Rainfall well optimal arrangement method and system based on digital twinning

Abstract

The application provides a kind of based on digital twinning's dewatering well optimization arrangement method and system, method includes: obtaining the geological survey data of target foundation pit dewatering area, foundation pit support boundary data, surrounding underground structure data, historical hydrological data and initial monitoring point arrangement data, establish groundwater seepage digital twin, and generate candidate dewatering well set and candidate monitoring point correlation table;According to candidate dewatering well set, select test well site, execute low-intensity perturbation pumping, collect water level delay, drawdown slope, water level recovery and surrounding structure micro-deformation data, generate perturbation pumping response data;The perturbation pumping response data is input into the groundwater seepage digital twin, response residual distribution map is generated and the probability field of underground recharge leakage is inverted;According to the probability field of underground recharge leakage, form well site repulsion boundary, configure interception well site, buffer well site and master control well site, generate truncated well group candidate scheme.

Description

Technical Field

[0001] This invention relates to the field of digital twin optimization technology, and in particular to a method and system for optimizing the layout of precipitation wells based on digital twins. Background Technology

[0002] Dewatering of foundation pits is a core aspect of underground engineering construction. Conventional designs typically determine the layout of dewatering wells based on known geological survey data and ideal boundary conditions. In typical scenarios with homogeneous strata and continuous water-stopping structures, existing methods are sufficient to meet basic dewatering control requirements.

[0003] However, in complex environments (such as old factory renovation sites or sites near underground structures), there are often hidden recharge channels such as defects in the joints of diaphragm walls, abandoned pipe trenches, or remnants of old wells. These channels are difficult to identify in advance using limited exploration boreholes and conventional pumping tests. Once the dewatering well is started, it can easily induce external groundwater to continuously flow in along these hidden channels, leading to sudden emergencies such as insufficient drawdown in some areas, a surge in pumping volume, and rapid recovery of water levels after pump shutdown.

[0004] Faced with the aforementioned abnormal recharge, existing dewatering optimization methods mostly rely on conventional means such as increasing the density of well locations or increasing the pumping intensity of individual wells. This approach not only fails to cut off the hidden recharge but also generates stronger hydraulic traction, introducing more external groundwater into the vicinity of the foundation pit, amplifying the abnormal recharge path, and thus falling into a vicious cycle of pumping more and more recharge, while also increasing the risk of uncontrolled micro-deformation of the surrounding structure. Therefore, this invention proposes a dewatering well optimization layout method and system based on digital twins. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a method and system for optimizing the layout of precipitation wells based on digital twins, thereby solving the technical problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The method for optimizing the layout of precipitation wells based on digital twins includes the following steps: S1. Obtain geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data of the target foundation pit dewatering area, generate the target foundation pit dewatering basic dataset, divide aquifer units, support boundary units, external recharge boundary units, and structurally sensitive units, and establish a groundwater seepage digital twin, candidate dewatering well location set, and candidate monitoring point association table; S2. Select test well locations based on the candidate precipitation well location set and candidate monitoring point association table, generate a micro-disturbance pumping execution table, execute low-intensity micro-disturbance pumping, collect and attribute water level delay, drawdown slope, water level recovery and surrounding structure micro-deformation data, and generate micro-disturbance pumping response data. S3. Input the perturbation pumping response data into the groundwater seepage digital twin, compare it with the predicted response results to generate a response residual distribution map, invert and calibrate the initial aquifer boundary, initial recharge boundary and support boundary unit, and generate the groundwater recharge vulnerability probability field. S4. Calculate the supply traction risk of candidate dewatering well locations based on the probability field of underground supply leakage, form a well location exclusion boundary and a set of candidate dewatering well locations after screening, configure interception well locations, buffer well locations and main control well locations, and generate candidate schemes for interception well groups. S5. Jointly verify the candidate schemes for the cut-off well group, optimize the strategy based on the joint verification results, generate a set of optimized well group schemes, select the target well group scheme, and output the optimized layout scheme for dewatering wells.

[0007] S1 specifically includes: acquiring geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data for the target foundation pit dewatering area; generating a target foundation pit dewatering basic dataset by unifying coordinates, elevations, times, and fields; dividing aquifer units, support boundary units, external recharge boundary units, and structurally sensitive units based on the target foundation pit dewatering basic dataset, and configuring permeability coefficients, initial water levels, boundary constraints, and monitoring correlations to generate an initial seepage modeling parameter table; establishing a groundwater seepage digital twin based on the initial seepage modeling parameter table, and generating a candidate dewatering well location set and a candidate monitoring point correlation table.

[0008] S2 specifically includes: reading the candidate set of precipitation well locations and the candidate monitoring point association table, eliminating well locations that do not meet the safety constraints of perturbation pumping, determining the priority of the trial based on the coverage of the aquifer unit, the proximity of the suspected recharge boundary, the integrity of the monitoring point, and the structural sensitivity risk, and generating a set of trial well locations and a perturbation pumping execution table; executing low-intensity perturbation pumping in sequence according to the perturbation pumping execution table, collecting data on water level delay, drawdown slope, water level recovery, and micro-deformation of the surrounding structure, and generating the original perturbation response sequence; and performing attribution processing on the original perturbation response sequence according to the candidate monitoring point association table, binding the response data to the corresponding aquifer unit, external recharge boundary unit, and structurally sensitive unit, and generating perturbation pumping response data.

[0009] S3 specifically includes: inputting perturbation pumping response data into a digital twin of groundwater seepage; establishing a step-by-step correspondence between measured and predicted responses based on test well locations, associated monitoring points, and corresponding units; calculating normalized response residuals and generating a response residual distribution map; performing inversion calibration on initial aquifer boundary, initial recharge boundary, and support boundary units based on the response residual distribution map; identifying a set of suspected recharge vulnerability units that meet the conditions of residual exceeding limits, abnormal water level delay, increased recharge intensity, and continuous hydraulic response path; and calculating contribution intensity, confidence level, and recharge vulnerability probability based on the set of suspected recharge vulnerability units, perturbation pumping response data, and response residual distribution map, thereby generating a groundwater recharge vulnerability probability field.

[0010] S4 specifically includes: reading the probability field of underground recharge vulnerabilities, the set of candidate dewatering well locations, and the node connection relationships of the digital twin of groundwater seepage; calculating the recharge traction risk of each candidate dewatering well location to the suspected recharge vulnerabilities; generating a candidate well location risk level table; based on the candidate well location risk level table and the probability field of underground recharge vulnerabilities, generating well location exclusion boundaries with high-probability suspected recharge vulnerabilities, high-risk well locations, and their hydraulic traction paths, and screening to form a set of selected candidate dewatering well locations; based on the well location exclusion boundaries, the set of selected candidate dewatering well locations, and the probability field of underground recharge vulnerabilities, configuring interception well locations, buffer well locations, and main control well locations, and generating candidate schemes for truncated well groups.

[0011] S5 specifically includes: inputting candidate cut-off well group schemes into a digital twin of groundwater seepage; jointly verifying the target drawdown, recharge attenuation rate, well group pumping volume, and micro-deformation constraints of the surrounding structure under baseline hydrological conditions, external recharge enhancement conditions, and structurally sensitive constraint conditions, generating joint verification results for the well group; optimizing strategies based on the non-compliant items in the joint verification results, adjusting the main control well location, cut-off well location, buffer well location, pumping intensity, and start-stop sequence, generating a set of optimized well group schemes; selecting target well group schemes that meet the target drawdown, recharge attenuation rate, well group pumping volume, and micro-deformation constraints of the surrounding structure from the set of optimized well group schemes, and outputting an optimized layout scheme for the dewatering wells.

