0015] Further technology of the present invention:

[0016] Preferably, step S1 is specifically as follows:

[0017] S11. Perform data screening and cleaning operations in the reservoir long-sequence water level monitoring database. In particular, handle the situation where multiple reservoir water level values ​​exist at the same time to eliminate data redundancy and ensure data consistency. If multiple reservoir water level values ​​are recorded at the same time, retain the reservoir water level value that is consistent with the piezometric water level time. If there is no completely consistent time record, retain the reservoir water level value that is closest to the piezometric water level time.

[0018] S12, the only reservoir water level value H after screening res And the pressure tube water level value H pieMatching is performed to ensure that there is a unique reservoir water level corresponding to the piezometer water level value at each time point, providing consistent and complete input data for subsequent data analysis and diagnosis.

[0019] Preferably, step S2 is specifically:

[0020] S21, for each reservoir water level and its corresponding osmotic pressure data, respectively calculate its mean (μ 水位 , μ 渗压 ) and standard deviation (σ 水位 , σ 渗压 ). The values of reservoir water level exceeding (μ 水位 -3σ 水位 , μ 水位 +3σ 水位 ) and osmotic pressure value exceeding (μ 渗压 -3σ 渗压 , μ 渗压 +3σ 渗压 ) are marked as abnormal points and excluded;

[0021] S22, divide the preprocessed data into training set and test set and input the prediction model, wherein the training set accounts for 80% of the effective database, and the test set accounts for 20% of the effective database.

[0022] Preferably, step S3 is specifically:

[0023] S31, generate basic features and derived features through feature engineering. Among them, the basic features include reservoir water level and piezometer water level, and the derived features include time features, rolling mean, difference features, lag features and polynomial features;

[0024] S32, normalize the features by Z-score standardization method, convert the data of each feature to standard normal distribution with mean of 0 and standard deviation of 1, so that different features have comparability. When Z-score standardization is performed on feature X, the following formula is applied to each data point in the feature:

[0025]

[0026] Where X i is a data point in feature X, μ X is the mean of feature X, and σ X is the standard deviation of feature X.

[0027] Preferably, step S4 is specifically:

[0028] S41, design cross-validation scheme, determine the number of folds and use mean square error and R 2As an index, and define the SVR parameter search space, including kernel function type (linear, radial basis function, polynomial) and the corresponding hyperparameter combination (penalty parameter C, error tolerance ∈, kernel parameter);

[0029] S42, perform cross-validation, evaluate different kernel functions and hyperparameter combinations, and select the best configuration for SVR model training.

[0030] Preferably, step S5 is specifically:

[0031] S51, determine the input variables and output variables of the model;

[0032] Input variables x i : reservoir water level, one or more features;

[0033] Output variables y i : piezometer water level;

[0034] S52, the optimization goal of the SVR model is to find a function y that fits the data points within the ∈ tolerance range and minimizes the complexity of the model. This optimization problem is defined as:

[0035]

[0036] s.t.

[0037]

[0038] Where represents the true target value of each data point in the training data; ω represents the weight vector of the model, which represents the coefficients that define the decision function; b is the bias term in the model, which determines the translation of the decision function; and ξ i are slack variables that allow the model to deviate in some cases, enabling the model to adapt to some noisy data; C represents the regularization parameter, which controls the trade-off between fitting error and model complexity on the training data; ∈ represents the tolerance error range, which defines the size of the error that the model can tolerate, and errors exceeding this range will be penalized; φ(x i ) represents the function that maps the input variable x i to a high-dimensional feature space, mapping the input features to a high-dimensional feature space; represents the model complexity, is the penalty term for errors exceeding the ∈ range, and the constraint condition represents that for each data point, if the difference between the model's predicted value ω·φ(x i )+b and is within ∈, then

[0039] S53, the original problem can be converted into a dual problem by the Lagrange multiplier method, the Lagrange function of the original problem is:

[0040]

[0041]

[0042] Where L represents the objective function of the dual problem transformed by the Lagrange multiplier method; α, α * is the Lagrange multiplier, corresponding to the relaxation variable ξ i and They are used to constrain conditions and control the punishment degree of error terms; η, η * Used to control the influence of relaxation variables on the objective function, the Lagrange function of the original problem is obtained by taking the derivative of ω, b, ξ i , Derivative and let it be 0 and after simplification get the dual problem of the original problem;

[0043] S54, by solving the dual problem of the original problem to get the final regression function:

[0044]

[0045] Where f(x) is a regression function, representing the prediction value of the final regression model obtained by solving the dual problem of the original problem, n is the total number of training samples, K(x i , x) is the kernel function, representing the inner product of x i and x in high-dimensional space; b is the bias term, obtained by the condition of support vector.

