Reservoir dam seepage pressure monitoring data processing method and system

Abstract

The application relates to the technical field of dam safety monitoring, and discloses a seepage pressure monitoring data processing method and system for a reservoir dam, which comprises the following steps: acquiring seepage pressure monitoring data and a reservoir water level; reconstructing and performing normalization processing based on the data to obtain a normalized potential energy field; constructing a topological network based on the normalized potential energy field and calculating seepage field topological entropy capable of representing the overall structural characteristics of a seepage field; identifying a hydraulic step event triggered by rapid change of the reservoir water level and collecting seepage field topological entropy response curves under the event; and extracting a real-time response feature vector from the curves. The application further provides a system for executing the method. The application can effectively distinguish between severe data fluctuation caused by normal working conditions (such as sudden change of the water level) and early weak abnormalities caused by internal structural diseases, and significantly improves the sensitivity and accuracy of seepage state abnormality diagnosis.

Description

Technical Field

[0001] This invention relates to the field of dam safety monitoring technology, specifically to a method and system for processing seepage pressure monitoring data for reservoir dams. Background Technology

[0002] Reservoirs and dams are vital infrastructure projects related to national welfare and people's livelihoods, and their safe operation is of paramount importance. Seepage is one of the key factors affecting the safety and stability of dams; therefore, long-term and effective monitoring and analysis of the seepage state inside the dam is a core task in ensuring the safe operation of the dam. Seepage pressure, as the most direct physical quantity reflecting the state of the seepage field inside the dam, is the primary basis for evaluating the health status of the dam through the analysis and interpretation of its monitoring data.

[0003] Currently, the analysis methods for dam seepage pressure monitoring data mainly rely on statistical analysis of data from a single or multiple monitoring points. These methods typically compare real-time monitoring values ​​with historical data or predicted values ​​from statistical regression models based on environmental quantities such as reservoir water levels, identifying anomalies by determining whether the deviation exceeds limits. However, such methods have inherent limitations in practical applications.

[0004] These traditional analytical methods often focus on the numerical changes at the measuring points themselves, which is essentially a localized assessment and makes it difficult to capture the structural evolution of the entire seepage field from a global perspective. Early defects inside the dam, such as the initiation and expansion of micro-cracks or localized blockage of the drainage body, may initially only cause slight distortions in the overall structure of the seepage field. These distortions are not significant in the readings of individual measuring points and are easily submerged in normal fluctuation noise, leading to missed or delayed diagnosis of defects.

[0005] Furthermore, when encountering drastic changes in hydraulic conditions such as floods or rapid rises and falls in reservoir water levels, the dam's seepage field generates a strong dynamic response, causing significant fluctuations in readings at all measuring points. Traditional diagnostic methods based on static or quasi-static thresholds struggle to effectively distinguish between the dam's normal dynamic response and abnormal dynamic responses caused by internal defects under such conditions, often leading to numerous false alarms or failing to identify genuine abnormal signals masked by the drastic dynamic processes. Therefore, existing technologies still fall short in their ability to provide a holistic and structural description of the seepage field and to perform accurate diagnosis under complex dynamic conditions. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a method and system for processing seepage pressure monitoring data for reservoir dams, aiming to solve the technical problem in existing technologies that make it difficult to effectively distinguish between drastic data fluctuations caused by changes in normal operating conditions and early, subtle anomalies caused by internal structural defects.

[0007] To address the aforementioned technical problems, this invention provides a novel method and system for processing seepage pressure monitoring data in reservoir dams.

[0008] The first aspect of this invention provides a method for processing seepage pressure monitoring data for a reservoir dam, the method comprising:

[0009] S1. Obtain seepage pressure monitoring data from multiple seepage pressure measuring points on the dam, along with their corresponding three-dimensional spatial coordinates and the current reservoir water level.

[0010] S2. Based on the seepage pressure monitoring data and the three-dimensional spatial coordinates, a continuous pressure field covering the monitoring area is reconstructed, and the continuous pressure field is normalized in combination with the reservoir water level to obtain a normalized potential energy field.

[0011] S3. Construct a topological network based on the normalized potential energy field, and calculate the topological entropy of the seepage field of the topological network;

[0012] S4. Identify hydraulic step events where the rate of change of water level in the reservoir exceeds a preset threshold.

[0013] S5. In response to the hydraulic step event, the seepage field topological entropy is collected during the event and for a period of time afterward to form a seepage field topological entropy response curve.

