An intelligent monitoring and maintenance system and method for water conservancy dams
By using intelligent monitoring systems and neural network models to correct water level information, the problem of geological disasters masking the true dangers in the monitoring of water conservancy dams has been solved, enabling accurate identification and early warning of water conservancy dam dangers and improving dam safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 浙江省水利科技推广服务中心
- Filing Date
- 2025-08-02
- Publication Date
- 2026-05-26
AI Technical Summary
Existing technologies are insufficient to accurately identify water level changes caused by geological disasters and actual hazards to the dam during long-term use of water conservancy dams. This results in monitoring data masking the true hazards and affecting the safety of the dams.
An intelligent monitoring system is adopted, including a water conservancy dam monitoring module, a data analysis module, and an early warning response unit. The system corrects water level information through a neural network model, separates the impact of geological disasters and dam body risks, and uses a joint probability density function to identify real risks.
It has enabled accurate correction of water level data for water conservancy dams, revealed the true dangers hidden by geological disasters, and improved the accuracy of dam safety monitoring.
Smart Images
Figure CN120912187B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of water conservancy engineering monitoring technology, specifically to an intelligent monitoring and maintenance system and method for water conservancy dams. Background Technology
[0002] Hydraulic dams typically have long service lives, usually designed for 50 to 100 years. During long-term use, geological disasters pose challenges to the monitoring of hydraulic dams, affecting the monitoring process. When there are real hazards in the dam structure, such as seepage causing a faster drop in the water level in the upstream wells and a faster rise in the water level in the downstream wells, the presence of geological hazards, such as karst collapse, may slow down or even cause a drop in the water level in the downstream wells. This can mask the water level changes caused by the real hazards in the monitoring data. In this case, the interference from geological hazards and the water level changes caused by the real hazards acting alone cancel each other out, making the rate of drop in the upstream wells appear normal or minimal. This makes it difficult to detect the real hazards in the dam structure, affecting the safety of the hydraulic dam. Summary of the Invention
[0003] The purpose of this invention is to provide an intelligent monitoring and maintenance system and method for water conservancy dams to solve the problems raised in the prior art.
[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring and maintenance system for water conservancy dams, comprising a water conservancy dam monitoring module, a data analysis module, a data storage module, and an early warning response unit; the water conservancy dam monitoring module is used to acquire water level data of the water conservancy dam and transmit it to the data processing module and the data storage module; the data storage module is used to store historical water level data of the water conservancy dam; the data analysis module is used to acquire geological disaster information and water level information, correct the water level information according to the geological disaster information, analyze the potential dangers of the water conservancy dam according to the corrected water level information, and obtain the probability of dam danger occurrence; the early warning response unit promptly issues an alarm to notify relevant personnel when it detects that the probability of dam danger occurrence reaches a threshold.
[0005] The water conservancy dam monitoring module also includes a geological disaster information acquisition unit and a water level information acquisition unit; the geological disaster information acquisition unit is used to acquire geological disaster information, and the water level information acquisition unit is used to acquire reservoir water level information, water level change rate information of wells in front of the dam, and water level change rate information of wells behind the dam.
[0006] The data processing module further includes a correction unit, an estimation unit, and a prediction unit; the correction unit is used to correct the water level information; the estimation unit is used to obtain the joint probability density function between the dam body hazard and the water level; and the prediction unit predicts the probability of the dam body hazard occurring based on the joint probability density function between the dam body hazard and the water level.
[0007] The correction unit takes reservoir water level data as input and well water level change data in front of the dam as output to train a first neural network model. It obtains reservoir water level data and well water level change data in front of the dam from historical water level data of the hydraulic dam at the time of the geological disaster. If the geological disaster did not affect the reservoir water level data, it obtains historical water level data of the hydraulic dam before the geological disaster as reference historical well water level change data in front of the dam. If the geological disaster affected the reservoir water level data, it inputs the reservoir water level data after the geological disaster into the first neural network model and obtains the reference historical well water level change data in front of the dam from the output of the first neural network. The impact of geological disasters on well water level data is obtained by subtracting historical well water level change data from reference historical well water level change data after a disaster. Geological disaster information is used as input, and the impact on well water level data is used as output to train a second neural network model. Current geological disaster information and water level information are obtained and input into the second neural network model to obtain the impact of current geological disaster information on well water level data. Corrected well water level change data is obtained based on the current well water level change data and the impact of current geological disaster information on well water level data.
