CN-121680088-B - Internet of things-based underground water pumping and injecting integrated intelligent management and control method and system

Abstract

The invention discloses an underground water pumping and injecting integrated intelligent management and control method and system based on the Internet of things, which comprises the steps of carrying out geological investigation on a target area and constructing a three-dimensional geological model, planning the layout of pumping and injecting well groups and a sensor network according to the geological investigation, deploying a monitoring network based on the layout scheme, collecting data, fusing the three-dimensional geological model to construct a polluted site digital twin body, extracting underground water state data through the digital twin body, predicting pollutant migration paths and water level dynamics by using an LSTM model, outputting pollution scene simulation information corrected by geological features, taking the simulation information as input, coupling geological parameters as constraint, adopting BLMFO double-layer multi-target optimization algorithm to generate pumping and injecting well group cooperative control strategy set, finally executing control instructions, monitoring real-time state situation in the restoration process, judging and pushing strategy change suggestions, including flow adjustment amplitude, medicament optimization scheme and equipment operation parameter correction, improving restoration efficiency and adapting to complex geological environment.

Inventors

• WANG YANWEI
• WANG LINA
• CAI LINYING
• ZHANG HAO
• YUAN WENCHAO
• WANG HENGQIN
• Song Zongzhong

Assignees

• 生态环境部土壤与农业农村生态环境监管技术中心

Dates

Publication Date

20260508

Application Date

20260205

Claims (10)

