Underground water level dynamic regulation and control method and system based on digital twinning

Abstract

The invention belongs to the technical field of hydrogeology and intelligent control, particularly relates to an underground water level dynamic regulation and control method and system based on digital twinning, and aims to solve the problems of one-sided perception, static model, extensive decision and single feedback of traditional regulation and control. The method comprises the following steps: collecting multi-source monitoring data and constructing a data set through cross-scale calibration; constructing a multi-field dynamic coupling digital twinborn model, and introducing a dynamic coupling coefficient to deduce a risk evolution field; calculating a variable coefficient of the risk value of each physical field based on the risk evolution field, and generating a comprehensive risk value; and matching the risk threshold to generate a deep and shallow layered regulation and control instruction. The multi-field dynamic coupling breaks the static limitation of the model, the dynamic weight based on the variable coefficient ensures that the regulation and control decision can preferentially respond to the key physical field with high absolute risk level and large spatial fluctuation, and the full-region visual tracking and accurate management and control of the underground water level and the geological risk are realized.

Description

Technical Field

[0001] This invention belongs to the field of hydrogeology and intelligent control technology, specifically relating to a method and system for dynamic regulation of groundwater level based on digital twins. Background Technology

[0002] The stability of groundwater levels is crucial for the sustainable use of regional water resources, the safety of urban underground engineering projects, and the ecological maintenance of specific geological environments. Traditional groundwater level regulation methods, such as pumping or recharge based on empirical models, face a series of technical bottlenecks.

[0003] Traditional control methods rely heavily on discretely deployed piezometers for data sensing, providing "point-like" data that fails to comprehensively reflect the spatiotemporal distribution characteristics of regional groundwater levels. Furthermore, they lack effective cross-scale calibration mechanisms for macroscopic and mesoscopic monitoring data such as satellite remote sensing and UAV thermal infrared data, resulting in incomplete and erroneous information. In model simulation, existing models often decouple multiple physical fields (hydraulic, mechanical, chemical, and thermal) or use fixed coupling parameters, failing to accurately reflect the real-time, dynamic interactions between these fields. This leads to a significant decrease in model prediction accuracy over time. Moreover, existing control strategies often employ conventional control algorithms that do not fully consider the inherent hysteresis and nonlinear characteristics of groundwater seepage. When dealing with the long hysteresis and nonlinear response unique to groundwater systems, their control effects sometimes fall short of ideal accuracy, failing to meet the precise control requirements under long-term, complex conditions.

[0004] Therefore, how to construct an intelligent closed-loop control system that can achieve dynamic modeling, forward-looking decision-making, and precise execution has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] To address the aforementioned problems in existing technologies, namely the issues of limited perception, static models, and coarse decision-making in traditional groundwater level control, the first aspect of this invention proposes a dynamic groundwater level control method based on digital twins, the method comprising the following steps:

[0006] Collect multi-source monitoring data of the groundwater system in the target area and construct a multi-dimensional dynamic dataset;

[0007] A multi-field dynamically coupled digital twin model is constructed, and a multi-dimensional dynamic dataset is used to drive the model to perform inferences in order to predict the future spatiotemporal distribution of multi-field parameters, thereby quantifying the evolution trend of key risk indicators and generating a risk evolution field.

[0008] The coefficient of variation of the risk value of each physical field is calculated based on the risk evolution field, and the decision weight is dynamically assigned to each physical field based on the coefficient of variation and the scene characteristics of the target area to generate a comprehensive risk value.

[0009] The comprehensive risk value is matched with the risk warning threshold to determine the risk assessment result; based on the risk assessment result, coordinated control instructions are generated for shallow and deep aquifers respectively.

[0010] In some preferred embodiments, multi-source monitoring data of the groundwater system in the target area are collected, including:

[0011] The target area is monitored using satellite remote sensing equipment, UAV thermal infrared equipment, ground sensor networks, and downhole intelligent monitoring equipment to acquire multi-source, multi-scale monitoring data. Specifically, satellite remote sensing equipment acquires raw data on surface deformation rate; UAV thermal infrared equipment acquires soil temperature and illumination deviation data; the ground sensor network acquires water level, structural stress, water quality parameters, formation temperature, and soil moisture content data; and the downhole intelligent monitoring equipment acquires pore water pressure and deep water level data. The water quality parameters include pH value and... concentration.

[0012] In some preferred embodiments, a multi-field dynamically coupled digital twin model is constructed using the following method:

[0013] A multi-field dynamic coupling model is constructed, incorporating hydraulic, mechanical, chemical, and thermal fields. A dynamic coupling coefficient is introduced to quantify and update the interaction between at least two physical fields in real time. The hydraulic field parameters include shallow and deep water levels; the mechanical field parameters include structural settlement and effective stress; the chemical field parameters include corrosion rate; and the thermal field parameters include thermal stress. The dynamic coupling coefficient includes… It affects the sensitivity coefficient, pore pressure sensitivity coefficient, temperature sensitivity coefficient, and temperature corrosion coefficient.

[0014] In some preferred embodiments, the interaction between two physical fields is updated by dynamically updating the key model parameters of the other physical field based on the simulation output of one physical field.

[0015] The update includes at least one of the following:

[0016] Based on the simulation results of solute concentration in the chemical field, using The sensitivity coefficient affects the updating of the permeability coefficient value of shallow aquifers;

[0017] Based on the simulation results of pore water pressure in the mechanical field, the permeability coefficient of the deep aquifer is updated using the pore pressure sensitivity coefficient and the temperature sensitivity coefficient.

[0018] Based on the temperature simulation results of the thermal field, the chemical reaction rate value of the chemical field is updated using the temperature corrosion coefficient.

[0019] In some preferred embodiments, generating the comprehensive risk value specifically includes:

[0020] For the states of the hydraulic field, mechanical field, chemical field and thermal field in the aforementioned risk evolution field, the independent risk assessment values ​​of each field at all monitoring points are statistically analyzed;

[0021] Calculate the mean and standard deviation of the independent risk assessment values ​​of each physical field, and obtain the coefficient of variation of each physical field based on the mean and standard deviation;

[0022] The basic decision weights are allocated proportionally based on the coefficient of variation of each physical field, with the physical field having a higher coefficient of variation receiving a greater basic decision weight.

[0023] The decision weights are adjusted based on the scene characteristics of the target area;

[0024] The average independent risk assessment value of each physical field is calculated, and the combined risk value is generated by weighted summation based on the modified decision weights.

[0025] In some preferred embodiments, the comprehensive risk value is matched with a risk warning threshold to determine the risk assessment result, and the method is as follows:

[0026] Based on the safety level of the target area project, at least two risk warning thresholds are dynamically preset in combination with the operating conditions or environmental conditions, and different risk levels are divided according to the risk thresholds.

[0027] The comprehensive risk value is compared with the risk threshold, and the risk level to which the comprehensive risk value belongs is determined based on the comparison result;

[0028] The risk level is used as the risk assessment result to trigger the corresponding control strategy.

[0029] In some preferred embodiments, based on the risk assessment results, coordinated control commands are generated for shallow and deep aquifers respectively, specifically as follows:

[0030] For shallow aquifers, pumping instructions are generated with the aim of maintaining dynamic water level stability in key areas or meeting engineering dewatering requirements.

[0031] For deep aquifers, a reinjection command is generated with the aim of providing regional pore water pressure support and mitigating land subsidence.

[0032] In some preferred embodiments, the system also includes collecting feedback data after the command is executed, and optimizing the system to form closed-loop control through a two-level closed-loop feedback mechanism.

[0033] In some preferred embodiments, the two-level closed-loop feedback mechanism specifically includes the following:

[0034] The first level is the command feedback loop: based on the real-time deviation between the measured physical quantity and the predicted data in the feedback data, the calculation parameters of the control command are quickly adjusted to correct the deviation in real time and ensure the accurate execution of the command; the calculation parameters include the lag compensation coefficient and the level coefficient.

[0035] The second level is the model evolution loop: it is activated when the overall deviation exceeds the preset deviation threshold. It periodically collects historical monitoring data within a control cycle and optimizes and updates the core physical parameters in the digital twin model through data inversion methods. The core physical parameters are at least one dynamic coupling coefficient to realize the periodic learning and evolution of the model.

