Karst groundwater intelligent early warning and collaborative control method and system

Abstract

The application relates to the technical field of hidden danger early warning, in particular to a karst groundwater intelligent early warning and collaborative control method and system, wherein the method comprises the following steps: cooperatively monitoring target pollutants and multiple physical and chemical parameters, collecting multi-source heterogeneous data; constructing a digital twin model, fusing the multi-source heterogeneous data, introducing a periodic active hydraulic disturbance diagnosis mechanism, dynamically updating the state of the model, realizing simulation and prediction; comparing the prediction result of the model with a dynamic threshold library to automatically trigger a graded early warning signal; in response to the graded early warning signal, using the adjoint equation method and combining the particle backtracking algorithm to reversely locate the position of a pollution source, comprehensively performing cross verification to generate a pollution source probability distribution map; according to the early warning level and the pollution source probability distribution map, calculating an optimal control scheme and generating specific execution instructions; and an execution mechanism receiving and executing the instructions. Therefore, the problems of pollution discovery lag, difficult source tracing and great treatment difficulty are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hidden danger early warning, in particular to a karst groundwater intelligent early warning and collaborative control method and system. BACKGROUND

[0002] Karst large springs are unique hydrogeological phenomena in carbonate rock distribution areas, not only providing important production and living water for the surrounding areas, but also maintaining a unique karst ecological system, with irreplaceable ecological, economic and social values. However, the management of karst large springs is a big problem, and the core problem stems from the extreme complexity of the karst aquifer itself. First, the aquifer is composed of multiple media such as dissolution pipes, fissures and pores, with a huge difference in permeability in space. Water flow preferentially selects large pipes, resulting in highly concentrated and unpredictable pollutant migration paths. Second, the water flow in karst pipes is extremely fast, and once pollutants enter the dominant runoff channel, they will quickly spread to the spring outlet, and the migration process occurs underground, making it difficult to detect. In addition, the large recharge area of the large spring is wide, and the pollution sources are scattered, making it extremely difficult to trace the pollution. The huge underground space and strong heterogeneity make it difficult for traditional management methods (such as injection of repair agents) to effectively distribute and contact pollutants, resulting in high management difficulty and easy recurrence. SUMMARY

[0003] The present application provides a karst groundwater intelligent early warning and collaborative control method and system to solve the problems of late pollution discovery, difficult pollution tracing and high management difficulty.

[0004] The first aspect embodiment of the present application provides a karst groundwater intelligent early warning and collaborative control method, comprising the following steps:

[0005] S1, deploying a pollution-triggered intelligent tracer capsule array, a distributed optical fiber sensing network and a multi-parameter sensor node group, and collaboratively monitoring target pollutants and multi-element physical and chemical parameters of karst groundwater, collecting multi-source heterogeneous data;

[0006] S2, constructing a digital twin model, fusing the multi-source heterogeneous data, and introducing a periodic active hydraulic disturbance diagnosis mechanism, dynamically updating the state of the digital twin model through data assimilation technology, realizing real-time simulation and multi-step advanced prediction of the water flow field and solute concentration field;

[0007] S3, comparing the prediction results of the digital twin model with a dynamic threshold library, and automatically triggering a hierarchical early warning signal including three levels of attention, early warning and alarm;

[0008] S4, in response to the hierarchical early warning signal, using the adjoint equation method combined with the particle backtracking algorithm to reversely locate the pollution source position, and cross-verifying the trigger signal of the intelligent tracer capsule array to generate a pollution source probability distribution map;

[0009] S5、According to the triggered early warning level and the pollution source probability distribution map, a model predictive control optimization algorithm is used for multi-objective optimization solution, and an optimal control scheme with gate dam group collaborative scheduling as the core is calculated, and specific execution instructions are generated;

[0010] S6, the execution mechanism receives the execution instruction, and accurately adjusts the gate opening or pump station speed to realize the governance goal of hydraulic collaborative control.

[0011] Optionally, the pollution trigger type intelligent tracer capsule array is composed of an environment responsive material, has a fluorescent tracer inside, and the degradation rate of the capsule shell is positively correlated with the concentration of the target pollutant, so as to realize the release triggered by pollution exceeding the standard; the distributed optical fiber sensing network is used to collect distributed physical field data; the multi-parameter sensor node group is used to collect water quality and hydrological data at the point position; the multi-source heterogeneous data includes the distributed physical field data and the water quality and hydrological data, wherein the distributed physical field data includes temperature field data and vibration field data; the water quality and hydrological data includes water level, pH (hydrogen ion concentration index), conductivity, dissolved oxygen, turbidity, and target ion concentration data.