1. 1. The underground water pumping and injecting integrated intelligent control method based on the Internet of things is characterized by comprising the following steps of: performing geological investigation on a target area, constructing an area three-dimensional geological model, and performing pumping and injecting well group and heterogeneous sensor monitoring network layout planning by using the area three-dimensional geological model to obtain an area pumping and injecting repairing layout scheme; the heterogeneous sensor monitoring network is deployed based on the regional pumping and injecting restoration layout scheme, regional data acquisition is carried out, spatial distribution data and water quality monitoring data of regional groundwater are obtained, and the regional groundwater is fused with a regional three-dimensional geological model to establish a digital twin body of a polluted site; extracting groundwater state data of a target area through the digital twin bodies of the pollution sites, setting geological condition constraint, carrying out dynamic prediction on a pollutant migration path and a water level by using an LSTM prediction model, and outputting pollution scene simulation information corrected by geological features; taking the pollution scenario simulation information as input, taking the coupling permeability coefficient and the fracture distribution parameter as constraint conditions, and carrying out multi-objective collaborative optimization calculation through BLMFO double-layer multi-objective optimization algorithm to generate a pumping and injecting well group collaborative control strategy set; Generating a control instruction sequence based on the pumping and injecting well group cooperative control strategy set, sending the control instruction sequence to a corresponding executing mechanism for executing, monitoring a real-time operation situation in the pumping and injecting repairing process, judging whether repairing strategy change is needed, and generating repairing strategy change suggestions for pushing; Integrating a solar energy storage module consisting of a photovoltaic panel array, a lithium ion storage battery pack and an intelligent bidirectional converter in a pumping and injecting integrated equipment system, optimizing the power generation efficiency of the photovoltaic panel in real time through a maximum power point tracking technology, monitoring the state of charge, the health state and the residual capacity parameters of the storage battery by using a battery management system, and establishing an energy data acquisition system comprising solar power generation prediction, energy storage state monitoring and power grid electricity price information; Acquiring meteorological data, irradiance monitoring values and photovoltaic power generation power data in a preset period through an energy data acquisition system to generate an initial energy data set, and fusing the meteorological data and the irradiance monitoring values based on the initial energy data set to generate a fusion characteristic sequence; Dividing the fusion characteristic sequence into a plurality of characteristic sequence segments according to a preset time step, extracting photovoltaic power generation power data through the initial energy data set, generating a photovoltaic power generation power sequence, correlating the photovoltaic power generation power sequence with the characteristic sequence segments to form a correlation characteristic sequence, and representing the correlation relation between the environment state and the power generation power in one monitoring period by the correlation characteristic sequence; Generating an associated feature sequence set by the associated feature sequences corresponding to the whole monitoring periods, dividing one monitoring period into a plurality of time nodes, constructing a state space by the associated feature sequence set, calculating state transition probabilities and state transition matrixes among different time nodes in the state space, combining the time nodes with the similarity of meteorological data and irradiance monitoring values by similarity calculation, and finally obtaining an energy feature sample set; Defining each time step with preset size as a single node based on the energy characteristic sample set, taking corresponding meteorological characteristics and irradiance monitoring values as node characteristics of the corresponding nodes, taking photovoltaic power generation power as node attributes, and generating a plurality of power generation situation paths taking one monitoring period as time length, wherein the power generation situation paths represent different photovoltaic power generation powers of different meteorological characteristics and irradiance monitoring values in one monitoring period; Building a power generation situation prediction model, training the power generation situation prediction model through the power generation situation path, and inputting meteorological data and irradiance monitoring values obtained through real-time monitoring after model training is completed to obtain a power generation situation prediction result at the current moment; And judging the power generation situation prediction result and a preset threshold value, if the power generation situation prediction result is larger than the preset threshold value, judging that the hybrid power supply strategy can be started, calculating deviation from the preset threshold value to generate dominant energy information, and carrying out hybrid power supply strategy formulation and regulation by utilizing a multi-objective optimization algorithm in combination with a load demand curve and real-time electricity price information.