[0012] A digital twin-based optimized layout system for precipitation wells includes: The twin creation module is used to acquire geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data of the target foundation pit dewatering area, and generate the target foundation pit dewatering basic dataset. The perturbation response acquisition module is used to select test well locations based on the candidate precipitation well location set and the candidate monitoring point association table, and generate a perturbation pumping execution table. The supply vulnerability inversion module is used to input the perturbation pumping response data into the groundwater seepage digital twin, compare it with the predicted response results item by item, and generate a response residual distribution map. The well group scheme generation module is used to calculate the recharge traction risk of candidate precipitation well locations to suspected recharge vulnerability units based on the underground recharge vulnerability probability field, and generate a candidate well location risk level table. The joint verification and optimization module is used to input candidate schemes for truncated well groups into the digital twin of groundwater seepage and conduct joint verification under baseline hydrological conditions, external recharge enhancement conditions, and structurally sensitive constraint conditions.

[0013] The beneficial effects of this invention are as follows: This invention collects data on water level delay, drawdown slope, water level recovery, and micro-deformation of surrounding structures by low-intensity micro-disturbance pumping, and inputs this data into a digital twin of groundwater seepage for response residual analysis. This enables the identification of underground recharge gaps formed by abandoned trenches, old well remnants, support joints, or sand layer lenses before formal well placement, reducing boundary misjudgments caused by relying solely on survey data.

[0014] This invention calculates the recharge traction risk of candidate dewatering well locations based on the probability field of underground recharge vulnerabilities and forms a well location exclusion boundary. This allows for the early exclusion of candidate well locations that might enhance abnormal external groundwater recharge, preventing dewatering wells from being placed on hidden recharge paths and reducing the risk of "the more you pump, the more recharged" from the source. By configuring intercepting well locations, buffer well locations, and main control well locations on both sides of the well location exclusion boundary, the well group no longer focuses solely on the target drawdown depth but also achieves internal pit water level control while weakening abnormal recharge paths, improving the adaptability of dewatering well layout to complex recharge sites.

[0015] This invention, through joint verification of candidate cutoff well group schemes under baseline hydrological conditions, external recharge enhancement conditions, and structurally sensitive constraint conditions, can simultaneously verify the target drawdown, recharge attenuation rate, well group pumping volume, and micro-deformation constraints of surrounding structures, reducing the possibility of achieving single water level targets but experiencing uncontrolled structural or recharge risks. Based on the joint verification results, strategy optimization is performed, adjusting the main control well locations, cutoff well locations, buffer well locations, pumping intensity, and start-stop sequence. This ensures that the final optimized layout scheme for dewatering wells has clearly defined well location roles, operating parameters, and monitoring feedback conditions, facilitating subsequent physical dewatering well construction, pumping operation, and on-site verification. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the optimized layout method for precipitation wells based on digital twins according to the present invention; Figure 2 This is a schematic diagram of the framework of the digital twin-based precipitation well optimization layout system of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1: As Figure 1 As shown, this embodiment provides a method for optimizing the layout of precipitation wells based on digital twins, including the following steps: S1. Obtain geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data of the target foundation pit dewatering area, generate the target foundation pit dewatering basic dataset, divide aquifer units, support boundary units, external recharge boundary units, and structurally sensitive units, and establish a groundwater seepage digital twin, candidate dewatering well location set, and candidate monitoring point association table; S2. Select test well locations based on the candidate precipitation well location set and candidate monitoring point association table, generate a micro-disturbance pumping execution table, execute low-intensity micro-disturbance pumping, collect and attribute water level delay, drawdown slope, water level recovery and surrounding structure micro-deformation data, and generate micro-disturbance pumping response data. S3. Input the perturbation pumping response data into the groundwater seepage digital twin, compare it with the predicted response results to generate a response residual distribution map, invert and calibrate the initial aquifer boundary, initial recharge boundary and support boundary unit, and generate the groundwater recharge vulnerability probability field. S4. Calculate the supply traction risk of candidate dewatering well locations based on the probability field of underground supply leakage, form a well location exclusion boundary and a set of candidate dewatering well locations after screening, configure interception well locations, buffer well locations and main control well locations, and generate candidate schemes for interception well groups. S5. Jointly verify the candidate schemes for the cut-off well group, optimize the strategy based on the joint verification results, generate a set of optimized well group schemes, select the target well group scheme, and output the optimized layout scheme for dewatering wells.

[0019] S1 specifically includes the following sub-steps: S110. Obtain geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data for the target foundation pit dewatering area. Then, perform coordinate unification, elevation unification, time unification, and field unification processing on the above data to generate the target foundation pit dewatering basic dataset.

[0020] Geological exploration data should include at least the coordinates of the exploration boreholes, soil layer numbers, top elevation of the soil layer, bottom elevation of the soil layer, aquifer type, permeability coefficient, and initial groundwater level; foundation pit support boundary data should include at least the plane boundary of the support structure, support depth, continuous record of the cutoff wall, and joint location; surrounding underground structure data should include at least the location, burial depth, and relative distance from the foundation pit boundary of existing pipelines, underground passages, old wells, abandoned pipe trenches, and adjacent building foundations; historical hydrological data should include at least historical groundwater level, rainfall, river and lake water levels, municipal recharge water level, and existing pumping test records; initial monitoring point layout data should include at least the coordinates of the monitoring points, monitoring type, sampling frequency, and corresponding monitoring objects.

[0021] When unifying coordinates, the exploration boreholes, support structures, underground structures, monitoring points, and foundation pit boundaries are all converted to the same engineering coordinate system, with the foundation pit design control points used as the plane coordinate reference. When unifying elevations, the top and bottom elevations of the soil layer, groundwater level, support depth, and structure burial depth are all converted to the same elevation reference. When unifying time, historical groundwater levels, rainfall, river and lake water levels, pumping test records, and monitoring point sampling records are aligned according to the same time interval, and missing data is filled in using adjacent valid sample values ​​or linear interpolation, while retaining missing data markers. When unifying fields, similar data from different source files are converted to a unified field name, unified unit, and unified data format.

[0022] The elevation shall be uniformly determined in the following manner:

[0023] in, This represents the relative elevation of the i-th data point after standardization. This represents the original elevation of the i-th data point. This indicates the unified elevation benchmark used for the dewatering area of ​​the target foundation pit.

[0024] The target foundation pit dewatering dataset obtained through the above processing should at least record the data object number, spatial coordinates, relative elevation, timestamp, data type, source file number, and quality identifier. For cases where abandoned pipe trenches exist in the foundation pit of an old factory renovation project, the surrounding underground structure data should record the centerline coordinates, bottom elevation, direction, cross-sectional dimensions, and closest distance to the foundation pit boundary of the abandoned pipe trench. This ensures that the abandoned pipe trench can be identified as a boundary object that may form an abnormal external groundwater recharge during subsequent unit division.

[0025] S120. Based on the target foundation pit dewatering basic dataset, divide the aquifer unit, support boundary unit, external supply boundary unit and structural sensitive unit, and configure the permeability coefficient, initial water level, boundary constraints and monitoring correlation for each unit to generate the initial seepage modeling parameter table.

[0026] Aquifer units are divided according to the top elevation of the soil layer, the bottom elevation of the soil layer, the type of aquifer, and the difference in permeability coefficient. When the difference in permeability coefficient between adjacent soil layers exceeds the preset stratification threshold, the adjacent soil layers are divided into different aquifer units. When there are sand lenses, local high-permeability interlayers or weakly permeable interlayers in the same soil layer, the corresponding spatial range is divided into aquifer units separately.

[0027] The difference in permeability coefficients is determined as follows:

[0028] in, Indicates the relationship between the k-th soil layer and the 1st soil layer. The difference in permeability coefficient between soil layers This represents the permeability coefficient of the k-th soil layer. Indicates the first The permeability coefficient of each soil layer. When When the value exceeds the preset stratification threshold, the k-th soil layer is compared with the 1st soil layer. The soil layers are divided into different aquifer units.