[0014] S6. Extract a real-time response feature vector from the topological entropy response curve of the seepage field;

[0015] S7. Compare the real-time response feature vector with a health response feature vector in a preset health response feature library that corresponds to the current hydraulic step event condition;

[0016] S8. Based on the comparison results, determine whether there is any abnormality in the dam's seepage state.

[0017] The technical solution provided by this invention transforms discrete monitoring data into a topological entropy index characterizing the overall structure of the seepage field. By using naturally occurring hydraulic step events during reservoir operation as system input and analyzing the dynamic response characteristics of the topological entropy, a quantitative assessment of the health status of the seepage system inside the dam is achieved, enabling the identification of structural anomalies that are difficult to detect using conventional static analysis methods.

[0018] In one specific embodiment, the step of constructing the topology network based on the normalized potential field includes:

[0019] A series of equipotential surfaces are extracted from the normalized potential energy field, and each equipotential surface is determined as a node in the topological network.

[0020] When two equipotential surfaces are spatially adjacent, an edge is established between the corresponding two nodes;

[0021] The reciprocal of the magnitude of the hydraulic gradient connecting the two equipotential surfaces is used as the weight of the corresponding edge.

[0022] In one specific embodiment, the step of calculating the topological entropy of the seepage field of the topological network includes: calculating the normalized importance of each node in the topological network. The topological entropy of the seepage field The information entropy formula is used to calculate the normalized importance of all nodes in the aforementioned topology network.

[0023] ;

[0024] in, This represents the total number of nodes in the network topology.

[0025] Preferably, the normalized importance of the node It is based on the comprehensive influence corresponding to the node. Determined by its proportion in the total combined influence of all nodes:

[0026] ;

[0027] Among them, the comprehensive influence Area of ​​the equipotential surface represented by the node Average connection weight of this node The formula for its calculation is jointly determined as follows: .

[0028] Preferably, the step of identifying the hydraulic step event specifically involves: calculating the rate of change of the reservoir water level in real time, and determining that a hydraulic step event has occurred when the absolute value of the rate of change of the reservoir water level is greater than a preset water level change rate threshold.

[0029] Preferably, the response feature vector includes at least one of the following:

[0030] The response delay is the time from the start of the hydraulic step event to the start of a significant change in the topological entropy response curve of the seepage field.

[0031] The response peak value is the maximum value of the topological entropy response curve of the seepage field;

[0032] Peak time is the duration from the start of the hydraulic step event to the occurrence of the peak response.

[0033] The decay time is the time required for the topological entropy response curve of the seepage field to decay from its peak value to a preset ratio.

[0034] The response impulse is the area enclosed by the topological entropy response curve of the seepage field and the baseline.

[0035] Preferably, the step of comparing the real-time response feature vector with the health response feature vector in the preset health response feature library...

[0036] include:

[0037] Calculate the real-time response feature vector With the health response feature vector Mahalanobis distance between :

[0038] ;

[0039] in, The covariance matrix of the health response feature vector; the Mahalanobis distance It is compared with a preset diagnostic threshold to determine abnormalities.

[0040] In one specific embodiment, the method further includes: when no hydraulic step event is detected, comparing the real-time calculated topological entropy of the seepage field with a preset static normal range under the current reservoir water level, and judging whether there is an abnormality in the dam seepage state based on the comparison result.

[0041] In one specific embodiment, the method for constructing the preset health response feature library includes:

[0042] Identify hydraulic step events that occur during the dam's healthy operating cycle;

[0043] For each hydraulic step event, the corresponding seepage field topological entropy response curve is extracted, and a healthy response feature vector is calculated.

[0044] The event conditions of the hydraulic step event and the health response feature vector are stored as paired data to form the preset health response feature library.

[0045] A second aspect of the present invention provides a data processing system for seepage pressure monitoring of a reservoir dam, the system comprising:

[0046] The data processing module is used to acquire seepage pressure monitoring data from multiple seepage pressure measuring points of the dam, their corresponding three-dimensional spatial coordinates, and the reservoir water level at the current moment. Based on the seepage pressure monitoring data and the three-dimensional spatial coordinates, a continuous pressure field is reconstructed. Then, the continuous pressure field is normalized by combining the reservoir water level to obtain a normalized potential energy field.

[0047] The topological entropy calculation module is used to construct a topological network based on the normalized potential energy field and calculate the topological entropy of the seepage field of the topological network.