[0008] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring and maintenance method for water conservancy dams, comprising the following steps:
[0009] Obtain historical water level and seepage data for water conservancy dams; derive water level relationships when no geological disasters have occurred based on historical water level data of water conservancy dams; and derive water level relationships when geological disasters have occurred based on historical water level data of water conservancy dams when no geological disasters have occurred and water level relationships when no geological disasters have occurred.
[0010] Obtain current geological disaster information and water level information, correct the water level information based on the geological disaster information, and analyze the potential dangers of water conservancy dams based on the corrected water level information.
[0011] Specifically, obtaining the water level relationship under conditions where no geological disasters have occurred based on historical water level data of water conservancy dams also includes the following steps:
[0012] Historical water level data of water conservancy dams are used to obtain reservoir water level data and water level change data of wells in front of the dam. These data are divided into training set and validation set. The reservoir water level data in the training set is used as input and the water level change data of wells in front of the dam is used as output to train the first neural network model. The first neural network model is then validated using the reservoir water level data and water level change data of wells in front of the dam in the validation set. After successful validation, the first model is obtained.
[0013] Specifically, obtaining the water level relationship during a geological disaster also includes the following steps:
[0014] The system obtains reservoir water level data and upstream well water level change data from historical water level data of the dam during geological disasters. If the geological disaster did not affect the reservoir water level, the system uses historical water level data of the dam before the geological disaster as reference historical upstream well water level change data. If the geological disaster affected the reservoir water level, the system inputs the reservoir water level data after the geological disaster into the first neural network model, and obtains the reference historical upstream well water level change data from the output of the first neural network. The system then calculates the difference between the historical upstream well water level change data after the geological disaster and the reference historical upstream well water level change data to obtain the impact of the geological disaster on the upstream well water level data. The system uses the geological disaster information as input and the impact on the upstream well water level data as output to train the second neural network model. The system then validates the second neural network model, and the second model is obtained after successful validation.
[0015] Specifically, correcting water level information based on geological disaster information and analyzing the potential dangers of water conservancy dams based on the corrected water level information also includes the following steps:
[0016] The system acquires current geological disaster information and water level information, inputs the current geological disaster information into the second neural network model, and obtains the impact of the current geological disaster information on the water level data of the well in front of the dam. Based on the current water level change data of the well in front of the dam and the impact of the current geological disaster information on the water level data of the well in front of the dam, the system obtains the corrected water level change data of the well in front of the dam. The system then calculates the difference between the corrected water level change data of the well in front of the dam and the reference water level change data of the well in front of the dam to obtain the deviation of the water level change data of the well in front of the dam.
[0017] Specifically, analyzing the potential dangers of water conservancy dams based on the corrected water level information also includes the following steps:
[0018] Following the same method used for the water level change data in the upstream wells, the water level change data in the downstream wells is corrected to obtain the deviations in both data. These deviations are then used as inputs to the joint probability density function f between dam risk and water level. X，Y In (x, y), the probability P of a dam failure is obtained; In the formula, X represents the random variable of the deviation of the water level change data in the well in front of the dam, Y represents the random variable of the deviation of the water level change data in the well behind the dam, x and y represent integral variables, x0 represents the deviation of the water level change data in the well in front of the dam, and y0 represents the deviation of the water level change data in the well behind the dam. If the probability P of the dam body danger is not less than the threshold, it is determined that there is a real danger in the dam body and maintenance is required.
