Water conservancy and hydropower engineering real-time monitoring management method and system

Abstract

The invention discloses a real-time monitoring management method and system for water conservancy and hydropower engineering, and relates to the field of water conservancy and hydropower. The method comprises the steps that environment data X and effect data Y are collected, and a correlation coefficient matrix between the environment data X and the effect data Y is constructed; constructing an anomaly detection algorithm library, and constructing an evaluation index matrix according to the correlation coefficient matrix and the anomaly detection algorithm library; according to the evaluation index matrix, selecting an optimal anomaly detection algorithm for each type of monitoring data; carrying out abnormal value processing on the collected monitoring data according to an optimal abnormal detection algorithm; grouping the monitoring data subjected to abnormal value processing according to seasons and working conditions, calculating a mean value and a standard deviation of each group of data through statistics in combination with a correlation coefficient matrix, and obtaining a mean value + / -2 sigma as a first-level early warning threshold value and a mean value + / -3 sigma as a second-level early warning threshold value; and performing monitoring management according to the first-level early warning threshold and the second-level early warning threshold. In view of low monitoring precision caused by lack of flexibility in hydraulic engineering threshold setting in the prior art, the monitoring precision is improved.

Description

Technical Field

[0001] This application relates to the field of water conservancy and hydropower treatment, and particularly to a real-time monitoring and management method and system for water conservancy and hydropower projects. Background Art

[0002] Water conservancy and hydropower projects are important national infrastructure, directly related to flood control safety, water resource utilization, power supply, and ecological environment protection. With the rapid development of China's water conservancy and hydropower undertakings, the number of large-scale water conservancy and hydropower projects is increasing continuously, and the safe operation management of the projects is facing more and more complex challenges. Water conservancy and hydropower projects are affected by various environmental factors such as water level changes, rainfall, and seasonal temperature differences, and the project status is reflected through effects such as dam deformation and seepage changes. These factors interact with each other to form a complex dynamic system.

[0003] Real-time monitoring is the key means to ensure the safe operation of water conservancy and hydropower projects. Traditional water conservancy and hydropower project monitoring systems collect environmental data such as water level and rainfall and effect data such as dam deformation and seepage through various monitoring devices, forming a large amount of monitoring data streams. After being processed and analyzed, these data provide a basis for project safety assessment and early warning decision-making. With the development of the Internet of Things and big data technologies, the automation of monitoring devices and the real-time nature of data collection have been significantly improved, creating conditions for more accurate project monitoring.

[0004] However, the effective utilization of monitoring data still faces many challenges. On the one hand, monitoring data is easily affected by factors such as equipment failures, transmission interference, and environmental changes, resulting in outliers; on the other hand, there are complex correlation relationships between environmental data and effect data, and this relationship varies significantly with seasonal changes and different project operation conditions.

[0005] For example, most existing systems set warning thresholds using fixed thresholds or simple statistical models. These thresholds are usually determined based on the overall statistical characteristics or empirical values of historical data, lacking differential considerations for different seasons and working conditions. For example, the normal data fluctuation ranges in the summer flood season and the winter dry season are significantly different, and using a unified threshold will inevitably lead to a decrease in monitoring accuracy. Summary of the Invention

[0006] Aiming at the low monitoring accuracy caused by the lack of flexibility in threshold setting in the prior art, this application provides a real-time monitoring and management method and system for water conservancy and hydropower projects. By comprehensively scoring to select the optimal anomaly detection algorithm and combining the season - working condition combination to generate differential warning thresholds, the monitoring accuracy is improved.

[0007] The purpose of this application is achieved through the following technical solutions.

[0008] One aspect of the present application provides a real-time monitoring and management method for water conservancy and hydropower projects, including: S1, collecting monitoring data of water conservancy and hydropower projects, where the monitoring data includes environmental data X and effect data Y. The environmental data X includes water level and rainfall; the effect data Y includes dam deformation and seepage; S2, using the multiple linear regression method to construct a correlation coefficient matrix between the environmental data X and the effect data Y; S3, constructing an anomaly detection algorithm library, which includes the mean method, median method, difference method, Hampel filtering method, and 3σ criterion; S4, constructing an evaluation index matrix based on the correlation coefficient matrix and the anomaly detection algorithm library; S5, selecting the optimal anomaly detection algorithm for each type of monitoring data according to the evaluation index matrix; S6, performing outlier processing on the collected monitoring data according to the optimal anomaly detection algorithm; S7, grouping the monitoring data after outlier processing by season and working conditions, and combining the correlation coefficient matrix, calculating the mean and standard deviation of each group of data through statistical calculation, and obtaining the mean ± 2σ as the first-level early warning threshold and the mean ± 3σ as the second-level early warning threshold; S8, performing monitoring and management according to the first-level early warning threshold and the second-level early warning threshold.

[0009] Further, in S2, using the multiple linear regression method to construct a correlation coefficient matrix between the environmental data X and the effect data Y includes: respectively processing the environmental data X and the effect data Y using Z-score. Among them, the environmental data X = {X1, X2}, X1 represents the water level data, and X2 represents the rainfall data; the effect data Y = {Y1, Y2}, Y1 represents the dam deformation data, and Y2 represents the seepage data; respectively constructing time-lag models Y1(t) and Y2(t) for the dam deformation data Y1 and the seepage data Y2; according to the time-lag models Y1(t) and Y2(t), through the sliding time window method, calculating the Pearson correlation coefficient r between each environmental data X = {X1, X2} and the effect data Y = {Y1, Y2} at different lag times,

[0010] r(X j (t - i), Y l (t)), i ∈ [0, k], j ∈ {1, 2}, l ∈ {1, 2}, where k is the maximum lag order; i and j represent positive integers};

[0011] For each combination of the environmental data X j and the effect data Y l , from the k + 1 Pearson correlation coefficients r(X j (t - i), Y l (t)), select the correlation coefficient r max with the largest absolute value and its corresponding optimal lag time t opt , and form a 2×2 time-delay correlation coefficient matrix R. The matrix element R j,l contains two parameter values {rmax ,t opt}, where \(j\in\{1,2\}\) represents the environmental data type and \(l\in\{1,2\}\) represents the effect data type; \(R\) 1,1 represents the maximum correlation coefficient between the water level \(X1\) and the dam deformation \(Y1\) and its corresponding lag time; \(R\) 1,2 represents the maximum correlation coefficient between the water level \(X1\) and the seepage \(Y2\) and its corresponding lag time; \(R\) 2,1 represents the maximum correlation coefficient between the rainfall \(X2\) and the dam deformation \(Y1\) and its corresponding lag time; \(R\) 2,2 represents the maximum correlation coefficient between the rainfall \(X2\) and the seepage \(Y2\) and its corresponding lag time;

[0012] Furthermore, the time-lag models \(Y1(t)\) and \(Y2(t)\): where \(t\) represents the current time point; \(k\) is the maximum lag order, determined according to the physical characteristics and historical data analysis of water conservancy and hydropower projects, and the value range is 1 - 30 days; \(\alpha1,\alpha2\) are the constant terms of each model; \(\beta\) 1i ,\(\beta\) 2i are the regression coefficients of the water level data \(X1\) at different lag times in the two effect data models respectively; \(\gamma\) 1i ,\(\gamma\) 2i are the regression coefficients of the rainfall data \(X2\) at different lag times in the two effect data models respectively; \(\varepsilon1,\varepsilon2\) are the random error terms of each model.

