Levee settlement anomaly early warning method and system based on water level change response

Abstract

The application provides a dike settlement anomaly early warning method and system based on water level change response, the method comprising: acquiring water level settlement time series data; identifying a water level turning point according to a change trend, event segmentation of the water level settlement time series data is performed in combination with a change amplitude and a change slope, and a water level excitation event formed by water level change is identified; a settlement response analysis window corresponding to each water level excitation event is established, and a settlement response fingerprint of a dike body feedback characteristic to water level excitation is acquired based on the settlement response analysis window; the spatial relationship between the settlement response fingerprint under the current water level excitation event and a preset corresponding dike section normal response envelope is calculated, the spatial relationship is judged in combination with water level excitation intensity information, when the settlement response fingerprint deviates from the allowable range of the normal response envelope, it is judged that the corresponding dike section has a settlement anomaly; the anomaly level is determined based on the distance between the feature distance and the dynamic envelope boundary, and the early warning action corresponding to the anomaly level is performed, and the application improves the reliability of dike operation state monitoring and anomaly early warning.

Description

Technical Field

[0001] This application relates to the field of early warning technology for the safety of dike structures, specifically a method and system for early warning of abnormal settlement of dikes based on water level change response. Background Technology

[0002] Dike projects are crucial infrastructure for flood control and disaster reduction. Their structural safety is directly related to regional flood control capacity and the safety of people's lives and property. To achieve continuous monitoring of dike operation, various monitoring devices are typically deployed along the dike line, such as water level gauges, hydrostatic levels, GNSS displacement monitoring equipment, and piezometers. These devices are used to acquire key parameters in real time, including river water level, dike settlement, and structural deformation. Analysis of the monitoring data allows for the identification of dike deformation response characteristics under high water levels or flood conditions, providing a basis for dike safety assessment and risk warning.

[0003] In existing technologies, levee safety monitoring systems typically acquire monitoring data from different monitoring devices in real time through a data acquisition platform, and then store and process the data uniformly. For example, by sorting or interpolating monitoring data from different sources according to timestamps, multi-source monitoring data can be analyzed on a unified time scale, forming a correlation data sequence between water level and levee settlement. Statistical analysis methods or predictive models are then used to analyze the levee settlement trend to determine whether there is abnormal deformation in the levee.

[0004] In practical engineering environments, the impact of water level changes on levee settlement typically manifests as an excitation-response relationship. When the river water level experiences a significant rise or fall, the soil within the levee will exhibit a certain settlement response under the influence of seepage and consolidation. This settlement response is usually correlated with the water level change process in terms of both time and amplitude. Existing methods mostly rely on continuous monitoring data for overall trend analysis or simple threshold judgments, lacking event-level classification of water level change processes. This makes it difficult to analyze the levee settlement response characteristics for specific water level change processes, which to some extent affects the accuracy of anomaly identification. Therefore, a technical solution is needed that can identify water level change processes and analyze corresponding settlement response characteristics to improve the reliability of levee operation status monitoring and anomaly early warning. Summary of the Invention

[0005] To overcome the aforementioned problems in the prior art, this application provides a method and system, which adopts the following technical solution:

[0006] Firstly, this application provides a method for early warning of abnormal settlement of dikes based on water level change response, including:

[0007] The system acquires real-time water level and settlement monitoring sequences for the target embankment section. It then aligns these sequences using resampling to obtain time-scaled data on a unified time axis. The system acquires the water level change trend, amplitude, and slope. Water level inflection points are identified based on the trend. The water level and settlement time-series data are segmented into events based on the amplitude and slope, identifying water level excitation events. A settlement response analysis window is established for each water level excitation event. Based on this window, the system acquires the settlement response fingerprint of the embankment's response to water level excitation. The spatial relationship between the settlement response fingerprint under the current water level excitation event and the preset normal response envelope for the corresponding embankment section is calculated. Constraints are applied to this spatial relationship based on water level excitation intensity information. When the settlement response fingerprint deviates from the allowable range of the normal response envelope, an anomaly is identified in the corresponding embankment section. The anomaly level is determined based on the distance between the feature distance and the dynamic envelope boundary, and corresponding early warning actions are executed.

[0008] Furthermore, before segmenting the event, the process includes: parsing the water level settlement time series data through a sliding window, calculating the trend direction of the water level change relative to the previous reference point at the current moment, the cumulative change amplitude of the water level, and the first derivative of the water level change as the slope of change; when the trend of change is rising or falling, and the absolute value of the slope of change exceeds the preset slope threshold within a preset time period, the current moment is marked as the start moment of the water level excitation event.

[0009] Furthermore, the trend, amplitude, and slope of water level changes are obtained. Water level inflection points are identified based on the trend. Water level settling time series data are segmented into events based on the amplitude and slope of the changes. Water level excitation events formed by water level changes are identified, including: based on the sign switching of the slope and the extreme point of the amplitude of the change at the start time, the complete fluctuation range of water level from rise and fall to stabilization is identified. When the slope of the change is continuously lower than the preset static threshold, the water level change trend tends to stabilize or the water level falls back to the preset deviation range of the start water level, the current time is marked as the end time of the water level excitation event, that is, the segmentation of a complete excitation event is completed.