[0012] Optionally, the periodic active hydraulic disturbance diagnosis mechanism includes:

[0013] Through periodic setting or emergency triggering, short-term and quantitative pumping or water injection experiments are carried out at the preset position downstream, and a hydraulic excitation is artificially created;

[0014] The digital twin model analyzes the water level response, water quality response and tracer response of each monitoring point in the upstream to the excitation, and inversely updates the permeability coefficient field and porosity field parameters of the karst aquifer, so as to realize model calibration;

[0015] The standard response curve measured during the healthy period disturbance is stored in the database, and in the subsequent periodic disturbance, the newly measured response curve is compared with the standard response curve, so as to identify abnormal points and realize early abnormal positioning.

[0016] Optionally, the early warning threshold in the dynamic threshold library is dynamically calculated and generated by the digital twin model according to the current water flow state, historical background concentration, and prediction uncertainty, and the core calculation formula is:

[0017]

[0018] Wherein, is the final dynamic threshold, is the background reference value, is the safety factor, is a standard deviation of predicted values, is a correlation coefficient, is a current simulated flow rate, is a historical average flow rate.

[0019] Optionally, the adjoint equation method obtains a sensitivity field by constructing and solving an adjoint equation of a concentration field of an observation point to a release intensity of a pollution source, and obtains a pollution source contribution probability distribution map by normalizing the sensitivity field; the particle backtracking algorithm simulates reverse migration of a large number of virtual particles released at an early warning point and a reverse flow field provided by the digital twin model, and counts a spatial distribution of all particles at a simulation termination time to generate a probability distribution map; and the cross verification generates a final high-confidence pollution source probability distribution map by consistency comparison between the pollution source contribution probability distribution map, the probability distribution map, and an actual triggering position of the intelligent tracer capsule.

[0020] Optionally, the model predictive control optimization algorithm includes multi-objective optimization solving.

[0021] A multi-objective cost function is constructed to minimize pollution concentration, gate operation amount and treatment economic cost at the same time, and a rolling optimization is performed within a limited time domain, wherein the multi-objective cost function is:

[0022]

[0023] wherein, is to minimize the comprehensive cost, is to minimize the deviation of pollution concentration from a target value, is to minimize the variation range of control variables, is to minimize the economic cost of treatment, , , is a weight coefficient.

[0024] Optionally, the step S6 further includes an effect feedback and adaptive optimization step: execution state of the execution mechanism and environmental response data after treatment are fed back to the digital twin model in real time, for evaluating treatment performance and incrementally learning and optimizing the model.

[0025] The second aspect embodiment of the present application provides a karst groundwater intelligent early warning and coordination system, comprising: an intelligent sensing module, a data fusion and communication module, a digital twin and early warning module, an intelligent tracing module, a coordination decision engine, an execution control module and an adaptive module, wherein,

[0026] The intelligent perception module includes an intelligent tracer capsule array, a distributed optical fiber sensing network, and a multi-parameter sensor node group, and is configured to perform an active hydraulic disturbance mechanism and collect multi-source heterogeneous data.

[0027] The data fusion and communication module is configured to clean, align, and transmit the multi-source heterogeneous data.

[0028] The digital twin and early warning module is internally provided with a digital twin model, and is configured to complete simulation, prediction, and trigger hierarchical early warning.

[0029] The intelligent tracing module is configured to run a concomitant equation method and a particle backtracking algorithm, and generate a pollution source probability distribution map.

[0030] The collaborative decision engine is configured to run a model predictive control optimization algorithm, and generate an optimal control scheme and an execution instruction.

[0031] The execution control module is configured to receive and execute the instruction, and control the execution mechanism to complete an action.

[0032] The adaptive module is configured to complete governance performance evaluation and model incremental learning.

[0033] The third aspect of the embodiments of the present application provides an electronic device, including: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are configured to execute the karst groundwater intelligent early warning and collaborative control method as described in the above embodiments.

[0034] The fourth aspect of the embodiments of the present application provides a computer readable storage medium having a computer program stored thereon, and the program is executed by a processor to implement the karst groundwater intelligent early warning and collaborative control method as described in the above embodiments.