[0036] A second aspect of the present invention proposes a dynamic groundwater level control system based on digital twins, the system comprising:

[0037] The data acquisition module is configured to collect multi-source monitoring data of the groundwater system in the target area and construct a multi-dimensional dynamic dataset containing hydraulic, mechanical, chemical and thermal parameters.

[0038] The twin simulation and prediction module is configured to construct a multi-field dynamically coupled digital twin model. It uses a multi-dimensional dynamic dataset to drive the multi-field dynamically coupled digital twin model to perform simulation, simulate the state response of the groundwater system under the influence of multiple fields, predict the future spatiotemporal distribution of multiple field parameters, and then quantify the evolution trend of key risk indicators to generate a risk evolution field.

[0039] The data processing module is configured to calculate the coefficient of variation of the risk values ​​of each physical field based on the risk evolution field, and dynamically allocate decision weights to each physical field based on the coefficient of variation and the scene characteristics of the target area to generate a comprehensive risk value.

[0040] The tiered control module is configured to match the comprehensive risk value with a preset risk warning threshold to determine the risk assessment result; and generate coordinated control instructions for shallow and deep aquifers based on the risk assessment result.

[0041] The feedback optimization module is configured to collect feedback data after the command is executed, and optimize the calculation parameters of the control command and the core physical parameters of the digital twin model through a two-level closed-loop feedback mechanism to form closed-loop control.

[0042] The beneficial effects of this invention are:

[0043] (1) By integrating multi-source data, constructing a multi-field coupled digital twin model, generating a three-dimensional risk dynamic evolution field, taking the dynamic permeability coefficient as the core physical parameter and the risk quantification closed loop as the decision core, breaking through the limitations of traditional point monitoring, and realizing the visualization tracking of the risk hot zone of groundwater level drop, water pollution diffusion, and land subsidence risk in the whole region.

[0044] (2) By breaking the static limitations of the model through multi-field dynamic coupling, the weights are dynamically allocated based on the coefficient of variation and scene characteristics, so that the comprehensive risk value can reflect the physical field interaction effect (such as the surge in geological stress induced by the drop in water level) and regional differences in a synchronous manner, thereby improving the accuracy of the control decision and the priority of response to key risks, and avoiding the control deviation that may be caused by static weights.

[0045] (3) The dual-loop mechanism transforms the feedback data deviation into a model calibration signal: by using the deviation between the monitoring data and the predicted value, key parameters such as the aquifer permeability coefficient are corrected in real time, and the prediction error is compensated to drive the continuous evolution of the model and ensure the long-term reliability of the system under environmental disturbances. Attached Figure Description

[0046] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0047] Figure 1 This is a flowchart of a method for dynamic control of groundwater level based on digital twin provided in one embodiment of this application;

[0048] Figure 2 This is a schematic diagram of the structure of a computer system used to implement the methods, systems, and electronic devices of this application. Detailed Implementation

[0049] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0050] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0051] To address the problems of limited perception, static models, and crude decision-making in traditional groundwater level control, this invention provides a dynamic groundwater level control method based on digital twins. Using dynamic permeability coefficient as the core physical parameter and risk quantification closed-loop as the decision-making core, the method is implemented in a four-stage progressive manner: data acquisition → model simulation → control execution → feedback optimization. It can be widely applied to sustainable management of regional groundwater resources (such as groundwater funnel control in the North China Plain and waterlogging prevention in southern paddy fields), safety control of urban underground engineering (such as anti-buoyancy of subway tunnels, seepage prevention of deep foundation pits, and deformation control of underground integrated pipe corridors), and groundwater system control in special environments (such as chemical-hydraulic coupling management in high-salinity and alkaline areas).

[0052] To more clearly explain the groundwater level dynamic control method based on digital twins of this invention, the following will be combined with... Figure 1 The steps in the embodiments of the present invention will be described in detail below.

[0053] Before proceeding with the specific implementation of this invention, some core terms involved in this invention will be defined as follows:

[0054] Risk evolution field: In this application, it refers to a four-dimensional dataset containing three-dimensional spatial coordinates and a future time axis, which stores the quantitative value or risk level of one or more key risk indicators predicted by the twin model at each spatiotemporal node.

[0055] Coefficient of variation: In this application, it refers to a standardized statistical index used to measure the relative dispersion of data. In this application, it is used as a comprehensive index to quantify the absolute level (mean) and spatial volatility (standard deviation) of a certain physical field risk. The higher the value, the higher the risk level of the physical field and the uneven spatial distribution, which requires priority intervention.

[0056] Scenario characteristics: In this application, it refers to the specific social, engineering or environmental attributes of the target area, such as whether the area is a "sensitive area for people's livelihood in the city", "ultra-deep foundation pit construction area" or "water source protection area", etc. These attributes will serve as the basis for modifying the control strategy.

[0057] A method for dynamic control of groundwater level based on digital twins according to the first embodiment of the present invention includes steps S100-S600, each step of which is described in detail below:

[0058] S100, the four-dimensional monitoring system collects multi-source monitoring data of the groundwater system in the target area:

[0059] Satellite remote sensing equipment InSAR (Interferometric Synthetic Aperture Radar): Acquires raw data on the rate of surface deformation. (mm / d), reflecting the settlement trend of the mechanical field structure; rainfall intensity P;

[0060] Drone thermal infrared equipment: to obtain soil temperature ( ) and illumination deviation coefficient (Dimensionless, 0~1), associated thermal field;

[0061] Ground sensor network: collecting shallow water level data (m), structural stress (kPa), ground concentration shallow formation temperature ( Shallow pore water pressure ; Soil moisture content ;

[0062] Downhole intelligent monitoring equipment: collecting deep water level data deep concentration Deep pore water pressure shallow formation temperature ( ), to supplement the hydraulic field parameters.

[0063] Preferably, the following basic data also need to be obtained:

[0064] 1) Obtain historical baseline data, such as geological borehole data (three-dimensional geological structure and deep lithology distribution) and hydrogeological test data (stratum compressibility coefficient). ,soil Saturation concentration Shallow initial permeability coefficient Initial permeability coefficient of deep layer Soil saturated moisture content Poisson's ratio of rock and soil We obtained initial values ​​for surface deformation and seepage lag time from water level fluctuation data and historical surface deformation data over the past 5-10 years, which were used for model initialization.

[0065] 2) Management constraint data for the target area, such as regional water resources planning documents (e.g., annual total extraction quota) and underground engineering safety standards (e.g., anti-buoyancy safety factor). ≥1.2, Water level fluctuation control threshold This yields the allowable drop threshold of deep water level, target value of water level control, allowable corrosion rate, and allowable effective stress of the formation, which are used as constraints for subsequent optimization decisions.

[0066] 3) Equipment parameter data: Deep extraction area Radius of influence of recharge Shallow drainage area Radius of influence of extraction Rated flow rate of water pump , Rated flow rate of the recharge pump .

[0067] Preferably, the acquired multi-source monitoring data is preprocessed, including the following:

[0068] Outlier cleaning: using " "Guidelines" to filter invalid values ​​such as sudden changes in water level and abnormal temperature caused by sensor malfunctions;

[0069] Missing value completion: Short-term missing values ​​(<2 hours) are filled by linear interpolation, while long-term missing values ​​(≥2 hours) are filled by prediction based on historical sequences using an LSTM (Long Short-Term Memory) model;

[0070] Spatiotemporal alignment: Unify all data to the WGS84 geographic coordinate system "UTC timestamp" is used to correlate quarterly deformation rate data from satellite remote sensing equipment, hourly thermal infrared data from drones, and minute-level water level data from smart wells through time interpolation, ensuring a one-to-one correspondence between spatial location and time node.