[0010] Secondly, this application also provides an early warning system for abnormal settlement of dikes based on water level change response, including:

[0011] The data acquisition module is used to acquire the water level monitoring sequence and settlement monitoring sequence of the target embankment section in real time. It aligns the water level monitoring sequence and settlement monitoring sequence with time scale through resampling to obtain water level and settlement time series data under a unified time axis.

[0012] The water level excitation event definition module is used to obtain the water level change trend, change amplitude, and change slope, identify water level inflection points based on the change trend, and perform event segmentation on water level settlement time series data by combining change amplitude and change slope to identify water level excitation events formed by water level changes.

[0013] The settlement response fingerprint extraction module is used to establish a settlement response analysis window corresponding to each water level excitation event, and to obtain the settlement response fingerprint of the dike body in response to water level excitation based on the settlement response analysis window.

[0014] The settlement anomaly judgment module is used to calculate the spatial relationship between the settlement response fingerprint under the current water level excitation event and the preset normal response envelope of the corresponding embankment section. It combines the water level excitation intensity information to constrain and judge the spatial relationship. When the settlement response fingerprint deviates from the allowable range of the normal response envelope, it is judged that there is a settlement anomaly in the corresponding embankment section.

[0015] The early warning action execution module is used to determine the anomaly level based on the distance between the feature distance and the dynamic envelope boundary, and to execute the early warning action corresponding to the anomaly level.

[0016] Thirdly, this application provides an electronic device, comprising:

[0017] One or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the device, cause the device to perform the method as described in the first aspect.

[0018] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the method described in the first aspect.

[0019] Fifthly, this application provides a computer program that, when executed by a computer, performs the method described in the first aspect.

[0020] In one possible design, the program in the fifth aspect can be stored wholly or partially on a storage medium packaged with the processor, or it can be stored wholly or partially on a memory not packaged with the processor.

[0021] This application has the following beneficial effects:

[0022] 1. This application achieves unified time scale alignment of multi-source monitoring data under different sampling frequencies and time bases by resampling the water level monitoring sequence and the settlement monitoring sequence, so that water level changes and settlement response can be correlated and analyzed on the same time axis, reducing the data mismatch problem caused by sampling frequency differences and timestamp deviations.

[0023] 2. This application analyzes the water level sequence based on the water level change trend, change amplitude and change slope, identifies water level turning points and performs event segmentation, divides the continuous water level change process into different water level excitation events, avoids the problem of mixing different change stages for analysis, makes the structural characteristics of the water level change process clearer, and is conducive to analyzing the response behavior of the levee.

[0024] 3. This application establishes a corresponding settlement response analysis window for each water level excitation event, extracts settlement response features within the window to form a settlement response fingerprint, and structures the response process of the embankment to different water level excitations, so that the settlement response is transformed from the original time series data into a comparable feature representation, thereby improving the ability to characterize the response law of the embankment.

[0025] 4. This application constructs the normal response envelope of the corresponding embankment segment, calculates the spatial relationship between the current settlement response fingerprint and the normal response envelope, and combines the water level excitation intensity for constraint judgment. Considering the differences in different excitation conditions, it adaptively evaluates the embankment response state, avoiding the misjudgment problem caused by using a uniform threshold under different water level conditions, and improving the accuracy and reliability of anomaly identification.

[0026] 5. This application aligns multi-source monitoring data to a unified time scale, divides the water level change process into water level excitation events, establishes a settlement response analysis window, extracts settlement response fingerprints, and combines the normal response envelope to constrain and determine the embankment response state. This achieves a shift from continuous time-series data analysis to event-driven response feature analysis. Compared with existing technologies, this application effectively characterizes the excitation-response relationship between water level changes and embankment settlement even when multi-source data has time differences and noise interference. It quantitatively assesses the degree to which the embankment response deviates from the normal state, improves the accuracy of embankment anomaly identification and the reliability of early warning, and has good engineering application value. Attached Figure Description

[0027] Figure 1 This is a flowchart of the levee settlement anomaly early warning method based on water level change response according to an embodiment of this application;

[0028] Figure 2 This is a flowchart illustrating the extraction of settlement response fingerprints according to an embodiment of this application.

[0029] Figure 3This is a flowchart illustrating the abnormal settlement judgment process for a levee section according to an embodiment of this application.

[0030] Figure 4 This is a schematic diagram of the multidimensional normal response envelope construction process in an embodiment of this application;

[0031] Figure 5 This is a system flowchart of an embodiment of this application. Detailed Implementation

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0033] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0034] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0035] Please refer to Figure 1 The following is a flowchart of a method for early warning of dike settlement anomalies based on water level change response, provided in an embodiment of this application. The specific implementation process is as follows:

[0036] Step 110: Real-time acquisition of water level monitoring sequence and settlement monitoring sequence of the target embankment section; time scale alignment of water level monitoring sequence and settlement monitoring sequence by resampling method to obtain water level and settlement time series data under a unified time axis.