[0029] The support boundary units are divided according to the diaphragm wall, water-stop curtain, support pile, support joint and soil penetration depth, and configured as water-proof boundary, weakly permeable boundary or suspected recharge boundary according to the water-stop continuity record; the external recharge boundary units are divided according to the river and lake boundary, the leakage area of ​​municipal pipeline, the old well residual hole, the abandoned pipe trench, the sand layer lens body and the recharge direction outside the foundation pit; the structurally sensitive units are divided according to the protection range of subway, pipe gallery, existing building foundation and important pipeline, the allowable deformation value and the distance from the foundation pit boundary.

[0030] The initial seepage modeling parameter table should include at least the unit number, unit type, spatial range, top elevation, bottom elevation, permeability coefficient, initial water level, boundary constraint type, adjacent unit number, associated monitoring point number, and structural sensitivity level. Among them, the unit type includes aquifer unit, support boundary unit, external recharge boundary unit, and structurally sensitive unit, and the boundary constraint type includes constant head boundary, weakly permeable boundary, impermeable boundary, and suspected recharge boundary.

[0031] If the abandoned pipe trench on the north side of the foundation pit intersects with the sand layer and its bottom elevation is lower than the initial groundwater level, then the corresponding unit of the abandoned pipe trench is marked as a suspected recharge boundary in the initial seepage modeling parameter table, and its adjacent aquifer unit number is recorded for subsequent micro-disturbance pumping response inversion.

[0032] In a specific embodiment of the present invention, the preset stratification threshold is preferably set to 10. That is, when the permeability coefficients of two adjacent soil layers differ by an order of magnitude or more, their groundwater hydraulic connection exhibits significant hysteresis, and dividing them into different aquifer units can significantly improve the accuracy of seepage simulation in the digital twin. Meanwhile, for the case where the support boundary unit is an impermeable boundary, its corresponding low connectivity coefficient value is preferably set to 1.0 × 10⁻⁶. -8 Up to 1.0×10 -7 between.

[0033] S130. Establish a digital twin of groundwater seepage based on the initial seepage modeling parameter table, and generate a set of candidate precipitation well locations and a table of candidate monitoring points in the digital twin of groundwater seepage.

[0034] When establishing a digital twin of groundwater seepage, aquifer units, support boundary units, external recharge boundary units, and structurally sensitive units are mapped as computational nodes or computational grids. Node connections are established based on adjacent unit numbers. Seepage parameters for each computational node are configured according to the permeability coefficient, initial water level, and boundary constraint type. Furthermore, micro-deformation constraints of the surrounding structure are bound to the corresponding computational nodes according to the structural sensitivity level. This enables the digital twin of groundwater seepage to receive pumping conditions from candidate dewatering well locations and output predicted water level response, predicted recharge response, and predicted structural micro-deformation values.

[0035] Water exchange between adjacent computing nodes is calculated in the following way:

[0036] in, This represents the amount of seepage exchange between the j-th computing node and its adjacent computing nodes per unit time. This represents the connectivity coefficient corresponding to the j-th computation node. This represents the permeability coefficient of the j-th computation node. This represents the water level at the j-th calculation node. This represents the water level of the computing node adjacent to the j-th computing node.

[0037] When the support boundary unit is a watertight boundary, the corresponding connectivity coefficient is set to a low connectivity coefficient value; when the support boundary unit has records of joints or construction defects, it is retained as a calibrable connectivity coefficient value so that hidden supply loopholes can be identified later through micro-disturbance pumping response.

[0038] The candidate set of dewatering well locations is generated based on the pit outline, aquifer unit distribution, support boundary unit location, external supply boundary unit location, and protection range of structurally sensitive units. Locations located within existing pipelines, subway protection zones, existing building foundation protection zones, and construction restricted areas are not included in the candidate set of dewatering well locations. Locations close to suspected supply boundaries but not falling within construction restricted areas are reserved as candidate locations for exploratory well locations.

[0039] The candidate monitoring point association table is established based on the spatial distance between the candidate dewatering well locations and the monitoring points, the aquifer unit to which they belong, the barrier relationship of the support boundary, and the influence range of the structurally sensitive unit. For each candidate dewatering well location, at least the in-pit water level monitoring point, the out-of-pit water level monitoring point, and the structural deformation monitoring point are associated, and the candidate dewatering well location number, associated monitoring point number, associated aquifer unit number, associated external recharge boundary unit number, associated structurally sensitive unit number, and associated purpose are recorded. The associated purpose includes water level delay acquisition, drawdown slope acquisition, water level recovery acquisition, and surrounding structural micro-deformation acquisition.

[0040] The resulting candidate precipitation well location set is used for subsequent selection of test well locations, and the candidate monitoring point association table is used for subsequent collection and attribution of micro-disturbance pumping response data and strategy optimization and verification.

[0041] S2 specifically includes the following sub-steps: S210. Read the candidate precipitation well location set and the candidate monitoring point association table, select the test well location for low-intensity micro-disturbance pumping from the candidate precipitation well location set, and generate the micro-disturbance pumping execution table.

[0042] When selecting test well locations, first eliminate candidate dewatering well locations that are located within construction restricted areas or within the protection range of structurally sensitive units and cannot meet the safety constraints of micro-disturbance pumping; then determine the test priority based on the coverage of the aquifer unit by the candidate dewatering well location, its proximity to the suspected recharge boundary, the completeness of the associated monitoring points, and the structural sensitivity risk.

[0043] The priority of probing is determined as follows:

[0044] in, This indicates the exploration priority of the a-th candidate precipitation well location. This represents the evaluation value of the aquifer unit covered by the a-th candidate precipitation well location. This represents the evaluation value of the a-th candidate precipitation well location being close to the suspected recharge boundary. This represents the evaluation value of the monitoring point that can be associated with the a-th candidate precipitation well location. This represents the structural sensitivity risk assessment value corresponding to the a-th candidate precipitation well location. , , and These represent the weighting coefficients of the corresponding evaluation values.

[0045] Multiple candidate precipitation well locations are selected as test well locations according to the test priority from high to low to form a set of test well locations; when two candidate precipitation well locations act on the same suspected recharge boundary and the associated monitoring points overlap, the candidate precipitation well location that can be associated with the pit water level monitoring point, the pit water level monitoring point and the structural deformation monitoring point is selected first.

[0046] The upper limit of the low-intensity micro-disturbance pumping flow rate is determined based on the reference flow rate allowed by existing pumping tests, the safe flow rate determined by the allowable micro-deformation of structurally sensitive units, and the detection flow rate required to identify water level changes.

[0047] in, This represents the upper limit of the perturbation pumping flow rate at the a-th test well location. This indicates that a pumping test has been conducted and the allowable reference flow rate has been achieved. This represents the safe flow rate determined based on the allowable micro-deformation of the structurally sensitive elements. This indicates the detection flow rate that enables associated monitoring points to generate detectable water level changes.

[0048] The perturbation pumping execution table should at least record the test well location number, pump start sequence, upper limit of perturbation pumping flow rate, pumping duration, pump stop recovery time, response acquisition frequency, associated monitoring point number, and conditions for eliminating the impact of previous pumping, and serve as the control basis for subsequent low-intensity perturbation pumping.

[0049] In a specific embodiment of the present invention, in order to fully activate the hidden supply channel while minimizing disturbance to the surrounding environment, the preferred value of the above-mentioned weighting coefficient is: =0.3、 =0.4、 =0.3、 =0.5. Due to its proximity to the suspected supply boundary ( The core purpose of identifying vulnerabilities is to identify them, therefore it is given the highest positive weight; while structurally sensitive risks ( As a safety control item, it is given a high penalty weight to ensure that the test well site avoids highly sensitive areas.