2. 2. The method for intelligent management and control of groundwater pumping and injection integration based on the internet of things according to claim 1, wherein the geological investigation is carried out on a target area and an area three-dimensional geological model is constructed, pumping and injection well groups and heterogeneous sensor monitoring network layout planning are carried out by using the area three-dimensional geological model, and an area pumping and injection repair layout scheme is obtained, and the method specifically comprises the following steps: acquiring undisturbed geotechnical samples with different depths through grid point distribution drilling, performing geophysical prospecting and interpretation, collecting remote sensing images and historical geological map pieces of a target area as supplementary data, and simultaneously acquiring layered hydrogeological parameters through a site pumping test, a water injection test and a pressurized water test to construct a regional geological investigation data set containing space positions, stratum attributes and the hydrogeological parameters; Based on a regional geological survey data set, performing spatial interpolation on discrete drilling lithology data by utilizing a Kriging interpolation algorithm, constructing a three-dimensional geological structure frame model comprising stratum lithology, structural faults and a weathering zone by combining a deterministic modeling method, performing lithology spatial distribution random simulation in the three-dimensional geological structure frame model by utilizing a sequential indication simulation algorithm, and finally obtaining a regional three-dimensional geological model for representing geometrical morphology of stratum interface and lithology parameter spatial heterogeneity; introducing the regional three-dimensional geological model into groundwater simulation software, performing unstructured grid discretization processing by adopting a Delaunay triangulation method to generate a finite volume calculation grid suitable for complex geological boundary conditions, endowing each grid unit with corresponding hydrogeological parameters, and setting hydraulic boundary conditions and source and sink items; And analyzing the space-time migration evolution process of the pollutants under the control of the complex geological structure by solving the convection-dispersion-reaction equation with the steady-state flow velocity field as a drive, and identifying the diffusion path and the potential high-concentration aggregation area of the pollution plume to obtain a pollutant migration simulation result.
3. 3. The method for intelligent management and control of groundwater pumping and injection integration based on the internet of things according to claim 1, wherein the geological survey is performed on a target area, an area three-dimensional geological model is constructed, pumping and injection well groups and heterogeneous sensor monitoring network layout planning are performed by using the area three-dimensional geological model, and an area pumping and injection repair layout scheme is obtained, and the method further comprises: Based on pollutant migration simulation results, an optimization model with the maximum pollution capturing rate and the minimum construction cost as targets is established, a well position population is randomly generated in a simulated pollution feather range by adopting a multi-target particle swarm optimization algorithm, the pollution capturing efficiency is estimated by calculating the overlapping volume of a well position capturing zone and the pollution feather, the drilling cost is estimated by combining well depth and stratum lithology, and an optimal solution set is output through iterative calculation to generate a pumping and injecting well swarm layout scheme; Extracting the current frontal surface position, the main migration path and the boundary of a hydraulic capture zone of the pollution plume through the pollutant migration simulation result, and carrying out differential deployment design of heterogeneous sensors by combining the pumping and injecting well group layout scheme; The distributed optical fiber sensor based on the Brillouin optical time domain reflectometer is adopted for laying along the line, the laying track of the distributed optical fiber sensor is preferentially selected along the pollution central axis predicted by simulation or the section vertical to the groundwater flow direction, and the continuous temperature or strain information along the sensing optical cable space is obtained by inversion through monitoring the frequency shift change of the optical fiber back to the Brillouin scattered light so as to interpret the continuous change condition of the water quality; For a potential pollutant aggregation area, a hydrogeological condition mutation area and a key verification point of a repair effect, deploying a miniature spectrum sensor, performing grid calculation on an analog concentration field by a spatial interpolation algorithm at a specific installation position, and selecting a grid node with the largest concentration gradient change as an optimal layout point; And finally, obtaining an area pumping and injecting repair layout scheme, and adjusting the communication distance between the sensors through network connectivity analysis to ensure that all sensor node data can be reliably transmitted to the convergent edge computing node and the cloud server through LoRaWAN or a wired mode.
4. 4. The method for intelligent management and control of groundwater pumping and injection integration based on the internet of things according to claim 1, wherein the deployment of heterogeneous sensor monitoring network based on the regional pumping and injection repair layout scheme and the regional data acquisition are performed to obtain regional groundwater spatial distribution data and water quality monitoring data, and the regional groundwater spatial distribution data and water quality monitoring data are fused with a regional three-dimensional geological model to establish a polluted site digital twin body, specifically comprising the following steps: acquiring an area pumping and injecting repairing layout scheme, performing physical deployment of a physical sensing layer based on sensor layout coordinates and communication paths determined in the area pumping and injecting repairing layout scheme, and paving a distributed optical fiber sensor along a planned main monitoring axis to form a closed loop; fixedly mounting a miniature spectrum sensor at a key monitoring node, accessing an independent power supply unit formed by a solar panel and a storage battery pack and a LoRaWAN wireless communication module, verifying the working state and network connectivity of the sensing unit through an initialization test, and constructing a physical perception layer infrastructure for forming stable operation; The method comprises the steps that regional data acquisition is carried out through a physical perception layer infrastructure, a