[0071] More preferably, cross-scale data mutual calibration is performed on the preprocessed multi-source monitoring data, specifically including the following:

[0072] Macro-layer (InSAR data) calibration: using measured deep water level increments The corresponding physical settlement model corrects the original surface deformation rate and removes noise caused by changes in non-confined aquifers. The specific method is as follows:

[0073] Based on the deep water level over a period of time Increment within Calculate the resulting settlement rate. :

[0074] ;

[0075] The calibrated surface deformation rate is:

[0076] ;

[0077] in, To calibrate the surface deformation rate ( ), This represents the collected surface deformation rate. t is the current time. The previous hour (m); The calibration coefficient (obtained by fitting historical InSAR data with measured deep water levels; the preferred value range in this embodiment) is... ); The thickness of the compression layer (m) is (where the shallow layer) deep ,) is the formation compressibility coefficient (kPa); ;

[0078] Mesoscopic (UAV thermal infrared data) calibration: Correcting thermal infrared temperature deviations caused by differences in illumination and surface moisture content.

[0079] ;

[0080] in, The soil temperature collected in the waterlogged area, To calibrate the soil temperature in the waterlogged area ( ), The illumination deviation coefficient (collected in real time by the illumination sensor, with a value range of...) (This represents the maximum temperature deviation caused by light exposure in a completely dry state). The measured value of soil moisture content (%). Soil saturated moisture content (%). Calibration improves the accuracy of seepage path identification.

[0081] In some embodiments, The calibration method is as follows: On a clear, cloudless day, select a dry, bare plot of land and a saturated, moist plot of land. At noon when the sunlight is strongest, simultaneously measure the UAV thermal infrared temperature and the actual contact thermometer temperature of both plots. This is the difference between the two temperatures on a dry plot of land.

[0082] In some embodiments, the cross-scale data after mutual calibration is used to construct a data structure containing the time dimension t and the three-dimensional spatial coordinates using the Kalman filter algorithm. The multi-parameter matrix is ​​used to generate a spatiotemporal four-dimensional dataset, specifically including:

[0083] Dynamic parameter field: Corrected surface deformation rate 2. Soil temperature in the waterlogged area after calibration Shallow water level Deep water level Shallow pore water pressure Deep pore water pressure Ground Cl⁻ concentration Deep concentration wait;

[0084] Static parameter library: water density Gravitational acceleration Elastic modulus Corrosion rate at room temperature Formation compressibility ,soil Saturation concentration Coefficient of thermal expansion Initial permeability coefficients of shallow and deep layers , Soil saturated moisture content wait.

[0085] S200. Construct a multi-field dynamic coupled digital twin model, use a multi-dimensional dynamic dataset to drive the model to perform inference, predict the future spatiotemporal distribution of multi-field parameters, and then quantify the evolution trend of key risk indicators to generate a risk evolution field.

[0086] Preferably, a multi-field dynamically coupled digital twin model is constructed, specifically including the following:

[0087] S201. Based on the fundamental mechanism of the groundwater system, construct the basic governing equations, specifically including:

[0088] Hydraulic field, saturated-unsaturated seepage equations constructed based on Darcy's law:

[0089] Darcy's Law: Continuity equation: ;

[0090] Where q is the seepage velocity ( ); k is the permeability coefficient ( ); Here, h represents the hydraulic gradient, and h represents the water level. The density of water ( ), where n is porosity and Q is the source / sink term (extraction / reinjection volume); in shallow conditions (Extraction); Deep subsurface, (Refilling);

[0091] In the mechanical field, a soil constitutive model (such as the Mohr-Coulomb model) is constructed based on the finite element method, and the soil deformation and structural stress are calculated using the effective stress principle.

[0092] Structural settlement: Effective stress ;

[0093] in, is the effective stress (kPa); s is the settlement (mm); H is the thickness of the compression layer (m); Compression modulus (kPa); Shallow layer: Deeper layers: ;

[0094] A chemical field was used to simulate the effects of Cl⁻ and other ions on soil permeability and structural corrosion rate through electrochemical corrosion kinetic equations.

[0095] Solute transport equation: ;

[0096] Where: C is the solute concentration ( D is the dispersion coefficient ( R represents the chemical reaction term (); ); shallow layer: Deeper layers: ;

[0097] A thermal field, based on the heat conduction equation, simulates the effect of water temperature changes on the viscosity coefficient of water flow and the rate of chemical reactions:

[0098] Heat conduction equation: ;

[0099] Thermal stress equation: ;

[0100] Where E is the elastic modulus of the soil and rock mass; For reference temperature (preferred in this embodiment) ); Poisson's ratio, This is the absolute value of the temperature deviation, ensuring that the thermal stress is positive. The volumetric heat capacity of rock and soil ( ); Thermal conductivity ( ); For heat source item ( ); shallow layer: Deeper layers: .

[0101] Preferably, the specific values ​​of the static data in this embodiment can be calibrated based on local soil test data. In this embodiment, only specific values ​​are selected for illustration.

[0102] S202. Introducing dynamic coupling coefficients to quantify real-time interaction between fields, multi-field dynamic coupling simulation, including the following:

[0103] Chemical-hydraulic coupling: shallow layers Concentration couples the hydraulic field by affecting the permeability coefficient, introducing the dynamic permeability coefficient of the shallow hydraulic field. ;

[0104] ;

[0105] in, dimensionless The sensitivity coefficient, which characterizes the permeability coefficient, is affected by the change in the sensitivity coefficient. Sensitivity to concentration changes; For the ground concentration, The initial permeability coefficient of shallow layers, For shallow background concentration; For soil Saturation concentration, which is the characteristic chloride ion concentration in shallow soil, can be obtained through soil sample testing or historical data statistics.

[0106] Coupling of mechanics, thermodynamics and hydraulics:

[0107] The deep lithology is mainly sandstone and limestone, which differs significantly from the seepage characteristics of the shallow silty clay; its dynamic permeability coefficient is independent of... The concentration is not determined by the dynamic influence of deep pore water pressure and formation temperature, but rather by the dynamic permeability coefficient introduced into the deep hydraulic field. ;

[0108] ;

[0109] in, The initial permeability coefficient at depth; The pore pressure sensitivity coefficient characterizes the degree to which permeability changes with effective stress. The temperature sensitivity coefficient characterizes the degree to which the permeability of a hydraulic field changes with temperature. , These are the real-time deep pore water pressure and temperature, respectively. , For reference pore water pressure and temperature.

[0110] Preferably, in the initial model construction stage, in this embodiment, the specific values ​​of the initial values ​​of the key parameters can be initially determined based on local soil test data. The specific method is as follows:

[0111] To ensure a high degree of consistency between the digital twin model and the real physical world, a method for determining multi-field dynamic coupling coefficients based on inversion of historical on-site monitoring data is proposed; the details are as follows:

[0112] Inversion objective: To determine key dynamic coupling coefficients, including but not limited to: Influence sensitivity coefficient ( ), pore pressure sensitivity coefficient ( Temperature sensitivity coefficient ( )wait;

[0113] Required data foundation: Historical monitoring data for at least one complete hydrological year for the target area, including hydrodynamic field data (such as... , ), chemical field data (such as ), mechanical field data (such as ) and thermal field data (such as );

[0114] The inversion method process is used to determine the sensitivity coefficient of Cl⁻ influence. For example:

[0115] Data filtering and matching: From the historical database, filter out N sets of data that correspond in time. , Data pairs. Among them, This is the equivalent permeability coefficient of the shallow aquifer during this period, which can be obtained by back-calculating Darcy's law using water level change data and pumping volume data during this period.

[0116] Establish an inversion optimization model: Transform the problem of determining the coupling coefficients into a nonlinear optimization problem, the goal of which is to find the optimal... This minimizes the overall error between the permeability coefficient calculated by the model and the equivalent permeability coefficient calculated on-site.

[0117] Define the objective function and construct the least squares objective function:

[0118] ;

[0119] Where n = 1...N; Shallow background concentration;

[0120] Finding the optimal solution: An optimization algorithm (such as the Levenberg-Marquardt algorithm or particle swarm optimization) is used to iteratively solve the objective function. When the objective function J reaches its minimum value, the corresponding... The value is the optimal value obtained through inversion from on-site data. Influence sensitivity coefficient.

[0121] Similarly, this inversion method can be applied to the determination of other coupling coefficients, and the inversion... , At that time, filter Data pairs, based on the formula:

[0122] Construct the objective function and solve it.

[0123] As an alternative, this embodiment selects a single numerical value as the initial value for illustration, shallow layer =0.5; =200mg / L, , .