[0048] The dynamic response analysis module is used to identify hydraulic step events in which the rate of change of water level in the reservoir exceeds a preset threshold, and in response to the hydraulic step event, controls the topological entropy calculation module to collect the seepage field topological entropy during and after the event to form a seepage field topological entropy response curve, and extracts a real-time response feature vector from the seepage field topological entropy response curve.

[0049] The anomaly diagnosis module is used to compare the real-time response feature vector with a health response feature vector in a preset health response feature library that corresponds to the current hydraulic step event condition, and to determine whether there is an anomaly in the dam seepage state based on the comparison result.

[0050] This invention provides a method and system for processing seepage pressure monitoring data in reservoir dams. It offers the following advantages:

[0051] 1. This invention constructs a topological network of the seepage field and calculates its topological entropy, transforming discrete, local monitoring data into a global index that characterizes the overall structural complexity and orderliness of the seepage field. Compared to methods that rely solely on single-point measurements or simple interpolation, this index is more sensitive to distortions in the overall structure of the seepage field caused by early internal structural defects (such as micro-cracks or local blockages), thereby improving the sensitivity and timeliness of anomaly identification.

[0052] 2. This invention utilizes naturally occurring hydraulic step events during reservoir operation as a trigger condition for dynamic diagnosis and captures the response curve of the seepage field topological entropy for analysis. This method can effectively distinguish between two situations: one is the normal system response caused by drastic changes in reservoir water level, and the other is the abnormal response caused by changes in the seepage characteristics inside the dam. By analyzing the response feature vector, the accuracy of diagnostic results under complex operating conditions is significantly improved, and the false alarm rate is reduced.

[0053] 3. This invention provides an objective and repeatable basis for judging the seepage state of dams by establishing a preset health response feature library and using statistical methods such as Mahalanobis distance for quantitative comparison. This data-driven benchmark comparison method avoids subjective judgments based on engineering experience or fixed, poorly universal static thresholds, making diagnostic conclusions more reliable and laying the foundation for realizing an automated and standardized health diagnosis process. Attached Figure Description

[0054] Figure 1 A flowchart of an embodiment of the method of the present invention;

[0055] Figure 2 This is a functional block diagram of a system embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of the seepage field reconstruction and normalization process of the present invention;

[0057] Figure 4 This is a schematic diagram illustrating the process of constructing a topological network from a normalized potential energy field and calculating the topological entropy according to the present invention.

[0058] Figure 5 This is a schematic diagram of the dynamic response curve of the topological entropy of the seepage field in this invention;

[0059] Figure 6 This is a schematic diagram of the abnormality diagnosis process of the present invention.

[0060] The module includes: 10. Data processing module; 20. Topological entropy calculation module; 30. Dynamic response analysis module; and 40. Anomaly diagnosis module. Detailed Implementation

[0061] The technical solutions in 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.

[0062] See attached document Figure 1 , Figure 1 This is a flowchart of a method for processing seepage pressure monitoring data for a reservoir dam according to an embodiment of the present invention. The method provided by the present invention may include the following steps:

[0063] S1. Obtain seepage pressure monitoring data and their corresponding three-dimensional spatial coordinates from multiple seepage pressure measuring points on the dam, as well as the current reservoir water level.

[0064] S2. Based on the seepage pressure monitoring data and the three-dimensional spatial coordinates, a continuous pressure field covering the monitoring area is reconstructed, and the continuous pressure field is normalized in combination with the reservoir water level to obtain a normalized potential energy field.

[0065] S3. Construct a topological network based on the normalized potential energy field, and calculate the topological entropy of the seepage field of the topological network.

[0066] S4. Identify hydraulic step events where the rate of change of the reservoir water level exceeds a preset threshold.

[0067] S5. In response to the hydraulic step event, the seepage field topological entropy is collected during the event and for a period of time thereafter, forming a seepage field topological entropy response curve.

[0068] S6. Extract a real-time response feature vector from the topological entropy response curve of the seepage field.

[0069] S7. Compare the real-time response feature vector with the health response feature vector corresponding to the current hydraulic step event condition in a preset health response feature library.

[0070] S8. Based on the comparison results, determine whether there is any abnormality in the dam's seepage state.

[0071] See attached document Figure 2 , Figure 2 This is a functional block diagram of a seepage pressure monitoring data processing system for a reservoir dam according to an embodiment of the present invention. The system is used to perform... Figure 1 The system, as shown, includes: a data processing module 10, a topological entropy calculation module 20, a dynamic response analysis module 30, and an anomaly diagnosis module 40.

