A method for monitoring and analyzing the dynamic of groundwater seepage field and hydrogeology

Abstract

The application discloses a kind of groundwater seepage field dynamic monitoring and hydrogeological analysis method, obtains monitoring well basic information, continuous monitoring data, external disturbance data and hydrogeological background data, converts water level into groundwater level elevation and carries out time alignment and basic anomaly mark;Adjacent well relationship table and disturbance event record are constructed, event response characteristics are extracted, and monitoring well reliability score is generated;Based on the matching relationship of monitoring well reliability score, abnormal source trigger condition, event response characteristics and disturbance event record, distorted anomaly monitoring well, real disturbance candidate monitoring well and to be on-site verification monitoring well are identified;Based on the identification result, the interpolation weight is corrected, the self-correcting groundwater level surface is reconstructed, the self-correcting groundwater seepage field and the review instruction table are generated.

Description

Technical Field

[0001] This invention relates to the field of groundwater monitoring and hydrogeological data processing technology, and in particular to a method for dynamic monitoring and hydrogeological analysis of groundwater seepage fields. Background Technology

[0002] Dynamic monitoring and hydrogeological analysis of groundwater seepage fields are widely used in groundwater resource assessment, engineering dewatering and drainage impact assessment, pollution migration risk identification, river and lake recharge and drainage relationship analysis, and underground space development safety monitoring. Existing technologies typically rely on monitoring wells equipped with devices such as water level gauges, water temperature and conductivity sensors to continuously collect data on groundwater level, water temperature, and conductivity. Groundwater flow direction and seepage intensity are then inferred through methods such as isostatic mapping, spatial interpolation, and hydraulic gradient calculation.

[0003] However, in actual engineering monitoring, monitoring well data is not entirely reliable. In practical applications, it is easily affected by factors such as well pipe blockage, mismatched filter pipe layers, sensor drift, communication interruptions, frozen data, and localized human-induced pumping and drainage disturbances. Individual monitoring wells may exhibit abnormal water levels, prolonged periods of no response, sudden changes, or inconsistencies with surrounding wells. Existing groundwater seepage field analysis methods often assume that all monitoring well data have equal reliability, directly incorporating all monitoring well data into groundwater level interpolation and hydraulic gradient calculations. This can easily lead to abnormal well points affecting groundwater level contour lines, thereby influencing groundwater flow direction, seepage intensity, and hydrogeological interpretation results.

[0004] Existing data quality control methods mostly focus on removing missing, outlier, or abrupt changes. However, groundwater systems exhibit significant hydrogeological spatial heterogeneity. Inconsistent responses between some monitoring wells and adjacent wells do not necessarily indicate data distortion. They may reflect real local pumping and drainage effects, changes in river replenishment and drainage relationships, or the development of fault zones or highly permeable channels, among other hydrogeological anomalies. If such data is removed solely based on outlier rules, real local seepage anomalies may be treated as bad data, weakening the monitoring system's ability to identify abnormal seepage changes. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] Therefore, in order to solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for dynamic monitoring and hydrogeological analysis of groundwater seepage field, which acquires basic information of monitoring wells, continuous monitoring data, external disturbance data and hydrogeological background data, converts water level into groundwater level elevation and forms a standardized monitoring well dataset; The standardized monitoring well dataset is time-aligned and basic anomaly markers are applied to generate time-aligned monitoring well time series. A table of relationships between adjacent wells is constructed based on the plane distance between monitoring wells, the degree of overlap of filter pipes, and the consistency of aquifers, and a record of disturbance events is established based on external disturbance data; Based on the adjacent well relationship table and disturbance event records, event response features are extracted, and data integrity score, abnormal mutation frequency score, response synchronization score, adjacent well consistency score and physical rationality score are calculated to generate a monitoring well reliability score. Based on the monitoring well reliability score, anomaly source triggering conditions, and the matching relationship between event response characteristics and disturbance event records, distorted anomaly monitoring wells, real disturbance candidate monitoring wells, and monitoring wells to be verified on-site are identified. Based on the reliability score of the monitoring wells, and the identification results of distorted and abnormal monitoring wells, real disturbance candidate monitoring wells, and monitoring wells to be verified on-site, the interpolation weights are corrected, the self-correcting groundwater level is reconstructed, and a self-correcting groundwater seepage field and verification instruction table are generated.

[0007] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method described in this invention, the standardized monitoring well dataset includes monitoring well number, plane coordinates, wellhead elevation, well depth, well diameter, top depth of filter pipe, bottom depth of filter pipe, aquifer type, well completion time, sensor model, sensor resolution, water level time series, water temperature time series, conductivity time series, sensor operating status data, external disturbance data, and hydrogeological background data. The water level time series is converted into a groundwater level elevation time series based on the wellhead elevation and the vertical distance from the water surface to the wellhead. The sensor resolution is used to set the minimum effective response threshold, and the hydrogeological background data is used to determine hydrogeologically sensitive areas.

[0008] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method of the present invention, the generation of the time-aligned monitoring well time series includes: resampling the groundwater level elevation time series, water temperature time series and electrical conductivity time series according to a preset sampling interval; When there are multiple sampled values ​​in the same sampling interval, the median is taken as the representative value; when there are no sampled values ​​in the same sampling interval, the sampling interval is marked as a missing value. Generate missing value markers, consecutive duplicate value markers, isolated mutation value markers, and sensor offline markers; wherein, the isolated mutation value marker is used to represent data points that exceed a physical threshold and do not form a continuous trend of change in the same direction within a time period in which no corresponding disturbance event is recorded.

[0009] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method of the present invention, the step of constructing the adjacent well relationship table includes: first determining whether the bottom depth of the filter pipe is greater than the top depth of the filter pipe; When the depth of the filter pipe is effective, the adjacent correlation coefficient is calculated based on the plane distance of the monitoring wells, the degree of overlap of the filter pipes, and the consistency of the aquifer. Among them, the aquifer consistency is used as a control term for the adjacent correlation coefficient, so that the monitoring wells corresponding to unconnected aquifers do not enter the adjacent well set; Among the candidate monitoring wells that meet the aquifer connectivity condition, monitoring wells with an adjacent correlation coefficient greater than the adjacent threshold are selected, or a preset number of candidate monitoring wells are selected according to the adjacent correlation coefficient to form an adjacent well relationship table.

[0010] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method described in this invention, the step of establishing the disturbance event record includes: extracting disturbance events from rainfall data, pumping data, drainage data, recharge data and river water level data, and recording the disturbance event type, disturbance start time, disturbance end time, disturbance intensity, response direction reference and disturbance event window; The disturbance event window includes a pre-event baseline period, an event response period, and an event recovery period; When the reference time period before the event is affected by other disturbances, a continuous time period that is not affected by other disturbances shall be selected as the reference time period; When the response periods of multiple disturbance events overlap, they are marked as composite disturbance events, and the event weights are reduced or the corresponding response synchronicity sub-scores are not calculated.

[0011] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method of the present invention, the step of extracting event response characteristics includes: determining the response amplitude based on the absolute deviation of the groundwater level elevation from the median groundwater level elevation of the pre-event reference period during the event response period. The moment when the absolute deviation reaches the preset number of sampling points consecutively and exceeds the water level response threshold is defined as the response start moment. The moment when the absolute deviation reaches its maximum value is defined as the peak moment; The recovery time is defined as the duration during which the groundwater level returns to the recovery range determined by the pre-event baseline groundwater level and water level response threshold after the peak time, and the time during which the groundwater level remains within the preset number of sampling points. A benchmark envelope for the event response of adjacent wells is constructed based on the median and interquartile range of the event response characteristics in the adjacent well set.

[0012] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method of the present invention, the step of generating a monitoring well reliability score includes: obtaining a data integrity score by taking the proportion of missing sampling points, continuous repeating value sampling points and offline sampling points to the sampling points that should be sampled in the current evaluation time window, respectively. The abnormal mutation frequency score is obtained by the proportion of isolated mutation points to valid sampling points; The event response characteristics of the target monitoring well are compared with the weighted event response characteristics of adjacent wells, and combined with the event response benchmark envelope of adjacent wells to obtain the response synchronicity score; The local groundwater level plane is fitted using adjacent monitoring wells other than the target monitoring well, and the consistency score of adjacent wells is obtained based on the residual of the target monitoring well. A physical rationality score is obtained based on water temperature, conductivity, hydraulic gradient direction changes, disturbance event response, and well structure constraints.

[0013] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method described in this invention, when generating the reliability score of the monitoring well, if the adjacent well benchmark is insufficient, resulting in the inability to calculate the response synchronization score and the adjacent well consistency score, or if there is no effective disturbance event within the current evaluation time window, resulting in the inability to calculate the response synchronization score, the weights of the remaining calculable scores are normalized. If the number of uncalculable scores exceeds the preset number of missing scores, the reliability score of the monitoring well is limited to not exceeding the first score limit. If the number of valid sampling points within the current evaluation time window is lower than the preset valid sampling ratio, the reliability score of the monitoring well is limited to not exceeding the second score limit. If the sensor's operating status data indicates that the system is offline at all times, the monitoring well reliability score will be set to the lowest possible value.

[0014] As a preferred embodiment of the groundwater seepage field dynamic monitoring and hydrogeological analysis method described in this invention, the steps of identifying distorted abnormal monitoring wells, real disturbance candidate monitoring wells, and monitoring wells to be verified on-site include: firstly, determining whether the identification of abnormal sources is triggered based on data integrity score, abnormal change frequency score, sensor offline ratio, adjacent well consistency score, response synchronization score, envelope boundary crossing mark, unrecovered mark, and physical rationality score. Once triggered, if the target monitoring well meets at least two of the following conditions: missing data, isolated mutation, offline, long-term no fluctuation, unrecoverable step mutation, or inconsistent filter pipe layer, it will be marked as a distortion anomaly monitoring well. When the abnormal response of a target monitoring well meets at least two of the following characteristics: disturbance event matching, synchronous change of multiple parameters, matching of hydrogeologically sensitive areas, or recurrence, it is marked as a candidate monitoring well for real disturbance. The remaining target monitoring wells that triggered the anomaly source identification were marked as monitoring wells awaiting on-site verification.