[0124] In this embodiment, a lower limit for the permeability coefficient is set. (Pick ),avoid Excessive concentration leads to The physical contradiction that affects the sensitivity coefficient to approach 0 (soil pores will not be completely closed, and the permeability coefficient has a minimum value).

[0125] Thermochemical coupling: Temperature couples with the chemical field by affecting the corrosion rate, and the introduction of shallow and deep temperatures accelerates the corrosion rate. , :

[0126] ;

[0127] ;

[0128] in, , The shallow and deep layers are respectively at the reference temperature ( The corrosion rate at room temperature under normal conditions; the shallow temperature-accelerated corrosion rate is related to the calibrated soil temperature monitored by the UAV, and the shallow temperature-accelerated corrosion rate is related to the deep formation temperature monitored downhole. Shallow temperature corrosion coefficient, The deep-layer temperature corrosion coefficient characterizes the effect of temperature on the chemical field reaction rate. The initial values ​​in this example are for shallow layers. =0.1 deep = .

[0129] Hydraulic field and mechanical field (coupling of pore water pressure and effective stress): Changes in pore water pressure directly alter the effective stress. Formula: ;

[0130] The effective stress change This refers to the change in pore water pressure. When a control operation (pumping / recharge) is performed, the pore water pressure will change, resulting in a change in pore water pressure, which in turn changes the effective stress, causing changes in settlement and altering the mechanical field, thus triggering new control decisions.

[0131] S203, Numerical Iteration and Risk Prediction: This method utilizes a multidimensional dynamic dataset to drive model derivation, employing finite element software (such as COMSOL) for spatiotemporal iterative calculations. With a time step of 1 hour, iterates at least 24 times to solve multiple field control equations, predicting the future spatiotemporal distribution of multiple field parameters. The predicted values ​​of physical field parameters for the next 24 hours are output (time resolution 1 hour, spatial resolution...). Hydraulic field: water level Mechanical field: effective stress Thermal field: thermal stress Chemical field: concentration Corrosion rate The above predicted values ​​include both shallow and deep layers.

[0132] During the iteration process, with The sensitivity coefficient, pore pressure sensitivity coefficient, temperature sensitivity coefficient, and temperature corrosion coefficient are the core coupled variables, which drive each other to update.

[0133] In this embodiment, the convergence condition is the change in water level over adjacent time steps. Stress change Convergence is determined at that time.

[0134] S204. Quantify the evolution trend of key risk indicators to generate a risk evolution field, specifically:

[0135] 1) Map the predicted physical field parameters to independent risk assessment values:

[0136] ;

[0137] in, Let J be the parameter value of the i-th field and j-th monitoring point; The safe interval for the i-th field parameter; The limiting threshold of the i-th field parameter (if exceeded) =1), setting safe ranges for shallow / deep parameters respectively.

[0138] Preferably, i=h represents the hydraulic field, i= i represents a mechanical field, i=v represents a chemical field. Representing a thermal field, when there are more than one risk indicator under the physical field, risk assessments are performed on each indicator separately, and the maximum value of the two indicators is taken as the independent risk assessment value of the physical field. For example, i=v corresponds to a chemical field, and the core risk indicator is... For concentration and corrosion rate, calculate the independent risk assessment values ​​for each parameter, and take the maximum value of the two indicators, i.e., as long as... If either the concentration or the corrosion rate exceeds the limit, the chemical field risk is high.

[0139] 2) Map the dimensionless independent risk assessment value to The risk levels are shown in Table 1 to address the problems of abstract evaluation values, poor visualization, and incomparability between fields in existing technologies. By using spatial interpolation (Kriging interpolation), the risk levels of discrete monitoring points are extended into a continuous spatiotemporal distribution field, forming a three-dimensional risk evolution model of "location-time-risk level".

[0140] Table 1

[0141]

[0142] S300. Calculate the coefficient of variation of each physical field based on the risk evolution field, and dynamically assign decision weights to each field based on the coefficient of variation and the scene characteristics of the target area to generate a comprehensive risk value. This includes the following steps:

[0143] S301. Calculation of Coefficient of Variation: For each physical field in the risk evolution field, calculate the independent risk assessment value for all monitoring points. Calculate the average value of the physical field risk. and standard deviation ;

[0144] = / ;

[0145] Obtain the coefficient of variation of the physical field. The high coefficient of variation indicates that the physical field has both a high average risk level and high spatial heterogeneity, making it a key field that requires priority intervention.

[0146] S302. Based on the coefficient of variation, the basic weights are dynamically allocated and then adjusted according to the characteristics of the scenario (such as urban areas and sensitive areas related to people's livelihoods) to ensure that high-risk fields and fields related to people's livelihoods are given priority in regulation.

[0147] Basic weight calculation (coefficient of variation → basic weight):

[0148] ;

[0149] Scenario Adjustment (Base Weight → Final Weight):

[0150] ;

[0151] In the formula, The scene correction coefficient for the i-th physics field, and the final weight. This information is then transferred to the comprehensive risk calculation.

[0152] If the target area is a coastal area. =1.2, based on coastal areas The corrosion accident rate has been increased by 20%, especially if the target area is an urban area and a sensitive area for people's livelihood (such as the old city). =1.2, = , =1, based on the risk of urban building settlement, the weights of the hydraulic field (preventing foundation settlement) and the mechanical field (preventing building cracking) are increased.

[0153] Preferably, a hierarchy importance coefficient is introduced. Such as shallow residential areas =1.2, deep engineering area .

[0154] S303. Calculate the comprehensive risk value across multiple events and determine the risk level:

[0155] ;

[0156] in, This represents the average independent risk assessment value for a single game. .

[0157] The coefficient of variation is a standardized, dimensionless measure of dispersion that can eliminate the incomparability of different physical fields (such as temperature and stress) due to differences in dimensions and numerical scales. This makes the weight allocation based on risk space fluctuations more fair and accurate, which cannot be achieved by using absolute dispersion indicators such as standard deviation or variance alone.

[0158] Those skilled in the art will understand that by using the average risk value simultaneously in the calculation of the coefficient of variation and the weighted summation of the comprehensive risk value, the method of this application can achieve a dynamic decision prioritization that is highly sensitive to risk patterns. This mechanism not only focuses on the average level of risk, but also prioritizes responses to physical fields with extremely uneven spatial distribution (high standard deviation) or local anomalies (low average value leading to high coefficient of variation) in an overall low-risk context. This enables early warning and precise intervention for potential risk points, avoiding decision lag or misjudgment that may occur with traditional weighted methods.

[0159] S400: Match the comprehensive risk value with the risk warning threshold to determine the risk assessment result; generate coordinated control instructions for shallow and deep aquifers based on the risk assessment result.

[0160] S401, Comprehensive Risk Value The risk assessment results are matched with the risk warning threshold to determine the overall risk level, risk area and key control indicators. If a high risk level or medium risk level is found, tiered regulation is initiated; otherwise, the status quo is maintained.

[0161] In some embodiments, three risk warning thresholds (0.3, 0.6, 0.8) are set according to the water level deviation and allowable deviation, and the comprehensive risk level is divided as shown in Table 2 to help locate high-risk areas (e.g., "the southeast corner of the foundation pit is judged to be a high-risk area due to the high concentration of Cl⁻ and the sudden increase in the hydraulic variation coefficient").

[0162] Table 2

[0163]

[0164] The overall risk value is the weighted sum of the risk assessment values ​​of each physical field. Due to the weighting, its threshold range differs from that of the individual field risk assessment value.

[0165] Preferably, before generating coordinated control commands for shallow and deep aquifers based on the risk assessment results, determining to initiate stratified control specifically includes the following:

[0166] The comprehensive risk value is matched with the risk warning threshold to determine the comprehensive risk level. Based on the comprehensive risk level, the overall severity of the risk across the entire region is determined, thereby confirming whether comprehensive regulation is needed and the corresponding regulatory intensity. The priority of regulation is determined based on the risk level of each individual event. Specific regulatory parameters are determined by combining the risk level of each individual event with the comprehensive risk level. For example... =0.7 (medium risk), hydraulic field level 3, mechanical field level 5, then regulation is required, and deep recharge should be initiated first (to prevent mechanical field level 5 subsidence), and then shallow pumping should be adjusted (to control the water level of hydraulic field level 3).