[0037] Specifically, water level monitoring equipment and settlement monitoring equipment are generally deployed on the target embankment section to collect water level data and settlement data. The water level monitoring equipment can be a water level gauge or a pressure water level sensor to obtain water level change information outside or inside the embankment. The settlement monitoring equipment can be a floating plate, a GNSS settlement monitoring device or a fiber optic settlement monitoring device to record settlement changes on or inside the embankment in real time. This application uploads the collected data to the monitoring system in real time through the water level monitoring equipment and settlement monitoring equipment to form a continuously updated water level monitoring sequence and settlement monitoring data sequence.

[0038] For example, for monitoring a certain section of a levee, water level data is collected every 5 minutes by water level monitoring equipment to obtain a water level monitoring sequence; settlement monitoring data is collected every 30 minutes by settlement monitoring equipment to obtain a settlement monitoring sequence. The collection frequency and collection time points of water level monitoring equipment and settlement monitoring equipment are usually inconsistent, so it is necessary to align the time scale of water level monitoring sequence and settlement monitoring sequence.

[0039] Specifically, a resampling method is used to uniformly process the water level monitoring sequence and the settlement monitoring sequence, mapping them to the same time axis. The resampling process can include operations such as time interpolation, time alignment, and missing value imputation. For example, when the sampling interval of the settlement monitoring data is greater than that of the water level monitoring data, the time node of the water level monitoring data can be used as a reference to perform linear interpolation or spline interpolation on the settlement monitoring data to obtain settlement data values ​​consistent with the time node of the water level data. For example, if the water level data sampling period is 1 minute and the settlement data sampling period is 10 minutes, the settlement data can be resampled to a 1-minute time interval using an interpolation method to achieve data time alignment.

[0040] Step 120: Obtain the trend, magnitude, and slope of water level changes; identify water level inflection points based on the trend; segment the water level settlement time series data into events based on the magnitude and slope of the changes; and identify water level excitation events caused by water level changes.

[0041] This application discretizes the continuous water level change process into multiple independent water level excitation events by identifying water level excitation events. Each water level excitation event includes a complete water level change cycle consisting of a rising phase, a stable phase, and a falling phase.

[0042] Specifically, by parsing the water level settlement time series data through a sliding window, the current water level change trend direction relative to the previous reference point, the cumulative water level change amplitude, and the first derivative of the water level change are calculated as the change slope. When the change trend is rising or falling, and the absolute value of the change slope exceeds the preset slope threshold within a preset time period, the current moment is marked as the start moment of the water level excitation event, and synchronous monitoring of the corresponding settlement sequence is started.

[0043] Determine whether the absolute value of the slope of change exceeds a preset slope threshold and whether the change amplitude exceeds a preset amplitude threshold. If both the absolute value of the slope of change and the change amplitude exceed the preset threshold, the data segment between the two candidate trend inflection points is divided into a water level excitation event.

[0044] In this embodiment of the application, obtaining the slope of water level change includes: calculating the ratio of the water level difference between adjacent sampling points in the water level sequence to the sampling interval time, and generating an instantaneous slope time sequence.

[0045] In this embodiment of the application, identifying water level inflection points based on the changing trend includes identifying critical moments in the instantaneous slope time series where the sign changes from positive to negative or from negative to positive, and marking the critical moments as candidate trend inflection points.

[0046] In this embodiment of the application, obtaining the change amplitude of water level includes: calculating the cumulative water level height difference between two adjacent candidate trend inflection points as the change amplitude.

[0047] It should be noted that water level excitation events are segments of water level changes with significant energy input, triggered by specific rainfall or flood control activities. The trend inflection point is the critical time coordinate point where the direction of water level evolution changes from rising to falling or vice versa.

[0048] For example, suppose the water level slope threshold is 10 mm / hour and the amplitude threshold is 100 mm; monitoring data records show that the water level started to rise at 12:00 and stopped rising at 18:00, with an average slope of 25 mm / hour and a cumulative amplitude of 150 mm during this period; 12:00 and 18:00 are identified as turning points. Since both the slope and amplitude meet the preset thresholds, the data segment from 12:00 to 18:00 is automatically segmented and marked as a water level rise excitation event. The settlement observation values ​​within this 6-hour period are then correlated for subsequent correlation analysis.

[0049] Step 130: Establish a settlement response analysis window corresponding to each water level excitation event, and obtain the settlement response fingerprint of the dike body in response to water level excitation based on the settlement response analysis window.

[0050] For details, please refer to Figure 2 The specific implementation process of step 130 is as follows:

[0051] Step 21: Establish a settlement response analysis window corresponding to each water level excitation event. Divide the settlement response analysis window into a pre-excitation baseline interval and a post-excitation response interval, with the time of occurrence of the water level excitation event as the center.

[0052] It should be noted that the settlement response analysis window is a time interval established around the water level excitation event, used to observe the changes in the settlement behavior of the dike before and after the water level change.