[0019] Specifically, the joint probability density function is determined through the following steps:
[0020] S10: Obtain historical water level data of the water conservancy dam; annotate historical data of the water conservancy dam when there are potential dangers; correct the historical water level change rates of the upstream and downstream wells of the water conservancy dam to remove the interference of geological disasters; obtain the deviation of the historical upstream well water level change data and the deviation of the downstream well water level change data based on the corrected data; generate a dataset based on the deviation data.
[0021] S20, set nodes a1, a2, ..., an for the deviation of the water level change rate in the well upstream of the dam, where n is the number of nodes; let Vb represent the random variable of the deviation of the water level change rate in the well upstream of the dam, and obtain the probability P{T≥a1} of the labeled data when the random variable Vb is greater than or equal to a1, where P{T≥a1}=n a1 / N a1 In the formula n a1 N represents the number of historical data points for hydraulic dikes in the dataset where the rate of change of water level in the well upstream of the dam deviates from a1 or greater and a dangerous situation exists. a1 This represents the number of wells in the dataset whose rate of change in water level is greater than or equal to a1. Following a similar approach to the node a1, we obtain the probabilities P{T≥a2}, ..., P{T≥an} of the labeled data when the random variable Vb of the rate of change in water level is greater than or equal to a2, ..., an. n data points are formed from a1 and P{T≥a1}, a2 and P{T≥a2}, ..., an and P{T≥an}, and these data points are denoted as x1, x2, ..., xn to obtain the dataset.
[0022] S20, the probability density function f between the probability of dam failure and the deviation of the rate of change of water level in the upstream well is obtained through density estimation. X(x); Select a kernel function that is nonnegative and symmetric and has an integral of 1 over the real number field R; Set a constant greater than zero as the bandwidth h of the kernel function K; Scale the kernel function according to the bandwidth to obtain Kh, where Kh(u) = 1 / h × K(u / h), and u is the input of the kernel function; Obtain the contribution rate Kh(x-xi) of the data point xi in the dataset to the estimated point x; Sum the contribution rates of all data points in the dataset to the estimated point x to obtain the kernel density estimate g(x) at the estimated point x; g(x) = 1 / n∑Kh(x-xi); Change the position of the estimated point x to obtain the kernel density estimate over the entire dataset; and select the bandwidth h with the best validation effect through cross-validation.
[0023] S30: Obtain the deviation data of the water level change rate of the wells behind the dam when the deviation of the water level change rate of the wells in front of the dam is x, where x∈{a1, a2, ..., an}. Following steps S10 and S20, obtain the probability density function f between the probability of dam failure and the deviation of the water level change rate of the wells behind the dam, given that the deviation of the water level change rate of the wells in front of the dam is x. Y|X (y|x); according to f Y|X (y|x) and f X (x) yields the joint probability density function f X，Y (x, y), f X，Y (x, y) = f Y|X (y|x)×f X (x).
[0024] Compared with the prior art, the beneficial effects of the present invention are: to correct the water level data of water conservancy dams, to reveal the true danger information of water conservancy dams that are covered by geological disasters, and to separate the influence of geological disaster interference and the true danger of the dam body by establishing a mathematical model, so as to achieve the effect of more accurately identifying the true danger. Attached Figure Description
[0025] Figure 1 This is a schematic diagram of the structure of an intelligent monitoring and maintenance system for water conservancy dams according to the present invention;
[0026] Figure 2 This is a flowchart of an intelligent monitoring and maintenance method for water conservancy dams according to the present invention. Detailed Implementation
[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0028] Example: Figure 1 As shown, this invention provides an intelligent monitoring and maintenance system for water conservancy dams, including a water conservancy dam monitoring module, a data analysis module, a data storage module, and an early warning response unit. The water conservancy dam monitoring module is used to acquire water level data of the water conservancy dam and transmit it to the data processing module and the data storage module. The data storage module is used to store historical water level data of the water conservancy dam. The data analysis module is used to acquire geological disaster information and water level information, correct the water level information according to the geological disaster information, and analyze the potential dangers of the water conservancy dam based on the corrected water level information to obtain the probability of dam danger occurrence. When the probability of dam danger occurrence is detected to reach a threshold, the early warning response unit promptly issues an alarm to notify relevant personnel.