[0072] The data processing module 10 is used to execute steps S1 and S2 in the above method. Specifically, this module is used to acquire seepage pressure monitoring data, three-dimensional spatial coordinates and reservoir water level information from multiple seepage pressure measuring points of the dam, and to reconstruct and normalize the continuous pressure field based on this information, and finally output the normalized potential energy field to the topological entropy calculation module 20.

[0073] The topological entropy calculation module 20 is used to perform step S3 in the above method. This module receives the normalized potential energy field generated by the data processing module 10, constructs a topological network based on the field data, and calculates the seepage field topological entropy, which characterizes the structural properties of the seepage field at the current moment. This module outputs the calculated seepage field topological entropy to the dynamic response analysis module 30 and the anomaly diagnosis module 40.

[0074] The dynamic response analysis module 30 is used to execute steps S4, S5, and S6 in the above method. This module is used to identify hydraulic step events, and when an event occurs, it continuously obtains the topological entropy of the seepage field from the topological entropy calculation module 20 to form a response curve, and extracts the real-time response feature vector from the curve. This module outputs the extracted response feature vector to the anomaly diagnosis module 40.

[0075] The anomaly diagnosis module 40 is used to execute steps S7 and S8 in the above method. This module receives the real-time response feature vector generated by the dynamic response analysis module 30 and performs a quantitative comparison with the healthy response feature vector in the preset healthy response feature library. Based on the comparison result, the module finally outputs a judgment conclusion on whether the dam seepage state is abnormal.

[0076] In one embodiment, the data processing module 10, the topology entropy calculation module 20, the dynamic response analysis module 30, and the anomaly diagnosis module 40 can be implemented as software program modules running on one or more processors, including but not limited to servers, workstations, or embedded computing devices. These modules exchange data and make function calls through predefined application programming interfaces (APIs).

[0077] See attached document Figure 3 , Figure 3 This is a schematic diagram of the seepage field reconstruction and normalization process according to an embodiment of the present invention. The specific implementation of steps S1 and S2 in the method of the present invention is as follows.

[0078] First, step S1 is executed to obtain the basic data required for the calculations. This step obtains the data at the current moment through an interface with the automated data acquisition unit of the dam safety monitoring system or a manual observation database. A dataset. This dataset includes: deployed within the dam. The seepage pressure monitoring values ​​of each seepage pressure measuring point and each measuring point Three-dimensional spatial coordinates in the dam coordinate system ,in At the same time, obtain the current upstream water level of the reservoir. After acquiring the data, a data cleaning operation can be performed to process invalid data or outliers caused by sensor malfunctions or transmission errors, and to unify the units of all data, such as unifying the pressure unit to kilopascals (kPa) and the coordinate and water level units to meters (m).

[0079] Next, step S2 is executed. Based on the discrete measurement point data obtained in step S1, a continuous field covering the entire monitoring area is generated and normalized. This step can be specifically divided into two sub-steps: continuous pressure field reconstruction and normalization. In the continuous pressure field reconstruction sub-step, a spatial interpolation algorithm is used, based on... The pressure values ​​and spatial coordinates of each measuring point are used to calculate the coordinates of any point within the monitoring area. The pressure value. In one specific embodiment, Kriging interpolation can be used. This method is an unbiased optimal estimation method based on spatial autocorrelation. At any location Pressure estimate at the location It can be obtained by a weighted linear combination of the pressure values ​​of known measuring points around it:

[0080] ;

[0081] in, For the first Measured pressure values ​​at each measuring point; To be assigned to the Weighting coefficients for each measurement point. The determination is based on the analysis of the spatial variation structure of the pressure field within the monitoring area (described by a variogram function), so that the estimated value... The estimated variance is minimized and the unbiasedness condition is satisfied. A continuous three-dimensional pressure field can be generated by calculating all grid points within the monitoring area.