[0050] S220. Perform low-intensity perturbation pumping sequentially on the test wells in the set of test wells according to the perturbation pumping execution table, and collect the original perturbation response sequence. Before execution, group the test wells according to spatial partitions, aquifer units, and corresponding suspected recharge boundaries; adjacent test wells in the same group should not pump water simultaneously in the same time period. After the pumping of the previous test well is stopped, the water level at its associated monitoring point should be restored to within the allowable deviation range of the reference water level, or the impact of the previous pumping should be recorded in the original perturbation response sequence before starting the next test well.

[0051] Before starting the pump at each test well, the baseline water level and baseline structural deformation value of the associated monitoring points are collected. After starting the pump, the start-up time, pumping flow rate, pumping duration and real-time water level changes of each associated monitoring point are recorded according to the micro-disturbance pumping execution table. After stopping the pump, the water level recovery process and structural micro-deformation recovery process of each associated monitoring point are collected.

[0052] Water level delay data is determined based on the time difference between the start-up time of the test well and the first identifiable water level change at the associated monitoring point:

[0053] in, This represents the water level delay time at the b-th associated monitoring point. This indicates the time when the b-th associated monitoring point first shows a identifiable water level change. This indicates the pump start-up time for the a-th exploratory well site.

[0054] Drawdown slope data are determined based on the water level change and corresponding time interval during the pumping response phase:

[0055] in, This represents the slope of the depth at the b-th associated monitoring point. This represents the baseline water level at the b-th associated monitoring point before pumping. This represents the water level at the b-th associated monitoring point at the end of the pumping response phase. This represents the baseline data collection time for the b-th associated monitoring point. This indicates the data collection time at the b-th associated monitoring point at the end of the pumping response phase.

[0056] The original perturbation response sequence includes at least the test well location number, test well location coordinates, pump start-up time, pump stop time, pumping flow rate, pumping duration, associated monitoring point number, monitoring point type, baseline water level, real-time water level, pump stop recovery water level, baseline structural deformation value, real-time structural micro-deformation value, sampling timestamp, data quality identifier, and prior pumping impact identifier. If rainfall, instrument offline, data spikes, or external construction disturbances occur during the data acquisition period, the corresponding data quality identifier is written into the original perturbation response sequence so that subsequent attribution processing can remove or downweight the disturbed data.

[0057] S230. Based on the candidate monitoring point association table, the original perturbation response sequence is assigned, and the water level delay data, drawdown slope data, water level recovery data and surrounding structural micro-deformation data corresponding to each test well location are bound to the corresponding aquifer unit, external recharge boundary unit and structural sensitive unit to generate perturbation pumping response data.

[0058] When assigning data, the test well location number is used as the primary attribution object, the associated monitoring point number is used as the secondary attribution object, and the associated aquifer unit number, associated external recharge boundary unit number, and associated structural sensitive unit number are used as spatial attribution objects. Continuous response data of the same test well location before pumping, during pumping, and after pumping stop are classified into the same response record.

[0059] For cases where the same monitoring point is associated with multiple test well locations, the source of the response is determined based on the execution time of the test well location, the impact of previous pumping, the water level recovery status, and the unit pumping response intensity. Data affected by rainfall, equipment offline, or residual impact of previous pumping are removed, downweighted, or marked as data to be verified and are not used as the master data for subsequent inversion calibration.

[0060] The unit pumping response intensity is determined as follows:

[0061] in, This represents the unit pumping response intensity of the a-th test well location to the b-th associated monitoring point. This represents the water level change at the b-th associated monitoring point during the pumping response phase. This represents the actual perturbation pumping flow rate performed at the a-th test well location (the value is equal to the upper limit of the perturbation pumping flow rate). This represents the duration of perturbation pumping at the a-th test well location. The impact of different test well locations on the same associated monitoring point is compared by unit pumping response intensity, avoiding the determination of response attribution solely based on planar distance.

[0062] The generated perturbation pumping response data includes at least the test well location number, associated aquifer unit number, associated external recharge boundary unit number, associated structurally sensitive unit number, water level delay data, drawdown slope data, water level recovery data, surrounding structural micro-deformation data, unit pumping response intensity, data quality level, and invertible flag. When the data quality level is lower than the preset level, or the impact of previous pumping has not been eliminated, the corresponding data is marked as non-invertible or pending verification. The perturbation pumping response data serves as input for subsequent generation of response residual distribution maps, inversion of suspected recharge vulnerability unit sets, and strategy optimization.

[0063] S3 specifically includes the following sub-steps: S310. Input the micro-disturbance pumping response data generated in S230 into the groundwater seepage digital twin and read the initial seepage modeling parameter table generated in S120. If necessary, call the candidate monitoring point association table generated in S130 to verify the attribution relationship between the test well location, associated monitoring point and corresponding unit, and establish the item-by-item correspondence between the measured response and the predicted response.

[0064] The digital twin of groundwater seepage, while maintaining the initial states of the initial aquifer boundary, initial recharge boundary, and support boundary units, performs simulation calculations based on the perturbation pumping flow rate, pumping duration, and start-stop time for each test well location, generating predicted response results for each test well location. The predicted response results include at least the predicted water level delay, predicted drawdown slope, predicted water level recovery curve, and predicted structural micro-deformation values, and are compared item by item with the water level delay data, drawdown slope data, water level recovery data, and surrounding structural micro-deformation data in the perturbation pumping response data.

[0065] For the same response type at the same associated monitoring point, the measured and predicted response values ​​are first converted into normalized response residuals. Then, based on the data quality level, it is determined whether to proceed to subsequent residual fusion. The normalized response residuals are determined as follows:

[0066] in, This represents the normalized response residual of the b-th associated monitoring point. This represents the measured response value of the b-th associated monitoring point. This represents the predicted response value output by the digital twin of groundwater seepage. This represents the smallest positive number used to avoid a denominator of 0. The measured and predicted response values ​​above correspond to water level delay, drawdown slope, water level recovery, or structural micro-deformation, respectively.

[0067] Based on the candidate monitoring point association table, the normalized response residuals of each associated monitoring point are mapped to the corresponding aquifer unit, external recharge boundary unit, and structurally sensitive unit. When the same unit corresponds to multiple associated monitoring points, a weighted fusion is performed based on the data quality level and unit pumping response intensity to generate the unit residual value.

[0068] in, This represents the element residual value of the u-th element. represents the data weight of the b-th associated monitoring point, and n represents the number of associated monitoring points participating in the residual fusion of the u-th unit.

[0069] A response residual distribution map is formed from the residual values ​​of each unit. This map is used to characterize the location and degree of inconsistency between the actual groundwater response and the predicted response of the groundwater seepage digital twin within the target pit dewatering area. If the measured water level delay at the monitoring point outside the pit is 3 minutes, while the predicted water level delay is 12 minutes, it indicates that the actual hydraulic connection between the monitoring point and the test well location is stronger than the initial model prediction. This difference is entered into the response residual distribution map through the normalized response residuals.

[0070] S320. Based on the response residual distribution map, perform inversion calibration on the initial aquifer boundary, initial recharge boundary, and support boundary units in the groundwater seepage digital twin to generate a set of suspected recharge vulnerability units (suspected recharge vulnerability units refer to units with abnormal recharge capacity or abnormal hydraulic connectivity characteristics identified from external recharge boundary units or support boundary units after inversion calibration using perturbation pumping response data and response residual distribution map).

[0071] The objects of inversion calibration include the permeability coefficient of aquifer units, the connectivity coefficient of support boundary units, the recharge intensity of external recharge boundary units, and the spatial range of suspected recharge boundaries. Among them, the permeability coefficient of aquifer units is used to characterize the flow capacity of groundwater within the corresponding aquifer unit, the connectivity coefficient of support boundary units is used to characterize the degree of restriction of groundwater passage by the support boundary, and the recharge intensity of external recharge boundary units is used to characterize the strength of external groundwater recharge to the target foundation pit dewatering area.