distributed optical fiber sensing system emits detection light pulses through a demodulator and acquires back Brillouin scattering signals, an original optical signal is processed through a frequency shift analysis algorithm, a frequency shift amount is converted into water quality parameter continuous distribution data by combining a pre-established calibration model, and meanwhile, a miniature spectrum sensor analyzes a water sample absorption spectrum and calculates a pollutant concentration value through a spectrum characteristic extraction algorithm; The edge node respectively carries out digital filtering and signal enhancement processing on the acquired original data, packages the processed data, the spatial position information and the time stamp into a transmission data unit, sends the transmission data unit to a cloud data processing center through a wireless communication network, and carries out space-time registration and data fusion after receiving the multi-source monitoring data by the cloud platform; The method comprises the steps of performing time synchronization processing on measurement data of heterogeneous sensors by adopting a time sequence alignment algorithm, eliminating time sequence deviation caused by sampling frequency difference, generating concentration measurement values of discrete points into a spatially continuous concentration distribution field by adopting a spatial interpolation algorithm, and inverting physical measurement values of distributed optical fibers into a water quality parameter distribution map by utilizing a calibration conversion relation; carrying out multi-source data fusion on the water quality parameter distribution map and the regional three-dimensional geological model, carrying out correlation deduction in the vertical direction on the two-dimensional water quality parameter field and the stratum structure and the permeability coefficient field in the three-dimensional geological model by a geostatistical analysis method, superposing and displaying the dynamic three-dimensional cloud map in the geological structure model by adopting a visualization technology, and simultaneously establishing a correlation mapping relation between a sensor network and monitoring data to realize interactive query and analysis of the monitoring data and the three-dimensional scene.
5. 5. The method for intelligent management and control of groundwater pumping and injection integration based on the internet of things according to claim 1, wherein the method for intelligent management and control of groundwater pumping and injection integration based on the internet of things is characterized in that the method for intelligent management and control of groundwater pumping and injection integration based on the internet of things extracts groundwater state potential data of a target area through the digital twin body of a polluted site and sets geological condition constraints, performs pollutant migration path and water level dynamic prediction by using an LSTM prediction model, and outputs pollution scenario simulation information corrected by geological features, and specifically comprises the following steps: Acquiring a pollutant concentration space distribution matrix updated by a sensor network, water level time sequence data of each monitoring point and instantaneous flow velocity field vector data obtained by numerical simulation calculation by inquiring a real-time data service interface of a digital twin body; simultaneously extracting key geological attribute parameters which are prestored in grid cells of the regional three-dimensional geological model and comprise osmotic coefficient tensor, effective porosity scalar and space geometric parameters and connectivity parameters of a fracture development zone of each grid cell; The method comprises fusing water level, pollutant concentration time series data with corresponding geological parameters by an attention mechanism based on the original data set, generating a mixed input vector containing time sequence features and static geological features, dividing into sample segments with fixed length, each sample containing the observed values of a plurality of time steps and the corresponding geological features, simultaneously, acquiring external environment factor time series data by using a big data network as additional feature dimension, adding the additional feature dimension into the input vector, constructing a multidimensional training sample set with time-space correlation, The multi-dimensional training sample set is input into a blank LSTM prediction model for supervised training, an input layer receives time sequence data of multi-dimensional sample characteristics, a hidden layer learns complex long-term dependence in a groundwater system through a forgetting gate, an input gate and an output gate, an output layer generates pollutant concentration and water level predicted values through a fully connected network, and finally the LSTM prediction model for training is obtained; The input of each time step comprises monitoring data at the current moment and corresponding static geological parameters, prediction output is calculated through forward propagation, gradient of a loss function to network weight is calculated through BPTT algorithm through reverse propagation, and the network parameters are updated through iteration of an adaptive moment estimation optimizer; performing multi-step recursion prediction by using the trained LSTM model, predicting the pollutant concentration distribution and the water level field of the next time step by taking the monitoring data at the current moment as an initial state, and then taking a prediction result as a new input, and performing recursion prediction to generate a pollutant concentration spatial distribution sequence and a water level dynamic change sequence of the future time step; Applying geological constraint correction to the generated pollutant concentration spatial distribution sequence and water level dynamic change sequence, performing spatial registration on a permeability coefficient field and a predicted concentration field by adopting an inverse distance weighted interpolation method based on the permeability coefficient spatial distribution tensor, identifying dominant migration channels of pollutants in a fracture network by adopting an optimal path analysis method based on Dijkstra algorithm by utilizing fracture development density field data, adjusting convection-dispersion parameters of the pollutants according to fracture density proportion in a fracture development area, and finally generating pollution scene simulation information corrected by geological features.