[0167] In this embodiment, the corresponding control intensity is determined based on the comprehensive risk level, specifically as follows:

[0168] A low overall risk level indicates that the overall risk is not exceeding the standard but vigilance is required, and full-capacity regulation is not necessary; a high overall risk level indicates that the overall risk is on the verge of getting out of control and full-capacity rapid intervention is required; a high overall risk level indicates that the overall risk is out of control and intervention beyond the designed capacity is required.

[0169] Based on the risk level of a single event, the priority of regulation is determined, and the specific rules are as follows:

[0170] Grade difference At level 1, the proportion of high-level field regulation resources ;

[0171] Grade difference At level 1, the proportion of high-level field regulation resources ;

[0172] Even if the risk level of a sensitive area (old city area, school) is one level lower, it can still share 50% of resources with high-risk areas.

[0173] Preferably, the control details are modified based on the characteristics of the target scenario, as shown in Table 3, adjusting the time, scope, emergency plan, etc.

[0174] Table 3

[0175]

[0176] S402. Generate coordinated control commands for shallow and deep aquifers respectively, specifically including:

[0177] 1) Shallow pumping command (hydraulic field control):

[0178]

[0179] in, The dynamic permeability coefficient, This is the current shallow water level; This is the target value for shallow water level; , This refers to the shallow drainage area and the drainage influence radius. The lag compensation coefficient is based on the feedback seepage lag time. Fitting, seepage hysteresis time The pumping flow rate is calculated by using the time difference between the issuance of the pumping command and the peak value of the water level response curve to actively offset the seepage hysteresis effect; It is necessary to combine the dynamic permeability coefficient of the shallow hydraulic field. Water level deviation and seepage hysteresis compensation To avoid over-adjustment of water level.

[0180] In some embodiments, ,like 2h, then =0.04. For the hydraulic field level coefficient, the system is based on The risk level corresponding to the physical field in Table 1 is matched, and an initial value is assigned to it from a preset mapping table. Specifically: The corresponding risk level is Level 1. =0.1; The corresponding risk level is level 2. =0.3; The corresponding risk level is level 3. =0.5; The corresponding risk level is level 4. =0.8; The corresponding risk level is level 5. =1.1.

[0181] 2) Deep reinjection command (mechanical field control):

[0182] ;

[0183] ;

[0184] in, The allowable drop threshold for deep water levels; Settlement compensation coefficient (based on fitting of reinjection data, empirical value 0.5~0.8, adjusted based on subsequent feedback settlement data); For deep pore water pressure, This represents the target value for deep pore water pressure. , The deep recharge area and the radius of influence of the recharge; the recharge volume. It is necessary to compensate for the drop in deep water level caused by shallow pumping to avoid ground subsidence; The vertical distance from the wellhead of the reinjection well to the center of the aquifer. Where L is the friction factor (0.02~0.03 for steel pipes), D is the pipe length (m), and V is the reinjection velocity (m / s). =0.025, L=100m, d=0.2m).

[0185] Calculated reinjection pressure Used to verify whether the head of the selected reinjection pump meets the requirements, and as the upper limit for safe monitoring of equipment operation.

[0186] In this embodiment, For mechanical field level coefficients, the system is based on The risk level corresponding to the physical field in Table 1 is matched, and an initial value is assigned to it from a preset mapping table. Specifically: The corresponding risk level is Level 1. =0.1; The corresponding risk level is level 2. =0.3; The corresponding risk level is level 3. =0.5; The corresponding risk level is level 4. =0.8; The corresponding risk level is level 5. =1.1.

[0187] Preferably, the control command is transmitted to the execution unit to control the water pump and the reinjection pump to regulate the water level.

[0188] Preferably, after generating the control command, a feedback step is also included, specifically including:

[0189] The feedback data after the collection command is executed is used to optimize the model parameters through a two-level closed-loop feedback mechanism to form a closed-loop control. The first loop calculates parameters through parameter calibration control commands, while the second loop optimizes and updates the core physical parameters in the digital twin model through field data inversion methods. When the deviation between the feedback data and the predicted data at the corresponding time node in the risk evolution field exceeds the preset deviation threshold, the second loop is activated.

[0190] S501. Collect feedback data after the execution of the command, including equipment control commands, control logs, on-site measured data (water level, pore water pressure, surface deformation, and operating energy consumption 1-24 hours after command execution) and equipment operating status feedback.

[0191] On-site measured data includes actual pumping flow rate. Actual reinjection volume Water level after regulation After regulation concentration Corrosion rate after adjustment Thermal stress after regulation Adjusting pore water pressure Based on the regulated thermal stress and regulated pore water pressure Effective stress after regulation .

[0192] After the control command is output, the generated specific command is added to the simulation calculation as a new boundary condition or source / sink term. The model then extrapolates how each physical field should respond within a future period (e.g., 72 hours) under strict execution of this new command, updating the predicted value as the target value.

[0193] Hydraulic field: water level Mechanical field: effective stress Thermal field: thermal stress Chemical field: concentration Corrosion rate The above predicted values ​​include both shallow and deep layers.

[0194] Equipment control instructions include three core types of instructions:

[0195] Pumping / recharge wells: target flow rate ( ), start and stop times (e.g., 8:00-16:00);

[0196] Control valve: target opening degree (%), adjustment frequency (e.g., fine-tuning once every 30 minutes), maximum allowable engineering error;

[0197] Emergency Command: If the risk level suddenly rises to "high", trigger the emergency command "maximum pumping + standby pump start".

[0198] The control log includes the time of instruction generation, the basis (e.g., "Based on the 8:00 risk evolution field, the water level in irrigation area A is below 25m, start recharge well No. 2"), and the initial state of the execution unit, which facilitates subsequent traceability and feedback optimization.

[0199] S502, Optimize model parameters through a two-level closed-loop feedback mechanism:

[0200] 1) Calculate the single-field deviation between the feedback data and the target data predicted at the corresponding time node in the risk evolution field.

[0201] Calculate one-dimensional bias:

[0202] Water level deviation: ; Concentration deviation: ;

[0203] Thermal stress deviation: Mechanical field deviation: .

[0204] Similarly, corresponding absolute errors can be defined for other physical fields. In multi-physics coupling control, if it is necessary to comprehensively evaluate the absolute errors of different dimensions, they can be standardized or weighted. However, in precision control, the focus is usually on directly comparing the absolute errors of each physical field with their respective engineering allowable error ranges.

[0205] The absolute error is converted into a dimensionless normalized value, representing the severity of the deviation relative to the allowable range:

[0206] ;

[0207] in: Normalization deviation of the i-th physical field The allowable engineering error for the i-th physical field (e.g.) =0.01m).

[0208] 2) Activate a two-level closed-loop feedback mechanism:

[0209] The first level is the instruction feedback loop (reverse control): This loop is designed to quickly respond to instantaneous deviations, directly adjusting the calculation parameters in the control instruction generation formula without changing the digital twin model itself. It has a high update frequency, typically evaluated in each control cycle (e.g., hourly).

[0210] 1) Update of the lag compensation coefficient:

[0211] The seepage lag time is updated by recalculating the new seepage lag time based on the water level change curve fed back from the field, using signal processing methods (such as cross-correlation analysis). This leads to the updating of the lag compensation coefficient:

[0212] ;in, This is an empirical coefficient, for example, 0.02. This is a direct update, overwriting the old value with the newly observed physical properties.

[0213] 2) Update of scene correction coefficients:

[0214] Poor control effects in specific scenarios (such as areas sensitive to people's livelihoods, where data on land subsidence and water level fluctuations fail to meet standards for more than one consecutive control cycle) indicate that the scenario correction coefficient... Improper settings require adjustment; for example, if the risk control of the mechanical field in sensitive areas is not ideal, the correction coefficient of the mechanical field scenario in sensitive areas should be adjusted. :

[0215] ;in, An empirical coefficient, such as 0.2, can be used to automatically increase the focus on key fields (such as mechanical fields) when the total risk prediction is too low.