[0072] During inversion calibration, the hard constraints determined by geological survey data and support design documents are retained, and the established construction prohibition boundaries, structural entity boundaries, and continuous water-proof boundaries without defect records are not extended without basis. For boundary units with joints, old well remnants, abandoned pipe trenches, sand layer lenses, or abnormal residual concentrations, the connectivity coefficient or replenishment intensity is allowed to be adjusted within a preset range.

[0073] The objective of inversion calibration is to reduce the difference between the measured response and the predicted response after calibration, and to limit the unfounded deviation of the parameter to be calibrated from its initial value. The objective function for inversion calibration is established as follows:

[0074] Where J represents the inversion calibration objective function, This represents the predicted response value output by the digital twin of groundwater seepage after parameter calibration. This represents the data weight of the b-th associated monitoring point. This represents the measured response value of the b-th associated monitoring point. This represents the r-th parameter to be calibrated. This represents the initial value of the r-th parameter to be calibrated. represents the parameter deviation constraint weight, m ​​represents the number of associated monitoring points participating in the inversion calibration, and s represents the number of parameters to be calibrated.

[0075] In the inversion calibration process of this invention, to prevent the model from overfitting, the parameters deviate from the constraint weights. The preferred setting is between 0.01 and 0.1. Within the preset range, the adjustment step size of the connectivity coefficient in a single iteration shall not exceed 50% of the initial value, and the maximum replenishment intensity after calibration shall not exceed 10 times the maximum allowable value of the geological exploration boundary. This serves as a physical hard constraint to ensure that the inversion results conform to the actual engineering hydrogeological laws.

[0076] When an external recharge boundary unit or support boundary unit simultaneously meets the following conditions: residual exceeding the limit, water level delay shorter than the predicted water level delay, calibrated connectivity coefficient or recharge intensity higher than the initial value, and a continuous hydraulic response path exists between it and at least one test well location, it is added to the suspected recharge vulnerability unit set. Suspected recharge vulnerability units are anomalous recharge units identified from external recharge boundary units or support boundary units after inversion calibration using perturbation pumping response data and response residual distribution maps.

[0077] For anomalous units supported only by a single low-quality monitoring point, they are not directly added to the suspected recharge vulnerability unit set, but are marked as units to be reviewed. If a monitoring point outside the pit near the abandoned trench on the north side of the foundation pit shows a water level change 2 minutes after the pump at the test well is started, and the groundwater seepage digital twin predicts a response after 10 minutes, and the recharge intensity of the corresponding unit of the abandoned trench increases after inversion calibration, then the corresponding unit of the abandoned trench can be added to the suspected recharge vulnerability unit set.

[0078] S330. Based on the set of suspected recharge vulnerability cells, perturbation pumping response data, and response residual distribution map, calculate the contribution intensity and confidence level of each suspected recharge vulnerability cell to the abnormal external groundwater recharge, and generate a groundwater recharge vulnerability probability field. The contribution intensity is determined based on the cell residual value, unit pumping response intensity, calibrated recharge intensity increment, and the degree of anomaly in water level recovery; the confidence level is determined based on the number of monitoring points, data quality level, repeated verification results from multiple test well locations, and evidence from underground structures.

[0079] The contribution intensity is determined as follows:

[0080] in, This represents the contribution strength of the u-th suspected supply vulnerability unit. This represents the absolute value of the element residual of the u-th element. This represents the average unit pumping response intensity corresponding to the u-th suspected supply vulnerability unit. This represents the increment of the supply strength after calibration for the u-th suspected supply vulnerability cell. This represents the abnormal water level recovery evaluation value corresponding to the u-th suspected supply vulnerability unit. , , and These represent the weight coefficients of the corresponding indicators.

[0081] Specifically, in calculating the intensity of contribution When the optimal values ​​for each weighting coefficient are: =0.4、 =0.3、 =0.2、 =0.1, meaning it primarily relies on the objective absolute value of the residual and the unit pumping response strength. This involves converting the contribution strength and trust level into a vulnerability probability. When the above parameters are selected, the preferred values ​​are: a=1.5, b=1.2, and c=3.0. This parameter combination utilizes the nonlinear mapping characteristics of the Sigmoid function to rapidly approximate the probability of boundary elements with strong hydraulic connections to 1.0, while suppressing the interference of low residual noise elements.

[0082] After obtaining the contribution strength and trust level, convert them into the probability of patching vulnerabilities:

[0083] in, This represents the probability of a supply vulnerability in the u-th suspected supply vulnerability unit. Let represent the credibility level evaluation value of the u-th suspected vulnerability unit, where a represents the contribution strength weight, b represents the credibility level weight, and c represents the probability transformation bias term.

[0084] The subsurface recharge vulnerability probability field includes at least the suspected recharge vulnerability cell number, spatial range, contribution intensity, confidence level, recharge vulnerability probability, and impact range. Adjacent suspected recharge vulnerability cells with similar hydraulic responses are merged into continuous high-probability regions based on spatial connectivity. Cells with insufficient confidence levels are retained as low-probability verification regions and are not directly used as strong exclusion boundaries. The subsurface recharge vulnerability probability field serves as input for subsequently generating a candidate well location risk level table, well location exclusion boundaries, truncated well group candidate schemes, and a strategy-optimized well group scheme set.

[0085] S4 specifically includes the following sub-steps: S410: Read the underground supply vulnerability probability field generated by S330, the candidate precipitation well location set generated by S130, and the node connection relationship in the groundwater seepage digital twin. Superimpose the underground supply vulnerability probability field onto the candidate precipitation well location set, calculate the supply traction risk of each candidate precipitation well location to the suspected supply vulnerability unit, and generate a candidate well location risk level table.

[0086] When calculating the risk of recharge traction, the probability of recharge vulnerability in suspected recharge vulnerability units, the hydraulic connectivity coefficient between candidate dewatering well locations and suspected recharge vulnerability units, the predicted pumping impact intensity of candidate dewatering well locations, and the hydraulic path distance between them are used as the calculation basis. Among them, the hydraulic connectivity coefficient is determined by the node connection relationship in the groundwater seepage digital twin, the connectivity coefficient of the support boundary unit, and the permeability coefficient of the aquifer unit. The hydraulic path distance is the path length from the calculation node where the candidate dewatering well location is located to the calculation node where the suspected recharge vulnerability unit is located along the connected nodes, rather than the simple planar distance.

[0087] The risk of resupply traction is determined in the following way:

[0088] in, This represents the recharge traction risk value for the i-th candidate precipitation well location. This represents the probability of a supply vulnerability in the u-th suspected supply vulnerability unit. This represents the hydraulic connectivity coefficient between the i-th candidate precipitation well location and the u-th suspected recharge vulnerability unit. This represents the predicted pumping impact intensity of the i-th candidate precipitation well location in the groundwater seepage digital twin. This represents the hydraulic path distance between the i-th candidate precipitation well location and the u-th suspected recharge vulnerability unit. This represents the smallest positive number used to avoid a denominator of 0, and N represents the number of suspected vulnerability units involved in the risk calculation.

[0089] Specifically, the normalized supply traction risk value Wells located in the range [0.8, 1.0] are defined as high-risk wells, those located in the range [0.4, 0.8) are defined as medium-risk wells, and those located in the range [0, 0.4) are defined as low-risk wells.

[0090] Based on the supply traction risk value and the preset risk threshold, candidate dewatering well locations are divided into high-risk, medium-risk, and low-risk well locations. For candidate dewatering well locations located within the protection range of structurally sensitive units or construction-restricted areas, the risk level is increased by 1 level or marked as a restricted well location based on structural sensitivity constraints.