[0082] In the normalization sub-step, to eliminate the direct impact of absolute water level fluctuations on the pressure field, the reconstructed continuous pressure field is converted into a normalized potential energy field. First, the pressure values ​​at each point are... Converted into piezometric head (or potential energy). Subsequently, the head field was normalized using the following formula to obtain the normalized potential energy. :

[0083] ;

[0084] in, For point Normalized potential energy at the location; The pressure head corresponding to this point; The current upstream water level of the reservoir serves as the upstream boundary condition. The downstream water level or a fixed downstream reference elevation is used as the downstream boundary condition. After this step, the normalized potential energy field is output. It is a dimensionless scalar field whose numerical distribution reflects only the hydraulic gradient distribution characteristics within the seepage field, and is unaffected by changes in the absolute elevation of the upstream water level. This normalized potential energy field will serve as the basis for subsequent construction of the topology network.

[0085] See attached document Figure 4 , Figure 4 This is a schematic diagram illustrating the process of constructing a topological network from a normalized potential field and calculating the topological entropy according to an embodiment of the present invention. The specific implementation of step S3 in the method of the present invention is as follows.

[0086] This step receives the normalized potential field output from the previous step. Based on this, a topology network is constructed and the topological entropy of the seepage field is calculated. This step can be divided into two sub-steps: topology network construction and seepage field topological entropy calculation.

[0087] In the topology network construction sub-step, the continuous normalized potential field is discretized and abstracted into a network model. First, the range of values ​​for the normalized potential (usually ) is divided into... A series of continuous, non-overlapping intervals. Each interval This corresponds to a set of equipotential surfaces. In three-dimensional space, this set is represented as a layer of equipotential surfaces or equipotential volumes. This equipotential surface (or equipotential volume layer) can be abstracted as a node in a topological network, denoted as a node. .

[0088] Then, edges are established between the defined nodes. When the equipotential surfaces corresponding to two nodes are directly adjacent in space, an edge is established between these two nodes. For example, representing a potential energy interval. nodes and representing the potential energy range nodes They are spatially adjacent, so an edge is created between them.

[0089] Next, calculate the weight of each established edge. Connect the nodes. and nodes weight of the edge It is defined as the reciprocal of the magnitude of the hydraulic gradient between the corresponding equipotential surfaces of the two nodes. In a specific embodiment, the magnitude of the hydraulic gradient... It can be calculated as the potential energy difference between two equipotential surfaces. and the average normal distance between them The ratio of the weights. The calculation formula is:

[0090] ;

[0091] in, Indicates weight; Indicates the magnitude of the hydraulic gradient; This represents the potential energy difference between the centers of two equipotential surfaces. This represents the average normal distance between two equipotential surfaces, which can be obtained through computational geometry algorithms. This weight value reflects the ease with which water flows through the region between these two equipotential surfaces.

[0092] In the sub-step of calculating the topological entropy of the seepage field, a single quantitative index is calculated based on the constructed weighted topological network. First, for each node in the network... Calculate its overall influence This influence is determined by the area of ​​the equipotential surface represented by the node. Average connection weight of this node Jointly determined:

[0093] ;

[0094] in, For the first Each equipotential surface at time... The spatial area can be obtained by summing the areas of the three-dimensional mesh cells that constitute the equipotential surface; For nodes At any moment The average connection weight is calculated as the arithmetic mean of the weights of all edges connected to the node.

[0095] Then, each node is calculated based on its overall influence. The importance of normalization This value is passed through the node. The proportion of the overall influence of all nodes in the network is used to determine the overall influence.

[0096] ;

[0097] in, This represents the total number of nodes in the network. This is the normalized importance set. This constitutes a probability distribution describing the spatial distribution characteristics of seepage energy. Finally, the current moment is calculated using the information entropy formula. Topological entropy of seepage field :

[0098] ;

[0099] The logarithmic operation here can be base-2 or use the natural logarithm. The calculated topological entropy of the seepage field... This is a scalar value whose magnitude reflects the complexity and uniformity of the entire dam seepage field structure. This entropy value will be output to subsequent steps for dynamic response analysis.

[0100] See attached document Figure 5 and attached Figure 6 , Figure 5 This is a schematic diagram of the dynamic response curve of the topological entropy of the seepage field according to an embodiment of the present invention. Figure 6 This is a schematic diagram of an anomaly diagnosis process according to an embodiment of the present invention. The specific implementation of steps S4 to S8 in the method of the present invention is as follows.

[0101] First, step S4 is executed to identify hydraulic step events. This step involves a monitoring process that continuously acquires the current reservoir water level. And calculate its rate of change per unit time. In one specific embodiment, the rate of change can be obtained by difference calculation: The absolute value of this rate of change is compared with a preset threshold for the rate of change of water level. When the value exceeds this threshold, the system determines that a hydraulic step event has been triggered and records the start time of the event. And event attributes, such as whether the water level is rising or falling, and the magnitude of the change.