[0029] The water conservancy dam monitoring module also includes a geological disaster information acquisition unit and a water level information acquisition unit; the geological disaster information acquisition unit is used to acquire geological disaster information, and the water level information acquisition unit is used to acquire reservoir water level information, water level change rate information of wells in front of the dam, and water level change rate information of wells behind the dam.
[0030] The data processing module further includes a correction unit, an estimation unit, and a prediction unit; the correction unit is used to correct the water level information; the estimation unit is used to obtain the joint probability density function between the dam body hazard and the water level; and the prediction unit predicts the probability of the dam body hazard occurring based on the joint probability density function between the dam body hazard and the water level.
[0031] The correction unit takes reservoir water level data as input and well water level change data in front of the dam as output to train a first neural network model. It obtains reservoir water level data and well water level change data in front of the dam from historical water level data of the hydraulic dam at the time of the geological disaster. If the geological disaster did not affect the reservoir water level data, it obtains historical water level data of the hydraulic dam before the geological disaster as reference historical well water level change data in front of the dam. If the geological disaster affected the reservoir water level data, it inputs the reservoir water level data after the geological disaster into the first neural network model and obtains the reference historical well water level change data in front of the dam from the output of the first neural network. The impact of geological disasters on well water level data is obtained by subtracting historical well water level change data from reference historical well water level change data after a disaster. Geological disaster information is used as input, and the impact on well water level data is used as output to train a second neural network model. Current geological disaster information and water level information are obtained and input into the second neural network model to obtain the impact of current geological disaster information on well water level data. Corrected well water level change data is obtained based on the current well water level change data and the impact of current geological disaster information on well water level data.
[0032] In another embodiment of the present invention, the present invention provides an intelligent monitoring and maintenance method for hydraulic dams, comprising the following steps:
[0033] Obtain historical water level and seepage data for water conservancy dams; derive water level relationships when no geological disasters have occurred based on historical water level data of water conservancy dams; and derive water level relationships when geological disasters have occurred based on historical water level data of water conservancy dams when no geological disasters have occurred and water level relationships when no geological disasters have occurred.
[0034] Obtain current geological disaster information and water level information, correct the water level information based on the geological disaster information, and analyze the potential dangers of water conservancy dams based on the corrected water level information.
[0035] Determining the water level relationship under conditions where no geological disasters have occurred based on historical water level data of water conservancy dams also includes the following steps:
[0036] Historical water level data of water conservancy dams are used to obtain reservoir water level data and water level change data of wells in front of the dam. These data are divided into training set and validation set. The reservoir water level data in the training set is used as input and the water level change data of wells in front of the dam is used as output to train the first neural network model. The first neural network model is then validated using the reservoir water level data and water level change data of wells in front of the dam in the validation set. After successful validation, the first model is obtained.
[0037] The water level change data in the upstream wells includes positive and negative signs; a positive sign indicates a rise in the water level, and a negative sign indicates a fall. The training and validation sets are selected from data collected over a period after the dam's maintenance and repair. This aims to determine the relationship between the reservoir water level and the upstream well water level when there are no potential dam-related emergencies. Since dam-related emergencies are a very lengthy process, selecting data from a period after maintenance and repair helps to eliminate interference from such emergencies. The upstream well water level change data and the reservoir water level data can be represented by average values within a preset time window.