[0091] The generated candidate well location risk level table includes at least the candidate precipitation well location number, well location coordinates, corresponding aquifer unit number, associated suspected recharge vulnerability unit number, recharge traction risk value, risk level, risk source, structural sensitive constraint identifier, and available roles; available roles include interception well locations, buffer well locations, main control well locations, and well locations to be verified.

[0092] S420. Generate well location exclusion boundaries based on the candidate well location risk level table and the underground recharge vulnerability probability field, and select candidate dewatering well locations from the candidate dewatering well location set that can be used for subsequent well group configuration, generating a selected candidate dewatering well location set.

[0093] When generating well site exclusion boundaries, based on high-probability suspected recharge vulnerability units, high-risk well sites, and the hydraulic traction path between them, continuous hydraulic connectivity nodes are extracted from the groundwater seepage digital twin. The continuous hydraulic connectivity nodes are then spatially expanded according to the exclusion boundary buffer distance, so that the well site exclusion boundary covers the continuous risk zone formed by suspected recharge vulnerability units, recharge traction paths, and high-risk well sites.

[0094] The well location exclusion boundary is determined in the following way:

[0095] in, Indicates the well location exclusion boundary. This represents the hydraulic traction path between the i-th high-risk well location and the u-th suspected recharge vulnerability unit. Indicates the exclusion boundary buffer distance. This indicates the spatial expansion of the hydraulic traction path according to the exclusion boundary buffer distance.

[0096] Based on the aquifer thickness and hydraulic conduction characteristics of the target foundation pit dewatering area, the repulsion boundary buffer distance The preferred setting is 15m to 30m. This extended range can effectively cover the hydraulic traction radiation zone caused by formation inhomogeneity, ensuring that the interception well is located outside the traction radiation zone.

[0097] When screening the candidate dewatering well site set, candidate dewatering well sites located within the well site exclusion boundary and with a high risk level are deleted from the candidate dewatering well site set; candidate dewatering well sites located at the edge of the well site exclusion boundary and with a medium risk level are not considered as primary control well sites and are marked as buffer well sites to be verified; candidate dewatering well sites located outside the well site exclusion boundary and with a low risk level are retained and included in the screened candidate dewatering well site set; candidate dewatering well sites located within the protection range of structurally sensitive units are marked as well sites with limited pumping intensity, even if the risk level is low.

[0098] After screening, the candidate dewatering well location set should at least record the well location number, well location coordinates, risk level, available role, allowable pumping intensity range, associated aquifer unit number, associated structural sensitive unit number, corresponding well location exclusion boundary number, and subsequent verification identifier. If there is a sand layer lens connecting the abandoned trench and the foundation pit, the well location exclusion boundary should cover the hydraulic traction path formed by the sand layer lens to prevent subsequent main control well locations from falling into this connecting path and continuing to attract abnormal external groundwater recharge.

[0099] S430. Based on the well location exclusion boundary, the set of candidate precipitation well locations after screening, and the probability field of underground recharge vulnerabilities, configure interception well locations, buffer well locations, and main control well locations to generate candidate schemes for interception well groups.

[0100] The interception well locations are selected from the set of candidate dewatering well locations after screening and are placed near the external recharge side or the recharge path side of the well location exclusion boundary. They are used to weaken the hydraulic transmission of suspected recharge vulnerability units towards the foundation pit, but must not be located within the core area of ​​high-probability recharge vulnerability. The buffer well locations are placed between the well location exclusion boundary and the target drawdown control area. They are used to reduce the sudden change in hydraulic gradient between the interception well locations and the main control well locations, and to limit the transmission of abnormal external groundwater recharge into the foundation pit. The main control well locations are placed in the target drawdown control area or low-risk aquifer units inside the foundation pit. They are used to achieve water level control in the target foundation pit dewatering area and do not directly undertake the strong pumping effect of attracting external recharge.

[0101] When configuring three types of well locations, the available roles, allowable pumping intensity range, and associated aquifer units of each candidate dewatering well location are combined, and the hydraulic gradient between interception well locations, buffer well locations, and main control well locations is predicted through the groundwater seepage digital twin.

[0102] The hydraulic gradient evaluation value is determined in the following way:

[0103] in, This represents the hydraulic gradient evaluation value between the i-th candidate precipitation well location and the u-th suspected recharge vulnerability unit. This represents the predicted water level of the calculation node where the i-th candidate precipitation well location is located. This represents the predicted water level of the computing node where the u-th suspected supply vulnerability unit is located. This represents the hydraulic path distance between the i-th candidate precipitation well location and the u-th suspected recharge vulnerability unit. This represents the smallest positive number used to avoid a denominator of 0. If a certain well group combination causes the hydraulic gradient evaluation value to exceed the preset gradient threshold, the location of the buffer well will be adjusted, the allowable pumping intensity of the corresponding well will be reduced, or the well group combination will be removed.

[0104] In specific engineering applications of the present invention, in order to avoid excessive groundwater flow velocity causing erosion or piping, the preset gradient threshold is preferably set to 0.5 (i.e., the ratio of water level difference to distance).

[0105] The generated candidate cutoff well group schemes should include at least the scheme number, cutoff well location number, buffer well location number, main control well location number, coordinates of each well location, corresponding aquifer unit, corresponding well location exclusion boundary, allowable pumping intensity, start-up and shutdown sequence, target drawdown control area, expected recharge attenuation range, and indicators to be verified. The indicators to be verified include target drawdown, recharge attenuation, well group pumping volume, and the impact of micro-deformation of surrounding structures. The candidate cutoff well group schemes serve as input for subsequent joint verification and strategy optimization.

[0106] S5 specifically includes the following sub-steps: S510. Input the candidate schemes of the cut-off well group generated in S430 into the digital twin of groundwater seepage, read the cut-off well location, buffer well location, main control well location, allowable pumping intensity, start-stop sequence, target drawdown control area and expected recharge weakening range in each scheme, and conduct joint verification under the benchmark hydrological conditions, external recharge enhancement conditions and structurally sensitive constraint conditions respectively.

[0107] The baseline hydrological condition is used to determine whether the candidate scheme of the cut-off well group can achieve the target drawdown depth; the external recharge enhancement condition is used to determine whether the candidate scheme of the cut-off well group weakens the abnormal recharge from the suspected recharge gap unit to the target pit dewatering area; the structurally sensitive constraint condition is used to determine whether the predicted structural microdeformation value of the candidate scheme of the cut-off well group for the structurally sensitive unit exceeds the allowable value.

[0108] During joint verification, the actual depth of each target depth control area is first calculated:

[0109] in, This represents the actual depth of the c-th target depth control area. This represents the initial water level in the c-th target drawdown control area before pumping. This represents the predicted water level in the c-th target drawdown control area after the operation of the cut-off well group candidate scheme.

[0110] Then calculate the drawdown deviation:

[0111] in, This represents the depth deviation of the depth control area for the c-th target. This represents the target depth of the c-th target depth control area; when When the value is less than 0, it indicates that the target depth control area has not reached the target depth.

[0112] For suspected supply vulnerability units, calculate the supply reduction amount and supply reduction rate:

[0113]

[0114] in, This represents the amount of supply reduction for the u-th suspected supply vulnerability unit. This represents the predicted recharge intensity of the u-th suspected recharge vulnerability cell when the truncated well group candidate scheme is not adopted. This represents the predicted recharge intensity of the u-th suspected recharge vulnerability unit after adopting the truncated well group candidate scheme. This represents the supply reduction rate of the u-th suspected supply vulnerability unit. This represents the smallest positive number used to avoid a denominator of 0.