[0013] Furthermore, in S4, according to the correlation coefficient matrix and the anomaly detection algorithm library, an evaluation index matrix is constructed, including: according to the optimal lag time \(t\) in the time-delay correlation coefficient matrix \(R\) opt , the environmental data \(X\) is lagged to form a lagged environmental data set \(X'\) that is most correlated with the effect data \(Y\); a data set is constructed based on the environmental data set \(X'\) and the effect data \(Y\), and the data set is divided into a training set and a test set; each anomaly detection algorithm in the anomaly detection algorithm library is used to detect anomalies in the training set data respectively; the evaluation indexes of the detection results of each anomaly detection algorithm are calculated using the test set, and the evaluation indexes include the detection accuracy \(A\), the false alarm rate \(B\), the miss rate \(C\), and the robustness index \(D\); a \(4\times5\) evaluation index matrix \(E\) is constructed, and the matrix element \(E\) i,j contains the anomaly detection evaluation results, where \(i\in\{1,2,3,4\}\) represents the monitoring data type (\(X1\) water level, \(X2\) rainfall, \(Y1\) dam deformation, \(Y2\) seepage), and \(j\in\{1,2,3,4,5\}\) represents the anomaly detection algorithm type (mean method, median method, difference method, Hampel filtering method, and 3\(\sigma\) criterion); each matrix element \(E\) i,j stores the four evaluation indexes \(\{A,B,C,D\}\) of the corresponding data type applying the corresponding anomaly detection algorithm.

[0014] Further, in S5, according to the evaluation index matrix, select the optimal anomaly detection algorithm for each type of monitoring data, including: setting the weight coefficients w A 、w B 、w C 、w D for each evaluation index in the evaluation index matrix; for each type of monitoring data, calculate the comprehensive score S i,j : S i,j = w A × A i,j - w B × B i,j - w C × C i,j + w D × D i,j ; where i ∈ {1, 2, 3, 4} represents the type of monitoring data, and j ∈ {1, 2, 3, 4, 5} represents the type of anomaly detection algorithm.

[0015] Further, in S5, according to the evaluation index matrix, selecting the optimal anomaly detection algorithm for each type of monitoring data further includes: for environmental data X1 and X2, adding a correlation coefficient correction term when calculating the comprehensive score: S i,j ' = S i,j × (1 + 0.2 × |r max |), where r ∈ {1, 2} represents environmental data, and |r max | is the absolute value of the maximum correlation coefficient corresponding to the environmental data in the time-delay correlation coefficient matrix R; for each type of monitoring data, select the anomaly detection algorithm with the highest comprehensive score as the optimal algorithm for this data type: represents effect data Y1 and Y2; i ∈ {1, 2} represents environmental data X1 and X2; where A opt (i) is the optimal anomaly detection algorithm corresponding to the monitoring data type i.

[0016] Further, in S7, obtain the mean ± 2σ as the first-level warning threshold and the mean ± 3σ as the second-level warning threshold, including: dividing the monitoring data after outlier processing into four groups by season; according to the water level state, dividing the monitoring data after outlier processing into groups by working conditions, where the working conditions include normal water storage conditions and flood season precipitation conditions; according to each season and working condition combination, form season - working condition combinations, including spring - normal water storage, spring - flood season precipitation, summer - normal water storage, summer - flood season precipitation, autumn - normal water storage, autumn - flood season precipitation, winter - normal water storage, and winter - flood season precipitation; for each season - working condition combination, calculate the statistical parameters of each type of monitoring data respectively, where the statistical parameters include mean μ, standard deviation σ, and skewness coefficient s; based on the data of each season - working condition combination, use the optimal lag time t opt, calculate the real-time correlation response coefficient between the environmental data X and the effect data Y in each group of data According to the statistical parameters of each season-operation condition combination, combined with the calculated real-time correlation response coefficient Calculate the warning threshold: For the effect data Y1 and Y2, the warning threshold calculation formula is: μ ± kσ, where k is the warning level coefficient; For the environmental data X1 and X2, when |r max | > 0.7, the warning threshold calculation formula is: Where: δ is the correlation adjustment coefficient, with a value of 0.2; when |r max | ≤ 0.7, the warning threshold calculation formula is: μ ± kσ, where k is the warning level coefficient; Generate the first-level warning threshold and the second-level warning threshold for each group of data; Integrate the warning thresholds under all season-operation condition combinations into the warning threshold matrix Θ, and the matrix element Θ i,j,m Represents the warning threshold interval of the monitoring data type i in season j and operation condition m,

[0017] Where: i ∈ {1, 2, 3, 4} respectively represent the monitoring data types X1 water level, X2 rainfall, Y1 dam deformation, Y2 seepage; j ∈ {1, 2, 3, 4} respectively represent spring, summer, autumn, winter; m ∈ {1, 2} respectively represent normal water storage condition, flood season precipitation condition; Represents the first-level warning lower limit of the monitoring data type i in season j and operation condition m; Represents the first-level warning upper limit of the monitoring data type i in season j and operation condition m; Represents the second-level warning lower limit of the monitoring data type i in season j and operation condition m; Represents the second-level warning upper limit of the monitoring data type i in season j and operation condition m;

[0018] Furthermore, the real-time correlation response coefficient The calculation formula is: Where: p ∈ {1, 2} represents the environmental data type (X1 water level, X2 rainfall); q ∈ {1, 2} represents the effect data type (Y1 dam deformation, Y2 seepage); t opt,p,q Is the optimal lag time of the corresponding combination in the time-delay correlation coefficient matrix R; μ p And σ p Are respectively the mean and standard deviation of the environmental data X p In the current season-operation condition combination; μ q And σ q Are respectively the mean and standard deviation of the effect data Y q In the current season-operation condition combination; n is the sample quantity in the current season-operation condition combination.