[0102] Next, step S5 is executed to collect and generate a seepage field topological entropy response curve in response to the identified hydraulic step event. Once the event is triggered, the system controls the topological entropy calculation module to continuously calculate and record the seepage field topological entropy at a high frequency (e.g., every 5 minutes) during the event and for a preset duration thereafter (e.g., for 24 hours or until the system reaches a stable state again). The values ​​of these entropy values, arranged in chronological order, constitute a sequence such as... Figure 5 The figure shows a complete topological entropy response curve for the seepage field.

[0103] Then, step S6 is executed to extract quantized features from the generated response curve. The system calculates and extracts one or more predefined response feature parameters from the curve and combines them into a real-time response feature vector. In one specific embodiment, the vector may include:

[0104] Response delay From the start of the event The duration for which the value of the response curve first exceeds a specific percentage (e.g., 5%) of its baseline fluctuation range.

[0105] Response peak The maximum entropy of the response curve over the entire response period.

[0106] Peak time From the start of the event To the peak response The duration of its occurrence.

[0107] decay time The response curve from reaching its peak value Then, it decays to a preset percentage of the difference between the peak value and the baseline (e.g., The time elapsed.

[0108] Response impulse The response curve is higher than its baseline value during the response period. The area enclosed by the portion and the time axis is calculated using the following formula: Combine these calculated scalar values ​​into a multidimensional vector: .

[0109] Subsequently, steps S7 and S8 are executed for comparison and diagnosis. The system first retrieves a health response feature vector matching the current hydraulic step event's attributes (such as the magnitude of the change and the initial water level) from a pre-set health response feature library. and its corresponding covariance matrix Then, the real-time response feature vector is calculated. With health response feature vector Mahalanobis distance between :

[0110] ;

[0111] Among them, superscript Indicates matrix transpose, superscript This represents the inversion of a matrix. The Mahalanobis distance calculation takes into account the correlation between various features and performs standardization.

[0112] Finally, the calculated Mahalanobis distance Compare with a preset diagnostic threshold. If If the seepage exceeds the diagnostic threshold, the system determines that the dam's seepage state is abnormal and generates an alarm signal; conversely, if... If the value is less than or equal to the diagnostic threshold, the current state is considered normal.

[0113] In an optional embodiment, the method further includes a static diagnostic mode. During smooth operation in which no hydraulic step events are detected, the system calculates the seepage field topological entropy in real time. With a water level based on the current reservoir level static normal range The system will make a comparison. If the real-time entropy value exceeds the normal range, the system will also make a preliminary judgment that there is an anomaly.

[0114] The aforementioned pre-defined health response feature library is constructed as follows: Historical operational data of the dam when it is confirmed to be in a healthy state are selected, and all hydraulic step events that have occurred are identified. For each healthy event, its complete topological entropy response curve is extracted, and the corresponding response feature vector is calculated. These healthy feature vectors are grouped according to event attributes, and the mean vector of each group is calculated (as...). ) and covariance matrix (as Finally, the event attributes, mean vector, and covariance matrix are stored as a single record in the library.

Claims

1. A method for processing seepage pressure monitoring data for reservoir dams, characterized in that, Includes the following steps: S1. Obtain seepage pressure monitoring data from multiple seepage pressure measuring points on the dam, along with their corresponding three-dimensional spatial coordinates and the current reservoir water level. S2. Based on the seepage pressure monitoring data and the three-dimensional spatial coordinates, a continuous pressure field covering the monitoring area is reconstructed, and the continuous pressure field is normalized in combination with the reservoir water level to obtain a normalized potential energy field. S3. Construct a topological network based on the normalized potential energy field, and calculate the topological entropy of the seepage field of the topological network. The step of calculating the topological entropy of the seepage field of the topological network includes: For each node in the aforementioned topology, calculate its normalized importance; The seepage field topological entropy is calculated by the information entropy formula based on the normalized importance of all nodes in the topological network. The normalized importance of a node is determined based on the proportion of the comprehensive influence of that node in the total comprehensive influence of all nodes. The overall influence is determined by the area of ​​the equipotential surface represented by the node and the average connection weight of the node. S4. Identify hydraulic step events where the rate of change of water level in the reservoir exceeds a preset threshold. S5. In response to the hydraulic step event, the seepage field topological entropy is collected during the event and for a period of time afterward to form a seepage field topological entropy response curve. S6. Extract a real-time response feature vector from the topological entropy response curve of the seepage field; S7. Compare the real-time response feature vector with a health response feature vector in a preset health response feature library that corresponds to the current hydraulic step event condition; S8. Based on the comparison results, determine whether there is any abnormality in the dam's seepage state.