[0115] Calculate the pumping rate of the well group under the constraints of well group operation:

[0116] in, This represents the pumping capacity of the well group in the candidate scheme for the cut-off well group. Let represent the pumping intensity of the i-th well, and p represent the number of wells involved in pumping.

[0117] For structural safety constraints, the predicted structural micro-deformation values ​​of each structurally sensitive unit are read, and the maximum value is taken as the maximum predicted structural micro-deformation value. This generates the well group joint verification results, which include at least the scheme number, depth reduction deviation, recharge attenuation amount, recharge attenuation rate, well group pumping volume, maximum predicted structural micro-deformation value, non-compliance items, and optimization trigger indicators. The optimization trigger indicators include target depth reduction optimization trigger indicators, recharge attenuation optimization trigger indicators, structural safety optimization trigger indicators, and pumping volume optimization trigger indicators, which serve as inputs for strategy optimization in S520.

[0118] S520. Based on the joint verification results of the well group, optimize the strategy and generate a set of optimized well group schemes: The strategy optimization is triggered by the non-compliance items: when the depth deviation of any target depth control area is less than the preset allowable depth deviation, target depth optimization is triggered; when the supply weakening rate of any high-probability suspected supply vulnerability unit is lower than the preset weakening rate threshold, supply weakening optimization is triggered; when the predicted structural micro-deformation value of any structurally sensitive unit exceeds the allowable micro-deformation value, structural safety optimization is triggered; when the pumping volume of the well group exceeds the preset total pumping volume limit, pumping volume optimization is triggered.

[0119] When the target drawdown is insufficient, prioritize adding or adjusting the main control well locations within low-risk aquifer units and adjust the pumping intensity of the main control well locations, without directly increasing the pumping intensity of interceptor well locations near the well location's rejection boundary; when recharge weakening is insufficient, adjust the position, spacing, and pump start-up sequence of interceptor well locations and buffer well locations to weaken the hydraulic transmission path from suspected recharge leakage units to the interior of the foundation pit; when structural micro-deformation exceeds limits, reduce the pumping intensity of well locations near structurally sensitive units, downgrade the corresponding main control well locations to buffer well locations, or move the main control well locations to low-risk aquifer units; when the pumping volume of the well group exceeds limits, delete well locations with lower contributions, or reduce the pumping intensity of well locations in overlapping coverage areas.

[0120] The objective value for strategy optimization is determined in the following way:

[0121] Where F represents the objective value of the policy optimization, This represents the target descent deviation evaluation value. This indicates an insufficient supply reduction rating. This indicates the evaluation value of the structural micro-deformation exceeding the limit. This indicates that the pumping volume of the well group exceeds the evaluation limit. , , and These represent the weighting coefficients of the corresponding evaluation values.

[0122] The strategy optimization, as a multi-objective constraint solution process for the layout of dewatering wells, has the following actual engineering parameters: the preset allowable deviation of the drawdown depth is preferably within 5% of the target drawdown depth, and the preset weakening rate threshold is preferably 60% or above.

[0123] The preferred weighting coefficients in the objective value F of the strategy optimization are: =0.35、 =0.35、 =0.20、 =0.10. This weight allocation effectively balances the target depth reduction within the foundation pit with the mitigation of abnormal recharge outside the pit, while prioritizing the safety of the surrounding structure over the economic efficiency of pumping (pumping volume). The preset convergence threshold for the decrease in the target value of the strategy optimization after two consecutive optimizations is set to 0.05, and the preset maximum number of optimizations is set to 50.

[0124] After each strategy optimization, the adjusted well group scheme is re-input into the groundwater seepage digital twin for joint verification. Strategy optimization stops when the target drawdown, recharge attenuation rate, well group pumping volume, and surrounding structural micro-deformation constraints all meet the requirements. Optimization stops when the decrease in the strategy optimization target value is less than the preset convergence threshold after two consecutive optimizations, or when the number of optimizations reaches the preset upper limit. Schemes that do not meet the constraints are marked as unusable. The resulting set of optimized well group schemes is used by S530 to select the target well group scheme.

[0125] S530. Select the target well group scheme that meets the constraints of target drawdown depth, recharge weakening rate, well group pumping volume and micro-deformation of surrounding structure from the strategy optimization well group scheme set, and output the optimized layout scheme of dewatering wells in the target foundation pit dewatering area.

[0126] When multiple well group optimization schemes satisfy the constraints exist, a comprehensive selection is made based on factors such as the weakening effect of recharge, the risk of recharge traction, the impact of structural micro-deformation, the pumping capacity of the well group, and the integrity of monitoring feedback:

[0127] in, This represents the comprehensive optimization score of the g-th strategy well group scheme. This represents the evaluation value of the supply weakening effect of the g-th strategy optimization well group scheme. This represents the resupply traction risk assessment value of the g-th optimized well group scheme. This represents the evaluation value of the structural micro-deformation impact of the g-th strategy optimization scheme for the well group. This represents the evaluation value of the pumping volume of the well group for the g-th optimized well group scheme. This represents the monitoring feedback integrity evaluation value of the g-th strategy optimization well group scheme. , , , and These represent the weighting coefficients of the corresponding evaluation values.

[0128] To select the optimal solution while satisfying all hard operating constraints, the weighting coefficients for the above comprehensive optimization score are preferably configured as follows: =0.35 (supply reduction weakens positive benefit), =0.25 (Negative penalty for traction risk) =0.20 (negative penalty for micro-deformation) =0.10 (negative penalty for pumping volume) =0.10 (Positive benefit from monitoring feedback).

[0129] The optimal well group scheme with the highest comprehensive evaluation score is selected as the target well group scheme. If the replenishment weakening effect of the two schemes is similar, the scheme with the main control well location far away from the well location exclusion boundary, complete coverage of the buffer well location, small impact of structural micro-deformation, and complete monitoring feedback conditions is given priority.

[0130] The output optimized layout scheme for dewatering wells should include at least the target well group scheme number, interception well location number and coordinates, buffer well location number and coordinates, main control well location number and coordinates, aquifer unit corresponding to each well location, allowable pumping intensity, pump start-up sequence, pump stop-down sequence, target drawdown control area, well location exclusion boundary, expected recharge weakening range, monitoring feedback conditions, and anomaly adjustment rules. The optimized layout scheme guides the construction, pumping operation, and monitoring feedback verification of actual dewatering wells. The measured operational data obtained from the monitoring feedback verification can be used to update the digital twin of groundwater seepage.

[0131] Example 2: Figure 2 As shown, this embodiment provides a digital twin-based precipitation well optimization layout system, including: The twin creation module is used to acquire geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data of the target foundation pit dewatering area, generate a basic dataset of target foundation pit dewatering, divide aquifer units, support boundary units, external recharge boundary units, and structurally sensitive units, establish a digital twin of groundwater seepage, and generate a set of candidate dewatering well locations and a candidate monitoring point association table; The perturbation response acquisition module is used to select test well locations based on the candidate precipitation well location set and the candidate monitoring point association table, generate a perturbation pumping execution table, control the test well locations to perform low-intensity perturbation pumping, collect water level delay, drawdown slope, water level recovery and micro-deformation data of surrounding structures, and perform attribution processing on the collected data to generate perturbation pumping response data. The recharge vulnerability inversion module is used to input the perturbation pumping response data into the groundwater seepage digital twin, compare it with the predicted response results item by item, generate a response residual distribution map, and perform inversion calibration on the initial aquifer boundary, initial recharge boundary and support boundary units based on the response residual distribution map, generating a set of suspected recharge vulnerability units and a groundwater recharge vulnerability probability field. The well group scheme generation module is used to calculate the supply traction risk of candidate dewatering well locations to suspected supply vulnerability units based on the underground supply vulnerability probability field, generate a candidate well location risk level table, form a well location exclusion boundary and a set of candidate dewatering well locations after screening based on the candidate well location risk level table, and configure interception well locations, buffer well locations and main control well locations to generate interception well group candidate schemes. The joint verification and optimization module is used to input candidate schemes of truncated well groups into the digital twin of groundwater seepage, and to conduct joint verification under the baseline hydrological conditions, external recharge enhancement conditions and structurally sensitive constraint conditions. It generates joint verification results of well groups, optimizes strategies based on the joint verification results, generates a set of strategy-optimized well group schemes, selects the target well group scheme, and outputs the optimized layout scheme of dewatering wells.