[0216] 3) Grade coefficient ( , The dynamic adaptive adjustment of ) eliminates steady-state error. The specific method is as follows:

[0217] Iterative fine-tuning is performed using the proportional-integral (PI) control approach:

[0218] ;

[0219] ;

[0220] in, These are the initial shallow hydraulic field level coefficients and mechanical field level coefficients. , The gain is adjusted proportionally to the preset value. It is the preset integral adjustment gain; where To adjust the time step of the control cycle, it is 1 hour in this embodiment. This represents the error in the past time step. The cumulative amount.

[0221] Output: Updated instruction calculation parameters , , , , The instructions will be immediately used to generate the next control cycle, enabling rapid closed-loop correction.

[0222] Preferably, if the current fine-tuning range of the grade coefficient is much smaller than the minimum flow rate adjustment accuracy of the water pump, it will not be updated by default and no adjustment will be output externally.

[0223] During the iterative fine-tuning process, only when "the single-dimensional deviation of two consecutive control cycles (1 cycle = 24h) is greater than or equal to..." The integral term is activated only when the condition is met; otherwise, the integral term is reset to zero, and only the proportional term is retained.

[0224] More preferably, in this embodiment, the values ​​of the proportional adjustment gain and integral adjustment gain are dynamically adjusted according to the risk level; the dynamic adjustment meets the following rules: for low risk or safe levels, smaller parameters are used (prioritizing stability and avoiding overshoot); for medium or high risk levels, larger parameters are used (prioritizing fast response and suppressing risk); and for sensitive areas, the parameters are 20% higher than those for ordinary areas at the same level (prioritizing prevention of subsidence or sudden surge).

[0225] The second level is the model evolution loop (data inversion): This loop aims to address the systematic biases of the model itself. Through periodic (such as weekly or monthly) data inversion, it optimizes the core physical parameters of the digital twin model and achieves the model's self-evolution.

[0226] The system pre-determines the engineering tolerances for each key physical field (e.g., water level tolerance). (±0.01 m, or 1 cm). When the absolute error of any key physics field continues or significantly exceeds its corresponding preset allowable error threshold... (or a portion of the allowable error threshold, such as 50%) When the system determines that the current model has systematic errors or that the external environment has changed significantly, it will initiate a model evolution loop. At this point, the system will use the latest monitoring data and an inversion algorithm to iteratively update and correct the model's core physical parameters (such as permeability coefficient, initial conditions, boundary conditions, etc.) to make the model predictions more realistic. If the absolute errors of all key physical fields are within the allowable error threshold, the model evolution loop will not be initiated, and the model parameters will remain unchanged.

[0227] Preferably, the model evolution loop is initiated, specifically as follows:

[0228] Historical feedback data for one cycle is collected. The specific inversion process adopts the method described in S202 of this application, which is based on historical monitoring data and least squares optimization. Measured data from a new cycle is used as input, and the solution is recalculated as follows: Influence sensitivity coefficient ( , , , , Core physical parameters such as permeability coefficients are used to achieve iterative evolution of the model. For example, if long-term monitoring reveals that the predicted deviation of the deep permeability coefficient exceeds 5%, the model is re-inverted. , Update the dynamic permeability coefficient of the deep hydraulic field The coefficients in the calculation formula are used to achieve a fundamental evolution of the model.

[0229] The following section uses a deep foundation pit project (excavation depth 28m) for a station on Metro Line 2 in a first-tier city as the target area, and combines actual engineering parameters and monitoring data to provide a detailed explanation of the dynamic groundwater level control method of this invention. The foundation pit is located in an old urban area, with sensitive buildings within 50m of it. The site has complex geology, with the shallow layer mainly consisting of… Layered fill and The first layer is silty clay with poor permeability and high sensitivity; the deeper layer is a confined aquifer with hydraulic connection to the foundation pit; the main risks facing the project are: excessive settlement of the surrounding ground and heave of the pit bottom caused by improper dewatering.

[0230] The control method in this embodiment is executed in five stages: "data acquisition → model construction (including coupling coefficient inversion) → risk quantification (including variation coefficient weighting) → control execution → two-level feedback". The specific data and operations for each step are as follows:

[0231] S100: Multi-source monitoring data acquisition and preprocessing;

[0232] Monitoring data was collected for 72 consecutive hours in June 2024 (high water season), as follows:

[0233] Satellite remote sensing: =0.8mm / d, P=8mm / d;

[0234] Drone thermal infrared: , ;

[0235] Ground sensor network (20 monitoring points deployed): 11.8-12.3m (average 12.05m) 120-150 kPa (foundation stress of surrounding old buildings) 330-370 mg / L (average 350 mg / L) 30-33 (mean 31.5) ), 118-125 kPa (average 121 kPa);

[0236] Downhole intelligent monitoring equipment (3 monitoring wells, depth 60m): 45-47m (average 46m) 480-520 mg / L (average 500 mg / L) 175-185 kPa (average 180 kPa) 23-24 (mean 23.5) ).

[0237] Data preprocessing: using The criteria exclude one set of abnormal water level data. =8.5m, determined to be a sensor fault), and the missing data from the past 2 hours was completed using an LSTM model. Concentration data are ultimately used to form a spatiotemporal four-dimensional dataset in the form of "WGS84 coordinate system + UTC timestamp".

[0238] S200: Construction of multi-field dynamic coupling model (including dynamic coupling coefficient inversion);

[0239] S201: Construction of basic governing equations;

[0240] The fundamental equations for the hydraulic, mechanical, chemical, and thermal fields remain within the same theoretical framework. Substituting parameters specific to this region (derived from geological survey reports and historical monitoring data), the following are applied: = m / d, =8× m / d, =8000kPa =600mg / L =180kPa, .

[0241] S202: Dynamic coupling coefficient inversion;

[0242] 1) Targeting key coupling coefficients , , Historical data from the complete hydrological year of 2023 (June 2023 to May 2024) were used for inversion:

[0243] Inversion : Filter 30 groups { , } data pairs, where Calculate using Darcy's Law (e.g., a certain period in August 2023:) The pumping depth decreased from 12.5m to 11.8m, and the pumping volume... =20 / d, reverse calculation =2.2× m / d);

[0244] Inversion , : Filter 25 groups { , , Data pairs, The range of the inverse calculation value is 7.5× ~8.5× m / d.

[0245] 2) Construction and solution of the objective function:

[0246] Inversion The objective function is:

[0247] ;

[0248] in, For the background concentration, the particle swarm optimization (PSO) algorithm is used iteratively to solve the problem. When it drops to the minimum value, we get =0.045L / mg;

[0249] Similarly, the inversion is obtained , .

[0250] 3) Dynamic coupling equation update:

[0251] Substitute shallow permeability coefficient =350mg / L:

[0252] ;

[0253] Substituting the deep permeability coefficient =180kPa 23.5 :

[0254] .

[0255] S203: Risk Evolution Field Generation:

[0256] Using COMSOL software with a time step of 1 hour and 24 iterations, predict the physical field parameters for the next 24 hours:

[0257] Hydraulic field: Shallow predicted water level =11.9−12.4m, deep predicted water level =45.8−46.2m; Mechanical field: Predicted effective stress =142kPa kPa, total stress Determined by geological conditions, it will remain unchanged in the short term; =322-180kPa), predicted settlement =0.3−0.5mm / d; Chemical field: Predicting shallow layers concentration =340−360mg / L, predicted corrosion rate =0.02−0.03mm / year; Thermal field: Predicting shallow layer temperature =31−32℃, predicted thermal stress =5−8kPa.

[0258] S300: Weighting of Coefficient of Variation and Calculation of Overall Risk Value:

[0259] 1) Calculate the independent risk assessment value and coefficient of variation;

[0260] Based on the safety thresholds of each physical field (e.g., the shallow water level safety range [10,13]m, and the ultimate threshold 8m), calculate the independent risk assessment values ​​for 20 monitoring points. Then calculate the mean. with standard deviation Calculate the coefficient of variation As shown in Table 4:

[0261] Table 4

[0262]

[0263] 2) Assign dynamic weights;

[0264] Basic weight calculation: ;

[0265] Hydraulic field: 0.286 / 1.242 0.230; Mechanical field: 0.429 / 1.242 0.345; Chemical field: 0227 / 1.242 0.183; Thermal field: 0.3 / 1.242 0.242;

[0266] Scenario Correction: Since the foundation pit is located in a sensitive area affecting people's livelihoods, a scenario correction factor is used for hydraulic (to prevent settlement caused by a sudden drop in water level). 1.25, mechanics (Preventing cracking in old buildings) is 1.2, chemical 1. Heat Set the value to 1, and substitute it into the final weight: ;

[0267] Hydraulic field:

[0268] Mechanical field: Chemical field: Thermal field: ;

[0269] Overall risk value calculation:

[0270] .