[0053] In this embodiment of the application, a settlement response analysis window corresponding to each water level excitation event is established, including: real-time monitoring of the water level change rate; when the water level change rate exceeds a preset threshold and the duration meets the set conditions, it is determined to be a water level excitation event, and the trigger time of the water level excitation event is recorded.

[0054] In this embodiment, the time span of the pre-excitation reference interval is determined based on the early settlement stabilization period of the dike, and the time span of the post-excitation response interval is determined based on the duration of water level changes and the response lag time of the dike.

[0055] Step 22: Extract baseline features reflecting the settlement stability of the embankment within the pre-excitation baseline interval, and extract response features reflecting the embankment's response to water level changes within the post-excitation response interval.

[0056] It should be noted that the settlement response analysis window includes a pre-excitation baseline interval and a post-excitation response interval. The pre-excitation baseline interval is used to record the settlement data of the embankment in a stable state, while the post-excitation response interval is used to record the settlement change process of the embankment after the water level changes.

[0057] In this embodiment of the application, extracting baseline features includes: obtaining a settlement observation sequence within the reference interval. The long-term consolidation trend and residual noise were separated by polynomial fitting. The slope of the trend term and residuals The variance is used as a baseline characteristic parameter. The slope is solved using the least squares method. Based on minimizing the sum of squared residuals The optimal solution is obtained by extracting the residual sequence. Calculate the variance of the residual term. .in Represents the mean of the residuals, and [ , [This serves as a baseline feature.] Indicates the first term within the reference interval Settlement observations at each sampling point This indicates the number of points within the reference interval. The timestamps corresponding to each sampling point, among which This represents the initial settlement displacement.

[0058] In this embodiment, the long-term consolidation trend refers to a slow, continuous, and directional settlement evolution process that occurs over time in water conservancy projects such as earth-rock dams and dikes under the combined effects of their own gravity, normal seepage, and the physical properties of soil consolidation. Essentially, it is structural compression caused by the discharge of pore water and the increase in effective stress. In monitoring data, it manifests as a monotonically decreasing (or stabilizing) low-frequency component, typically fitted by a linear or logarithmic function. Residual noise is the random fluctuation signal remaining in the original settlement monitoring data after removing the long-term consolidation trend. Its components include instrument noise, electromagnetic interference from the sensor itself, and quantization errors. The long-term consolidation trend, before water level changes, indicates the rate at which the dike is sinking, preventing normal consolidation settlement from being misjudged as water level-induced subsidence. If the variance is extremely small, it indicates stable data and high reliability for small responses; if the variance is large, it indicates strong environmental interference, requiring a higher warning threshold. The health status of the embankment is constructed by using the slope of the trend term and the variance of the residual term, providing a data reference for abnormal response analysis.

[0059] In this embodiment of the application, extracting response features includes: calculating the increment of settlement caused by a unit change in water level within the response interval. The time offset between the water level curve and the settlement curve was calculated through cross-correlation analysis. Obtain the displacement when the water level curve and settlement curve show the highest similarity. Calculate the time offset The increment of settlement and the time offset [ This serves as a response characteristic parameter. The calculation of the settlement increment eliminates random fluctuations by obtaining the ratio of covariance to variance, thus determining the contribution intensity of water level to settlement. This represents the sedimentation time series sampled synchronously within the response interval. This represents the water level time series sampled synchronously within the response interval. Represents the cross-correlation function, evaluating the shift of two sequences at a given time. Waveform consistency This represents the average settlement. Indicates the average water level. Indicates the sampling interval. This indicates that the settlement curve has shifted relative to the water level curve on the time axis. The waveform overlap between the two curves is highest when there are only one sampling point.

[0060] It should be noted that the settlement increment is the additional settlement caused solely by water level changes after a water level-excited event, excluding the contribution of long-term consolidation trends. It is typically expressed as settlement per unit water level change, quantifying the mapping strength between water level and settlement. The time offset is the time difference between the point in time when the water level changes and the point in time when the levee responds with settlement. Essentially, it represents the process of water pressure propagating deeper into the soil pores. Due to the soil permeability coefficient and skeletal stiffness, the settlement response lags behind water level fluctuations. This lag time reflects the permeability and structural integrity within the levee. A sudden increase in the settlement increment of a levee indicates a decrease in the structural stiffness of the corresponding levee, weakening its ability to resist water level disturbances. A sudden decrease in the time offset indicates possible early piping within the levee, with an abnormally accelerated signal transmission speed.

[0061] Step 23: Combine the baseline features and response features through feature vector processing to generate a settlement response fingerprint representing the water level feedback characteristics of the embankment.

[0062] Settlement response fingerprint refers to the feature expression formed by combining multiple settlement response feature parameters, which is used to describe the stability response mode of the embankment under water level excitation. The settlement response fingerprint features can reflect the health status and change trend of the embankment structure.