[0015] As a preferred embodiment of the dynamic monitoring and hydrogeological analysis method for groundwater seepage field described in this invention, the step of reconstructing the self-correcting groundwater level and generating the self-correcting groundwater seepage field includes: generating a reliability correction coefficient based on the reliability score of the monitoring well; Specifically, when the reliability score of the monitoring well is not lower than the first reliability threshold, the reliability correction coefficient takes the first weight value; when the reliability score of the monitoring well is lower than the first reliability threshold but not lower than the second reliability threshold, the reliability correction coefficient decreases as the reliability score of the monitoring well decreases; when the reliability score of the monitoring well is lower than the second reliability threshold, the reliability correction coefficient takes the second weight value, and the first weight value is greater than the second weight value. Distortion anomaly monitoring wells are set to not participate in regional interpolation, the reliability correction coefficient of monitoring wells to be verified on-site is limited to a preset upper limit, real disturbance candidate monitoring wells are set to not participate in regional smooth interpolation, and local anomaly constraint weights are generated. The reliability correction coefficient is multiplied by the inverse distance basic spatial weight and normalized in the monitoring wells involved in the interpolation to obtain the corrected interpolation weight. The self-correcting groundwater level is then reconstructed based on the corrected interpolation weight. The direction of groundwater flow is determined based on the negative gradient of the self-correcting groundwater level, the intensity of seepage is determined based on the hydraulic gradient, and a real anomaly candidate area and a verification instruction table are generated based on the local anomaly constraint weight.

[0016] The beneficial effects of this invention are: 1. This invention constructs an adjacent well relationship table by monitoring the plane distance between wells, the degree of overlap of filter pipes, and the consistency of aquifers, so that the reliability evaluation of the target monitoring well is combined with the common response law of adjacent wells under similar hydrogeological conditions. The consistency of aquifers, as a control item of the adjacent correlation coefficient, can reduce the situation where monitoring wells of unconnected aquifers are included in the adjacent well set, and improve the hydrogeological rationality of the comparison benchmark between adjacent wells.

[0017] 2. This invention establishes disturbance event records based on external disturbance data and extracts event response features such as response start time, peak time, response amplitude, and recovery time. By comparing the event response features of the target monitoring well with the weighted event response features of adjacent wells and the benchmark envelope of the event response of adjacent wells, it can be determined whether the monitoring well conforms to the disturbance response pattern of adjacent wells. This invention is applicable to various disturbance scenarios such as rainfall, pumping, drainage, recharge, and river water level changes.

[0018] 3. This invention generates a monitoring well reliability score by comprehensively considering data integrity score, abnormal mutation frequency score, response synchronization score, adjacent well consistency score, and physical rationality score. These scores evaluate the monitoring well data quality from the perspectives of acquisition link stability, isolated mutation risk, disturbance response consistency, adjacent well spatial consistency, and hydrogeological and physical constraints, thus reducing bias caused by judging based on a single indicator.

[0019] 4. This invention sets trigger conditions for anomaly sources and classifies anomaly monitoring wells into distorted anomaly monitoring wells, real disturbance candidate monitoring wells, and monitoring wells awaiting on-site verification. For data distortion caused by sensor drift, well pipe blockage, abnormal acquisition link, or mismatch in filter pipe layer, this invention reduces or cancels their weight in regional interpolation. For anomaly responses that match disturbance events, synchronous changes in multiple parameters, hydrogeologically sensitive areas, or recurring characteristics, this invention retains them as a basis for identifying real anomaly candidate areas through local anomaly constraint weights, thereby balancing low-reliability data suppression and the preservation of real anomaly clues.

[0020] 5. This invention generates a reliability correction coefficient based on the reliability score of the monitoring well, and combines the reliability correction coefficient with the inverse distance basic spatial weight to obtain the corrected interpolation weight, thereby reconstructing the self-correcting groundwater level. This method can reduce the impact of low-reliability monitoring wells on the spatial distribution of groundwater level, thereby improving the stability of the judgment of groundwater flow direction, hydraulic gradient, and seepage intensity level.

[0021] 6. This invention determines the groundwater flow direction based on the negative gradient direction of the self-calibrated groundwater level and generates candidate areas for real anomalies and a verification instruction table by combining local anomaly constraint weights. The verification instruction table can provide verification methods for siltation risk, sensor drift risk, structural anomaly risk, data acquisition link anomaly risk, and candidate monitoring wells for real disturbances, realizing closed-loop management from data evaluation to on-site verification. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating the overall workflow of the present invention.

[0023] Figure 2 This is a flowchart of the reliability scoring and anomaly source identification of the present invention.

[0024] Figure 3 This is a sub-flowchart of the adjacent well relationship and disturbance response of the present invention.

[0025] Figure 4 This is a sub-flowchart of the weight correction and verification closed loop of the present invention. Detailed Implementation

[0026] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0027] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0028] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention; the phrase "in an embodiment" appearing in different places in this specification does not all refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0029] Example Reference Figures 1-4 This is the first embodiment of the present invention, which provides a method for dynamic monitoring and hydrogeological analysis of groundwater seepage field, wherein the method is executed by a processor; the processor is preferably deployed in a groundwater monitoring platform, an edge computing gateway or a server.

[0030] The specific steps are as follows: S1. Acquire basic information of monitoring wells, continuous monitoring data, external disturbance data and hydrogeological background data, and generate a standardized monitoring well dataset; First, the static attributes of each monitoring well are collected. Then, continuous monitoring data, external disturbance data, and hydrogeological background data are collected. Finally, the water level time series is uniformly converted into a groundwater level elevation time series to form a standardized monitoring well dataset. The standardized monitoring well dataset includes basic information of the monitoring well, groundwater level elevation time series after elevation conversion, water temperature time series, conductivity time series, sensor operating status data, external disturbance data, and hydrogeological background data.

[0031] S11. Collect basic information on monitoring wells; The following information is collected for the i-th monitoring well: well number, plane coordinates, wellhead elevation, well depth, well diameter, top depth of the filter pipe, bottom depth of the filter pipe, aquifer type, well completion time, sensor model, and sensor resolution. The plane coordinates are used to calculate the plane distance between monitoring wells. The wellhead elevation is used to convert the water level monitoring results into groundwater level elevation. The top depth, bottom depth, and aquifer type of the filter pipe are used to construct a table of adjacent well relationships. The sensor resolution is used to set the minimum effective response threshold. The top depth and bottom depth of the filter pipe are both positively correlated with the direction downwards from the wellhead.

[0032] S12. Collect continuous monitoring data; The system collects water level time series, water temperature time series, conductivity time series, and sensor operating status data for the i-th monitoring well. The water level time series is collected by a pressure level gauge or water level sensor; the water temperature and conductivity time series are collected by an in-well water temperature and conductivity sensor; and the sensor operating status data is uploaded by a field data acquisition terminal, including online status, offline status, communication interruption status, and abnormal voltage status.

[0033] S13. Collect external disturbance data and hydrogeological background data; The external disturbance data includes at least one of rainfall data, pumping data, drainage data, recharge data, and river water level data; wherein, the rainfall data includes the start time of rainfall, the end time of rainfall, and the cumulative rainfall; the pumping and drainage data include the well number, start and stop times, and pumping / drainage flow rates; the recharge data includes the recharge well number, start and stop times, and recharge flow rates; and the river water level data includes the river monitoring section number, sampling time, and river water level elevation.

[0034] Collect hydrogeological background data; the hydrogeological background data includes at least one of the following: geological survey data, aquifer distribution maps, fault structure maps, riverbank data, pumping and drainage project location data, recharge project location data, and existing hydrogeological zoning data. The hydrogeological background data is used to subsequently determine hydrogeologically sensitive areas and to determine whether abnormal responses in monitoring wells may be caused by river replenishment / drainage, pumping and drainage effects, recharge effects, fault zones, sand interlayers, karst development zones, or highly permeable channels.

[0035] Pumping and drainage typically correspond to a decrease in the surrounding groundwater level, while recharge typically corresponds to an increase in the surrounding groundwater level. Rainfall and river level changes are used to determine changes in natural recharge or river recharge-drainage relationships. The external disturbance data is used to establish disturbance event records, rather than being stored as isolated data.

[0036] S14. Convert the water level time series to groundwater level elevation; When the water level sensor outputs the vertical distance from the water surface to the wellhead, the groundwater level elevation is calculated using the following formula: ;in, Let be the groundwater level elevation of the i-th monitoring well at time t, in meters; The elevation of the wellhead of the i-th monitoring well is given in meters. The vertical distance from the water surface to the wellhead of the i-th monitoring well at time t is expressed in meters.

[0037] It should be noted that if the water level sensor outputs water column pressure, the pressure value is first converted into water column height based on water density and gravitational acceleration, and then converted into groundwater level elevation.

[0038] S2. Perform time alignment and basic anomaly labeling on the standardized monitoring well dataset to generate time-aligned monitoring well time series; Since different monitoring wells and different sensors may have different sampling frequencies, it is necessary to first unify the time reference and then mark missing, duplicate, offline and isolated mutation data so that subsequent scoring can be carried out on the same time scale; the time-aligned monitoring well time series refers to the groundwater level elevation, water temperature and conductivity series after resampling according to the unified sampling interval.