[0038] Obtaining the water level relationship during a geological disaster also includes the following steps:
[0039] The system obtains reservoir water level data and upstream well water level change data from historical water level data of the dam during geological disasters. If the geological disaster did not affect the reservoir water level, the system uses historical water level data of the dam before the geological disaster as reference historical upstream well water level change data. If the geological disaster affected the reservoir water level, the system inputs the reservoir water level data after the geological disaster into the first neural network model, and obtains the reference historical upstream well water level change data from the output of the first neural network. The system then calculates the difference between the historical upstream well water level change data after the geological disaster and the reference historical upstream well water level change data to obtain the impact of the geological disaster on the upstream well water level data. The system uses the geological disaster information as input and the impact on the upstream well water level data as output to train the second neural network model. The system then validates the second neural network model, and the second model is obtained after successful validation.
[0040] Geological disasters such as karst collapses affect the water levels in wells upstream and downstream of the dam, easily leading to the erroneous conclusion that the dam is in danger and leakage is exacerbated. However, the impact of karst collapses on water levels is usually self-recovering. By subtracting the data from the dam's condition, the interference of the dam's state on the results is removed, retaining only the impact of the geological disaster, and the second neural network model is trained. Historical data showing that the dam has not been damaged by geological disasters are selected to train the second neural network model. Since the change in the dam's condition is a long process, if the dam itself is not damaged by the geological disaster within a short period, the dam's state is considered... The impact on the water level can be considered as no change. If the reservoir water level does not change, the water level data of the wells in front of the dam for a period of time before the geological disaster is obtained. This is the impact of the state of the water conservancy dam on the water level data of the wells in front of the dam. It is considered that there is no change in the short period of time during the geological disaster (which is short relative to the change in the danger of the water conservancy dam). Then, the impact of the geological disaster on the water level data of the wells in front of the dam is obtained by subtracting the water level data of the wells in front of the dam after the geological disaster and the impact of the state of the water conservancy dam on the water level data of the wells in front of the dam. If the reservoir water level changes, the impact of the state of the water conservancy dam on the water level data of the wells in front of the dam is obtained through the first neural network model.
[0041] Correcting water level information based on geological disaster information, and then analyzing the potential dangers of water conservancy dams based on the corrected water level information, also includes the following steps:
[0042] The system acquires current geological disaster information and water level information, inputs the current geological disaster information into the second neural network model, and obtains the impact of the current geological disaster information on the water level data of the well in front of the dam. Based on the current water level change data of the well in front of the dam and the impact of the current geological disaster information on the water level data of the well in front of the dam, the system obtains the corrected water level change data of the well in front of the dam. The system then calculates the difference between the corrected water level change data of the well in front of the dam and the reference water level change data of the well in front of the dam to obtain the deviation of the water level change data of the well in front of the dam.
[0043] The method for obtaining reference dam-front well water level change data is the same as the training process of the second neural network model. If the geological disaster does not affect the reservoir water level data, the historical water level data of the hydraulic dam before the geological disaster is obtained as the reference historical dam-front well water level change data. If the geological disaster affects the reservoir water level data, the reservoir water level data after the geological disaster is input into the first neural network model, and the reference historical dam-front well water level change data is obtained from the output of the first neural network. After correction, the influence of the dam body status on the dam-front well water level change data is obtained. The difference is calculated with the dam-front well water level change data when there is no danger in the dam body. The difference value is used to analyze whether there is a danger in the dam body.
[0044] Analyzing the potential dangers of water conservancy dams based on the corrected water level information also includes the following steps:
[0045] Following the same method used for the water level change data in the upstream wells, the water level change data in the downstream wells is corrected to obtain the deviations in both data. These deviations are then used as inputs to the joint probability density function f between dam risk and water level. X，Y In (x, y), the probability P of a dam failure is obtained; In the formula, X represents the random variable of the deviation of the water level change data in the well in front of the dam, Y represents the random variable of the deviation of the water level change data in the well behind the dam, x and y represent integral variables, x0 represents the deviation of the water level change data in the well in front of the dam, and y0 represents the deviation of the water level change data in the well behind the dam. If the probability P of the dam body danger is not less than the threshold, it is determined that there is a real danger in the dam body and maintenance is required.