[0063] In this embodiment, the feature vectorization process includes: arranging baseline features and response features according to a preset dimensional order to form a high-dimensional feature vector, which is the settlement response fingerprint. For example, for a dam in a certain dam water level rise event, 240 hours of data are extracted from the baseline interval, and the slope is obtained by fitting using the least squares method. ,variance In the subsequent response period, the water level rose from 10m to 13m, and the settlement increased from 5mm to 9.5mm; the increase in settlement was obtained through regression calculation. Cross-correlation analysis was performed, and the correlation coefficient reached a peak of 0.97 when the sedimentation sequence was shifted 6 units backward relative to the water level sequence (sampling interval of 1 hour). This determined the time shift in hours. After hours, the final settlement response fingerprint vector generated is [0.01, 0.002, 1.5, 6].

[0064] Step 140: Calculate the spatial relationship between the settlement response fingerprint under the current water level excitation event and the preset normal response envelope of the corresponding embankment segment. Combine the water level excitation intensity information to constrain and determine the spatial relationship. When the settlement response fingerprint deviates from the allowable range of the normal response envelope, it is determined that there is an abnormal settlement in the corresponding embankment segment.

[0065] For details, please refer to Figure 3 The specific implementation process of step 140 is as follows:

[0066] Step 31: Obtain the embankment settlement monitoring sequence within a preset time period before and after the current water level excitation event. Extract the settlement response fingerprint based on the settlement time series. Use the baseline features of the settlement response fingerprint as static features and the response features of the settlement response fingerprint as dynamic features.

[0067] Step 32: Synchronously collect water level monitoring data within the corresponding time period, and analyze the water level excitation intensity parameter set, wherein the water level excitation intensity parameter set includes at least one of the following: water level change amplitude, water level change rate, and water level continuous change duration.

[0068] Step 33: Take the historical health sample set of the corresponding embankment section, perform boundary fitting based on the topological distribution of the historical health sample set in the multidimensional feature space, and construct a multidimensional normal response envelope describing the boundary of normal settlement response behavior.

[0069] It should be noted that the multidimensional normal response envelope is fitted using a support vector data description algorithm to form a minimal closed hypergeometry that covers the distribution pattern of historical healthy sample sets.

[0070] In this embodiment, a historical health sample set for the corresponding embankment section is retrieved. Boundary fitting is performed based on the topological distribution of the historical health sample set in the multidimensional feature space to construct a multidimensional normal response envelope describing the boundary of normal settlement response behavior. Please refer to [reference needed]. Figure 4 The specific content includes:

[0071] Step 331, obtain the historical health sample set X= of the geocoding matching of the embankment section to be analyzed. ,in This represents a four-dimensional feature vector that includes baseline slope, residual variance, settlement increment, and time offset. This indicates the number of historical healthy samples. The historical healthy sample set is selected from historical operating intervals where the dike structure is intact and there is no risk of leakage.

[0072] Step 332: Project the historical healthy sample set onto a multi-dimensional feature space, and perform boundary fitting of the topological distribution using a support vector data description algorithm. Specifically, find a minimum hypersphere with center 'a' and radius 'R' that contains most of the normal samples. Solve the objective function... The minimum value of , where This represents the penalty coefficient, used to balance the hypersphere volume and sample coverage. For slack variables that allow sample points to lie outside the boundary, the constraints are: , Represents sample vector The expression after mapping to a high-dimensional feature space using a kernel function is used to transform the nonlinear distribution relationships that may exist in the original feature space into a linearly separable structure in the high-dimensional space. The problem is transformed into a dual form using the Lagrange multiplier method, and a radial basis function kernel function is introduced to achieve nonlinear mapping of samples, obtaining the optimal hypersphere boundary function. The surface of the hypersphere in the multidimensional feature space is defined as the multidimensional normal response envelope. Commonly used kernel functions include... ,in Indicates sample With sample Similarity measurement in high-dimensional space The kernel function width parameter controls the smoothness of the sample distribution in the mapping space. Through the kernel function, the distance calculation of the feature space can be completed in the case of high-dimensional mapping, and the optimal hypersphere center a and radius R can be obtained.

[0073] Step 34: Perform normalization mapping on the settlement response fingerprint and calculate the spatial relationship parameters of the normalized settlement response fingerprint relative to the multidimensional normal response envelope in the multidimensional feature space. The spatial relationship parameters include the minimum distance from the settlement response fingerprint to the envelope boundary and the feature distance to the centroid of the envelope.

[0074] In this embodiment, the normalization mapping of the settlement response fingerprint includes: normalizing the static features and dynamic features of the settlement response fingerprint to obtain standard static features and standard dynamic features; concatenating the standard static features and standard dynamic features to obtain the coordinates of the settlement response fingerprint in the multidimensional feature space. This application transforms physical quantities in a time series into location points in a geometric space by obtaining the coordinates of settlement response fingerprints in a multidimensional feature space. Mapping to a multidimensional feature space, each water level and settlement event is transformed into a cluster point within this space, providing a data carrier for anomaly detection. By comparing the highest point with historical healthy operating areas, when the point is located at the edge or outside of the coordinate system, the corresponding dike section is identified as being in an unstable operating state.