[0035] Therefore, the present application has at least the following beneficial effects:

[0036] The embodiments of the present application realize 24-hour real-time monitoring and minute-level data acquisition by deploying pollution-triggered intelligent tracer capsule arrays, distributed optical fiber sensing networks and multi-parameter sensor node groups, and innovates the early warning mode from "exceeding the standard and then alarming" to "predictive early warning before pollution occurs or in the early stage of diffusion", which provides valuable time for response; through a periodic active hydraulic disturbance diagnosis mechanism, the hydraulic conductivity field of the aquifer is inverted and updated, the channel patency is evaluated, and early diagnosis of system health is realized; through the fusion of multiple tracing technologies such as concomitant equation method and particle backtracking algorithm, and combined with the field evidence of intelligent tracers, the pollution source is quickly, accurately and reliably located, and the great difficulty of pollution tracing in karst channels is overcome; through digital twinning and model predictive control optimization algorithm for multi-objective optimization solution, the treatment effect and economic cost of different schemes can be automatically simulated, and the scientific optimal solution is given, which fundamentally improves the scientificity and accuracy of decision-making; abandoning the extensive management mode, the precise management strategy of "hydraulic coordination control" is proposed, the intelligent scheduling of gates and dams and pump stations in the spring area is used to occupy the channel with clean water flow, block the pollution group, and guide it to the preset treatment area, and the low-cost, high-efficiency and environmentally friendly management goal is achieved.

[0037] Thus, the technical problems of pollution discovery lag, tracing difficulty and treatment difficulty are solved.

[0038] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0039] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the following description, including the accompanying drawings, wherein:

[0040] Figure 1 A flow chart of a karst groundwater intelligent early warning and coordinated control method according to an embodiment of the present application;

[0041] Figure 2 A structural schematic diagram of a karst groundwater intelligent early warning and coordinated control system according to an embodiment of the present application;

[0042] Figure 3 A structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0043] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference signs represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as limiting the present application.

[0044] A karst groundwater intelligent early warning and collaborative control method and system according to an embodiment of the present application is described below with reference to the accompanying drawings. In view of the problems of late pollution discovery, difficulty in tracing, and great difficulty in treatment mentioned in the background art, the present application provides a karst groundwater intelligent early warning and collaborative control method. In the method, a pollution-triggered intelligent tracer capsule array, a distributed optical fiber sensing network, and a multi-parameter sensor node group are deployed to realize 24-hour real-time monitoring and minute-level data acquisition, and the early warning mode is innovated from "alarm after exceeding the standard" to "predictive early warning before pollution occurs or in the early stage of diffusion", which provides valuable time for response; through a periodic active hydraulic disturbance diagnosis mechanism, the aquifer permeability coefficient field is inverted and updated, the channel patency is evaluated, and early diagnosis of system health is realized; by fusing multiple tracing technologies such as the concomitant equation method and the particle backtracking algorithm, and combining with the field evidence of intelligent tracers, the rapid, accurate, and reliable positioning of pollution sources is realized, and the great difficulty of tracing pollution in karst channels is overcome; through digital twinning and model predictive control optimization algorithm for multi-objective optimization solution, the treatment effect and economic cost of different schemes can be automatically simulated, and a scientific optimal solution is given, which fundamentally improves the scientificity and accuracy of decision-making; the extensive treatment mode is abandoned, and a precise treatment strategy of "hydraulic collaborative control" is proposed, which occupies the channel with clean water flow, blocks the pollution mass, and guides it to the preset treatment area through intelligent scheduling of gates, dams, and pump stations in the spring area, so as to realize the treatment goal of low cost, high efficiency, and environmental friendliness. Thus, the technical problems of late pollution discovery, difficulty in tracing, and great difficulty in treatment are solved.

[0045] Specifically, Figure 1 A flowchart of a karst groundwater intelligent early warning and collaborative control method provided by an embodiment of the present application is shown.

[0046] As Figure 1 shown, the karst groundwater intelligent early warning and collaborative control method includes the following steps:

[0047] In step S101, a pollution-triggered intelligent tracer capsule array, a distributed optical fiber sensing network, and a multi-parameter sensor node group are deployed to collaboratively monitor target pollutants and multi-element physical and chemical parameters of karst groundwater, and to collect multi-source heterogeneous data.