[0019] Furthermore, the first-level warning threshold and the second-level warning threshold for each group of data are generated, and the formula is:

[0020] First-level warning threshold:

[0021] Second-level warning threshold:

[0022] Among them, μ i,j,m is the mean value of the monitoring data type i under season j and working condition m; σ i,j,m is the standard deviation of the monitoring data type i under season j and working condition m; i ∈ {1, 2, 3, 4} respectively represent the monitoring data types X1 water level, X2 rainfall, Y1 dam deformation, Y2 seepage; j ∈ {1, 2, 3, 4} respectively represent spring, summer, autumn, winter; m ∈ {1, 2} respectively represent normal water storage working condition, flood season precipitation working condition.

[0023] C i,j,m is the correction coefficient, which is determined according to the data characteristics:

[0024] For effect data (i ∈ {3, 4}) and the data is normally distributed |s i,j,m | < 0.5, C i,j,m = 1; For effect data (i ∈ {3, 4}) and the data is skewed |s i,j,m | ≥ 0.5, the upper warning limit takes C i,j,m = 1 + 0.1 × s i,j,m and the lower warning limit takes C i,j,m = 1 - 0.1 × s i,j,m ;

[0025] For environmental data (i ∈ {1, 2}) and |r max | > 0.7 (strong correlation):

[0026] When the upper warning limit takes the lower warning limit takes

[0027] When the upper warning limit takes the lower warning limit takes

[0028] For environmental data (i ∈ {1, 2}) and |r max | ≤ 0.7 (weak correlation), C i,j,m = 1;

[0029] is the defined real-time correlation response coefficient; p ∈ {1, 2} represents the environmental data type, where p = 1 when i = 1 and p = 2 when i = 2; q ∈ {1, 2} represents the effect data type with the greatest correlation with the environmental data p; s i,j,m is the skewness coefficient of the monitoring data type i under season j and working condition m; δ is the correlation adjustment coefficient, with a fixed value of 0.2; r max is the maximum correlation coefficient corresponding to the environmental data i and the effect data in the time-delay correlation coefficient matrix R.

[0030] Another aspect of the present application also provides a real-time monitoring and management system for water conservancy and hydropower projects, which is used to execute a real-time monitoring and management method for water conservancy and hydropower projects of the present application.

[0031] Compared with the prior art, the advantages of the present application are as follows:

[0032] Due to the complex influence of various environmental factors (such as seasonal changes, water level fluctuations, rainfall changes) and engineering conditions in water conservancy and hydropower projects, there is a characteristic of an obvious time-delay relationship between environmental data and effect data. The prior art generally uses fixed thresholds or simple statistical methods for monitoring, resulting in low warning accuracy, high false alarm rates, and inability to adapt to changes in different seasons and working conditions. The present application realizes the accurate identification and processing of abnormal values of monitoring data, the precise quantification of the time-delay relationship between environmental-effect data, and the dynamic adaptive adjustment of warning thresholds by establishing a time-delay correlation model for environmental and effect data, implementing an optimal anomaly detection algorithm selection based on comprehensive scoring, and constructing a differential warning threshold matrix for season-working condition combinations, thereby significantly improving the accuracy and reliability of real-time monitoring of water conservancy and hydropower projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The present application will be further described in the form of exemplary embodiments, and these exemplary embodiments will be described in detail through the drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where:

[0034] Figure 1 is an exemplary flowchart of a real-time monitoring and management method for water conservancy and hydropower projects shown in some embodiments of the present application;

[0035] Figure 2 is an exemplary flowchart of constructing the time-delay correlation coefficient matrix R shown in some embodiments of the present application;

[0036] Figure 3 is an exemplary flowchart of constructing the evaluation index matrix E method shown in some embodiments of the present application;

[0037] Figure 4 is an exemplary flowchart of generating warning thresholds shown in some embodiments of the present application. Detailed implementation manners

[0038] The methods and systems provided by the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0039] As Figure 1 shown, collect the monitoring data of water conservancy and hydropower projects. The monitoring data includes environmental data X and effect data Y. The environmental data X includes water level and rainfall; the effect data Y includes dam deformation and seepage; use the multiple linear regression method to construct the correlation coefficient matrix between the environmental data X and the effect data Y; construct an anomaly detection algorithm library, and the anomaly detection algorithm library includes the mean method, the median method, the difference method, the Hampel filtering method and the 3σ criterion; construct an evaluation index matrix according to the correlation coefficient matrix and the anomaly detection algorithm library; select the optimal anomaly detection algorithm for each type of monitoring data according to the evaluation index matrix; perform outlier processing on the collected monitoring data according to the optimal anomaly detection algorithm; group the monitoring data after outlier processing by season and working conditions, and combine the correlation coefficient matrix, and obtain the mean ± 2σ as the first-level warning threshold and the mean ± 3σ as the second-level warning threshold by statistically calculating the mean and standard deviation of each group of data; perform monitoring and management according to the first-level warning threshold and the second-level warning threshold.

[0040] S1. Collect the monitoring data of water conservancy and hydropower projects; install automatic monitoring equipment at key positions of the reservoir dam, including water level gauges, rain gauges, deformation measuring instruments and seepage monitoring equipment. The water level gauges are installed at representative positions in the reservoir area and the downstream river channel; the rain gauges are arranged around the reservoir area and the upstream catchment area; the deformation monitoring equipment (such as GNSS reference stations, total stations, displacement gauges, etc.) is installed on the dam body, and the dam crest and dam slope are mainly monitored; the seepage monitoring equipment (such as piezometers, seepage flow measurement weirs, etc.) is arranged in the downstream and foundation areas of the dam.

[0041] Set different collection frequencies according to the project characteristics and monitoring requirements. Usually, the water level data is collected once every 1 hour; the rainfall data is accumulated once every 1 hour; the dam deformation data is collected once every 6 - 12 hours; the seepage data is collected once every 6 hours. During the flood season or under special weather conditions, the collection frequency can be increased, such as the water level and rainfall are adjusted to be collected once every 30 minutes.

[0042] Adopt the Internet of Things technology to realize the real-time transmission of monitoring data, and transmit the data to the monitoring center through wired network, wireless network or satellite communication. The data adopts a hierarchical storage strategy. The original data is stored in the local database, and the data after preliminary processing is transmitted to the cloud data center to ensure data security and efficient access.

[0043] Preprocess the collected raw data, including unit conversion, numerical correction, and basic error screening. For example, convert seepage data from milliliters per minute to liters per day, convert dam deformation data from instrument readings to actual displacements, and at the same time eliminate error data that significantly exceeds the range.

[0044] Classify and store the preprocessed data into two major categories: environmental data (X) and effect data (Y). Environmental data X includes water level X1 and rainfall X2; effect data Y includes dam deformation Y1 and seepage Y2. Each type of data is archived according to the collection time, monitoring point location, and data type for subsequent analysis.