[0075] In this embodiment, the minimum distance from the sedimentation response fingerprint to the envelope boundary can be expressed as: The characteristic distance from the sedimentation response fingerprint to the centroid of the envelope can be expressed as: ,in Let be the covariance matrix of the historical health sample set. Represents a high-dimensional feature map. This represents a real-time state point in a high-dimensional space. This represents a scalar value. A large scalar value indicates that the current monitoring point is moving away from the healthy center, and the levee may have experienced structural weakening, abnormal seepage, or other abnormalities. A small scalar value indicates that the current monitoring point is very close to the healthy center, and the levee is in a stable condition. This represents the displacement vector of the current state point relative to the reference center point. The minimum distance... Distance from features Encapsulated as spatial parameters .

[0076] The minimum distance is the shortest geometric distance from the target point to the nearest boundary hypersurface in the feature space, used to determine whether it has gone out of bounds; the feature distance is the weighted distance from the target point to the reference center point calculated based on the probability distribution and feature correlation, used to determine whether it is abnormal; the envelope centroid is the geometric center or statistical centroid of the multidimensional normal response envelope in the feature space, representing the most ideal and healthiest operating state point of the embankment; the spatial relationship parameters are a set of numerical indicators describing the target fingerprint's position relative to the standard healthy distribution, providing richer topological information than a single alarm threshold.

[0077] For example, the standardized coordinates are [0.4, 0.2, 1.8, 0.5]. Calculations show that the minimum distance from these coordinates to the boundary of the normal response envelope is... A positive value indicates that the fingerprint has slightly exceeded the security boundary; simultaneous calculation of feature distance (Mahanobis distance) reveals... The characteristic distances of historical health points are generally between [0, 2.0]; a larger characteristic distance indicates that, although the out-of-bounds distance... Although the risk level is not very high, the combination of various characteristic dimensions deviates significantly from historical patterns. For example, an abnormal combination of low settlement increment and long lag time appears. Therefore, it is determined that the evolution logic of the dike section has changed and the risk level is high.

[0078] Step 35: Couple the spatial relationship parameters with the water level excitation intensity parameters for analysis. Based on the preset excitation constraint rules, determine whether the deviation of the current settlement response matches the water level excitation intensity parameters. If the deviation exceeds the boundary of the multidimensional normal response envelope and still exceeds the allowable range after correction by the water level excitation intensity parameters, it is determined that the corresponding embankment section has abnormal settlement response output abnormal identification information.

[0079] It should be noted that the incentive constraint rules include: establishing a nonlinear mapping function between the water level incentive intensity and the allowable deviation threshold, and dynamically adjusting the radial width of the multidimensional normal response envelope in the multidimensional feature space according to the water level incentive intensity parameter.

[0080] Specifically, the amplitude H of water level change, the rate v of water level change, and the duration t of water level change under the current water level excitation event are obtained to form a set of water level excitation intensity parameters, which are then used based on a preset nonlinear constraint function. Calculate the radial width correction factor corresponding to the current load level. ,in , , The original radial width of the multidimensional normal response envelope is adjusted by pre-defined weighting coefficients. Compensation is performed to generate a dynamic envelope boundary. By comparing feature distances With dynamic envelope boundary The relationship, when At that time, it is determined that the deviation of the current settlement response matches the intensity of the water level excitation, even if That is, coordinate points outside the original envelope are still considered reasonable physical fluctuations caused by the load. When If the deviation still exceeds the preset allowable range after correction by the water level excitation intensity, it is determined that the deviation belongs to an abnormal settlement response. At this time, the number of the embankment section is obtained, and identification information containing the deviation, excitation intensity level and abnormality level is generated.

[0081] It should be noted that the nonlinear mapping function is a model that describes the non-proportional relationship between water level load and embankment deformation, simulating the nonlinear response of the structure under high load. The radial width is the geometric length from the centroid of the envelope to the boundary in the multidimensional feature space, representing the allowable error of structural feature fluctuation. Dynamic adjustment is the real-time contraction or expansion of the envelope boundary based on the real-time changes of external excitation. Abnormal settlement response is anomalies such as piping and voiding caused by damage to the internal structure of the embankment after excluding the contribution of external load.

[0082] Step 150: Determine the anomaly level based on the distance between the feature distance and the dynamic envelope boundary, and execute the warning action corresponding to the anomaly level.

[0083] In this embodiment, the anomaly level is determined based on the distance between the feature distance and the dynamic envelope boundary, including obtaining the radial difference based on the difference between the feature distance and the dynamic envelope boundary. The radial difference is mapped to a preset distance threshold range, when 0 When the deviation is 1, it is considered a level 1 anomaly, indicating a slight deviation; when If it is, it is considered a level 2 anomaly, meaning there is a moderate deviation; when If the anomaly is significant, it is classified as a Level 3 anomaly, indicating a severe deviation. Based on the anomaly level, a corresponding early warning action is matched from the preset strategy library. For Level 1 anomalies, data encryption monitoring is performed, automatically shortening the sampling interval of the front-end sensors to obtain more densely packed change features. For Level 2 anomalies, multi-source verification is performed, calling pore water pressure or deep displacement data from the surrounding area of ​​the dam section for logical cross-verification. For Level 3 anomalies, an immediate early warning action is performed, pushing a high-gain alarm message to the dam safety management platform via the BeiDou / 5G link.