[0039] S21. Resample according to the preset sampling interval; According to the preset sampling interval Resampling is performed on the time series of groundwater level elevation, water temperature, and conductivity; the preset sampling interval The time interval can be set from 5 minutes to 24 hours depending on actual needs. When the target area is significantly affected by pumping, drainage or recharge and the groundwater level changes rapidly, it is preferable to set the time interval from 5 minutes to 1 hour. When the target area is a long-term monitoring scenario for regional groundwater and the groundwater level changes relatively slowly, it is preferable to set the time interval from 6 hours to 24 hours. The preferred value in this embodiment is 1 hour. This interval can usually balance the accuracy of disturbance response identification and the cost of data storage.

[0040] S22. Determine the representative value for each sampling interval; If multiple sampled values ​​exist within the same sampling interval, the median (which is not sensitive to individual spike noise) is taken as the representative value for that sampling interval. If no sampled value exists within the same sampling interval, that sampling interval is marked as a missing value and is not directly interpolated to avoid the interpolated value masking the sensor's offline or communication interruption issues.

[0041] S23. Generate basic exception markers; Generate missing value markers, continuous repeat value markers, isolated abrupt value markers, and sensor offline markers for time-aligned monitoring well time series; The continuous repeating value marker refers to the monitoring values ​​of groundwater level elevation, water temperature or conductivity being exactly the same or having a change range less than twice the corresponding sensor resolution within a continuous period of 6 to 24 hours. If the continuous duration is less than 6 hours, there is a risk of misjudging the natural steady state as data freezing. If the continuous duration exceeds 24 hours, it may delay the detection of data freezing.

[0042] The isolated mutation value marker refers to a data point that simultaneously meets the following conditions: the rate of change within adjacent sampling intervals exceeds the corresponding physical threshold; no consecutive 2-3 sampling points before and after the data point show the same trend of change; and no rainfall, pumping, drainage, recharge, or river water level change event occurred at the sampling time of the data point.

[0043] S3. Construct an adjacent well relationship table and establish a disturbance event record; like Figure 3 As shown, after time alignment is completed, the comparison objects for each target monitoring well are further determined, and the external disturbance data are converted into computable disturbance event records.

[0044] S31. Calculate the horizontal distance between monitoring wells; The planar distance between the i-th monitoring well and the j-th monitoring well is calculated using the following formula: ;in, The horizontal distance between the i-th monitoring well and the j-th monitoring well is expressed in meters. , , , These are the plane coordinates of the two monitoring wells, in meters.

[0045] S32. Calculate the overlap coefficient of the filter pipes; Before calculating the filter pipe overlap coefficient, first determine whether the bottom depth of the filter pipe of the i-th monitoring well and the j-th monitoring well is greater than the top depth of the filter pipe. If the bottom depth of the filter pipe of the i-th monitoring well is less than or equal to its top depth, then the well is marked as a structural data anomaly monitoring well. If the bottom depth of the filter pipe of the j-th monitoring well is less than or equal to its top depth, then the well is also marked as a structural data anomaly monitoring well. For any monitoring well marked as having structural data anomalies, it will not be used for the construction of the adjacent well relationship table.

[0046] When the top and bottom depths of the filter pipes of both the i-th and j-th monitoring wells are valid, the overlap coefficient of their filter pipes is calculated according to the following formula: ;in, The overlap coefficient of the water filter pipes is dimensionless and its value range is [value range missing]. ; and These are the top and bottom depths of the filter pipe in the i-th monitoring well, respectively, in meters. and These are the top and bottom depths of the filter pipe in the j-th monitoring well, respectively, in meters.

[0047] S33. Calculate the adjacent correlation coefficient; To prevent unrelated monitoring wells from being mistakenly included in the adjacent relationships due to overlapping filter pipe depths but disconnected aquifers, this step calculates the adjacent correlation coefficient using the following formula. The calculation formula is: ; in, Let be the adjacency correlation coefficient between the i-th monitoring well and the j-th monitoring well, which is dimensionless; The preset influence distance is in meters. For aquifer consistency parameters; This is the overlap coefficient of the water filter pipes.

[0048] When the two monitoring wells belong to the same aquifer When two monitoring wells belong to different aquifers but are connected to each other, When the two monitoring wells belong to different aquifers, .

[0049] Preferably, the preset influence distance The value should be 1 to 3 times the average nearest neighbor distance of the monitoring wells in the target area. If the value is too small, there will be insufficient number of adjacent wells to be associated. If the value is too large, monitoring wells from different hydrogeological units may be mixed in. This range helps to achieve a balance between sample size and hydrogeological similarity.

[0050] In satisfying Select from candidate monitoring wells Greater than adjacent thresholds The monitoring wells that satisfy the condition are recorded as adjacent wells of the i-th monitoring well, and added to the set of adjacent wells; or, from the wells that satisfy the condition... Select from candidate monitoring wells The largest 3-8 monitoring wells are considered as adjacent wells. Fewer than 3 wells make it difficult to establish a stable statistical benchmark, while more than 8 wells easily introduce monitoring wells that are far apart or have significantly different hydrogeological backgrounds. Among these, the adjacent threshold... It is used to screen candidate monitoring wells that have spatial proximity to the target monitoring well, comparability of the filter pipe layer, and aquifer connectivity.

[0051] Preferably, adjacent thresholds The preferred value range is 0.35~0.60; if adjacent thresholds Monitoring wells with a threshold below 0.35, that are far apart, or have low overlap in their filter pipes, may be included in the neighboring well set, reducing the representativeness of the neighboring well comparison benchmark; if the adjacent threshold... A value higher than 0.60 can easily lead to an insufficient number of adjacent wells in sites with sparsely distributed monitoring wells, affecting the calculation of response synchronization scores and neighbor well consistency scores; therefore, in this embodiment, the adjacent threshold... The preferred value range is set to 0.35~0.60; however, in practical applications, in sites where monitoring wells are densely distributed and aquifers are clearly distributed, a preferred value of 0.50~0.60 can be considered; in sites where monitoring wells are sparsely distributed but aquifer connectivity is clear, a preferred value of 0.35~0.50 can be considered.

[0052] S34. Establish a disturbance event log; Disturbance events are extracted from rainfall data, pumping data, drainage data, recharge data, and river water level data to generate disturbance event records. The disturbance event records include disturbance event number, disturbance event type, disturbance start time, disturbance end time, disturbance intensity, response direction reference, and disturbance event window. For rainfall events, the disturbance intensity is expressed as the cumulative rainfall in mm; for pumping and drainage events, the disturbance intensity is expressed as the product of flow rate and duration in m³; for recharge events, the disturbance intensity is expressed as the product of recharge flow rate and duration in m³; and for river level change events, the disturbance intensity is expressed as the amplitude of river level change in m.

[0053] The response direction reference is used to record the groundwater level change direction that usually corresponds to the disturbance event, where pumping and drainage correspond to a decrease in groundwater level, and recharge and effective rainfall replenishment correspond to an increase in groundwater level.

[0054] The aforementioned disturbance event window includes a pre-event baseline period, an event response period, and an event recovery period; wherein, the pre-event baseline period is used to determine the groundwater level elevation baseline before the disturbance occurs, the event response period is used to extract the response start time, peak time, and response amplitude, and the event recovery period is used to determine whether the groundwater level elevation has recovered to the recovery determination range; Preferably, for shallow high-permeability aquifers or sites significantly affected by engineering pumping and drainage, shorter event response and recovery periods are preferred; for confined aquifers, low-permeability aquifers, or sites with delayed monitoring well responses, longer event response and recovery periods are preferred. Therefore, the pre-event baseline period is 6-72 hours before the disturbance begins, the event response period is 1-7 days after the disturbance begins, and the event recovery period is 6-14 days after the disturbance ends. The above time ranges are used to cover the rapid response of shallow aquifers and the slow response of confined aquifers.

[0055] If other events such as rainfall, pumping, drainage, recharge, or river level changes occur within the baseline period before the event, that baseline period is marked as the disturbed baseline period, and the continuous period closest to the current disturbance event and not affected by other disturbances is selected as the baseline period. If the event response periods of multiple disturbance events overlap, they are marked as a composite disturbance event, and the weight of this event is reduced or the response synchronicity sub-score of this event is not calculated in the response synchronicity score.

[0056] S4. Extract event response features and construct a baseline envelope for event responses in adjacent wells; like Figure 3 As shown, based on the obtained neighbor well relationship table and disturbance event records, each target monitoring well is evaluated in the context of neighbor well response, avoiding the reliance on fixed thresholds for anomaly judgment, thus better adapting to dynamic monitoring scenarios with different sites, different aquifers and different disturbance intensities.

[0057] S41. Extract event response features; Based on the time-aligned monitoring well time series, the event response features corresponding to the k-th disturbance event for the i-th monitoring well are extracted. The event response features include the response start time. Peak time Response amplitude and recovery time ; The above response amplitude The calculation formula is: ; in, Let be the response amplitude of the i-th monitoring well to the k-th disturbance event, in meters; This represents the event response period for the k-th disturbance event. Let be the groundwater level elevation of the i-th monitoring well at time t, in meters; The median groundwater level elevation during the preset reference period before the disturbance event occurs, in meters.

[0058] The above response start time Expressed in hours relative to the start of the disturbance; when the groundwater level elevation is relative to The absolute change exceeds the water level response threshold for 2-3 consecutive sampling points. At that time, it will exceed the water level response threshold for the first time. The time is determined as the response start time; the water level response threshold The preferred depth is 0.02m to 0.10m, or 2 to 3 times the standard deviation of the groundwater level elevation during the baseline period before the event. It should be noted that the range of 0.02m to 0.10m is used to make the water level response threshold higher than the resolution and short-term noise level of common water level sensors, while retaining the ability to identify small water level responses in shallow aquifers; when the background fluctuations in the target area are large, 2 to 3 times the standard deviation of the groundwater level elevation during the baseline period before the event is used as the threshold to adapt to the natural fluctuation amplitude of different sites.