[0045] As Figure 2 shown in S2, establish the time-lag correlation relationship between environmental data and effect data through the multiple linear regression method. First, collect the historical data sets of environmental data X and effect data Y for a certain period (such as the recent 3 years); for the water level data X1, calculate its historical mean μ1 and standard deviation σ1, and apply the formula for standardization; for the rainfall data X2, calculate its historical mean μ2 and standard deviation σ2, and apply the formula for standardization; for the dam deformation data Y1, calculate its historical mean μ3 and standard deviation σ3, and apply the formula for standardization; for the seepage data Y2, calculate its historical mean μ4 and standard deviation σ4, and apply the formula for standardization.

[0046] Form the standardized environmental data set X' = {X1', X2'} and the standardized effect data set Y' = {Y1', Y2'}; the standardized data all have the characteristics of a mean of 0 and a standard deviation of 1, eliminating the dimension difference; the standardization process does not change the correlation relationship between the data, but makes different types of data comparable.

[0047] Determine the appropriate maximum lag order k according to the physical characteristics of the water conservancy and hydropower project; for concrete dams, usually choose k = 7 to 14 days; for earth-rock dams, usually choose k = 14 to 30 days; the k value that can best reflect the environment-effect relationship can be selected through repeated verification of historical data.

[0048] For the standardized data set, prepare a historical observation value sequence with sufficient length; construct a data set for each time point t, including the environmental data at the current time point t and the past k time points, as well as the effect data at the current time point t; form a sliding data window: {X1'(t), X1'(t - 1),......, X1'(t - k), X2'(t), X2'(t - 1),....., X2'(t - k), Y1'(t), Y2'(t)}.

[0049] The time-lag model of the dam deformation data Y1 expresses the relationship between the current deformation value and the current and historical environmental factors: Use the least squares method to estimate the model parameters α1, β 1i , γ 1i ; The β coefficient in the model 1i characterizes the influence intensity of the water level data X1 on the current dam deformation i days ago; the γ coefficient in the model 1i characterizes the influence intensity of the rainfall data X2 on the current dam deformation i days ago.

[0050] The time-lag model of the seepage data Y2 expresses the relationship between the current seepage value and the current and historical environmental factors: Similarly, use the least squares method to estimate the model parameters α2, β 2i , γ 2i ; The seepage lag model focuses on capturing the time process of the changes in water level and rainfall being transmitted to the seepage system.

[0051] Select an appropriate time window length w (such as 90 days or 180 days), and this window length should be sufficient to include the characteristics of seasonal variations; set the initial window [t - w + 1, t] on the time series, which contains the data of w consecutive time points; for each pair of environment-effect combinations (X j , Y l ), j ∈ {1, 2}, l ∈ {1, 2}, calculate the Pearson correlation coefficients at different lag times; for the lag time i ∈ [0, k], calculate r(X j (t - i), Y l (t)); The formula for calculating the Pearson correlation coefficient:

[0052] where μ x , μ y are the means of X j (t - i) and Y l (t) within the window respectively, σ x , σ y are the standard deviations respectively, and n is the number of samples within the window.

[0053] Slide the window forward by a certain step size (such as 1 day) to form a new window [t - w + 2, t + 1]; repeat the calculation of the correlation coefficients at each lag time within the new window; continue sliding until the entire data range is covered; for each lag time i, take the average of the calculation results of all windows as the final correlation coefficient r(X j (t - i), Y l (t)).

[0054] For each pair of environment-effect combinations (X j , Y l), check the correlation coefficients at k + 1 different lag times; select the correlation coefficient with the largest absolute value as the maximum correlation coefficient r of this combination max ; record the corresponding lag time as the optimal lag time t of this combination opt ; for example, if the correlation coefficient between water level and seepage is 0.78 at a lag of 2 days, and this value is the maximum among all lag times, then r of this combination max = 0.78, t opt = 2.

[0055] Construct a 2×2 time-lag correlation coefficient matrix R; each element R j,l in the matrix contains the parameter pair {r max , t opt}; R 1,1 represents the maximum correlation coefficient and the optimal lag time between water level X1 and dam deformation Y1; R 1,2 represents the maximum correlation coefficient and the optimal lag time between water level X1 and seepage Y2; R 2,1 represents the maximum correlation coefficient and the optimal lag time between rainfall X2 and dam deformation Y1; R 2,2 represents the maximum correlation coefficient and the optimal lag time between rainfall X2 and seepage Y2.

[0056] Through the matrix R, the influence intensity and time-lag characteristics of each environmental factor on the effect index can be intuitively understood; for example, if R 2,2 = {0.85, 3}, it indicates that the influence of rainfall on seepage is very significant (correlation coefficient 0.85), and there is a time lag of 3 days. If the correlation coefficients corresponding to multiple lag times are very close (difference < 0.05), multiple candidate optimal lag times can be recorded for further analysis.

[0057] S3. In this embodiment, an algorithm library containing 5 typical anomaly detection algorithms is constructed to process the outliers in the monitoring data of water conservancy and hydropower projects. The algorithm library is designed as an extensible architecture, and each algorithm plays a role in different types of anomaly characteristics and data distribution characteristics to effectively identify various anomaly patterns.

[0058] Specifically, for the mean method, an appropriate time window length w (such as 7 days) is selected. For each monitoring data point x(t), a window is formed by taking a total of w data points before and after it. Calculate the mean μ of all data points in the window except the current point, where μ = (x(t - w / 2) +... + x(t - 1) + x(t + 1) +... + x(t + w / 2)) / (w - 1). Calculate the deviation d between the current data point and the mean, where d = |x(t) - μ|. Set a threshold θ (usually 2 - 3 times the standard deviation within the window). If d > θ, then x(t) is determined to be an outlier. Window length w: It is determined according to the data collection frequency. For daily collected data, it is usually set to 5 - 9 days. Judgment threshold θ: It adopts an adaptive method, where θ = c × σ, c is a coefficient (usually 2.5), and σ is the standard deviation within the window.

[0059] For the median method, select a time window length w (usually an odd number, such as 7 days), and take a total of w data points before and after each data point x(t). Calculate the median of all data points in the window. Calculate the deviation d between the current point and the median, where d = |x(t) - med|. Calculate the median absolute deviation of the deviations within the window. Set a threshold θ = c × mad (c is usually 3.0). If d > θ, then x(t) is determined to be an outlier. Window length w: An odd value, usually 5 to 11 days. Judgment coefficient c: It is adjusted according to the data fluctuation characteristics, generally 3.0 to 4.5.

[0060] For the difference method, detect mutant outliers by calculating the change rate of differences between adjacent data points. Calculate the first-order difference sequence for the time series: diff(t) = x(t) - x(t - 1). Select a window length w and calculate the statistical characteristics (mean μ diff and standard deviation σ diff ) of the historical differences. For the current difference value diff(t), calculate its standardized score Set a threshold θ (usually 2.5 - 3.5). If z > θ, then x(t) is determined to be an outlier. Difference calculation method: First-order difference or relative change rate can be selected. Statistical characteristic calculation period: It is determined according to the seasonal characteristics of the data, usually 30 - 90 days. Judgment threshold θ: It is adjusted according to the project's sensitivity to mutations, usually 2.5 - 3.5.