6. 6. The method for intelligent management and control of integrated pumping and injecting of groundwater based on the internet of things according to claim 1, wherein the pollution scenario simulation information is used as input, the coupling permeability coefficient and the fracture distribution parameter are used as constraint conditions, and the multi-objective collaborative optimization calculation is performed through BLMFO double-layer multi-objective optimization algorithm, so as to generate a pumping and injecting well group collaborative control strategy set, which specifically comprises the following steps: Constructing an upper-layer optimization model of BLMFO double-layer multi-objective optimization algorithm, taking pollution feather range minimization, system operation energy consumption minimization and equipment use load balancing as three mutually conflicting optimization targets, taking a flow distribution scheme, an operation time sequence and a medicament dosing concentration of each pumping and injecting well as decision variables, and constructing an upper-layer optimization problem mathematical model comprising multiple objective functions and basic constraint conditions; Based on the established upper-layer optimization problem model, constructing a decision variable space according to a pumping and injecting well group real-time operation control strategy and operation monitoring data of a target area through a preset three-level coding strategy, introducing an improved artificial bee colony optimization algorithm to perform upper-layer solving calculation, combining random selection and strategy selection to perform honey source initialization, and constructing an initial Pareto optimal front through iterative optimization analysis to generate a candidate control strategy; Transmitting the candidate control strategies to a lower-layer optimization model, establishing a response surface model under geological constraint conditions in the lower-layer model based on the osmotic coefficient spatial variation characteristics and fracture network connectivity parameters, performing geological feasibility verification on each candidate strategy transmitted by an upper layer, and feeding back verification results to an upper-layer optimization process in a constraint violation mode; Based on a geological constraint verification result fed back by a lower layer model, an upper layer optimization algorithm executes co-evolution operation, the speed and the position of particles are updated by adopting dynamically adjusted inertia weight and learning factors, inferior solutions which do not meet geological conditions are eliminated through a tournament selection mechanism, and after double-layer iterative computation for preset times, an optimal solution set which reaches optimal balance among a plurality of optimization targets and meets geological constraint conditions is output, so that a pumping and injecting well group cooperative control strategy set is generated.
7. 7. The method for intelligent management and control of groundwater pumping and injection integration based on the internet of things according to claim 6, wherein the method is characterized in that a decision variable space is built through a preset three-level coding strategy, an improved artificial bee colony optimization algorithm is introduced to perform upper-layer solving calculation, honey source initialization is performed by combining random selection and strategy selection, an initial Pareto optimal front is built through iterative optimization analysis, and a candidate control strategy is generated, and specifically comprises the following steps: the three-level coding strategy comprises pumping and injecting well codes, procedure codes and monitoring codes, wherein the pumping and injecting well codes comprise position attributes and function attributes of each pumping and injecting well, the procedure codes comprise operation time sequences and operation strategies corresponding to the pumping and injecting well, and the monitoring codes are operation monitoring situation characteristics corresponding to the pumping and injecting well; In the bee employment stage, carrying out neighborhood search on a current honey source by adopting a uniform crossing strategy, randomly selecting a solution from a Pareto archive set to carry out the exchange operation of a coding segment with the current solution, generating a new candidate solution by adopting a strategy with shortest accumulated processing time at a machine coding layer, and selecting a better solution from the new solution and the old solution and updating the archive set by using a Pareto domination principle; In the following bee stage, calculating the following probability of all honey sources based on the non-dominant grade and crowding degree of the honey sources, selecting the honey sources needing to be searched further in a roulette mode, wherein the searching strategy is consistent with that of the employment bee stage, and deep mining is ensured around a high-quality solution; When the honey source reaches the upper search limit but the quality is not improved, the scout bees reconstruct well group distribution codes by adopting a random exchange strategy, randomly select part of well groups to be redistributed to other processing units, simultaneously keep the balance of the overall processing capacity, and regenerate scheduling sequences and equipment parameters according to an initialization scheme so as to increase population diversity; In the iterative optimization process, non-dominant solutions are recorded in a mode of updating the archive set to serve as Pareto optimal solution sets, after each iteration, quick non-dominant sorting is carried out on the current population, solutions ranked in a preset range are added into the archive set, individuals subjected to new solution domination in the archive set are removed, after iterative computation of a preset algebra, initial Pareto optimal leading edges are output, and candidate control strategies are generated.