[0271] S400: Generate control commands:

[0272] Based on risk threshold (safety) (Low risk 0.3-0.6) =0.302 is determined to be a low-risk level, requiring the maintenance of stable shallow water levels to prevent the risk of sudden surges during the high-water season. A control instruction is generated:

[0273] Shallow extraction instructions:

[0274] ;

[0275] Substitute parameters: =0.3 (The risk level of the water academy is level 2). =500 (Extraction area) , =11.5m (target water level) =20m (radius of influence) =0.05 (lag compensation coefficient) ;

[0276] (Right now Conventional high-flow pumps could not meet the fine-tuning requirements, so a small precision water pump with an accuracy of [insert accuracy here] was used instead. Only when the flow changes Only then can the pump operate stably.

[0277] Deep reinjection command: Maintain stable deep pore water pressure, reinjection volume = .

[0278] , , (and actual measurement) Consistent) ; ;

[0279] S500: Dual-level closed-loop feedback optimization;

[0280] 1) Collect measured data 24 hours after the command is executed:

[0281] Within 72 hours of the command execution, the system continuously monitors the changes in each physical field at a high frequency (e.g., every 15 minutes). Because the target soil layer is silty clay with extremely poor permeability, the hydraulic field response is very slow, which is an expected phenomenon consistent with physical laws.

[0282] based on Based on the parameters and the multi-field coupling model prediction, the water level should drop to 12.043 m 72 hours after the command is executed (predicted drop of 7 mm).

[0283] 72 hours later, the system aggregated the data and found that:

[0284] shallow water level The pressure of the 144th group decreased from 12.05m to 12.045m (the peak response occurred 3 hours after the command was issued). At the same time, the mechanical field monitoring showed that the deformation rate of the retaining structure slowed down by 0.02 mm / d, which proved that the micro-drainage played a positive role in stabilizing the soil pressure.

[0285] The system calculates the error between the predicted and actual values, specifically the water level deviation.

[0286] =12.045-12.043=0.002m; =0.2;

[0287] 2) Two-level feedback optimization:

[0288] in conclusion: =0.002m =0.01m) (maximum permissible deviation), the trigger adjustment threshold is not reached, and the "model evolution loop" is not started (no need to invert and update the coupling coefficient).

[0289] Adjustment to the "Command Feedback Loop":

[0290] Update the lag compensation coefficient (based on cross-correlation analysis) =3h, =0.06), no change in reinjection volume (still 0) (original) =0.05, because τ is extended, the compensation coefficient needs to be increased to avoid water level overshoot due to lag).

[0291] The control measures have achieved the desired effect; the scenario remains unchanged; no updates are needed. ;

[0292] As an example, in this specific embodiment, the settings are as shown in Table 5.

[0293] Table 5

[0294]

[0295] More preferably, the above learning parameters are all based on conventional values ​​from engineering practice and can be adjusted according to the needs of the target area.

[0296] Assuming the overall risk value and risk level remain unchanged, and given that this is the first deviation to occur within this adjustment cycle, the integral term is simplified to the current value. Calculations show that:

[0297] ;

[0298] The purpose of PI fine-tuning is to "optimize the control precision," but it must meet the requirement that "the adjustment has a perceptible impact on the actual control effect." The current fine-tuning magnitude should reflect this change. Based on changes Change Less than the minimum flow rate adjustment accuracy of the water pump (the accuracy of a conventional water pump is...). No update; the reinjection volume remains 0, and the change is less than the minimum flow rate adjustment accuracy of the pump, so no update is needed. .

[0299] Implementation effect verification: Through dynamic coupling coefficient inversion, coefficient of variation weighting, and two-level feedback, the following was achieved: the shallow water level was stabilized at 11.5-11.7m, with fluctuation range... The settlement of surrounding old buildings is controlled at 0.3-0.35 mm / d, which is far below the standard limit of 1 mm / d. The corrosion rate is maintained below 0.02 mm / year, meeting the engineering durability requirements.

[0300] Ultimately, it achieves the goal of "ensuring maximum engineering safety with minimal control costs," providing a feasible technical paradigm for groundwater level control in complex scenarios such as deep foundation pits in old urban areas and sensitive areas affecting people's livelihoods.

[0301] Although the steps in the above embodiments are described in the above order, those skilled in the art will understand that in order to achieve the effect of this embodiment, different steps do not need to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple variations are all within the protection scope of this invention.

[0302] The second embodiment of the present invention is a digital twin-based dynamic groundwater level control system, the system comprising:

[0303] The data acquisition module is configured to collect multi-source monitoring data of the groundwater system in the target area and construct a multi-dimensional dynamic dataset containing hydraulic, mechanical, chemical and thermal parameters.

[0304] The twin simulation and prediction module is configured to construct a multi-field dynamically coupled digital twin model. It uses a multi-dimensional dynamic dataset to drive the multi-field dynamically coupled digital twin model to perform simulation, simulate the state response of the groundwater system under the influence of multiple fields, predict the future spatiotemporal distribution of multiple field parameters, and then quantify the evolution trend of key risk indicators to generate a risk evolution field.

[0305] The data processing module is configured to calculate the coefficient of variation of the risk values ​​of each physical field based on the risk evolution field, and dynamically allocate decision weights to each physical field based on the coefficient of variation and the scene characteristics of the target area to generate a comprehensive risk value.

[0306] The stratified control module is configured to match the comprehensive risk value with the preset risk warning threshold to determine the risk assessment result; and generate coordinated control instructions for shallow and deep aquifers respectively based on the risk assessment result.

[0307] Preferably, the system further includes a feedback optimization module, configured to collect feedback data after the command is executed, and optimize the calculation parameters of the control command and the core physical parameters of the digital twin model through a two-level closed-loop feedback mechanism to form closed-loop control.

[0308] More preferably, the feedback optimization module specifically includes:

[0309] The physical model calibration unit is configured to start when the overall deviation exceeds a preset deviation threshold, periodically collect historical monitoring data within a control cycle, and optimize and update the core physical parameters in the digital twin model through data inversion methods. The core physical parameters are at least one dynamic coupling coefficient to realize the periodic learning and evolution of the model.

[0310] The prediction model optimization unit is configured to quickly adjust the calculation parameters of the control command based on the real-time deviation between the measured physical quantity and the prediction data in the feedback data, so as to correct the deviation in real time and ensure the accurate execution of the command; the calculation parameters include the hysteresis compensation coefficient and the grade coefficient.

[0311] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process and related explanations of the methods described above can be found in the corresponding processes in the foregoing system embodiments, and will not be repeated here.

[0312] It should be noted that the above embodiments of the digital twin-based groundwater level dynamic control method and system are merely illustrative examples of the functional module divisions. In practical applications, the functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be further decomposed or combined. For example, the modules in the above embodiments can be merged into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are merely for distinguishing the various modules or steps and are not considered as an improper limitation of the present invention.

[0313] A device according to a third embodiment of the present invention includes:

[0314] At least one processor;

[0315] and a memory communicatively connected to at least one of the processors;

[0316] The memory stores instructions that can be executed by the processor to implement the aforementioned method for dynamic control of groundwater levels based on digital twins.

[0317] A computer-readable storage medium according to a fourth embodiment of the present invention stores computer instructions, which are executed by the computer to implement the above-described method for dynamic control of groundwater level based on digital twin.