2. The method for processing seepage pressure monitoring data for reservoir dams according to claim 1, characterized in that, The steps for constructing the topology network based on the normalized potential field include: A series of equipotential surfaces are extracted from the normalized potential energy field, and each equipotential surface is determined as a node in the topological network. When two equipotential surfaces are spatially adjacent, an edge is established between the corresponding two nodes; The reciprocal of the magnitude of the hydraulic gradient connecting the two equipotential surfaces is used as the weight of the corresponding edge.

3. The method for processing seepage pressure monitoring data for reservoir dams according to claim 1, characterized in that, The steps for identifying the hydraulic step event are as follows: The rate of change of the reservoir water level is calculated in real time. When the absolute value of the rate of change of the reservoir water level is greater than the preset water level change rate threshold, it is determined that a hydraulic step event has occurred.

4. The method for processing seepage pressure monitoring data for reservoir dams according to claim 1, characterized in that, The response feature vector includes at least one of the following: The response delay is the time from the start of the hydraulic step event to the start of a significant change in the topological entropy response curve of the seepage field. The response peak value is the maximum value of the topological entropy response curve of the seepage field; Peak time is the duration from the start of the hydraulic step event to the occurrence of the peak response. The decay time is the time required for the topological entropy response curve of the seepage field to decay from its peak value to a preset ratio. The response impulse is the area enclosed by the topological entropy response curve of the seepage field and the baseline.

5. The method for processing seepage pressure monitoring data for reservoir dams according to claim 1, characterized in that, The step of comparing the real-time response feature vector with the health response feature vector in the preset health response feature library includes: Calculate the Mahalanobis distance between the real-time response feature vector and the health response feature vector; The Mahalanobis distance is compared with a preset diagnostic threshold to determine abnormalities.

6. The method for processing seepage pressure monitoring data for reservoir dams according to claim 1, characterized in that, The method also includes: If no hydraulic step event is detected, the real-time calculated topological entropy of the seepage field is compared with the preset static normal range under the current reservoir water level, and the dam seepage state is judged to be abnormal based on the comparison result.

7. The method for processing seepage pressure monitoring data for reservoir dams according to claim 1, characterized in that, The method for constructing the preset health response feature library includes: Identify hydraulic step events that occur during the dam's healthy operating cycle; For each hydraulic step event, the corresponding seepage field topological entropy response curve is extracted, and a healthy response feature vector is calculated. The event conditions of the hydraulic step event and the health response feature vector are stored as paired data to form the preset health response feature library.

8. A data processing system for seepage pressure monitoring of reservoir dams, characterized in that, A method for processing seepage pressure monitoring data for a reservoir dam as described in any one of claims 1-7, comprising: The data processing module is used to acquire seepage pressure monitoring data from multiple seepage pressure measuring points of the dam, their corresponding three-dimensional spatial coordinates, and the reservoir water level at the current moment. Based on the seepage pressure monitoring data and the three-dimensional spatial coordinates, a continuous pressure field is reconstructed. Then, the continuous pressure field is normalized by combining the reservoir water level to obtain a normalized potential energy field. The topological entropy calculation module is used to construct a topological network based on the normalized potential energy field and calculate the topological entropy of the seepage field of the topological network. The steps of calculating the topological entropy of the seepage field of the topological network include: For each node in the aforementioned topology, calculate its normalized importance; The topological entropy of the seepage field is calculated by the information entropy formula based on the normalized importance of all nodes in the topological network. The dynamic response analysis module is used to identify hydraulic step events in which the rate of change of water level in the reservoir exceeds a preset threshold, and in response to the hydraulic step event, controls the topological entropy calculation module to collect the seepage field topological entropy during and after the event to form a seepage field topological entropy response curve, and extracts a real-time response feature vector from the seepage field topological entropy response curve. The anomaly diagnosis module is used to compare the real-time response feature vector with a health response feature vector in a preset health response feature library that corresponds to the current hydraulic step event condition, and to determine whether there is an anomaly in the dam seepage state based on the comparison result.

Images