8. 8. The method for intelligent management and control of integrated pumping and injecting of groundwater based on the internet of things according to claim 1, wherein the generating a control instruction sequence based on the pumping and injecting well group cooperative control strategy set is transmitted to a corresponding executing mechanism for executing, and in the pumping and injecting repairing process, real-time running situation is monitored and whether repairing strategy change is needed is judged, and the generating of repairing strategy change suggestion for pushing specifically comprises: based on the operation parameters and target set values of all wells contained in the pump-injection well group cooperative control strategy set, converting continuous control parameters in the strategy set into instruction sequences with time sequence relations through time discretization, wherein each instruction contains an equipment identifier, an execution time stamp, a control parameter set value and an expected effect index, and verifying the executability and time sequence rationality of each instruction in the equipment capacity range through an instruction logic verification algorithm to form a verified control instruction sequence; issuing the verified control instruction sequence to a corresponding executing mechanism through an Internet of things communication protocol, immediately returning a confirmation signal after each executing mechanism receives the instruction, starting an instruction executing flow, and simultaneously acquiring actual operating parameters of equipment in real time through a built-in sensor to form an execution feedback data stream comprising instantaneous flow, actual medicament adding amount and equipment operating state; in the process of executing the instruction, synchronously starting multidimensional operation situation monitoring, acquiring underground water dynamic data including a water level change sequence, pollutant concentration distribution data and equipment operation state parameters in real time through a heterogeneous sensor network deployed in a pumping and injecting well group, and generating a high-quality real-time operation situation data set after data preprocessing; Performing dynamic assessment of the repair effect based on a real-time running situation data set, performing comparative analysis on the actually-monitored pollutant concentration distribution and the repair track expected in the strategy set, calculating the deviation degree of the actual value and the expected value of the repair performance index, and simultaneously, assessing the workload and the running efficiency of each execution unit by analyzing the running state data of the equipment, identifying the performance attenuation or abnormal running situation of the equipment, and forming an area repair situation assessment report; Starting a strategy change necessity judging flow according to an area restoration situation evaluation report, establishing a judging logic based on a multidimensional threshold, and judging that strategy change is needed when the fact that the actual reduction rate of the pollutant concentration is lower than a set proportion of an expected value or the pollutant concentration of a key point position has a rebound trend or the running efficiency of equipment is continuously lower than a rated efficiency threshold is monitored; when the strategy change is judged to be needed, according to the real-time operation situation characteristics and the effect evaluation result, the effective adjustment strategy under the similar working condition is retrieved from the historical case base based on the case reasoning method, and the generated flow adjustment amplitude, the medicament proportion optimization scheme and the device operation parameter correction value are combined with the current device operation state and the geological condition constraint to obtain the generated restoration strategy change suggestion to be pushed to the management terminal.
9. 9. The underground water pumping and injecting integrated intelligent management and control system based on the Internet of things is characterized by comprising a geological perception and digital twin unit, an intelligent prediction and decision unit, an execution and closed-loop control unit, a data communication and edge calculation unit and an energy management and guarantee unit: the geological perception and digital twin unit is used for fusion construction and continuous updating of the digital twin body of the polluted site through geological investigation data and monitoring data of a real-time sensor network, so that comprehensive perception of the underground environment is provided; the intelligent prediction and decision unit predicts pollutant migration and water level dynamics based on data provided by the digital twin body, and is coupled with geological constraints to correct and formulate an optimal pump-injection well group cooperative control strategy set; The execution and closed-loop control unit converts the decision into action and adjusts the action in real time according to feedback, drives the on-site execution mechanism to carry out repairing operation, continuously monitors the running state and repairing effect of the equipment, and rapidly judges and generates strategy adjustment suggestions; The data communication and edge calculation unit is responsible for transmission and preliminary processing of all data and instructions, and adopts a LoRaWAN and 5G hybrid communication network to connect all sensors, an actuating mechanism and a cloud platform; the energy management and guarantee unit integrates a solar photovoltaic panel, an energy storage battery and commercial power to form a hybrid power supply system, is responsible for providing power supply and dynamically optimizing an energy distribution strategy, and guarantees normal operation of equipment under abnormal power supply conditions.
10. 10. A computer readable storage medium, wherein the computer readable storage medium comprises an internet of things-based groundwater pumping and injecting integrated intelligent control method program, and when the internet of things-based groundwater pumping and injecting integrated intelligent control method program is executed by a processor, the steps of the internet of things-based groundwater pumping and injecting integrated intelligent control method according to any one of claims 1 to 8 are realized.