[0059] The above peak times During the event response period Inside The time when the maximum value is reached is expressed in hours relative to the start of the disturbance.

[0060] The above recovery time To from the peak moment Initially, until the groundwater level returned to normal. Within the range, the duration of 2-3 consecutive sampling points is recorded, in hours; if the recovery conditions are not met within the event recovery period, the recovery time will be extended. Record the maximum duration of the event recovery period and generate an unrecovered flag.

[0061] If no valid response is detected in the i-th monitoring well within the disturbance event window, the disturbance event is marked as a no-response event, and the response synchronicity sub-score corresponding to the disturbance event is set to 0 in the subsequent response synchronicity scoring.

[0062] S42. Calculate the weighted event response characteristics of adjacent wells; Based on the adjacent well relationship table, read the set of adjacent wells corresponding to the i-th monitoring well, and then determine the adjacent well set based on the adjacent correlation coefficient. The event response characteristics of each adjacent well in this set are weighted; the weighted response start time is calculated according to the following formula. The calculation formula is: ; in, The weighted response start time of the set of adjacent wells corresponding to the i-th monitoring well under the k-th disturbance event, in hours; Let i be the set of adjacent wells of the i-th monitoring well; The starting time of the response of the j-th monitoring well to the k-th disturbance event is expressed in hours.

[0063] Similarly, the weighted peak time of adjacent wells is calculated using the same method described above. Weighted response amplitude of adjacent wells Weighted recovery time of adjacent wells .

[0064] like If the weighted event response characteristics of adjacent wells are not calculated, the i-th monitoring well will be marked as a monitoring well with insufficient benchmark among adjacent wells.

[0065] If the number of effective adjacent wells is less than 3, the preset influence distance will be used. Expand to 1.5 times the original value, and based on the expanded value. Re-execute the adjacent correlation coefficient calculation and adjacent well screening in steps S33 and S34 to update the adjacent well set corresponding to the i-th monitoring well; It should be noted that the preset influence distance will be... The expansion to 1.5 times the original value is intended to appropriately broaden the candidate range when the number of adjacent wells is insufficient, and to recalculate the adjacent correlation coefficient to update the adjacent well set. The expansion should not be too large to avoid introducing monitoring wells from different hydrogeological units into the adjacent well set. If the expansion still fails to obtain the required number of adjacent wells, it is marked as insufficient adjacent well benchmarks, and the weight of the computable score is normalized in the subsequent monitoring well reliability score.

[0066] If the number of valid adjacent wells is still less than 3 after re-screening, or If the i-th monitoring well is marked as an adjacent well with insufficient benchmark, the adjacent well event response benchmark envelope, response synchronization score and adjacent well consistency score are not calculated, but the data integrity score, abnormal mutation frequency score and physical rationality score are still calculated.

[0067] S43. Construct the benchmark envelope for event responses in adjacent wells; Based on event response characteristics , For any one of the response start time, peak time, response amplitude, or recovery time, calculate the median of this characteristic in the set of adjacent wells. and interquartile range The benchmark envelope for adjacent well event responses is constructed according to the following calculation formula. : ; in, Event response characteristics The corresponding adjacent well event response baseline envelope; Event response characteristics in adjacent well sets the median; Event response characteristics in adjacent well sets Interquartile range; This is the envelope relaxation factor.

[0068] Preferred, The value range is 1.5 to 3.0, where, Suitable for sites with high monitoring well density and relatively uniform aquifer; Suitable for sites with strong heterogeneity and large differences in response between adjacent wells.

[0069] It should be noted that this embodiment uses the median and interquartile range to reduce the impact of individual abnormal adjacent wells on the reference range.

[0070] S5. Calculate the data integrity score and the abnormal mutation frequency score; This step is used to evaluate whether the monitoring well data has the basic quality to participate in the reliability assessment. It is mainly used to identify data missing, data frozen, sensor offline, and isolated mutations in a undisturbed environment.

[0071] S51. Calculate the data integrity score; In the current evaluation time window Within, calculate the data missing rate of the i-th monitoring well. Proportion of consecutive repeating values Offline ratio of sensors ; The data missing rate The ratio of the number of missing sampling points to the number of sampling points that should be sampled; the proportion of consecutive repeated values. The ratio of the number of consecutive repeating value sampling points to the total number of sampling points required; the sensor offline ratio. This is the ratio of the number of offline sampling points to the number of sampling points that should be sampled.

[0072] It should be noted that the current evaluation time window The sampling points within the preset sampling interval refer to the sampling points within the preset sampling interval. In the current evaluation time window The sampling location should be formed within the time frame; the number of sampling points should be based on the current evaluation time window. Duration and preset sampling interval Sure.

[0073] Preferably, the current evaluation time window in this embodiment The value ranges from 7d to 30d. Values ​​less than 7d may not cover typical disturbance events, while values ​​more than 30d will reduce the sensitivity to recent anomalies.

[0074] The formula for calculating the data integrity score is: ; in, The data integrity score for the i-th monitoring well is given, with a value ranging from 0 to 100. , , All are proportional values, ranging from 0 to 1; The baseline value for the missing rate is 0.15, with a preferred value of 0.15. The preferred value for the proportion of consecutive repeating values ​​is 0.20. The offline proportional reference value for the sensor is preferably 0.10; if the calculated result is less than 0, then take... .

[0075] It should be noted that when the missing value rate exceeds 15%, the stability of event response identification decreases significantly, so the missing value rate is given a high weight in the data integrity score; when the proportion of consecutive duplicate values ​​exceeds 20%, there is a risk of data freezing or sensor malfunction, so it is given the second highest weight; when the offline proportion exceeds 10%, the stability of the acquisition link is insufficient, but the offline status can usually be directly identified by the sensor operating status data, so it is given a relatively low weight; the sum of the above weights is 1, so that the data integrity score is kept within the score range of 0 to 100.

[0076] S52. Calculate the abnormal mutation frequency score; Read the current evaluation time window The time series of groundwater level elevation, water temperature and conductivity of the i-th monitoring well are collected, and the rate of change of adjacent samples is calculated only for the sampling segment where there are valid sample values ​​at two adjacent sampling times.

[0077] For the groundwater level elevation time series, the processor calculates the rate of change of groundwater level elevation according to the following formula: ;in, Let i be the rate of change of groundwater level elevation at time t for monitoring well i, in m / h. Let be the groundwater level elevation of the i-th monitoring well at time t, in meters; The preset sampling interval is expressed in hours (h).

[0078] For a water temperature time series, the rate of change of water temperature is calculated using the following formula: ;in, Let be the rate of change of water temperature in the i-th monitoring well at time t, in °C / h; Let be the water temperature of the i-th monitoring well at time t, in °C.

[0079] For a conductivity time series, the processor calculates the rate of change of conductivity using the following formula: ;in, Let be the rate of change of conductivity of the i-th monitoring well at time t, in units of . ; Let be the conductivity of the i-th monitoring well at time t, in units of . .

[0080] The rate of change of groundwater level elevation, rate of change of water temperature, and rate of change of electrical conductivity are compared with the corresponding thresholds. When any type of rate of change exceeds the corresponding threshold, and there are no events such as rainfall, pumping, drainage, recharge, or sudden changes in river water level at that moment, and the monitoring data of this type does not show a trend of change in the same direction within 2 to 3 consecutive sampling points, an isolated mutation marker for the parameter is generated for this type of monitoring data.

[0081] If at least one type of monitoring data, including groundwater level elevation, water temperature, and electrical conductivity, generates an isolated mutation marker for a parameter at the same sampling time, then that sampling time is marked as an isolated mutation point. If multiple types of isolated mutation markers exist at the same sampling time, they are still counted as a single isolated mutation point. This is to avoid multiple parameter anomalies at the same time being counted repeatedly.

[0082] The formula for calculating the abnormal mutation frequency score is as follows: ;in, The abnormal mutation frequency score for the i-th monitoring well is given, with a value ranging from 0 to 100. This is the ratio of the number of isolated mutation points to the number of valid sampling points. The baseline value for the isolated mutation ratio is preferably 0.03; the effective sampling time refers to the time when at least one of the monitoring data (groundwater level elevation, water temperature, conductivity) has a valid sampling value, and the sensor operating status data does not represent the sampling time when the entire period is offline; if the calculation result is less than 0, then take... .

[0083] The threshold values ​​for water level change rate, water temperature change rate, and conductivity change rate are preferably determined based on actual field test data. When actual field test data is lacking, historical monitoring data of no less than 30 days prior to the current evaluation time window are used to calculate the 95th percentile values ​​of the groundwater level change rate, water temperature change rate, and conductivity change rate, respectively, and three times the corresponding 95th percentile value is used as the corresponding physical threshold value.

[0084] It should be noted that the baseline value for the proportion of isolated mutations The preferred value is 0.03 because continuous monitoring data allows for a small amount of occasional noise, but when isolated mutation points exceed 3% of the number of effective sampling times, they can usually no longer be considered as occasional measurement errors. If there is less than 30 days of historical monitoring data before the current evaluation time window, the processor uses the existing historical monitoring data to calculate the corresponding rate of change quantile value and marks the rate of change threshold as a temporary threshold. It will be updated again after the historical monitoring data reaches 30 days.