[0318] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process and related descriptions of the storage device and processing device described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0319] The following is for reference. Figure 2 It shows a schematic diagram of the structure of a computer system for implementing embodiments of the systems, methods, and electronic devices of this application. Figure 2 The server shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0320] like Figure 2 As shown, the computer system includes a Central Processing Unit (CPU) 201, which can perform various appropriate actions and processes based on programs stored in Read Only Memory (ROM) 202 or programs loaded from storage section 208 into Random Access Memory (RAM) 203. The RAM 203 also stores various programs and data required for system operation. The CPU 201, ROM 202, and RAM 203 are interconnected via a bus 204. An Input / Output (I / O) interface 205 is also connected to the bus 204.

[0321] The following components are connected to I / O interface 205: an input section 206 including a keyboard, mouse, etc.; an output section 207 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 208 including a hard disk, etc.; and a communication section 209 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 209 performs communication processing via a network such as the Internet. A drive 210 is also connected to I / O interface 205 as needed. Removable media 211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 210 as needed so that computer programs read from them can be installed into storage section 208 as needed.

[0322] Specifically, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 209, and / or installed from removable medium 211. When the computer program is executed by central processing unit (CPU) 201, it performs the functions defined in the methods of this application. It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0323] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0324] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0325] The terms “first”, “second”, etc., are used to distinguish similar objects, not to describe or indicate a specific order or sequence.

[0326] The term "comprising" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent in such process, method, article, or apparatus / device.

[0327] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A digital twin-based underground water level dynamic regulation method, characterized in that, The method comprises the following steps: Collecting multi-source monitoring data of the groundwater system in the target area, and constructing a multi-dimensional dynamic data set; Constructing a multi-field dynamic coupling digital twin model, and driving the model with the multi-dimensional dynamic data set to predict the future spatio-temporal distribution of multi-field parameters, and then quantize the evolution trend of key risk indicators to generate a risk evolution field; Based on the risk evolution field, calculate the coefficient of variation of the risk value of each physical field, and based on the coefficient of variation, combine the scene characteristics of the target area to dynamically allocate decision weights for each physical field to generate a comprehensive risk value; Matching the comprehensive risk value with the risk warning threshold to determine the risk assessment result; according to the risk assessment result, generate a collaborative control instruction for the shallow and deep aquifers respectively.

2. The groundwater level dynamic regulation method based on digital twinning according to claim 1, characterized in that, Collecting multi-source monitoring data of the groundwater system in the target area, including: The target area is monitored by a satellite remote sensing device, a UAV thermal infrared device, a ground sensor network and an underground intelligent monitoring device to obtain multi-source and multi-scale monitoring data of the target area; wherein the satellite remote sensing device obtains original surface deformation rate data, the UAV thermal infrared device obtains soil temperature and illumination deviation data, the ground sensor network obtains water level, structural stress, water quality parameter, formation temperature and soil moisture content data, and the underground intelligent monitoring device obtains pore water pressure and deep water level data; the water quality parameter includes pH value and concentration. Concentration.

3. The groundwater level dynamic regulation method based on digital twinning according to claim 1, characterized in that, The method for constructing a multi-field dynamic coupling digital twin model is as follows: A multi-field dynamic coupling model including a hydraulic field, a mechanical field, a chemical field and a thermal field is constructed, and the mutual influence between at least two physical fields is quantified and updated in real time by introducing a dynamic coupling coefficient; wherein the hydraulic field parameters include shallow water level and deep water level, the mechanical field parameters include structure settlement and effective stress, the chemical field parameters include corrosion rate, and the thermal field parameters include thermal stress; the dynamic coupling coefficient includes influence sensitive coefficient, pore pressure sensitive coefficient, temperature sensitive coefficient and temperature corrosion coefficient.

4. The groundwater level dynamic regulation method based on digital twinning according to claim 3, characterized in that, Update the mutual influence between the two physical fields, specifically: according to the simulation output result of one physical field, dynamically update the key model parameters of another physical field; The update includes at least one of the following: According to the simulation results of solute concentration in the chemical field, the permeability coefficient value of the shallow aquifer is updated by using The sensitive coefficient update influences the permeability coefficient value of the shallow aquifer; According to the simulation result of the pore water pressure of the mechanical field, update the permeability coefficient value of the deep aquifer using the pore pressure sensitivity coefficient and the temperature sensitivity coefficient; According to the simulation result of the temperature of the thermal field, update the chemical reaction rate value of the chemical field using the temperature corrosion coefficient.

5. The groundwater level dynamic regulation method based on digital twinning according to claim 1, characterized in that, The generation of the comprehensive risk value specifically includes: For the state of the hydraulic field, the mechanical field, the chemical field and the thermal field in the risk evolution field, the independent risk evaluation value of each field at all monitoring points is respectively counted; Calculate the mean and standard deviation of the independent risk evaluation value of each physical field, and obtain the coefficient of variation of each physical field based on the mean and standard deviation; According to the coefficient of variation of each physical field, the basic decision weight is allocated in proportion, and the physical field with higher coefficient of variation is allocated with larger basic decision weight; Combine the scene characteristics of the target area to correct the decision weight; Calculate the average independent risk evaluation value of each physical field, and generate the comprehensive risk value by weighted summation based on the corrected decision weight.

6. The groundwater level dynamic regulation method based on digital twinning according to claim 1, characterized in that, Matching the comprehensive risk value with the risk warning threshold to determine the risk assessment result, the method is as follows: According to the engineering safety level of the target area, combine the operation condition or environmental condition to dynamically preset at least two risk warning thresholds, and divide different risk levels according to the risk threshold; Compare the comprehensive risk value with the risk threshold, and determine the risk level to which the comprehensive risk value belongs according to the comparison result; The risk level is taken as the risk assessment result, which is used to trigger the control strategy corresponding to the risk level.

7. The groundwater level dynamic regulation method based on digital twinning according to claim 6, characterized in that, According to the risk assessment result, generate a collaborative control instruction for the shallow and deep aquifers respectively, specifically: For the shallow aquifer, generate pumping instructions to maintain the dynamic water level stability of the key area or meet the engineering dewatering demand; For the deep aquifer, generate recharge instructions to provide regional pore water pressure support and slow down land subsidence.

8. The groundwater level dynamic regulation method based on digital twinning according to claim 1, characterized in that, It also includes feedback data after the execution of the instruction, and optimizes the system to form a closed-loop control through a double-level closed-loop feedback mechanism.

9. The groundwater level dynamic regulation method based on digital twinning according to claim 8, characterized in that, The double-level closed-loop feedback mechanism specifically includes the following contents: The first level is an instruction feedback loop: according to the real-time deviation of the measured physical quantity in the feedback data and the predicted data, the calculation parameters of the regulation and control instruction are quickly adjusted to real-time correct and ensure accurate execution of the instruction; the calculation parameters include a lag compensation coefficient and a level coefficient; The second level is a model evolution loop: when the comprehensive deviation exceeds the preset deviation threshold, it is started, the historical monitoring data in a regulation and control period is periodically collected, the core physical parameters in the digital twin model are optimized and updated through a data inversion method, the core physical parameters are at least one dynamic coupling coefficient, to realize the periodic learning and evolution of the model.

10. A digital-twin-based underground water level dynamic regulation system, characterized in that, The system comprises: A data acquisition module configured to acquire multi-source monitoring data of a target area groundwater system, and construct a multi-dimensional dynamic data set containing hydraulic, mechanical, chemical and thermal parameters; A twin simulation and prediction module configured to construct a multi-field dynamic coupling digital twin model, drive the multi-field dynamic coupling digital twin model using the multi-dimensional dynamic data set for simulation, simulate the state response of the groundwater system under the influence of multi-field interaction, predict the future spatio-temporal distribution of multi-field parameters, and then quantify the evolution trend of key risk indicators to generate a risk evolution field; A data processing module configured to calculate the variation coefficient of the risk value of each physical field based on the risk evolution field, and dynamically allocate decision weights for each physical field based on the variation coefficient and the scene characteristics of the target area to generate a comprehensive risk value; A hierarchical regulation and control module configured to match the comprehensive risk value with a preset risk warning threshold to determine a risk assessment result, and generate a collaborative regulation and control instruction for the shallow and deep aquifers according to the risk assessment result; A feedback optimization module configured to collect feedback data after the execution of the instruction, and optimize the calculation parameters of the regulation and control instruction and the core physical parameters of the digital twin model through a double-level closed-loop feedback mechanism to form a closed-loop regulation and control.

Images