Description

Internet of things-based underground water pumping and injecting integrated intelligent management and control method and system Technical Field The invention relates to the technical field of environmental protection and intelligent control, in particular to an underground water pumping and injecting integrated intelligent control method and system based on the Internet of things. Background With the acceleration of the industrialization process, the problem of groundwater pollution is increasingly serious. The traditional underground water pumping and injecting system has the obvious defects of insufficient real-time performance, namely the system depends on manual regular monitoring and manual regulation and control and cannot respond to water quality mutation (such as chemical leakage) in time. The energy consumption and the cost are high, the operation parameters of the equipment are fixed, a dynamic optimization mechanism is lacked, and the operation and maintenance cost accounts for more than 40% of the total repair cost. Poor geological adaptability, and low pumping and injecting efficiency under complex geological conditions such as fracture water, karst areas and the like. In the prior art, a double-layer multi-objective optimization framework is proposed, but only coupling of a focusing simulation-optimization algorithm layer is realized, and the full-flow closed-loop execution capability driven by real-time monitoring data is lacking. For example, the optimization result is carried out by manual leading-in equipment, the second-level response of 'water quality mutation-parameter adjustment-execution feedback' cannot be realized, the modularized hardware integrated design and the solar hybrid power supply system are not involved, and the rapid deployment is difficult under the condition of no external power supply or complex terrain. In the prior art, the application of the Internet of things and the deep learning stays in single equipment control, and multi-objective collaborative optimization and full-flow closed-loop management are lacked. Therefore, an underground water pumping and injecting system integrating real-time monitoring, intelligent decision making and efficient execution is urgently needed, so that the repairing efficiency is improved, the cost is reduced, and the system is suitable for a complex geological environment. Disclosure of Invention The invention overcomes the defects of the prior art, and provides an integrated intelligent management and control method and system for groundwater pumping and injection based on the Internet of things, which have the important purposes of improving the repair efficiency, reducing the cost and adapting to complex geological environments. In order to achieve the above object, the first aspect of the present invention provides an integrated intelligent management and control method for pumping and injecting groundwater based on the internet of things, comprising: performing geological investigation on a target area, constructing an area three-dimensional geological model, and performing pumping and injecting well group and heterogeneous sensor monitoring network layout planning by using the area three-dimensional geological model to obtain an area pumping and injecting repairing layout scheme; the heterogeneous sensor monitoring network is deployed based on the regional pumping and injecting restoration layout scheme, regional data acquisition is carried out, spatial distribution data and water quality monitoring data of regional groundwater are obtained, and the regional groundwater is fused with a regional three-dimensional geological model to establish a digital twin body of a polluted site; extracting groundwater state data of a target area through the digital twin bodies of the pollution sites, setting geological condition constraint, carrying out dynamic prediction on a pollutant migration path and a water level by using an LSTM prediction model, and outputting pollution scene simulation information corrected by geological features; taking the pollution scenario simulation information as input, taking the coupling permeability coefficient and the fracture distribution parameter as constraint conditions, and carrying out multi-objective collaborative optimization calculation through BLMFO double-layer multi-objective optimization algorithm to generate a pumping and injecting well group collaborative control strategy set; and generating a control instruction sequence based on the pumping and injecting well group cooperative control strategy set, sending the control instruction sequence to a corresponding executing mechanism for executing, monitoring a real-time operation situation in the pumping and injecting repairing process, judging whether repairing strategy change is needed, and generating a repairing strategy to be changed into a suggestion for pushing. In this scheme, carry out geological survey and build regional three-dimensiona