[0085] S6. Calculate the response synchronicity score, adjacent well consistency score, and physical rationality score; After obtaining the data integrity score and the abnormal mutation frequency score, the adjacent well relationship table, disturbance event records, event response characteristics and time-aligned monitoring well time series are further used to evaluate whether the hydrological response of the i-th monitoring well is consistent with the response pattern of the adjacent monitoring wells, and whether the data of the monitoring well conforms to the constraints of groundwater movement and well structure.

[0086] S61. Calculate the response synchronicity score; Determine whether there are any valid disturbance events within the current evaluation time window; the valid disturbance events include rainfall events, pumping events, drainage events, recharge events, or river water level change events.

[0087] If there are no valid disturbance events within the current evaluation time window, a disturbance event insufficiency flag is generated, and the response synchronicity score for the i-th monitoring well is not calculated. After generating a disturbance event insufficiency marker, the current evaluation time window is extended forward by one evaluation time window, and the total extended duration is preferably no more than 60 days. If there are still no valid disturbance events after the extension, the weights of the remaining calculable scores are normalized according to step S64. The disturbance event insufficiency marker does not directly indicate an anomaly in the monitoring well, but only indicates that there is a lack of external disturbance evidence for evaluating the synchronicity of hydrological response within the current time range.

[0088] In the presence of a valid disturbance event, the event response characteristics of the i-th monitoring well to the k-th disturbance event are read, including the response start time. Peak time Response amplitude and recovery time Simultaneously, based on the adjacent well relationship table, the weighted event response characteristics of adjacent monitoring wells to the same disturbance event are calculated, including the start time of the weighted response of adjacent wells. Weighted peak times of adjacent wells Weighted response amplitude of adjacent wells Weighted recovery time of adjacent wells .

[0089] The weighted event response characteristics of adjacent wells are obtained by weighting the adjacent correlation coefficients. When the number of adjacent monitoring wells participating in the calculation is insufficient, or the sum of the adjacent correlation coefficients is 0, the response synchronicity sub-score corresponding to the disturbance event is not calculated, and an insufficient adjacent well benchmark is generated.

[0090] When there are sufficient references from adjacent wells, the formula for calculating the response synchronicity score is: ; in, The response synchronization score of the i-th monitoring well to the k-th disturbance event is given, with a value ranging from 0 to 1. The response start time is in hours (h). The starting time of the weighted response of adjacent wells, in hours; The peak time is expressed in hours (h). The weighted peak time for adjacent wells, in hours; The response amplitude is expressed in meters (m). The weighted response amplitude of adjacent wells, in meters; Recovery time is measured in hours (h). Weighted recovery time for adjacent wells, in hours; The time scale parameter is in hours, and the preferred value is... Up to 24 hours; This is the resolution of the water level sensor, measured in meters (m), used to avoid the inability to calculate the amplitude ratio when the response amplitude is close to 0.

[0091] If any event response feature of the i-th monitoring well falls outside the corresponding adjacent well event response baseline envelope, an envelope boundary crossing marker is recorded. The envelope boundary crossing marker is not directly equivalent to distortion anomaly, but is used to distinguish distortion anomaly monitoring wells from real disturbance candidate monitoring wells.

[0092] When multiple disturbance events exist within the current evaluation time window, the synchronicity of the multiple responses is scored according to the weight of the disturbance events. Perform a weighted average and multiply by 100 to obtain the response synchronicity score. .

[0093] The formula for calculating the weight of the k-th perturbation event is: in, Let be the perturbation event weight for the k-th perturbation event; This is the normalized value of the intensity of the k-th disturbance event, ranging from 0 to 1; The effective sampling ratio within the window of the k-th disturbance event, with a value ranging from 0 to 1; This represents the sum of the normalized intensity values ​​of all valid disturbance events within the current evaluation time window and the effective sampling ratio; if there is only one valid disturbance event within the current evaluation time window, then the weight of that disturbance event is 1.

[0094] S62. Calculate the consistency score of adjacent wells; Using the groundwater level elevations of adjacent monitoring wells (excluding the i-th monitoring well) in the adjacent well relationship table, a local groundwater level plane is fitted at each sampling time; the local groundwater level plane is used to represent the spatial change trend of the local groundwater level formed by adjacent monitoring wells around the i-th monitoring well.

[0095] After obtaining the local groundwater level, the plane coordinates of the i-th monitoring well are substituted into the local groundwater level to obtain the predicted groundwater level elevation at the location of the i-th monitoring well. And will predict the groundwater level elevation. The measured groundwater level elevation of the i-th monitoring well By comparison, the residual of the target monitoring well is obtained: ; in, The residual of the target monitoring well at time t for the i-th monitoring well is expressed in meters. Let be the measured groundwater level elevation of the i-th monitoring well at time t, in meters; The predicted groundwater level elevation is obtained by fitting data from adjacent monitoring wells, in meters.

[0096] The formula for calculating the adjacent well consistency score is as follows: ; in, The consistency score between adjacent wells for the i-th monitoring well is given, with a value ranging from 0 to 100; This represents the 90th percentile of the absolute value of the residuals of the target monitoring wells within the current evaluation time window, in meters. The robust standard deviation of the fitted residuals of adjacent monitoring wells is expressed in meters. This represents the resolution of the water level sensor, measured in meters (m).

[0097] Robust standard deviation of the fitting residuals of adjacent monitoring wells The MAD method is preferred for calculation. Specifically, cross-validation fitting is performed on adjacent monitoring wells. One adjacent monitoring well is used as a validation well in turn, and the local groundwater level is fitted using the remaining adjacent monitoring wells. The fitting residuals of the validation wells are then calculated to obtain the set of cross-validation fitting residuals of adjacent monitoring wells. Then calculate according to the following formula. : ; in, The fitting residuals for cross-validation of adjacent monitoring wells are expressed in meters. This indicates median calculation; if there are fewer than 3 adjacent monitoring wells, or if the spatial positions of adjacent monitoring wells are collinear, causing the local groundwater level to be unstable and unable to fit, then the neighboring well consistency score is not calculated. And generate a missing rating marker.

[0098] S63. Calculate the physical rationality score; A physical rationality score is generated based on water temperature, conductivity, hydraulic gradient direction changes, disturbance event response, and well structure constraints. The physical rationality score The initial value is 100, and the following fixed preferred deduction values ​​are applied. If multiple deduction items are triggered simultaneously in the same monitoring well, the deductions are accumulated, but the minimum value is 0.

[0099] If the water temperature changes by less than twice the resolution of the water temperature sensor within 7 to 30 consecutive days, and there are water temperature changes in adjacent monitoring wells, deduct 15 points.

[0100] When the conductivity shows a single point jump, and the two or three consecutive sampling points do not show a continuous trend, 15 points will be deducted.

[0101] If the angle between the directions of the local hydraulic gradient exceeds the direction difference threshold before and after the addition of the i-th monitoring well, and this difference exists within two or more consecutive evaluation time windows, 20 points will be deducted. The preferred direction difference threshold is 30° to 60°, 30° to 45° when the monitoring well density is high, and 45° to 60° when the monitoring well density is low.

[0102] When there is a clear pumping, drainage, recharge, or heavy rainfall event, and the i-th monitoring well does not respond at all within a reasonable lag time, the processor deducts 20 points; the reasonable lag time is preferably 1 hour to 7 days; for shallow unconfined aquifers, it is preferably 1 hour to 24 hours; for confined aquifers or low-permeability aquifers, it is preferably 1 day to 7 days.

[0103] When the filter pipe layer is inconsistent with the target aquifer, or when the filter pipe overlap coefficient is consistently below 0.2 and the adjacent well consistency score is consistently below 50, the processor will deduct 25 points.

[0104] It should be noted that the above deduction values ​​are set according to the degree of impact of anomalies on the reconstruction results of the groundwater seepage field; no fluctuation in water temperature and single-point jump in conductivity mainly reflect sensor jamming or instantaneous noise, and have a relatively small direct impact on the reconstruction of the groundwater level, so a deduction of 15 points is preferred; abnormal local hydraulic gradient direction or no response to a clear disturbance event will directly affect the judgment of groundwater flow direction, so a deduction of 20 points is preferred; inconsistency between the filter pipe layer and the target aquifer will lead to the wrong aquifer object represented by the monitoring well, which has a greater impact on the reconstruction of the seepage field, so a deduction of 25 points is preferred.

[0105] The heavy rainfall event refers to a rainfall event in which the cumulative rainfall exceeds the 75th percentile of the historical single rainfall in the target area, or a rainfall event in which the cumulative rainfall is not less than 25 mm; the terms "long-term below 0.2" and "consistently below 50" mean that the corresponding conditions are met in two or more consecutive evaluation time windows.

[0106] For abnormal responses related to disturbance events, multi-parameter synchronous change markers and time correspondence markers are further recorded as the basis for subsequent identification of candidate monitoring wells for actual disturbances. The multi-parameter synchronous change markers refer to the starting time difference of changes in at least two of the following data: groundwater level elevation, water temperature, and conductivity, not exceeding a synchronization time threshold. The synchronization time threshold is preferably set to... Up to 24 hours, when the preset sampling interval is short and the target area is significantly affected by the engineering drainage, it is preferable to take... The time correspondence marker refers to the response start time falling within the disturbance event window. The time correspondence marker refers to the time when the response start time falls within the disturbance event window.

[0107] If at least two types of data, including groundwater level elevation, water temperature, and electrical conductivity, of the i-th monitoring well change synchronously within the same disturbance event window, and this change has a temporal correspondence with the disturbance event record, then the monitoring well will not be directly judged as having low physical rationality simply because it deviates from the response pattern of adjacent monitoring wells. Instead, it will be retained as a candidate monitoring well for a real disturbance.