[0061] For the Hampel filtering method, select a window length w (such as 11 days), and take a total of w points before and after each data point x(t). Calculate the median within the window. Calculate the deviation of each point in the window from the median: Calculate the median absolute deviation madm = 1.4826 × median(dev); set the determination threshold θ = k × madm (k is usually 3). If |x(t) - med| > θ, then determine x(t) as an outlier. Window length w: Select according to the data frequency and fluctuation characteristics, usually 9 - 15 days; Scaling factor 1.4826: Make MADM equivalent to the standard deviation under the normal distribution; Determination coefficient k: Adjust according to the requirements of anomaly sensitivity, usually 3 - 5.

[0062] For the 3σ criterion, calculate the overall mean μ and standard deviation σ of the monitoring data sequence; for each data point x(t), calculate its standardized score z = |x(t) - μ| / σ; if z > 3, then determine x(t) as an outlier; for different types of monitoring data, the determination threshold can be adjusted to 2σ or 4σ. Statistical calculation period: Determine according to the data nature, and global calculation or sliding window calculation can be selected; Anomaly determination threshold: The standard setting is 3, and it can be flexibly adjusted to 2 - 4 according to different data types.

[0063] As Figure 3 shown in S4, construct an evaluation index matrix according to the correlation coefficient matrix and the anomaly detection algorithm library. Specifically, extract the optimal lag time t of each pair of environment - effect combinations from the time - lag correlation coefficient matrix R opt ; apply the optimal lag time t opt,1,1 and t opt,1,2 to the water level data X1 to form two lag sequences X1(t - t opt,1,1 ) and X1(t - t opt,1,2 ); apply the optimal lag time t opt,2,1 and t opt,2,2 to the rainfall data X2 to form two lag sequences X2(t - t opt,2,1 ) and X2(t - t opt,2,2 ).

[0064] Integrate the processed lagged environmental data into a new dataset X'; for each type of effect data, select the lagged environmental data most relevant to it: select X1(t - t opt,1,1 ) and X2(t - t opt,2,1 ) for Y1; select X1(t - t opt,1,2 ) and X2(t - t opt,2,2 ) for Y2; form a lagged environmental dataset X' that is time - aligned with the effect data Y to ensure the correct expression of the causal relationship.

[0065] Combine the lagged environmental dataset X' and the effect data Y to construct a comprehensive dataset D that includes all monitoring indicators; the dataset D contains the original environmental data, lagged environmental data, and effect data; add timestamps and monitoring location identifiers for each data type. Divide the dataset D into a training set Dtrain and a test set Dtest according to a ratio of 7:3

[0066] On the training set, for each monitoring data type (water level, rainfall, dam deformation, seepage), adjust the parameters of each anomaly detection algorithm; use grid search or Bayesian optimization methods to find the optimal parameter combination for each algorithm; use the F1 score (the harmonic mean of accuracy and recall) as the objective function for parameter optimization.

[0067] Calculate the evaluation metrics, detection accuracy A: the number of correctly detected anomaly points / (the number of correctly detected anomaly points + the number of false alarm points); false alarm rate B: the number of false alarm points / the total number of detected anomaly points; miss rate C: the number of missed anomaly points / the total number of actual anomaly points; robustness metric D: the degree to which the algorithm performance remains stable after adding random noise, calculate the consistency ratio between the original result and the result after adding noise.

[0068] Construct an evaluation metric matrix, construct a 4×5 evaluation metric matrix E; the row index i ∈ {1, 2, 3, 4} corresponds to the four monitoring data types respectively: water level X1, rainfall X2, dam deformation Y1, seepage Y2; the column index j ∈ {1, 2, 3, 4, 5} corresponds to the five anomaly detection algorithms respectively: mean method, median method, difference method, Hampel filter method, and 3σ criterion.

[0069] Each matrix element E i,j Store the four evaluation metrics {A, B, C, D}; for example, E 1,3 represents the evaluation result of detecting the water level data X1 using the difference method; the four metrics in each element are stored in a structured form: {accuracy A, false alarm rate B, miss rate C, robustness D}

[0070] S5. According to the evaluation metric matrix, select the optimal anomaly detection algorithm for each monitoring data. Use a combination of the analytic hierarchy process (AHP) and the Delphi method to determine the weight coefficients. The weight w A of the detection accuracy A = 0.4, reflecting the basic ability of the algorithm to correctly identify anomalies; the weight w B of the false alarm rate B = 0.25, reflecting the impact degree of the algorithm's false alarms; the weight w C of the miss rate C = 0.25, reflecting the risk cost of the algorithm missing real anomalies; the weight w D of the robustness metric D = 0.1, reflecting the adaptability of the algorithm to data fluctuations.

[0071] According to the actual risk level of the project, the weights can be adjusted; for high-risk projects, the weight of the false negative rate C can be increased (e.g., w C = 0.3 - 0.35); for monitoring systems with frequent false positives, the weight of the false positive rate B can be increased (e.g., w B = 0.3); keep the sum of all weights equal to 1: w A + w B + w C + w D = 1.

[0072] For each type of monitoring data i and anomaly detection algorithm j, calculate the comprehensive score S i,j ,

[0073] S i,j = w A × A i,j - w B × B i,j - w C × C i,j + w D × D i,j , where i ∈ {1, 2, 3, 4} represents the type of monitoring data, and j ∈ {1, 2, 3, 4, 5} represents the type of anomaly detection algorithm; standardize the evaluation indicators: A i,j and D i,j are positive indicators, with a value range of [0, 1], the larger the better; B i,j and C i,j are negative indicators, with a value range of [0, 1], the smaller the better, so a negative sign is taken in the formula.

[0074] For example, the evaluation results of the mean method applied to water level data are {A = 0.85, B = 0.12, C = 0.08, D = 0.75}; calculate the comprehensive score S 1,1 = 0.4 × 0.85 - 0.25 × 0.12 - 0.25 × 0.08 + 0.1 × 0.75 = 0.365; repeat the above calculation for all algorithm-data type combinations to form a comprehensive score matrix.

[0075] Extract the absolute value of the maximum correlation coefficient |r max | of the environmental data from the time-delay correlation coefficient matrix R; for the water level data X1, take |r max,1 | = max(|r 1,1 |, |r 1,2 |), that is, the maximum correlation coefficient between the water level and the dam deformation and seepage; for the rainfall data X2, take |r max,2 | = max(|r 2,1 |, |r 2,2 |), that is, the maximum correlation coefficient between the rainfall and the dam deformation and seepage.