[0046] The joint probability density function is determined through the following steps:
[0047] S10: Obtain historical water level data of the water conservancy dam; annotate historical data of the water conservancy dam when there are potential dangers; correct the historical water level change rates of the upstream and downstream wells of the water conservancy dam to remove the interference of geological disasters; obtain the deviation of the historical upstream well water level change data and the deviation of the downstream well water level change data based on the corrected data; generate a dataset based on the deviation data.
[0048] S20, set nodes a1, a2, ..., an for the deviation of the water level change rate in the well upstream of the dam, where n is the number of nodes; let Vb represent the random variable of the deviation of the water level change rate in the well upstream of the dam, and obtain the probability P{T≥a1} of the labeled data when the random variable Vb is greater than or equal to a1, where P{T≥a1}=n a1 / N a1 In the formula n a1N represents the number of historical data points for hydraulic dikes in the dataset where the rate of change of water level in the well upstream of the dam deviates from a1 or greater and a dangerous situation exists. a1 This represents the number of wells in the dataset whose rate of change in water level is greater than or equal to a1. Following a similar approach to the node a1, we obtain the probabilities P{T≥a2}, ..., P{T≥an} of the labeled data when the random variable Vb of the rate of change in water level is greater than or equal to a2, ..., an. n data points are formed from a1 and P{T≥a1}, a2 and P{T≥a2}, ..., an and P{T≥an}, and these data points are denoted as x1, x2, ..., xn to obtain the dataset.
[0049] S20, the probability density function f between the probability of dam failure and the deviation of the rate of change of water level in the upstream well is obtained through density estimation. X (x); Select a kernel function that is nonnegative and symmetric and has an integral of 1 over the real number field R; Set a constant greater than zero as the bandwidth h of the kernel function K; Scale the kernel function according to the bandwidth to obtain Kh, where Kh(u) = 1 / h × K(u / h), and u is the input of the kernel function; Obtain the contribution rate Kh(x-xi) of the data point xi in the dataset to the estimated point x; Sum the contribution rates of all data points in the dataset to the estimated point x to obtain the kernel density estimate g(x) at the estimated point x; g(x) = 1 / n∑Kh(x-xi); Change the position of the estimated point x to obtain the kernel density estimate over the entire dataset; and select the bandwidth h with the best validation effect through cross-validation.
[0050] S30: Obtain the deviation data of the water level change rate of the wells behind the dam when the deviation of the water level change rate of the wells in front of the dam is x, where x∈{a1, a2, ..., an}. Following steps S10 and S20, obtain the probability density function f between the probability of dam failure and the deviation of the water level change rate of the wells behind the dam, given that the deviation of the water level change rate of the wells in front of the dam is x. Y|X (y|x); according to f Y|X (y|x) and f X (x) yields the joint probability density function f X，Y (x, y), f X，Y (x, y) = f Y|X (y|x)×f X (x).
[0051] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.
Claims
1. A method for intelligent monitoring and maintenance of water conservancy dams, characterized in that, Includes the following steps: Obtain historical water level and seepage data for water conservancy dams; derive water level relationships when no geological disasters have occurred based on historical water level data of water conservancy dams; and derive water level relationships when geological disasters have occurred based on historical water level data of water conservancy dams when no geological disasters have occurred and water level relationships when no geological disasters have occurred. Obtain current geological disaster information and water level information, correct the water level information based on the geological disaster information, and analyze the potential dangers of water conservancy dams based on the corrected water level information; The process of obtaining the water level relationship during a geological disaster also includes the following steps: The system obtains reservoir water level data and upstream well water level change data from historical water level data of the dam during geological disasters. If the geological disaster did not affect the reservoir water level, the system uses historical water level data of the dam before the geological disaster as reference historical upstream well water level change data. If the geological disaster affected the reservoir water level, the system inputs the reservoir water level data after the geological disaster into the first neural network model, and obtains the reference historical upstream well water level change data from the output of the first neural network. The system then calculates the difference between the historical upstream well water level change data after the geological disaster and the reference historical upstream well water level change data to obtain the impact of the geological disaster on the upstream well water level data. The system uses the geological disaster information as input and the impact on the upstream well water level data as output to train the second neural network model. The system then validates the second neural network model, and the second model is obtained after successful validation.