[0048] The pollution trigger type intelligent tracer capsule array is composed of an environment responsive material, has a fluorescent tracer inside, and the degradation rate of the capsule shell is positively correlated with the concentration of the target pollutant, thereby triggering release when the pollution exceeds the standard; a distributed optical fiber sensing network is used to collect distributed physical field data; a multi-parameter sensor node group is used to collect water quality and hydrological data at the point location; the multi-source heterogeneous data includes distributed physical field data and water quality and hydrological data, wherein the distributed physical field data includes temperature field data and vibration field data; and the water quality and hydrological data includes water level, pH, conductivity, dissolved oxygen, turbidity, and target ion concentration data.

[0049] The pollution trigger type intelligent tracer capsule array mainly plays a role in targeted early warning and source tracing, and is a "chemical signal bomb" deployed in front of the risk point. It solves the problem that traditional sensors cannot be deployed at all potential risk points and cannot provide early warning before pollution occurs. Because the shell material of the capsule is composed of an environment responsive material, for example, an environment responsive hydrogel, the crosslinking degree of the gel will chemically react with the concentration of a specific target pollutant. When the target pollutant diffuses around the capsule and reaches a threshold concentration, ion exchange or integration occurs with the gel network, causing the crosslinking bonds of the gel network to break or the structure to become loose, thereby causing the capsule shell to fail. The high-concentration, easily detectable fluorescent tracer wrapped in the capsule shell will be immediately released into the groundwater. The released tracer moves with the water flow. At the downstream spring or monitoring well, the sudden appearance of the signal can be easily captured by a fluorescence sensor.

[0050] In an implementable manner, the environment responsive hydrogel selection includes: pH sensitive hydrogel, ion sensitive hydrogel, and enzyme responsive hydrogel. The pH sensitive hydrogel can be used to monitor acid mine drainage or alkaline industrial wastewater. The ion sensitive hydrogel specifically responds to target pollutants such as nitrate, chromate, and ammonium. The enzyme responsive hydrogel can respond to specific organic pollutants. The internal tracer selects a fluorescent substance with high fluorescence intensity, low background value, good stability, and environmental friendliness, for example, fluorescein sodium and rhodamine WT. The fluorescein sodium is a green fluorescent substance with low cost and the widest application. The rhodamine WT is a red fluorescent substance with better stability and strong anti-interference ability.

[0051] It can be understood that the pollution trigger type intelligent tracer capsule array is in a dormant state at ordinary times. Once a specific pollutant exceeds the standard, the capsule will "self-destruct" and release easily traceable "dye", telling the downstream: "something happened here!", to provide the most direct and non-delayed pollution event alarm signal, thereby achieving early warning. The capsule burial location itself is the most direct pollution source suspect point, and the concentration curve of the released tracer provides verification data for subsequent source tracing.

[0052] It should be noted that the target pollutant trigger threshold of the environment-responsive material is determined according to the water quality standard of the protection target, and the degradation threshold of the capsule shell can be customized in cooperation with the material supplier.

[0053] In an embodiment, the deployment of the intelligent tracer capsule array is to fix the intelligent tracer capsule at the karst fissure or pipe inlet downstream of the potential pollution source by using the drilling implantation technology, to ensure that it is in full contact with the groundwater, and to arrange 3-5 intelligent tracer capsules downstream of each key source according to the risk level to form a monitoring array to prevent omission.

[0054] In an embodiment, the deployment of the distributed optical fiber sensing network is to bury the optical cable at the bottom of the riverbed or underground in the key section where leakage is likely to occur to monitor the surface water-groundwater exchange.

[0055] In an embodiment, the deployment of the multi-parameter sensor node group is to arrange sensors at different depths (shallow, medium, and deep) in important monitoring wells to reveal the vertical water quality stratification, and to arrange nodes at the spring outlet, water intake of the water source, junction of different aquifers, and upstream and downstream of the pollution domain. In addition, a star network topology is used for networking to ensure reliable data return.

[0056] In step S102, a digital twin model is constructed, multi-source heterogeneous data is fused, and a periodic active hydraulic disturbance diagnosis mechanism is introduced. The state of the digital twin model is dynamically updated through data assimilation technology to realize real-time simulation and multi-step advanced prediction of the water flow field and solute concentration field.