[0076] Perform correlation weighted correction on the comprehensive scores of environmental data X1 and X2; S i,j ' = S i,j ×(1 + 0.2×|r max |), i ∈ {1, 2}; The correction coefficient 0.2 represents the influence weight of correlation on the score, which can be adjusted according to the specific engineering situation (generally taking values from 0.1 to 0.3); The stronger the correlation, the more significant the correction, reflecting the importance of environmental data for predicting effect data.

[0077] For example, the maximum correlation coefficient |r max,1 | between water level data X1 and effect data is 0.75; The original comprehensive score S 1,1 of water level data using the mean method is 0.365; The corrected score S 1,1 ' = 0.365×(1 + 0.2×0.75) = 0.42.

[0078] For dam deformation data Y1 and seepage data Y2, directly select the optimal algorithm based on the original comprehensive score, Select the optimal algorithm for dam deformation data Y1; Select the optimal algorithm for seepage data Y2.

[0079] For water level data X1 and rainfall data X2, select the optimal algorithm based on the corrected comprehensive score;

[0080] Select the optimal algorithm for water level data X1; Select the optimal algorithm for rainfall data X2.

[0081] Perform further comparison and verification on algorithms with similar scores (score difference < 0.05); Execute candidate algorithms on actual data subsets to confirm their performance stability; Considering the algorithm execution efficiency, preferentially select algorithms with lower computational complexity when the performance is comparable.

[0082] S6. Perform outlier processing on the collected monitoring data according to the optimal anomaly detection algorithm; Perform further comparison and verification on algorithms with similar scores (score difference < 0.05); Execute candidate algorithms on actual data subsets to confirm their performance stability; Considering the algorithm execution efficiency, preferentially select algorithms with lower computational complexity when the performance is comparable; Apply the corresponding optimal anomaly detection algorithm to the newly collected monitoring data; Water level data X1 uses A opt (1) algorithm to detect outliers; Rainfall data X2 uses A opt (2) algorithm to detect outliers; Dam deformation data Y1 uses A opt (3) algorithm to detect outliers; Seepage data Y2 uses A opt(4) The algorithm detects anomalies. Calculate the anomaly score value for each monitored data point, indicating the degree of anomaly of that point; set a basic threshold θ0 and an advanced threshold θ1 (θ1 > θ0) to achieve anomaly grading; data points with an anomaly score less than θ0 are marked as "normal"; data points with an anomaly score between θ0 and θ1 are marked as "slightly abnormal"; data points with an anomaly score greater than θ1 are marked as "severely abnormal"; save the correspondence between the original data points and the anomaly marks.

[0083] For slightly abnormal data, adopt the strategy of marking but retaining the original value; for severely abnormal data, adopt the replacement strategy for processing; select different replacement methods according to the data type: for environmental data (X1 and X2): use the predicted value of the time series model (such as ARIMA) for replacement; for effect data (Y1 and Y2): use the predicted value of the multiple regression model based on the lagged environmental data for replacement; the replaced data is marked as "corrected", and at the same time, the original value and the basis for correction are retained.

[0084] As Figure 4 shown in S7, group the monitored data after anomaly value processing by season and working condition, combine the correlation coefficient matrix, and obtain the mean ± 2σ as the first-level warning threshold and the mean ± 3σ as the second-level warning threshold through statistical calculation of the mean and standard deviation of each group of data.

[0085] First, divide the monitored data after anomaly value processing according to the calendar date into seasons; Spring: monitored data from March 1 to May 31; Summer: monitored data from June 1 to August 31; Autumn: monitored data from September 1 to November 30; Winter: monitored data from December 1 to February 28 / 29 of the following year; establish a season mark field and add the corresponding season identifier (1 - 4) to each monitoring record.

[0086] Based on the comprehensive determination of the working condition status by the water level data X1 and the rainfall data X2; normal water storage working condition (marked as 1): the water level is lower than the warning water level and the water level change rate within 72 hours is less than the set threshold (such as 0.5 m / day), and there is no large amount of rainfall recently; flood season precipitation working condition (marked as 2): the water level exceeds the warning water level or the water level rising rate within 72 hours exceeds the set threshold, or the cumulative rainfall within 24 hours exceeds 50 mm; add a working condition mark field to all monitored data to ensure that it matches the current hydraulic state.

[0087] By combining the season markers (1 - 4) and the operating condition markers (1 - 2), eight season - operating condition combinations are formed; add a combination type identifier (1 - 8) to each monitoring record, corresponding to the eight combinations: Combination 1: Spring - Normal Water Storage (j = 1, m = 1); Combination 2: Spring - Flood Season Precipitation (j = 1, m = 2); Combination 3: Summer - Normal Water Storage (j = 2, m = 1); Combination 4: Summer - Flood Season Precipitation (j = 2, m = 2); Combination 5: Autumn - Normal Water Storage (j = 3, m = 1); Combination 6: Autumn - Flood Season Precipitation (j = 3, m = 2); Combination 7: Winter - Normal Water Storage (j = 4, m = 1); Combination 8: Winter - Flood Season Precipitation (j = 4, m = 2); For combinations that are not common at a specific time (such as Winter - Flood Season Precipitation), the number of samples may be small. In this case, historical data from the past 3 to 5 years can be integrated to increase the sample size.

[0088] Grouped statistical calculation: For each type of monitoring data (i) under each season - operating condition combination (j, m), calculate the following statistical parameters respectively: Mean where x is the data value within the group and n is the number of samples within the group; Standard Deviation Skewness Coefficient Measuring the degree of asymmetry of the data distribution. Handling of insufficient sample size: Set a minimum sample size threshold, such as 30 data points; for combinations with insufficient sample size, use approximate season - operating condition combination data for supplementation; for example, if the "Winter - Flood Season Precipitation" sample is insufficient, the "Autumn - Flood Season Precipitation" or the data from the same period of the previous year can be borrowed for supplementation.

[0089] Statistical parameter verification: Use a normality test (such as the Shapiro - Wilk test) to verify the data distribution characteristics; for data with an obviously non - normal distribution (test p - value < 0.05), mark its skewness characteristics to prepare for subsequent threshold correction; generate a statistical parameter report, including the sample size, mean, standard deviation, skewness coefficient, and the results of the normal distribution test for each combination.

[0090] Extract the optimal lag time t for each pair of environment - effect combinations (p, q) from the time - lag correlation coefficient matrix R opt,p,q ; The corresponding relationships are: (p = 1, q = 1) represents water level and dam deformation; (p = 1, q = 2) represents water level and seepage; (p = 2, q = 1) represents rainfall and dam deformation; (p = 2, q = 2) represents rainfall and seepage. For the data of each season - operating condition combination (j, m), according to the optimal lag time t opt,p,q Process the environmental data; for example, if the optimal lag time between the water level and seepage is 2 days, then shift the water level data sequence backward by 2 days to align X1(t - 2) with Y2(t).