[0046] Preferably, step S6 is specifically:

[0047] S61, calculate MSE and R 2 Score, the specific calculation formula is as follows:

[0048]

[0049] Where n is the number of data points, is the actual value of the i-th piezometer water level, y i is the predicted value of the i-th piezometer water level, is the mean value of the actual value of the piezometer water level. R 2 The value range of score is [0, 1], the closer to 1 indicates that the model is better fitted to the data. MSE measures the average squared difference between the predicted value and the actual value, the closer to 0, the better the prediction effect of the model;

[0050] S62, scatter plot is drawn to visualize the difference between actual value and predicted value.

[0051] Preferably, step S7 is specifically:

[0052] S71, the residual of each data point, i.e. the difference between the actual piezometer water level value and the corresponding predicted value, is calculated, and the specific calculation formula is as follows:

[0053]

[0054] S72, the standard deviation of the residual is calculated, which is used to estimate the distribution of the model prediction error, and the specific calculation formula is as follows:

[0055]

[0056] Where σ res is the standard deviation, represents the difference between the actual piezometer water level value and the corresponding predicted value, is the average value of the residual, and n is the number of data points.

[0057] S73, the warning range is calculated. By analyzing the historical abnormal data and comprehensively considering the risk control and actual demand, 97% confidence level is selected to improve the identification accuracy of abnormal water level; then the upper and lower limits of the warning range are calculated according to the confidence level. The calculation formula of the warning range is as follows:

[0058]

[0059] Where is the lower limit of the warning range, i.e. the water level exceeds will start the corresponding warning, and for 97% confidence level, the critical value z≈2.17.

[0060] Preferably, step S8 is specifically:

[0061] S81, the newly collected osmotic pressure monitoring data is plotted on the warning range curve, and by comparing and analyzing the actual osmotic pressure data with the warning range , it is determined whether the data is out of the warning range. When the actual osmotic pressure data is within the warning range, the data is automatically archived into the database; when the actual osmotic pressure data exceeds the upper limit of the interval or is lower than the lower limit of the interval , the warning mechanism is started.

[0062] The beneficial effects of the present application are that, by strictly screening high-quality sample data, eliminating logical error data, removing outliers using the 3Sigma principle, and utilizing feature engineering, Z-score normalization and support vector regression technology, the accuracy of data processing and diagnosis is further enhanced, making the osmotic pressure monitoring and prediction more accurate and reliable. The present application is particularly suitable for safety monitoring of key infrastructures such as reservoir dams, ensuring the efficiency and robustness of data processing, and improving the reliability and response capability of the overall monitoring system. DETAILED DESCRIPTION

[0063] In order to make the technical means, creative features, purposes and effects realized by the present application easy to understand, the present application will be further illustrated below in combination with specific embodiments.

[0064] The present application provides a reservoir seepage pressure data intelligent analysis and early warning method, comprising:

[0065] Step S1: In the long-term continuous water level monitoring data of the reservoir, the seepage pressure value corresponding to each water level value at each time is extracted.

[0066] S11, in the long sequence water level monitoring database of the reservoir, data screening and cleaning operation is performed. In particular, the case of multiple reservoir water level values existing at the same time is processed to eliminate data redundancy and ensure data consistency. For the case of multiple reservoir water level values recorded at the same time, the reservoir water level value consistent with the piezometric tube water level time is retained. If there is no completely consistent time record, the reservoir water level value closest to the piezometric tube water level time is retained;

[0067] S12, match the unique reservoir water level value H res with the piezometric tube water level value H pie , to ensure that there is a unique reservoir water level and corresponding piezometric tube water level value at each time point, providing consistent and complete input data for subsequent data analysis and diagnosis.

[0068] Step S2: Use the calibration data and other information of the sensor to preliminarily evaluate the extracted seepage pressure value, and exclude data with logical errors, and apply the 3Sigma principle to process outliers.

[0069] S21, for each reservoir water level and its corresponding seepage pressure data, calculate its mean (μ 水位 , μ 渗压 ) and standard deviation (σ 水位 , σ 渗压 ) respectively. The reservoir water level exceeds (μ 水位 -3σ 水位 , μ 水位 +3σ 水位 ), and the seepage pressure value exceeds (μ 渗压-3σ 渗压 , μ 渗压 +3σ 渗压 ) range are marked as outliers and removed;

[0070] S22, divide the preprocessed data into training set and test set and input the prediction model, wherein the training set accounts for 80% of the effective database, and the test set accounts for 20% of the effective database.

[0071] Step S3: generate basic features and derived features through feature engineering, and normalize these features using Z-score standardization method.

[0072] S31, basic features and derived features are generated through feature engineering. Among them, the basic features include reservoir water level and piezometric tube water level, and the derived features include time features, rolling mean, difference features, lag features and polynomial features;

[0073] S32, normalize the features by Z-score standardization method, convert the data of each feature to standard normal distribution with mean of 0 and standard deviation of 1, so that different features have comparability. When Z-score standardization is performed on feature X, the following formula is applied to each data point in the feature:

[0074]

[0075] Where X i is a data point in feature X, μ X is the mean of feature X, and σ X is the standard deviation of feature X.