[0108] S64. Handling missing scoring boundary conditions; When the i-th monitoring well experiences a response synchronicity score due to insufficient benchmarks from adjacent wells... Consistency score with adjacent wells The response synchronicity score cannot be calculated, or there are no valid disturbance events within the current evaluation time window. If a score cannot be calculated, the weights of the remaining calculable scores are normalized to generate a temporary monitoring well reliability score.

[0109] If more than two scores cannot be calculated, the reliability score of the monitoring well will be limited to no more than 60. This process is used to avoid giving an excessively high reliability score when there is insufficient basis for key comparisons, and also to avoid misjudging the monitoring well as abnormal due to the absence of disturbance events in the short term.

[0110] S7. Generate a reliability score for the monitoring well and identify distorted, abnormal monitoring wells and candidate monitoring wells with genuine disturbances; like Figure 2 As shown, this step combines the above scores into a monitoring well reliability score and further distinguishes the sources of anomalies.

[0111] S71. Calculate the reliability score of the monitoring well; The formula for calculating the reliability score of a monitoring well is: ; in, The reliability score of the i-th monitoring well is given, with a value ranging from 0 to 100. Score the data integrity. Scoring based on the frequency of abnormal mutations; In response to synchronization scoring; Scoring for consistency between adjacent wells; The physical plausibility is scored; the sum of the weights is 1.

[0112] When a certain score cannot be calculated due to insufficient benchmarks from adjacent wells, the remaining score weights are normalized according to step S64; if the number of valid sampling points in the current evaluation time window is less than 30% of the total number of sampling points, the monitoring well reliability score is limited to no more than 40; if the sensor operating status data indicates that the system is offline for all time periods, the monitoring well reliability score is set to 0.

[0113] S72. Determine the basic state according to the reliability threshold; Determine whether the i-th monitoring well meets the anomaly source triggering condition; Specifically, the anomaly source triggering conditions include: data integrity score. Abnormal mutation frequency score Sensor offline ratio Consistency score of adjacent wells Response synchronicity score The following are possible issues: presence of envelope out-of-bounds markers, presence of unrecovered markers, or physical plausibility score. ; When the i-th monitoring well meets any of the abnormal source triggering conditions, it is sent to the abnormal source identification process in subsequent steps S73~S75; when the i-th monitoring well does not have any abnormal source triggering conditions, the basic state is determined according to the monitoring well reliability score.

[0114] Set the first reliability threshold. Second reliability threshold ; Specifically, the first reliability threshold is used to screen monitoring wells that can normally participate in seepage field reconstruction, and is therefore set in a higher scoring range; the second reliability threshold is used to distinguish between monitoring wells with significantly low reliability and those that can be downgraded, and is therefore set in a medium-low scoring range; in this embodiment, the first reliability threshold... The preferred value range is 75~85, with 75 used as an example; the second reliability threshold The preferred value range is 40 to 55, and the example value is 45.

[0115] When there is no abnormal source triggering condition and When the i-th monitoring well is marked as a reliable monitoring well, and when there is no anomaly triggering condition, the i-th monitoring well is marked as a reliable monitoring well; When the i-th monitoring well is marked as a downweighted monitoring well; when there is no abnormal source triggering condition and When the i-th monitoring well is marked as a monitoring well to be verified on-site.

[0116] S73, Identify and monitor distorted anomalies in monitoring wells; After the anomaly source identification is triggered in the i-th monitoring well, the data integrity score, anomaly mutation frequency score, sensor offline ratio, water temperature time series, conductivity time series, groundwater level elevation time series, filter pipe layer, target aquifer and adjacent well consistency score of the monitoring well are read, and it is determined whether the monitoring well has data distortion caused by acquisition link, sensor status, well structure or monitoring well response capability.

[0117] For each monitoring well of the i-th type, a distortion evidence marker is generated; the distortion evidence marker includes: A low data integrity flag indicates a data integrity score below 50. Frequent abnormal mutation markers are used to indicate an abnormal mutation frequency score below 50; The offline ratio anomaly flag indicates that the sensor offline ratio corresponding to the sensor's operating status data exceeds 10%. The water temperature and conductivity non-response marker is used to indicate that the change in water temperature or conductivity within 7 to 30 days is less than twice the resolution of the corresponding sensor, and there is a change in water temperature or conductivity in adjacent monitoring wells. The step mutation non-recovery marker is used to indicate that after a step mutation occurs in the groundwater level elevation, it does not return to the recovery judgment range determined by the pre-event benchmark groundwater level elevation and the water level response threshold within 2 to 3 consecutive sampling points, and the step mutation does not correspond to any disturbance event record. The well structure mismatch marker indicates that the filter pipe layer is inconsistent with the target aquifer, and the consistency score of adjacent wells is below 50 in more than two consecutive evaluation time windows.

[0118] If the i-th monitoring well meets at least two of the above distortion evidence markers and does not meet the identification conditions for a real disturbance candidate monitoring well, the i-th monitoring well will be marked as a distortion anomaly monitoring well. If the i-th monitoring well meets both the distortion evidence marker and the identification conditions for a real disturbance candidate monitoring well, it will be marked as a monitoring well to be verified on-site, so as to avoid directly eliminating real local seepage anomalies as distorted data.

[0119] Anomaly types are generated based on triggered distorted evidence markers. When the groundwater level elevation response amplitude is lower than that of adjacent monitoring wells, and there are also no response markers for water temperature and conductivity, the anomaly type is marked as siltation risk. When groundwater level elevation, water temperature, or conductivity shows continuous unidirectional drift, and there is no unidirectional change in adjacent monitoring wells, the anomaly type is marked as sensor drift risk. When there is a well structure mismatch marker, the anomaly type is marked as structural anomaly risk. When there is an offline ratio anomaly marker or a low data integrity marker, the anomaly type is marked as acquisition link anomaly risk. When there are frequent anomaly mutation markers and the mutation does not correspond to any disturbance event record, the anomaly type is marked as local human disturbance or acquisition noise risk. The anomaly types are used to generate a review instruction table later.

[0120] It should be noted that the scoring threshold of 50 is used to indicate that the corresponding single item score is less than half of the full score, indicating that the data quality or hydrological response reliability of that item is insufficient; the offline ratio of 10% is consistent with the sensor offline ratio benchmark value in the data integrity score; 7 consecutive days to 30 days are used to identify long-term monitoring without response, and 2 to 3 consecutive sampling points are used to identify whether short-term step anomalies can be recovered; the above thresholds and time scales correspond to different identification objects of long-term anomalies, short-term anomalies and acquisition link anomalies, respectively.

[0121] S74. Identify candidate monitoring wells for real disturbances; When the neighbor consistency score of the i-th monitoring well is less than 50, the response synchronicity score is less than 50, or there is an envelope out-of-bounds marker, but at least two of the following conditions are met, it is marked as a candidate monitoring well for real disturbance: A1. There is a temporal correlation between groundwater level response and changes in rainfall, pumping, drainage, recharge, or river water level. A2. At least two types of data, including groundwater level, water temperature, and electrical conductivity, change synchronously within the same disturbance event window; A3. The i-th monitoring well is located in a hydrogeologically sensitive area; A4. The abnormal response occurs repeatedly within two or more consecutive evaluation time windows.

[0122] The hydrogeologically sensitive areas are derived from at least one of the following: geological survey data, aquifer distribution map, fault structure map, riverbank data, pumping and drainage project location data, or recharge project location data. Specifically, they include riverbank zones, pumping and drainage affected areas, recharge affected areas, fault zones, sand interlayers, karst development zones, or areas with highly permeable channels.

[0123] If the i-th monitoring well simultaneously meets the conditions for both distortion and anomaly monitoring wells and actual disturbance candidate monitoring wells, it is prioritized as a monitoring well to be verified on-site and is not directly removed as a distortion and anomaly monitoring well within the current evaluation time window; whether to treat it as a local anomaly constraint point is determined according to the local anomaly constraint weight in subsequent step S85.

[0124] S75. Identify the monitoring wells to be verified on-site; If the i-th monitoring well enters the anomaly source identification process, but does not meet the conditions for both distorted anomaly monitoring wells and real disturbance candidate monitoring wells, it is marked as a monitoring well to be verified on-site. The monitoring well to be verified on-site is treated as a downgraded monitoring well within the current evaluation time window and written into the review instruction table.

[0125] S8. Generate reliability correction coefficients based on the monitoring well reliability score and anomaly source identification results, and reconstruct the self-correcting groundwater level. like Figure 4 As shown, after completing the reliability evaluation and anomaly source identification of the monitoring wells, the monitoring well reliability score is converted into a reliability correction coefficient in the interpolation calculation. The weights of the monitoring wells participating in the interpolation are then corrected based on the identification results of distorted anomaly monitoring wells, candidate monitoring wells with real disturbances, and monitoring wells awaiting on-site verification. The reliability correction coefficient refers to the interpolation correction coefficient obtained from the monitoring well reliability score; the self-correcting groundwater level refers to the spatial distribution of the groundwater level reconstructed based on the reliability correction coefficient and the corrected interpolation weights. This embodiment preferably uses an inverse distance-based spatial weight to avoid the problem of negative Kriging weights potentially amplifying the negative impact on low-reliability monitoring wells.

[0126] S81, Generate reliability correction coefficients; Based on the monitoring well reliability score of the i-th monitoring well First reliability threshold Second reliability threshold Generate the reliability correction coefficient for the i-th monitoring well. ;in, , It is a dimensionless number, with a value range from 0 to 1.

[0127] when hour, ; when hour, ; when hour, .

[0128] Among them, the first reliability threshold The second reliability threshold is used to distinguish monitoring wells that can normally participate in seepage field reconstruction. Used to distinguish between monitoring wells with obviously low reliability and monitoring wells that can be downgraded for use.