[0091] For each season - operating condition combination (j, m), calculate the real - time correlation response coefficient between the environmental data and the effect data Among them, μ p and σ p are the mean and standard deviation of the environmental data X p in the current combination; μ q and σ q are the mean and standard deviation of the effect data Y q in the current combination; n is the number of valid paired samples in the current combination.

[0092] For each type of environmental data (p), determine the type of effect data (q) with the greatest correlation; for example, if the correlation coefficient between water level and dam deformation is greater than the correlation coefficient between water level and seepage, the maximum correlated effect data of water level is dam deformation.

[0093] For the effect data (i ∈ {3, 4}): If the data is approximately normally distributed If the data is skewed (|s i,j,m | ≥ 0.5): Early warning upper limit correction coefficient: C i,j,m = 1 + 0.1 × s i,j,m ; Early warning lower limit correction coefficient: C i,j,m = 1 - 0.1 × s i,j,m .

[0094] If it is strongly correlated with the effect data (|r max | > 0.7): When : The early warning upper limit takes The early warning lower limit takes When : The early warning upper limit takes The early warning lower limit takes If it is weakly correlated with the effect data (|r max | ≤ 0.7): C i,j,m = 1.

[0095] Primary early warning threshold (k = 2): Lower limit: (lower limit); Upper limit: (upper limit).

[0096] Secondary early warning threshold (k = 3): Lower limit: (lower limit); Upper limit: (upper limit).

[0097] For the water level data (X1), the secondary early warning upper limit shall not exceed the design flood level; for the dam deformation data (Y1), the secondary early warning upper limit shall not exceed the design allowable deformation value; if the calculated value exceeds the physical safety limit, the safety limit shall be used as the final threshold.

[0098] Construction of the early warning threshold matrix Θ: Construct a four-dimensional early warning threshold matrix Θ; the matrix dimension is 4×4×2×4, corresponding to the monitoring data type (4)×season (4)×operating condition (2)×threshold type (4); each element Θ i,j,m contains four early warning thresholds: For example, Θ 1,2,1 represents the early warning threshold set of water level data under normal water storage conditions in summer.

[0099] Convert the early warning threshold matrix into an early warning threshold table for engineering applications; the table includes: monitoring data type, season, operating condition, lower limit of the first-level early warning, upper limit of the first-level early warning, lower limit of the second-level early warning, upper limit of the second-level early warning; add information on the threshold effective time and the next update time.

[0100] At the conversion points of seasons or operating conditions, set a threshold transition period of 7 - 10 days; within the transition period, the threshold gradually transitions from one season-operating condition combination to the next at a linear ratio; avoid false alarms caused by sudden changes in the threshold at the conversion point.

[0101] S8. Conduct monitoring and management according to the first-level early warning threshold and the second-level early warning threshold. Collect monitoring data at the frequency specified in S1; perform outlier processing using the optimal anomaly detection algorithm determined in S6; determine the corresponding season-operating condition combination based on the current date and water level status.

[0102] Extract the early warning thresholds of each monitoring data type under the current season-operating condition combination from the early warning threshold matrix Θ; compare the real-time monitoring data with the early warning thresholds; for environmental data (X1 and X2), dynamically fine-tune the early warning thresholds according to the real-time relevant response coefficient φ_{p,q}.

[0103] Normal state: The monitoring data is within the range of the first-level early warning threshold; First-level early warning: The monitoring data exceeds the first-level early warning threshold but does not exceed the second-level early warning threshold; Second-level early warning: The monitoring data exceeds the second-level early warning threshold; Comprehensive early warning: When multiple related monitoring indicators trigger early warnings simultaneously, raise the early warning level.

[0104] First-level early warning response: Automatically generate first-level early warning alarm information and display it through the system interface; notify the on-duty monitoring personnel to pay attention and increase the data collection frequency of relevant monitoring points; start the short-term prediction model to estimate the future 24 - 72-hour change trend of the indicators; record the details of the early warning event, including the trigger time, monitoring point location, early warning value, threshold range, etc.

[0105] Level 2 early warning response: Automatically generate level 2 early warning alarm information and notify through multiple channels such as system interface, mobile phone text messages, and emails; notify the project management person in charge and professional and technical personnel, and organize an emergency response team; increase the data collection frequency of all relevant monitoring points to the highest level; activate the emergency plan and prepare necessary protective measures; conduct on-site inspections of relevant areas to verify the reasons for abnormal monitoring data.

[0106] Warning upgrade and downgrade: Continuous warning conditions: If the monitoring data remains in the warning state for 24 consecutive hours, the warning level will be automatically upgraded; Cross-validation warning: If warnings appear at multiple related monitoring points at the same time, the warning level will be raised; Warning release conditions: The monitoring data returns to the normal range for 48 consecutive hours and no abnormalities are found; Warning downgrade process: The second-level warning must meet the 24-hour stability condition to be downgraded to the first-level warning, and the first-level warning must meet the 48-hour stability condition to be released.

Claims

1. A real-time monitoring and management method for water conservancy and hydropower projects, characterized in that, Including: S1, collecting the monitoring data of water conservancy and hydropower projects, where the monitoring data includes environmental data X and effect data Y. The environmental data X includes water level and rainfall; the effect data Y includes dam deformation and seepage; S2, using the multiple linear regression method to construct the correlation coefficient matrix between the environmental data X and the effect data Y; S3, constructing an anomaly detection algorithm library, which includes the mean method, median method, difference method, Hampel filtering method, and 3σ criterion; S4, constructing an evaluation index matrix according to the correlation coefficient matrix and the anomaly detection algorithm library; S5, selecting the optimal anomaly detection algorithm for each type of monitoring data according to the evaluation index matrix; S6, performing outlier processing on the collected monitoring data according to the optimal anomaly detection algorithm; S7, grouping the monitoring data after outlier processing by season and working conditions, and combining the correlation coefficient matrix. By statistically calculating the mean and standard deviation of each group of data, the mean ± 2σ is obtained as the first-level early warning threshold and the mean ± 3σ as the second-level early warning threshold; S8, performing monitoring and management according to the first-level early warning threshold and the second-level early warning threshold.