[0084] It should be noted that the radial difference is the radial geometric length of the feature vector point in the multidimensional feature space that exceeds the dynamic envelope boundary after load correction; multi-source verification is to cross-validate the same abnormal event for monitoring indicators of different physical types.

[0085] Please refer to Figure 5 The present application provides a module diagram of a dike settlement anomaly early warning system based on water level change response, specifically including: a data acquisition module 510, a water level excitation event definition module 520, a settlement response fingerprint extraction module 530, a settlement anomaly judgment module 540, and an early warning action execution module 550. Wherein:

[0086] The data acquisition module 510 is used to acquire the water level monitoring sequence and settlement monitoring sequence of the target embankment section in real time. It aligns the water level monitoring sequence and settlement monitoring sequence with time scale through resampling to obtain water level and settlement time series data under a unified time axis.

[0087] The water level excitation event definition module 520 is used to obtain the water level change trend, change amplitude and change slope, identify water level inflection points based on the change trend, and perform event segmentation on water level settlement time series data by combining change amplitude and change slope to identify water level excitation events formed by water level changes.

[0088] The settlement response fingerprint extraction module 530 is used to establish a settlement response analysis window corresponding to each water level excitation event, and to obtain the settlement response fingerprint of the embankment's response to water level excitation based on the settlement response analysis window.

[0089] The settlement anomaly judgment module 540 is used to calculate the spatial relationship between the settlement response fingerprint under the current water level excitation event and the preset normal response envelope of the corresponding embankment section. It combines the water level excitation intensity information to constrain and judge the spatial relationship. When the settlement response fingerprint deviates from the allowable range of the normal response envelope, it is judged that there is a settlement anomaly in the corresponding embankment section.

[0090] The early warning action execution module 550 is used to determine the anomaly level based on the distance between the feature distance and the dynamic envelope boundary, and to execute the early warning action corresponding to the anomaly level.

[0091] This application improves the reliability of dike operation status monitoring and anomaly warning by acquiring water level settlement time series data; identifying water level inflection points based on the trend of change; segmenting the water level settlement time series data into events based on the change amplitude and slope; identifying water level excitation events formed by water level changes; establishing a settlement response analysis window corresponding to each water level excitation event; obtaining the settlement response fingerprint of the dike body's response to water level excitation based on the settlement response analysis window; calculating the spatial relationship between the settlement response fingerprint under the current water level excitation event and the preset normal response envelope of the corresponding dike section; constraining and judging the spatial relationship based on the water level excitation intensity information; and judging that there is a settlement anomaly in the corresponding dike section when the settlement response fingerprint deviates from the allowable range of the normal response envelope; determining the anomaly level based on the distance between the feature distance and the dynamic envelope boundary; and executing the early warning action corresponding to the anomaly level.

[0092] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A method for early warning of abnormal settlement of dikes based on water level change response, characterized in that, include: Real-time acquisition of water level monitoring sequences and settlement monitoring sequences of the target embankment section; time scale alignment of water level monitoring sequences and settlement monitoring sequences through resampling to obtain water level and settlement time series data under a unified time axis. Identify water level inflection points based on the changing trend, and segment water level settlement time series data into events by combining the change amplitude and change slope to identify water level excitation events formed by water level changes. Establish a settlement response analysis window corresponding to each water level excitation event, and obtain the settlement response fingerprint of the dike body in response to water level excitation based on the settlement response analysis window; Calculate the spatial relationship between the settlement response fingerprint under the current water level excitation event and the preset normal response envelope of the corresponding embankment segment. Combine the water level excitation intensity information to constrain and determine the spatial relationship. When the settlement response fingerprint deviates from the allowable range of the normal response envelope, it is determined that there is an abnormal settlement in the corresponding embankment segment. The anomaly level is determined based on the distance between the feature distance and the dynamic envelope boundary, and the corresponding early warning action is executed.

2. The method for early warning of dike settlement anomalies based on water level change response according to claim 1, characterized in that, Before performing event segmentation, the following steps are included: The water level settlement time series data is analyzed by sliding window, and the trend direction of water level change relative to the previous reference point, the cumulative change amplitude of water level, and the first derivative of water level change are calculated as the change slope. When the change trend is rising or falling, if the absolute value of the change slope exceeds the preset slope threshold within a preset time period, the current moment is marked as the start moment of the water level excitation event.

3. The method for early warning of abnormal levee settlement based on water level change response according to claim 2, characterized in that, Identify water level excitation events caused by water level changes, including: Determine whether the absolute value of the slope of change exceeds a preset slope threshold and whether the change amplitude exceeds a preset amplitude threshold. If both the absolute value of the slope of change and the change amplitude exceed the preset threshold, the data segment between the two candidate trend inflection points is divided into a water level excitation event.

4. The method for early warning of abnormal levee settlement based on water level change response according to claim 3, characterized in that, Identifying water level turning points based on changing trends includes identifying critical moments in the instantaneous slope time series where the sign changes from positive to negative or from negative to positive, and marking these critical moments as candidate trend turning points.