[0132] All the above formulas are performed using dimensionless numerical calculations; the relevant formulas are based on empirical models that approximate the real situation, obtained through extensive data collection and software simulation fitting. The preset parameters and thresholds involved in the formulas can be conventionally set and adjusted by those skilled in the art according to the physical constraints of the actual application scenario.

[0133] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0134] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0135] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing the layout of precipitation wells based on digital twins, characterized in that, Includes the following steps: S1. Obtain geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data of the target foundation pit dewatering area, generate the target foundation pit dewatering basic dataset, and establish a groundwater seepage digital twin, candidate dewatering well location set, and candidate monitoring point association table. S2. Select test well locations based on the candidate precipitation well location set and candidate monitoring point association table, generate a micro-disturbance pumping execution table, execute low-intensity micro-disturbance pumping, collect and attribute water level delay, drawdown slope, water level recovery and surrounding structure micro-deformation data, and generate micro-disturbance pumping response data. S3. Input the perturbation pumping response data into the groundwater seepage digital twin, compare it with the predicted response results to generate a response residual distribution map, invert and calibrate the initial aquifer boundary, initial recharge boundary and support boundary unit, and generate the groundwater recharge vulnerability probability field. S4. Calculate the supply traction risk of candidate dewatering well locations based on the probability field of underground supply vulnerabilities, form a well location exclusion boundary and a set of candidate dewatering well locations after screening, configure interception well locations, buffer well locations and main control well locations, and generate candidate schemes for interception well groups.

2. The method for optimizing the layout of precipitation wells based on digital twins according to claim 1, characterized in that, Also includes: S5. Jointly verify the candidate schemes for the cut-off well group, optimize the strategy based on the joint verification results, generate a set of optimized well group schemes, select the target well group scheme, and output the optimized layout scheme for dewatering wells.

3. The method for optimizing the layout of precipitation wells based on digital twins according to claim 1, characterized in that, S1 specifically includes: Geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data of the target foundation pit dewatering area are obtained. After processing with unified coordinates, unified elevations, unified time, and unified fields, the target foundation pit dewatering basic dataset is generated. Based on the target foundation pit dewatering basic dataset, aquifer units, support boundary units, external recharge boundary units, and structurally sensitive units are divided, and permeability coefficients, initial water levels, boundary constraints, and monitoring correlations are configured to generate an initial seepage modeling parameter table. A digital twin of groundwater seepage is established based on the initial seepage modeling parameter table, and a set of candidate precipitation well locations and a table of candidate monitoring points are generated.

4. The method for optimizing the layout of precipitation wells based on digital twins according to claim 1, characterized in that, S2 specifically includes: Read the candidate precipitation well location set and candidate monitoring point association table, remove well locations that do not meet the safety constraints of perturbation pumping, determine the test priority according to the aquifer unit coverage, the proximity of the suspected recharge boundary, the integrity of the monitoring point and the structural sensitivity risk, and generate the test well location set and perturbation pumping execution table; Low-intensity perturbation pumping is performed sequentially according to the perturbation pumping execution table. Data on water level delay, drawdown slope, water level recovery, and micro-deformation of surrounding structures are collected to generate the original perturbation response sequence.

5. The method for optimizing the layout of precipitation wells based on digital twins according to claim 4, characterized in that, Also includes: The original perturbation response sequence is assigned based on the candidate monitoring point association table, and the response data is bound to the corresponding aquifer unit, external recharge boundary unit and structural sensitive unit to generate perturbation pumping response data.

6. The method for optimizing the layout of precipitation wells based on digital twins according to claim 1, characterized in that, S3 specifically includes: Input the perturbation pumping response data into the groundwater seepage digital twin, establish a one-to-one correspondence between the measured response and the predicted response according to the test well location, associated monitoring point and corresponding unit, calculate the normalized response residual and generate the response residual distribution map. Based on the response residual distribution map, the initial aquifer boundary, initial recharge boundary and support boundary units are inverted and calibrated to identify a set of suspected recharge vulnerability units that meet the conditions of residual exceeding limits, abnormal water level delay, increased recharge intensity and continuous hydraulic response path. Based on the set of suspected supply vulnerability units, perturbation pumping response data, and response residual distribution map, the contribution intensity, confidence level, and supply vulnerability probability are calculated to generate an underground supply vulnerability probability field.

7. The method for optimizing the layout of precipitation wells based on digital twins according to claim 1, characterized in that, S4 specifically includes: Read the node connection relationships of the underground recharge vulnerability probability field, the candidate precipitation well location set, and the groundwater seepage digital twin, calculate the recharge traction risk of each candidate precipitation well location to the suspected recharge vulnerability unit, and generate a candidate well location risk level table; Based on the candidate well location risk level table and the underground recharge vulnerability probability field, well location exclusion boundaries are generated using high-probability suspected recharge vulnerability units, high-risk well locations and their hydraulic traction paths, and a set of candidate precipitation well locations is formed after screening. Based on the well location exclusion boundary, the set of candidate precipitation well locations after screening, and the probability field of underground recharge vulnerabilities, interception well locations, buffer well locations, and main control well locations are configured to generate candidate schemes for interception well groups.

8. The method for optimizing the layout of precipitation wells based on digital twins according to claim 2, characterized in that, S5 specifically includes: The candidate schemes for cut-off well groups are input into the digital twin of groundwater seepage. Under the benchmark hydrological conditions, external recharge enhancement conditions, and structurally sensitive constraint conditions, the target drawdown, recharge weakening rate, well group pumping volume, and micro-deformation constraints of the surrounding structure are jointly verified to generate joint verification results for the well group. Based on the non-compliance items in the joint verification results of the well group, the strategy is optimized by adjusting the main control well location, interception well location, buffer well location, pumping intensity and start-stop sequence, and generating a set of optimized well group schemes.

9. The method for optimizing the layout of precipitation wells based on digital twins according to claim 8, characterized in that, Also includes: The target well group scheme that meets the constraints of target drawdown, recharge weakening rate, well group pumping volume and micro-deformation of surrounding structure is selected from the strategy optimization well group scheme set, and the optimized layout scheme of dewatering wells is output.

10. A digital twin-based precipitation well optimization layout system, employing the digital twin-based precipitation well optimization layout method as described in any one of claims 1 to 9, characterized in that, include: The twin creation module is used to acquire geological survey data, foundation pit support boundary data, surrounding underground structure data, historical hydrological data, and initial monitoring point layout data of the target foundation pit dewatering area, and generate the target foundation pit dewatering basic dataset. The perturbation response acquisition module is used to select test well locations based on the candidate precipitation well location set and the candidate monitoring point association table, and generate a perturbation pumping execution table. The supply vulnerability inversion module is used to input the perturbation pumping response data into the groundwater seepage digital twin, compare it with the predicted response results item by item, and generate a response residual distribution map. The well group scheme generation module is used to calculate the recharge traction risk of candidate precipitation well locations to suspected recharge vulnerability units based on the underground recharge vulnerability probability field, and generate a candidate well location risk level table. The joint verification and optimization module is used to input candidate schemes for truncated well groups into the digital twin of groundwater seepage and conduct joint verification under baseline hydrological conditions, external recharge enhancement conditions, and structurally sensitive constraint conditions.

Images