[0129] After obtaining the reliability correction coefficient, the reliability correction coefficient or interpolation participation status is corrected according to the anomaly source identification results: when the i-th monitoring well is marked as a distorted anomaly monitoring well, it is set to not participate in regional interpolation; when the i-th monitoring well is marked as a monitoring well to be verified on-site, its reliability correction coefficient is limited to no more than 0.2; when the i-th monitoring well is marked as a real disturbance candidate monitoring well, it is set to not participate in regional smooth interpolation, and the local anomaly constraint weight is recorded separately for subsequent identification of real anomaly candidate areas.

[0130] S82. Calculate the inverse distance fundamental space weights; At spatial location x and time t, the inverse distance basic spatial weight of the i-th monitoring well is calculated according to the following formula: ; in, The inverse distance basic spatial weight of the i-th monitoring well at spatial location x; is the planar distance between the i-th monitoring well and spatial location x, in meters; To prevent a smooth distance of zero, a distance of 1m to 5m is preferred; p is the distance attenuation index, preferably 1 to 3; this range is used to reflect the spatial relationship of the weakening of the influence of groundwater monitoring wells as the distance increases, and to avoid excessive concentration of weights on near-well points.

[0131] S83. Calculate the corrected interpolation weights; Multiply the inverse distance base spatial weights by the reliability correction coefficients and normalize them in the monitoring wells involved in the interpolation to obtain the corrected interpolation weights: ; in, The corrected interpolation weight for the i-th monitoring well at spatial location x and time t is dimensionless; Let be the reliability correction factor for the i-th monitoring well; Let be the inverse distance basic spatial weight of the i-th monitoring well; This represents the sum of the product of the reliability correction coefficients of all monitoring wells participating in the interpolation and the inverse distance basic spatial weights.

[0132] when If the value is 0, or if there are fewer than 3 valid monitoring wells around spatial location x participating in the interpolation, the self-correcting groundwater level elevation for that spatial location will not be generated, and an insufficient interpolation confidence marker will be output.

[0133] S84. Reconstruct the self-correcting groundwater level; Calculate the self-correcting groundwater level using the following formula: ; in, The self-corrected groundwater level elevation at spatial location x at time t, in meters; The corrected interpolation weight for the i-th monitoring well at spatial location x and time t is dimensionless; Let be the groundwater level elevation of the i-th monitoring well at time t, in meters. Since the interpolation weight after correction is a dimensionless number, the groundwater level elevation is in meters. Therefore, the unit of the self-corrected groundwater level elevation is still meters.

[0134] S85. Generate local anomaly constraint weights; For candidate monitoring wells with real disturbances, the local anomaly constraint weights are generated according to the following formula: ; in, The local anomaly constraint weight for the i-th real disturbance candidate monitoring well, with a value ranging from 0 to 1; The marker for disturbance events is set to 1 when there is a time correlation between the groundwater level elevation response and changes in rainfall, pumping, drainage, recharge, or river water level; otherwise, it is set to 0. For multi-parameter synchronization, the value is 1 when at least two of the following data—groundwater level elevation, water temperature, and conductivity—change synchronously within the same disturbance event window; otherwise, it is 0. This is used to mark hydrogeologically sensitive areas. The value is 1 when the monitoring well is located in a hydrogeologically sensitive area, and 0 otherwise. The value is set to 1 if the abnormal response occurs repeatedly within two or more consecutive evaluation time windows, and 0 otherwise.

[0135] when When, the i-th real disturbance candidate monitoring well is used as a local anomaly constraint point; when When the time comes, it will be transferred to the monitoring well awaiting on-site verification; the threshold of 0.60 is used to require that the marker corresponding to the disturbance event and the multi-parameter synchronous marker exist at the same time, or that at least three types of markers jointly support the judgment of local anomalies, so as to avoid a single abnormal signal directly forming a real anomaly candidate area.

[0136] S9. Generate a self-correcting groundwater seepage field based on the self-correcting groundwater level and output a verification instruction table; like Figure 4As shown, after obtaining the self-corrected groundwater level and local anomaly constraint points, the groundwater flow direction, hydraulic gradient, and seepage intensity level are further calculated, and a verification instruction table is generated by combining the real anomaly candidate area and anomaly monitoring well markers; the self-corrected groundwater seepage field refers to the distribution of groundwater flow direction, hydraulic gradient, and seepage intensity based on the self-corrected groundwater level.

[0137] S91. Calculate the self-correcting groundwater seepage field; Calculate the hydraulic gradient based on the self-calibrated groundwater level: ; in, Let x be the hydraulic gradient at spatial location x at time t, which is dimensionless; To self-correct the spatial gradient of the groundwater level.

[0138] The direction of groundwater flow is determined by the negative gradient direction of the self-correcting groundwater level. The direction of groundwater flow is perpendicular to the groundwater level contour lines and points in the direction of decreasing groundwater level elevation.

[0139] The intensity of seepage is classified according to the magnitude of the hydraulic gradient. When no historical data is available, a hydraulic gradient less than 0.001 is marked as a weak seepage zone, a hydraulic gradient between 0.001 and 0.005 is marked as a moderate seepage zone, and a hydraulic gradient greater than 0.005 is marked as a strong seepage zone. If the target area has historical hydraulic gradient data of no less than 30 days, the processor calculates the 33rd and 66th percentile values ​​of the historical hydraulic gradient. Areas below the 33rd percentile value are marked as weak seepage zones, areas between the 33rd and 66th percentile values ​​are marked as moderate seepage zones, and areas above the 66th percentile value are marked as strong seepage zones. Alternatively, the average of the above initial threshold and the historical quantile threshold can be used as the correction threshold.

[0140] Given the hydraulic conductivity coefficient K and effective porosity Then, the average linear velocity of groundwater was further calculated: ; in, V(x,t) is the average linear velocity of groundwater; K is the hydraulic conductivity coefficient. When K is taken as m / d, the unit of v(x,t) is m / d. Effective porosity, dimensionless.

[0141] S92. Generate real anomaly candidate regions; Based on the local anomaly constraint points, generate real anomaly candidate regions; for the i-th local anomaly constraint point, establish a candidate influence radius centered on that monitoring well. : ; in, Let be the candidate influence radius of the i-th local anomaly constraint point, in meters; The preset influence distance is in meters. The distance from the i-th local anomaly constraint point to the nearest valid adjacent monitoring well is in meters; this candidate influence radius is used to limit the spatial range of the real anomaly candidate area, so as to avoid the infinite extrapolation of single-well anomalies.

[0142] Spatial location x is marked as a true anomaly candidate region when it simultaneously meets the following conditions: B1. Spatial location x is located within the candidate influence radius of any local anomaly constraint point; B2. Local anomaly constraint weights corresponding to local anomaly constraint points ; B3. There is at least one reliable monitoring well or a deweighted monitoring well in the vicinity of spatial location x, and the changes in groundwater level, water temperature or conductivity of the reliable monitoring well or the deweighted monitoring well are synchronous with the local anomaly constraint point within the same disturbance event window.

[0143] The real anomaly candidate area refers to a localized abnormal seepage area supported by real disturbance candidate monitoring wells and surrounding effective monitoring wells; the real anomaly candidate area is not directly equivalent to the final hydrogeological anomaly conclusion, but serves as a key area for subsequent on-site verification and intensified monitoring.

[0144] S93. Generate self-calibration analysis results; Based on the self-calibrated groundwater level, groundwater flow direction, hydraulic gradient, seepage intensity level, interpolation confidence insufficient marker, and real anomaly candidate area marker, a self-calibration analysis result is generated; the self-calibration analysis result includes at least one of spatial raster data, vector layer data, monitoring well attribute table, and visualization display result.

[0145] The spatial raster data includes self-corrected groundwater level elevation, hydraulic gradient, and groundwater flow direction at each spatial location; the vector layer data includes the boundaries of real anomaly candidate areas, the boundaries of seepage intensity zones, and the boundaries of areas with insufficient interpolation confidence; the monitoring well attribute table includes monitoring well reliability score, anomaly source identification results, reliability correction coefficient, and verification suggestion fields.

[0146] S94. Generate a review instruction table and form a closed-loop update; A review instruction table is generated based on the markers of distorted anomaly monitoring wells, real disturbance candidate monitoring wells, monitoring wells awaiting on-site verification, and real anomaly candidate areas. The review instruction table includes the monitoring well number, anomaly type, anomaly basis, suggested review method, review priority, and review period.

[0147] For the risk of siltation, the recommended verification methods include well washing, pumping recovery tests, and water level response retesting; for the risk of sensor drift, the recommended verification methods include sensor calibration, manual water level retesting, and sensor replacement; for the risk of structural anomalies, the recommended verification methods include well pipe integrity checks, filter pipe layer verification, and well depth retesting; for the risk of data acquisition link anomalies, the recommended verification methods include data acquisition terminal checks, communication link checks, and power supply status checks; for candidate monitoring wells with real disturbances or candidate areas with real anomalies, the recommended verification methods include surrounding pumping, drainage, or recharge surveys, verification of river replenishment and drainage relationships, increased sampling, and synchronous monitoring of adjacent monitoring wells.

[0148] The verification data includes at least one of the following: artificial water level remeasurement value, sensor calibration results, post-well flushing response data, well structure verification results, surrounding engineering disturbance investigation results, and encrypted sampling data. After the verification data is returned, it is written into the standardized monitoring well dataset for the next evaluation time window, and steps S2 to S9 are re-executed.

[0149] When a certain distortion anomaly monitoring well reaches the first reliability threshold within two or more consecutive evaluation time windows... When the processor restores the monitoring well to its normal weight, it will do so if the weight remains below the second reliability threshold for an extended period. At that time, it will continue to be treated as a distortion anomaly monitoring well.