[0076] Step S4: based on cross-validation, select the SVR kernel function and corresponding hyperparameter combination that performs best on the current dataset.

[0077] S41, design cross-validation scheme, determine the number of folds and use mean squared error and R 2 as indicators, and define SVR parameter search space, including kernel function type (linear, radial basis function, polynomial) and corresponding hyperparameter combination (penalty parameter C, error tolerance ∈, kernel parameter);

[0078] S42, perform cross-validation, evaluate different kernel functions and hyperparameter combinations, and select the best configuration for SVR model training.

[0079] Step S5: solve the optimization problem and its dual problem of SVR model, and construct the regression function for prediction.

[0080] S51, determine the input variables and output variables of the model;

[0081] input variable x i is: reservoir water level (one or more features);

[0082] output variable y i is: piezometer water level;

[0083] S52, the optimization goal of the SVR model is to find a function y that fits the data points within the ∈ tolerance range and minimizes the complexity of the model. This optimization problem is defined as:

[0084]

[0085] s.t.

[0086]

[0087] S52, the optimization goal of the SVR model is to find a function y that fits the data points within the ∈ tolerance range and minimizes the complexity of the model. This optimization problem is defined as:

[0088]

[0089] s.t.

[0090]

[0091] where represents the true target value of each data point in the training data; ω represents the weight vector of the model, which defines the coefficients of the decision function; b is the bias term in the model, which determines the translation of the decision function; and ξ i are slack variables that allow the model to make deviations in some cases, enabling the model to adapt to some noisy data; C represents the regularization parameter, which controls the trade-off between fitting error and model complexity on the training data; ∈ represents the tolerance error range, which defines the size of the error that the model can tolerate, and errors beyond this range will be penalized; φ(x i ) represents the function that maps the input features x i into a high-dimensional feature space; represents the model complexity, is the penalty term for errors beyond the ∈ range. The constraint condition indicates that for each data point, if the predicted value of the model ω·φ(x i )+b and the true value are within ∈, then

[0092] S53, by the method of Lagrange multipliers, the original problem can be converted into a dual problem, and the Lagrange function of the original problem is For:

[0093]

[0094] where L represents the objective function of the dual problem transformed by the Lagrange multiplier method; α, α * is the Lagrange multiplier, corresponding to the relaxation variable ξ i and which are used to constrain the conditions and control the punishment degree of the error term; η, η * are used to control the influence of the relaxation variable on the objective function. The Lagrange function of the original problem is obtained by taking the derivative of ω, b, ξ i , and setting it to 0 and simplifying;

[0095] S54, the final regression function is obtained by solving the dual problem of the original problem:

[0096]

[0097] where f(x) is a regression function, representing the predicted value of the final regression model obtained by solving the dual problem of the original problem. n is the total number of training samples, K(x i , x) is the kernel function, representing the inner product of input data points x i and x in high-dimensional space; b is the bias term, which can be obtained by the condition of support vector.

[0098] Step S6: Evaluate the model performance by MSE M ean- S quare E rror, MSE) and R 2 square.

[0099] S61, calculate MSE and R 2 square, the specific calculation formula is as follows:

[0100]

[0101] where n is the number of data points, is the actual value of the i-th piezometer water level, y i is the predicted value of the i-th piezometer water level, is the mean value of the actual value of the piezometer water level. R 2 square ranges from 0 to 1, and the closer to 1 indicates the better fitting degree of the model to the data. MSE measures the average squared difference between the predicted value and the actual value, and the closer to 0, the better the prediction effect of the model;

[0102] S62, scatter plot is drawn to visualize the difference between actual and predicted values.

[0103] Step S7: generate the early warning range by calculating the residual of the SVR model and combining the standard deviation of the residual.

[0104] S71, calculate the residual of each data point, i.e. the difference between the actual piezometer water level value and the corresponding predicted value, the specific calculation formula is as follows:

[0105]

[0106] S72, calculate the standard deviation of the residual, which is used to estimate the distribution of the model prediction error, the specific calculation formula is as follows:

[0107]

[0108] where is the average value of the residual, n is the number of data points;

[0109] S73, calculate the early warning range. By analyzing historical abnormal data and considering risk control and actual demand, select the confidence level of 97% to improve the identification accuracy of abnormal water level; then calculate the upper and lower limits of the early warning range according to the confidence level. The calculation formula of the early warning range is as follows:

[0110]

[0111] where for the confidence level of 97%, the critical value z≈2.17.

[0112] Step S8: compare and analyze the actual osmotic pressure data with the early warning range, and start the early warning mechanism according to the difference result.