[0057] The periodic active hydraulic disturbance diagnosis mechanism includes: through periodic setting or emergency triggering, a short-term and quantitative pumping or injection experiment is carried out at a preset position downstream to artificially create a hydraulic excitation; the digital twin model analyzes the water level response, water quality response, and tracer response of each monitoring point upstream to this excitation, and inversely updates the permeability coefficient field and porosity field parameters of the karst aquifer to realize model calibration; the standard response curve measured during the disturbance in the healthy period is stored in the database, and in subsequent periodic disturbances, the newly measured response curve is compared with the standard response curve to identify abnormal points and realize early abnormal positioning.

[0058] It should be noted that the periodic setting mode can be set to a suitable time according to the needs, for example, in an embodiment, it is preset to be carried out once every quarter; in another embodiment, it is preset to be carried out once a month during the stable period of water flow to establish a "healthy baseline". The emergency triggering mode refers to manually triggering an additional disturbance by the administrator during the high-risk period, such as after a rainstorm or after a nearby pollution accident.

[0059] Since the karst groundwater system is like an extremely complex and invisible "black box", and traditional monitoring can only passively wait for the black box to produce changes (such as pollution events) and then react afterwards. The periodic active hydraulic disturbance diagnosis mechanism is not passive waiting for natural changes, but regularly and actively pumping water at the downstream spring or main discharge point for a short time and in a quantitative manner, artificially creating a water level fluctuation of known intensity and regularity, which will propagate upstream in the form of a water pressure wave. Among them, the disturbance time and pumping flow rate are artificially set, and the pumping flow rate needs to be large enough to produce a measurable water level response that can propagate upstream, but not too large to cause problems such as ground subsidence. Usually determined through preliminary tests, for example, the pumping flow rate is set between 50-100m³ / h. The disturbance time is usually set to 6-24 hours to ensure that the water level depression cone can fully develop and propagate. For example, pumping water at a flow rate of 1000m³ / h for 6 hours at the main spring downstream.

[0060] Due to the downstream hydraulic disturbance excitation, the upstream will excite corresponding water level response, water quality response and tracer response to the excitation, wherein the water level response refers to the time, amplitude and shape of the water level drop at each monitoring point in the upstream; the water quality response refers to the synchronous change of the water quality parameters (such as conductivity, temperature) at each point in the upstream; the tracer response refers to the mutation of the tracer signal if there is a tracer signal during the disturbance period, which can be observed.

[0061] According to the above responses, all monitoring points in the upstream synchronously collect multi-source heterogeneous data at a high frequency (such as once every minute) during and after the pumping period, forming a full-field "water level-time" drawdown curve, a "water quality-time" change curve, etc.

[0062] The digital twin model receives all data of the disturbance test, including the pumping amount, the "water level-time" drawdown curve at each point, the "water quality-time" change curve, etc., and through the inversion algorithm, continuously and automatically adjusts the permeability coefficient, porosity coefficient and other parameters of each region in the digital twin model until the water level response curve simulated by the model is highly consistent with the actually observed water level response curve, so as to realize the calibration of the model. After multiple disturbance calibrations, when the water level response curve simulated by the model is completely consistent with the actually observed water level response curve, the application embodiment obtains a permeability coefficient field and a porosity field that are infinitely close to the actual situation, thereby completing high-precision calibration.

[0063] After high-precision calibration, the high-fidelity digital twin model has more accurate early warning capability and more reliable traceability capability.

[0064] It can be understood that the active hydraulic disturbance is an innovation of data acquisition means, which provides high-value "dynamic physical examination data" for calibrating the model that traditional passive monitoring cannot obtain; the digital twin is an innovation of data analysis and modeling means, which uses disturbance data to greatly improve the simulation accuracy of itself, thereby providing more reliable early warning, traceability and decision support.

[0065] Specifically, the state of the digital twin model is dynamically updated by the data assimilation technology, realizing real-time simulation and multi-step advanced prediction of the water flow field and solute concentration field. It can be understood that the purpose of real-time simulation is to make the state of the digital twin model infinitely close to the current state of the real world, which is mainly realized by the data assimilation technology; after the data assimilation ensures that the current state is as accurate as possible, the model can be used for multi-step prediction. The specific steps of real-time simulation include: starting from the current time t, the digital twin model calculates a very short time step (for example, 10 minutes in the future) based on its internal physical equations (such as groundwater flow equation, solute transport equation) and the current best estimate state, to obtain a predicted state value; at time t+1, new real-time observation data are transmitted from sensors (water level, water quality, etc.) all over the place; the data assimilation algorithm starts to work, calculates the difference between the predicted state value and the new real-time observation data, and then according to the respective uncertainties (model error and observation error) of the two, optimally "absorbs" or "injects" the new information of the observation value into the model; output a new, corrected optimal estimate state value. This optimal estimate state value is closer to the true situation than the model prediction alone or the observation alone, which completes the update of the model state and realizes real-time simulation.