[0150] Through the above steps, this method generates a monitoring well reliability score based on data integrity, abnormal mutation frequency, response synchronization, neighbor well consistency, and physical rationality. The monitoring well reliability score is then converted into a reliability correction coefficient for use in the calculation of self-correcting groundwater level and self-correcting groundwater seepage field.

[0151] Meanwhile, this method further distinguishes between distorted anomaly monitoring wells and genuine disturbance candidate monitoring wells. For distorted anomaly monitoring wells caused by sensor drift, well pipe blockage, filter pipe layer mismatch, or abnormal acquisition link, their weight in seepage field reconstruction is reduced or canceled. For genuine disturbance candidate monitoring wells that have a time correspondence with changes in rainfall, pumping, drainage, recharge, or river water level, and whose groundwater level elevation, water temperature, and conductivity change synchronously, they are retained as local anomaly constraint points. Thus, the interference of anomaly well points on the judgment of groundwater flow direction, hydraulic gradient, and seepage strength can be reduced, and the situation where genuine local seepage anomalies are directly rejected as distorted data can be reduced.

[0152] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for dynamic monitoring and hydrogeological analysis of groundwater seepage field, characterized in that, Includes the following steps: Acquire basic information of monitoring wells, continuous monitoring data, external disturbance data and hydrogeological background data, convert water levels into groundwater elevation and form a standardized monitoring well dataset; The standardized monitoring well dataset is time-aligned and basic anomaly markers are applied to generate time-aligned monitoring well time series. A table of relationships between adjacent wells is constructed based on the plane distance between monitoring wells, the degree of overlap of filter pipes, and the consistency of aquifers, and a record of disturbance events is established based on external disturbance data; Based on the adjacent well relationship table and disturbance event records, event response features are extracted, and data integrity score, abnormal mutation frequency score, response synchronization score, adjacent well consistency score and physical rationality score are calculated to generate a monitoring well reliability score. Based on the monitoring well reliability score, anomaly source triggering conditions, and the matching relationship between event response characteristics and disturbance event records, distorted anomaly monitoring wells, real disturbance candidate monitoring wells, and monitoring wells to be verified on-site are identified. Based on the reliability score of the monitoring wells, and the identification results of distorted and abnormal monitoring wells, real disturbance candidate monitoring wells, and monitoring wells to be verified on-site, the interpolation weights are corrected, the self-correcting groundwater level is reconstructed, and a self-correcting groundwater seepage field and verification instruction table are generated.

2. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 1, characterized in that: The standardized monitoring well dataset includes monitoring well number, plane coordinates, wellhead elevation, well depth, well diameter, top depth of filter pipe, bottom depth of filter pipe, aquifer type, well completion time, sensor model, sensor resolution, water level time series, water temperature time series, conductivity time series, sensor operating status data, external disturbance data, and hydrogeological background data. The water level time series is converted into a groundwater level elevation time series based on the wellhead elevation and the vertical distance from the water surface to the wellhead. The sensor resolution is used to set the minimum effective response threshold, and the hydrogeological background data is used to determine hydrogeologically sensitive areas.

3. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 2, characterized in that: Generating the time-aligned monitoring well time series includes: resampling the groundwater level elevation time series, water temperature time series, and conductivity time series according to a preset sampling interval; When there are multiple sampled values ​​in the same sampling interval, the median is taken as the representative value; when there are no sampled values ​​in the same sampling interval, the sampling interval is marked as a missing value. Generate missing value markers, consecutive duplicate value markers, isolated mutation value markers, and sensor offline markers; wherein, the isolated mutation value marker is used to represent data points that exceed a physical threshold and do not form a continuous trend of change in the same direction within a time period in which no corresponding disturbance event is recorded.

4. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 3, characterized in that: The steps for constructing the adjacent well relationship table include: first determining whether the bottom depth of the filter pipe is greater than the top depth of the filter pipe; When the depth of the filter pipe is effective, the adjacent correlation coefficient is calculated based on the plane distance of the monitoring wells, the degree of overlap of the filter pipes, and the consistency of the aquifer. Among them, the aquifer consistency is used as a control term for the adjacent correlation coefficient, so that the monitoring wells corresponding to unconnected aquifers do not enter the adjacent well set; Among the candidate monitoring wells that meet the aquifer connectivity condition, monitoring wells with an adjacent correlation coefficient greater than the adjacent threshold are selected, or a preset number of candidate monitoring wells are selected according to the adjacent correlation coefficient to form an adjacent well relationship table.

5. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 4, characterized in that: The steps for establishing the disturbance event record include: extracting disturbance events from rainfall data, pumping data, drainage data, recharge data and river water level data, and recording the disturbance event type, disturbance start time, disturbance end time, disturbance intensity, response direction reference and disturbance event window; The disturbance event window includes a pre-event baseline period, an event response period, and an event recovery period; When the reference time period before the event is affected by other disturbances, a continuous time period that is not affected by other disturbances shall be selected as the reference time period; When the response periods of multiple disturbance events overlap, they are marked as composite disturbance events, and the event weights are reduced or the corresponding response synchronicity sub-scores are not calculated.

6. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 5, characterized in that: The steps for extracting event response features include: determining the response amplitude based on the absolute deviation of the groundwater level elevation from the median groundwater level elevation during the event response period; The moment when the absolute deviation reaches the preset number of sampling points consecutively and exceeds the water level response threshold is defined as the response start moment. The moment when the absolute deviation reaches its maximum value is defined as the peak moment; The recovery time is defined as the duration during which the groundwater level returns to the recovery range determined by the pre-event baseline groundwater level and water level response threshold after the peak time, and the time during which the groundwater level remains within the preset number of sampling points. A benchmark envelope for the event response of adjacent wells is constructed based on the median and interquartile range of the event response characteristics in the adjacent well set.

7. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 6, characterized in that: The steps for generating a monitoring well reliability score include: obtaining a data integrity score by taking the proportions of missing sampling points, continuous duplicate sampling points, and offline sampling points to the total number of sampling points to be sampled within the current evaluation time window; The abnormal mutation frequency score is obtained by the proportion of isolated mutation points to valid sampling points; The event response characteristics of the target monitoring well are compared with the weighted event response characteristics of adjacent wells, and combined with the event response benchmark envelope of adjacent wells to obtain the response synchronicity score; The local groundwater level plane is fitted using adjacent monitoring wells other than the target monitoring well, and the consistency score of adjacent wells is obtained based on the residual of the target monitoring well. A physical rationality score is obtained based on water temperature, conductivity, hydraulic gradient direction changes, disturbance event response, and well structure constraints.

8. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 7, characterized in that: When generating a monitoring well reliability score, if the response synchronization score and the neighboring well consistency score cannot be calculated due to insufficient benchmarks of adjacent wells, or if there are no effective disturbance events within the current evaluation time window, resulting in the inability to calculate the response synchronization score, the weights of the remaining calculable scores are normalized. If the number of uncalculable scores exceeds the preset number of missing scores, the reliability score of the monitoring well is limited to not exceeding the first score limit. If the number of valid sampling points within the current evaluation time window is lower than the preset valid sampling ratio, the reliability score of the monitoring well is limited to not exceeding the second score limit. If the sensor's operating status data indicates that the system is offline at all times, the monitoring well reliability score will be set to the lowest possible value.

9. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 8, characterized in that: The steps for identifying distorted anomaly monitoring wells, candidate monitoring wells for real disturbances, and monitoring wells awaiting on-site verification include: first, determining whether anomaly source identification has been triggered based on data integrity score, anomaly mutation frequency score, sensor offline ratio, neighbor well consistency score, response synchronization score, envelope boundary crossing marker, unrecovered marker, and physical rationality score; Once triggered, if the target monitoring well meets at least two of the following conditions: missing data, isolated mutation, offline, long-term no fluctuation, unrecoverable step mutation, or inconsistent filter pipe layer, it will be marked as a distortion anomaly monitoring well. When the abnormal response of a target monitoring well meets at least two of the following characteristics: disturbance event matching, synchronous change of multiple parameters, matching of hydrogeologically sensitive areas, or recurrence, it is marked as a candidate monitoring well for real disturbance. The remaining target monitoring wells that triggered the anomaly source identification were marked as monitoring wells awaiting on-site verification.

10. The method for dynamic monitoring and hydrogeological analysis of groundwater seepage field as described in claim 9, characterized in that: The steps for reconstructing the self-correcting groundwater level and generating the self-correcting groundwater seepage field include: generating a reliability correction coefficient based on the reliability score of the monitoring well; Specifically, when the reliability score of the monitoring well is not lower than the first reliability threshold, the reliability correction coefficient takes the first weight value; when the reliability score of the monitoring well is lower than the first reliability threshold but not lower than the second reliability threshold, the reliability correction coefficient decreases as the reliability score of the monitoring well decreases; when the reliability score of the monitoring well is lower than the second reliability threshold, the reliability correction coefficient takes the second weight value, and the first weight value is greater than the second weight value. Distortion anomaly monitoring wells are set to not participate in regional interpolation, the reliability correction coefficient of monitoring wells to be verified on-site is limited to a preset upper limit, real disturbance candidate monitoring wells are set to not participate in regional smooth interpolation, and local anomaly constraint weights are generated. The reliability correction coefficient is multiplied by the inverse distance basic spatial weight and normalized in the monitoring wells involved in the interpolation to obtain the corrected interpolation weight. The self-correcting groundwater level is then reconstructed based on the corrected interpolation weight. The direction of groundwater flow is determined based on the negative gradient of the self-correcting groundwater level, the intensity of seepage is determined based on the hydraulic gradient, and a real anomaly candidate area and a verification instruction table are generated based on the local anomaly constraint weight.

Images