2. The intelligent monitoring and maintenance method for water conservancy dams according to claim 1, characterized in that, The process of obtaining the water level relationship under conditions where no geological disasters have occurred based on historical water level data of water conservancy dams also includes the following steps: Historical water level data of water conservancy dams are used to obtain reservoir water level data and water level change data of wells in front of the dam. These data are divided into training set and validation set. The reservoir water level data in the training set is used as input and the water level change data of wells in front of the dam is used as output to train the first neural network model. The first neural network model is then validated using the reservoir water level data and water level change data of wells in front of the dam in the validation set. After successful validation, the first model is obtained.
3. The intelligent monitoring and maintenance method for water conservancy dams according to claim 1, characterized in that, The process of correcting water level information based on geological disaster information and analyzing the potential dangers of water conservancy dams based on the corrected water level information also includes the following steps: The system acquires current geological disaster information and water level information, inputs the current geological disaster information into the second neural network model, and obtains the impact of the current geological disaster information on the water level data of the well in front of the dam. Based on the current water level change data of the well in front of the dam and the impact of the current geological disaster information on the water level data of the well in front of the dam, the system obtains the corrected water level change data of the well in front of the dam. The system then calculates the difference between the corrected water level change data of the well in front of the dam and the reference water level change data of the well in front of the dam to obtain the deviation of the water level change data of the well in front of the dam.
4. The intelligent monitoring and maintenance method for water conservancy dams according to claim 3, characterized in that, The analysis of the potential dangers of water conservancy dams based on the corrected water level information also includes the following steps: Following the same method used for the water level change data in the upstream wells, the water level change data in the downstream wells is corrected to obtain the deviations in both data. These deviations are then used as inputs to the joint probability density function f between dam risk and water level. X，Y In (x, y), the probability P of a dam failure is obtained; In the formula, X represents the random variable of the deviation of the water level change data in the well in front of the dam, Y represents the random variable of the deviation of the water level change data in the well behind the dam, x and y represent integral variables, x0 represents the deviation of the water level change data in the well in front of the dam, and y0 represents the deviation of the water level change data in the well behind the dam. If the probability P of the dam body danger is not less than the threshold, it is determined that there is a real danger in the dam body and maintenance is required.
5. The intelligent monitoring and maintenance method for water conservancy dams according to claim 4, characterized in that, The joint probability density function is determined through the following steps: S10: Obtain historical water level data of the water conservancy dam; annotate historical data of the water conservancy dam when there are potential dangers; correct the historical water level change rates of the upstream and downstream wells of the water conservancy dam to remove the interference of geological disasters; obtain the deviation of the historical upstream well water level change data and the deviation of the downstream well water level change data based on the corrected data; generate a dataset based on the deviation data. S20, set nodes a1, a2, ..., an for the deviation of the water level change rate in the well upstream of the dam, where n is the number of nodes; let Vb represent the random variable of the deviation of the water level change rate in the well upstream of the dam, and obtain the probability P{T≥a1} of the labeled data when the random variable Vb is greater than or equal to a1, where P{T≥a1}=n a1 / N a1 In the formula n a1 N represents the number of historical data points for hydraulic dikes in the dataset where the rate of change of water level in the well upstream of the dam deviates from a1 or greater and a dangerous situation exists. a1 This represents the number of wells in the dataset whose rate of change in water level is greater than or equal to a1. Following a similar approach to the node a1, we obtain the probabilities P{T≥a2}, ..., P{T≥an} of the labeled data when the random variable Vb of the rate of change in water level is greater than or equal to a2, ..., an. n data points are formed from a1 and P{T≥a1}, a2 and P{T≥a2}, ..., an and P{T≥an}, and these data points are denoted as x1, x2, ..., xn to obtain the dataset. S20, the probability density function f between the probability of dam failure and the deviation of the rate of change of water