5. The method for early warning of abnormal settlement of dikes based on water level change response according to claim 1, characterized in that, Establish a settlement response analysis window corresponding to each water level excitation event, and obtain the settlement response fingerprint of the dike body in response to water level excitation based on the settlement response analysis window, including: Establish a settlement response analysis window corresponding to each water level excitation event, and divide the settlement response analysis window into a pre-excitation baseline interval and a post-excitation response interval with the time of occurrence of the water level excitation event as the center. Baseline features reflecting the settlement stability of the embankment are extracted within the pre-excitation baseline interval, and response features reflecting the embankment's response to water level changes are extracted within the post-excitation response interval. The baseline features and response features are combined through feature vector processing to generate a settlement response fingerprint representing the water level feedback characteristics of the embankment.

6. The method for early warning of abnormal settlement of dikes based on water level change response according to claim 5, characterized in that, Baseline features reflecting the settlement stability of the embankment are extracted within the pre-excitation reference interval, including: obtaining the settlement observation sequence within the reference interval, separating the long-term consolidation trend and residual noise using the moving average method or polynomial fitting method, and using the slope of the trend term and the variance of the residual term as baseline feature parameters.

7. The method for early warning of abnormal settlement of dikes based on water level change response according to claim 5, characterized in that, The response characteristics reflecting the levee's response to water level changes are extracted within the post-excitation response interval, including: Within the response interval, the increment of settlement caused by a unit change in water level is calculated. The time offset between the water level curve and the settlement curve is calculated through cross-correlation analysis. The increment of settlement and the time offset are used as response characteristic parameters.

8. The method for early warning of abnormal settlement of dikes based on water level change response according to claim 1, characterized in that, Determining that the corresponding embankment section has abnormal settlement includes: Obtain the levee settlement monitoring sequence within a preset time period before and after the current water level excitation event, and extract the settlement response fingerprint based on the settlement time sequence; Simultaneously collect water level monitoring data within the corresponding time period, and analyze the set of water level excitation intensity parameters; The settlement response fingerprint is transformed into a multidimensional feature space. The historical health sample set of the corresponding embankment section is retrieved. The boundary is fitted according to the topological distribution of the historical health sample set in the multidimensional feature space to construct a multidimensional normal response envelope describing the boundary of normal settlement response behavior. Perform a normalization mapping on the settlement response fingerprint and calculate the spatial relationship parameters of the normalized settlement response fingerprint relative to the multidimensional normal response envelope in the multidimensional feature space. The spatial relationship parameters are coupled with the water level excitation intensity parameters for analysis. Based on the preset excitation constraint rules, it is determined whether the deviation of the current settlement response matches the water level excitation intensity parameters. If the deviation exceeds the boundary of the multidimensional normal response envelope and still exceeds the allowable range after correction by the water level excitation intensity parameters, it is determined that the corresponding embankment section has abnormal settlement response output abnormal identification information.

9. The method for early warning of abnormal settlement of dikes based on water level change response according to claim 8, characterized in that, Retrieve the historical health sample set for the corresponding embankment section, and perform boundary fitting based on the topological distribution of the historical health sample set in the multidimensional feature space to construct a multidimensional normal response envelope describing the boundary of normal settlement response behavior, including: Obtain the historical health sample set of geocoding matching for the embankment section to be analyzed; The historical health sample set is projected onto a multidimensional feature space, and the boundary of the topological distribution is fitted by the support vector data description algorithm. That is, a minimum hypersphere with center a and radius R is found, and the surface of the minimum hypersphere in the multidimensional feature space is defined as the multidimensional normal response envelope.

10. A dike settlement anomaly early warning system based on water level change response, used to implement the method of claims 1-9, characterized in that, include: The data acquisition module is used to acquire the water level monitoring sequence and settlement monitoring sequence of the target embankment section in real time. The water level monitoring sequence and settlement monitoring sequence are aligned in time scale by resampling to obtain water level and settlement time series data under a unified time axis. The water level excitation event definition module is used to obtain the water level change trend, change amplitude and change slope, identify water level inflection points based on the change trend, and perform event segmentation on water level settlement time series data by combining change amplitude and change slope to identify water level excitation events formed by water level changes. The settlement response fingerprint extraction module is used to establish a settlement response analysis window corresponding to each water level excitation event, and to obtain the settlement response fingerprint of the dike body in response to water level excitation based on the settlement response analysis window. The settlement anomaly judgment module is used to calculate the spatial relationship between the settlement response fingerprint under the current water level excitation event and the preset normal response envelope of the corresponding embankment section. It combines the water level excitation intensity information to constrain and judge the spatial relationship. When the settlement response fingerprint deviates from the allowable range of the normal response envelope, it is judged that there is a settlement anomaly in the corresponding embankment section. The early warning action execution module is used to determine the anomaly level based on the distance between the feature distance and the dynamic envelope boundary, and to execute the early warning action corresponding to the anomaly level.

Images