[0113] S81, plot the newly collected osmotic pressure monitoring data on the early warning range curve, compare and analyze the actual osmotic pressure data with the early warning range , and determine whether the data is out of the early warning range. When the actual osmotic pressure data is within the early warning range, the data is automatically archived into the database; when the actual osmotic pressure data exceeds the upper limit or is lower than the lower limit , the early warning mechanism is started.

[0114] By comprehensively using sample preliminary screening, logic error identification, 3Sigma principle, feature engineering, Z-score normalization and support vector regression technology, the ability of osmotic pressure data processing and diagnosis is significantly improved. This method not only enhances the accuracy and stability of data analysis, but also improves the practical application value of osmotic pressure monitoring and prediction, especially suitable for application in the field of reservoir safety and other engineering safety.

[0115] The above merely describes preferred embodiments of the present application, and is not intended to limit the present application in any form; any person skilled in the art, without departing from the technical scheme of the present application, can make many possible changes and modifications to the technical scheme of the present application, or modify equivalent embodiments with equivalent changes. Therefore, any simple modification, equivalent replacement, equivalent change and modification made to the above embodiments according to the technical essence of the present application, without departing from the technical scheme of the present application, still falls within the scope of protection of the technical scheme of the present application.

Claims

1. A reservoir seepage pressure data intelligent analysis and early warning method, characterized in that: The steps include: Step S1: extracting the seepage pressure value corresponding to each water level value at each moment from the long-term continuous water level monitoring data of the reservoir; Step S1 is specifically as follows: S11. Perform data screening and cleaning operations in the reservoir long-sequence water level monitoring database; In S11: The situation where there are multiple reservoir water level values ​​at the same time is handled to eliminate data redundancy and ensure data consistency. If multiple reservoir water level values ​​are recorded at the same time, the reservoir water level value that is consistent with the piezometer water level time is retained. If there is no completely consistent time record, the reservoir water level value that is closest to the piezometer water level time is retained; S12, the only reservoir water level value H after screening res And the pressure tube water level value H pie Matching ensures that at each time point there is a unique reservoir water level and corresponding piezometric pipe water level value, providing consistent and complete input data for subsequent data analysis and diagnosis; Step S2: Using the sensor calibration data information, the extracted osmotic pressure values ​​are preliminarily evaluated, and data with logical errors are excluded. Outliers are processed using the 3Sigma principle. Step S3: Generate basic features and derived features through feature engineering, and normalize these features using the Z-score standardization method; Step S4: Based on cross-validation, select the SVR kernel function and corresponding hyperparameter combination that performs best on the current dataset; Step S5: Solve the optimization problem of the SVR model and its dual problem, and construct a regression function for prediction; Step S6: Through MSE and R 2 The score evaluates the model performance; Step S7: Generate a warning range by calculating the residual of the SVR model and combining the standard deviation of the residual; Step S7 is specifically as follows: S71. Calculate the residual of each data point, that is, the difference between the actual pressure tube water level value and the corresponding predicted value. The specific calculation formula is as follows: in is the actual value of the water level of the ith pressure gauge, y i is the predicted value of the water level of the i-th piezometer; S72. Calculate the standard deviation of the residuals to estimate the distribution of the model prediction error. The specific calculation formula is as follows: where σ res is the standard deviation, Represents the difference between the actual pressure tube water level value and the corresponding predicted value, is the mean of the residuals, n is the number of data points; S73. Calculate the warning range. By analyzing historical abnormal data and taking into account risk control and actual needs, a 97% confidence level is selected to improve the accuracy of identifying abnormal water levels. Then, the upper and lower limits of the warning range are calculated based on the confidence level. The calculation formula for the warning range is as follows: in Is the lower limit of the warning range, that is, the water level exceeds The corresponding warning will be activated. For a 97% confidence level, the critical value z≈2.17; Step S8: Compare and analyze the actual seepage pressure data with the warning range, and activate the warning mechanism based on the difference results.

2. According to the method for intelligent analysis and early warning of reservoir seepage pressure data described in claim 1, it is characterized in that: Step S2 is specifically as follows: S21. Calculate the mean (μ) of the water level and its corresponding seepage pressure data of each reservoir. 水位 、μ 渗压 ) and standard deviation (σ 水位 , σ 渗压 ), the reservoir water level exceeds (μ 水位 -3σ 水位 、μ 水位 +3σ 水位 ), osmotic pressure value exceeds (μ 渗压 -3σ 渗压 、μ 渗压 +3σ 渗压 ) range are marked as outliers and removed; S22. Divide the preprocessed data into a training set and a test set and input them into the prediction model.

3. According to the method for intelligent analysis and early warning of reservoir seepage pressure data as described in claim 2, it is characterized in that: In S22, the training set accounts for 80% of the effective database, and the test set accounts for 20% of the effective database.

Images