2. The real-time monitoring and management method for water conservancy and hydropower projects according to claim 1, characterized in that: S2, using the multiple linear regression method to construct the correlation coefficient matrix between the environmental data X and the effect data Y, including: Using Z-score to process the environmental data X and the effect data Y respectively. Among them, the environmental data X = {X1, X2}, X1 represents the water level data, and X2 represents the rainfall data; the effect data Y = {Y1, Y2}, represents the dam deformation data, and Y2 represents the seepage data; Respectively constructing the time-lag models Y1(t) and Y2(t) of the dam deformation data Y1 and the seepage data Y2; According to the time-lag models Y1(t) and Y2(t), through the sliding time window method, calculate the Pearson correlation coefficient r of each environmental data X = {X1, X2} and effect data Y = {Y1, Y2} at different lag times, Y1 r(X j (t - i), Y l (t)), i ∈ [0, k], j ∈ {1, 2}, l ∈ {1, 2}, where k is the maximum lag order; i and j represent positive integers; For each environmental data X j and effect data Y l Combine, from k + 1 Pearson correlation coefficients r(X j (t - i), Y l (t)), select the correlation coefficient r with the largest absolute value max and its corresponding optimal lag time t opt , to form a 2×2 time-delay correlation coefficient matrix R.

3. The real-time monitoring and management method for water conservancy and hydropower projects according to claim 2, characterized in that: The time-lag models Y1(t) and Y2(t), the expressions are: where t represents the current time point; k is the maximum lag order; α1 and α2 are the constant terms of each model; β 1i , β 2i are the regression coefficients of the water level data X1 at different lag times in the two-effect data models respectively; γ 1i , γ 2i are the regression coefficients of the rainfall data X2 at different lag times in the two-effect data models respectively; ε1 and ε2 are the random error terms of each model.

4. The real-time monitoring and management method for water conservancy and hydropower projects according to claim 2, characterized in that: S4, constructing an evaluation index matrix, including: According to the optimal lag time t in the time-delay correlation coefficient matrix R opt , the environmental data X is lag-processed to form a lagged environmental data set X' that is most correlated with the effect data Y; Constructing a data set according to the environmental data set X' and the effect data Y, and dividing the data set into a training set and a test set; Using each anomaly detection algorithm in the anomaly detection algorithm library to perform anomaly detection on the data in the training set respectively; Using the test set to calculate the evaluation indexes of the detection results of each anomaly detection algorithm. The evaluation indexes include detection accuracy A, false alarm rate B, miss rate C, and robustness index D; Construct a 4×5 evaluation index matrix E, where the matrix element E i,j includes the evaluation results of anomaly detection.

5. The real-time monitoring and management method for water conservancy and hydropower projects according to claim 4, characterized in that: S5, selecting the optimal anomaly detection algorithm for each type of monitoring data, including: Set the weight coefficients \(w\) of each evaluation index in the evaluation index matrix A , \(w\) B , \(w\) C , \(w\) D ; For each type of monitoring data, calculate the comprehensive score S of each anomaly detection algorithm i,j : S i,j = w A × A i,j - w B × B i,j - w C × C i,j + w D × D i,j 。 6. The real-time monitoring and management method for water conservancy and hydropower projects according to claim 5, characterized in that: S5, selecting the optimal anomaly detection algorithm for each type of monitoring data, further including: For environmental data X1 and X2, a correlation coefficient correction term is added when calculating the comprehensive score: S i,j ' = S i,j ×(1 + 0.2×|r max |) where r ∈ {1, 2}|r max is the absolute value of the maximum correlation coefficient corresponding to the environmental data in the time-delay correlation coefficient matrix R; For each type of monitoring data, the anomaly detection algorithm with the highest comprehensive score is selected as the optimal algorithm for this data type: Among them, A opt (i) is the optimal anomaly detection algorithm corresponding to the monitoring data type i.

7. The real-time monitoring and management method for water conservancy and hydropower projects according to any one of claims 2 to 6, characterized in that: S7, obtaining the mean ± 2σ as the first-level warning threshold and the mean ± 3σ as the second-level warning threshold, including: The monitored data after outlier processing is divided into four groups according to seasons: Spring Festival, Summer, Autumn, and Winter; According to the water level status, the monitored data after outlier processing is grouped by working conditions, and the working conditions include normal water storage conditions and flood season precipitation conditions; According to each season and working condition combination, a season - working condition combination is formed, including Spring - Normal Water Storage, Spring - Flood Season Precipitation, Summer - Normal Water Storage, Summer - Flood Season Precipitation, Autumn - Normal Water Storage, Autumn - Flood Season Precipitation, Winter - Normal Water Storage, and Winter - Flood Season Precipitation; For each season - working condition combination, the statistical parameters of each type of monitored data are calculated respectively, and the statistical parameters include mean μ, standard deviation σ, and skewness coefficient s; Based on the data for each season - operating condition combination, using the optimal lag time t in the time - lag correlation coefficient matrix R opt , calculate the real - time correlation response coefficient between the environmental data X and the effect data Y in each group of data Based on the statistical parameters of each season-operation condition combination and combined with the calculated real-time relevant response coefficients Calculate the warning threshold: For effect data Y1 and Y2, the warning threshold calculation formula is: μ ± kσ, where k is the warning level coefficient; For environmental data X1 and X2, when |r max | > 0.7, the early warning threshold calculation formula is: Where: δ is the relevant adjustment coefficient, with a value of 0.2; When |r max | ≤ 0.7, the early warning threshold calculation formula is: μ ± kσ, where k is the early warning level coefficient; Generate the first-level warning threshold and the second-level warning threshold for each group of data; Integrate the warning thresholds under all season - operating condition combinations into a warning threshold matrix Θ, and the matrix element Θ i,j,m represents the warning threshold interval of the monitoring data type i under season j and operating condition m, Wherein: represents the lower limit of the first-level warning for the monitoring data type i under season j and working condition m; represents the upper limit of the first-level warning for the monitoring data type i under season j and working condition m; represents the lower limit of the second-level warning for the monitoring data type i under season j and working condition m; represents the upper limit of the second-level warning for the monitoring data type i under season j and working condition m.

8. The real-time monitoring and management method for water conservancy and hydropower projects according to claim 7, characterized in that: Real-time correlation response coefficient The calculation formula is as follows: where: p and q represent positive integers; t opt,p,q is the optimal lag time of the corresponding combination in the time-delay correlation coefficient matrix R; μ p and σ p are the mean and standard deviation of the environmental data X p in the current season-operation condition combination respectively; μ q and σ q are the mean and standard deviation of the effect data Y q in the current season-operation condition combination respectively; n is the number of samples in the current season-operation condition combination.

9. The real-time monitoring and management method for water conservancy and hydropower projects according to claim 8, characterized in that: Generate the first-level warning threshold and the second-level warning threshold for each group of data, and the formula is: First-level warning threshold: Secondary warning threshold: Among them, μ i,j,m is the mean value of the monitoring data type i under season j and working condition m; σ i,j,m is the standard deviation of the monitoring data type i under season j and working condition m; C i,j,m is the correction coefficient.

10. A system based on the real-time monitoring and management method for water conservancy and hydropower projects according to any one of claims 1 to 9.

Images