level in the upstream well is obtained through density estimation. X (x); Select a kernel function that is nonnegative and symmetric and has an integral of 1 over the real number field R; Set a constant greater than zero as the bandwidth h of the kernel function K; Scale the kernel function according to the bandwidth to obtain Kh, where Kh(u) = 1 / h × K(u / h), and u is the input of the kernel function; Obtain the contribution rate Kh(x-xi) of the data point xi in the dataset to the estimated point x; Sum the contribution rates of all data points in the dataset to the estimated point x to obtain the kernel density estimate g(x) at the estimated point x; g(x) = 1 / n∑Kh(x-xi); Change the position of the estimated point x to obtain the kernel density estimate over the entire dataset; and select the bandwidth h with the best validation effect through cross-validation. S30: Obtain the deviation data of the water level change rate of the wells behind the dam when the deviation of the water level change rate of the wells in front of the dam is x, where x∈{a1, a2, ..., an}. Following steps S10 and S20, obtain the probability density function f between the probability of dam failure and the deviation of the water level change rate of the wells behind the dam, given that the deviation of the water level change rate of the wells in front of the dam is x. Y|X (y|x); according to f Y|X (y|x) and f X (x) yields the joint probability density function f X，Y (x, y), f X，Y (x, y) = f Y|X (y|x)×f X (x).
6. An intelligent monitoring and maintenance system for water conservancy dams, characterized in that, The system includes a water conservancy dam monitoring module, a data analysis module, a data storage module, and an early warning response unit. The water conservancy dam monitoring module acquires water level data of the water conservancy dam and sends it to the data processing module and the data storage module. The data storage module stores historical water level data of the water conservancy dam. The data analysis module acquires geological disaster information and water level information, corrects the water level information based on the geological disaster information, and analyzes the potential dangers of the water conservancy dam based on the corrected water level information to obtain the probability of dam danger occurring. When the probability of dam danger occurring reaches a threshold, the early warning response unit promptly issues an alarm to notify relevant personnel.
7. The intelligent monitoring and maintenance system for water conservancy dams according to claim 6, characterized in that, The monitoring module further includes a geological disaster information acquisition unit and a water level information acquisition unit; the geological disaster information acquisition unit is used to acquire geological disaster information, and the water level information acquisition unit is used to acquire reservoir water level information, water level change rate information of wells in front of the dam, and water level change rate information of wells behind the dam.
8. The intelligent monitoring and maintenance system for water conservancy dams according to claim 6, characterized in that, The data processing module further includes a correction unit, an estimation unit, and a prediction unit; the correction unit is used to correct the water level information; the estimation unit is used to obtain the joint probability density function between the dam body hazard and the water level; and the prediction unit predicts the probability of the dam body hazard occurring based on the joint probability density function between the dam body hazard and the water level.
9. The intelligent monitoring and maintenance system for water conservancy dams according to claim 8, characterized in that, The correction unit takes reservoir water level data as input and well water level change data in front of the dam as output to train a first neural network model. It obtains reservoir water level data and well water level change data in front of the dam from historical water level data of the hydraulic dam at the time of the geological disaster. If the geological disaster did not affect the reservoir water level data, it obtains historical water level data of the hydraulic dam before the geological disaster as reference historical well water level change data in front of the dam. If the geological disaster affected the reservoir water level data, it inputs the reservoir water level data after the geological disaster into the first neural network model and obtains the reference historical well water level change data in front of the dam from the output of the first neural network. The impact of geological disasters on well water level data is obtained by subtracting historical well water level change data from reference historical well water level change data after a disaster. Geological disaster information is used as input, and the impact on well water level data is used as output to train a second neural network model. Current geological disaster information and water level information are obtained and input into the second neural network model to obtain the impact of current geological disaster information on well water level data. Corrected well water level change data is obtained based on the current well water level change data and the impact of current geological disaster information on well water level data.