[0066] It should be noted that the data assimilation algorithm aims to optimally fuse the observation data (measured values of the real world) and the dynamic model (such as the digital twin model) to obtain the most accurate and complete estimate of the current state of the system, and the core idea is neither to completely trust the model nor to completely trust the data.

[0067] Optionally, in some embodiments, the data assimilation algorithm is Kalman filter method, and the algorithm process is as follows: starting from the current time step, each member in the model set (for example, 100) is independently run the model to predict forward to the next time step with observation data. At this time, each member will obtain a predicted value, and the distribution of all members represents the uncertainty range of model prediction. When new observation data arrives, assimilation begins: compare the predicted value of each member with the actual observation value, and calculate the difference. Based on the statistical relationship of all members, an optimal gain matrix is calculated, which determines how much the model prediction value should be corrected by the observation value. If the observation is very reliable (small error) and the model prediction is uncertain, the matrix value is large, and the correction amplitude is large; if the observation noise is large (large error) and the model prediction is confident, the matrix value is small, and the correction amplitude is small. For each member in the set, correction is performed, and the correction formula is:

[0068] ;

[0069] wherein, is the optimal gain matrix value.

[0070] After the state of all model members is corrected, the average value of them is the optimal estimate of the current state of the system, and the dispersion of them represents the uncertainty of the updated state. The updated state is taken as the new initial field, and the above steps are repeated to continue the cycle.

[0071] After ensuring that the current state is as accurate as possible, the model can start multi-step ahead prediction. The optimal estimated state value after data assimilation correction is taken as the initial field for prediction. This is the premise of accurate prediction. When running the prediction model, the digital twin model will no longer receive new observation data, but will purely calculate forward based on internal physical mechanisms and initial state. In this process, it will incorporate known or predicted external driving conditions, such as including future 72-hour rainfall, evaporation, etc. meteorological forecast data, and known agricultural and industrial water intake plans, reservoir scheduling plans, etc. human activity plans, river water level, exchange amount of other aquifers, etc. boundary conditions. The model will calculate the water head value and pollutant concentration value at each time step (such as every hour) in the entire simulation area at each location in the future, and finally generate a "spatiotemporal evolution animation of water flow field and solute concentration field", which intuitively shows how the pollution plume moves, spreads, and dilutes.

[0072] It can be understood that through the data assimilation technology, the digital twin model can be continuously synchronized with the real world, ensuring that its virtual state "in the body" is consistent with reality. On this basis, it can scientifically and quantitatively predict the future water flow and pollutant transport, thus realizing a revolutionary change from "passive monitoring" to "active early warning" and providing valuable advance for decision-makers.

[0073] In step S103, the prediction results of the digital twin model are compared with the dynamic threshold library, and a graded early warning signal containing three levels of attention, early warning and alarm is automatically triggered.

[0074] Among them, the early warning threshold in the dynamic threshold library is dynamically calculated and generated by the digital twin model according to the current water flow state, historical background concentration, and the uncertainty of the prediction itself. The core calculation formula is:

[0075]

[0076] Among them, is the final dynamic threshold, is the background reference value, is the safety factor, is the standard deviation of the prediction value, is the correlation coefficient, is the current simulation flow rate, is the historical average flow rate.

[0077] It should be noted that the background reference value represents the natural background concentration or historical normal level of a certain pollutant in the water body under no human interference or low-level human activity. It is usually the higher quantile (such as the 95% quantile) of the historical concentration data of the same place and the same period (such as the same quarter) in the past 3-5 years. The background reference value is the baseline of the dynamic threshold, and the background reference value itself can be adjusted according to the season and hydrological period to form a more detailed background library. is the uncertainty increment, which represents the uncertainty of the model prediction. It measures the degree of confidence of the model in predicting the concentration at a certain time in the future. The larger the value is, the lower the credibility of the prediction result and the higher the uncertainty. The value is obtained through ensemble prediction. For example, run the digital twin model 100 times (each time use slightly different parameters or initial conditions), get 100 prediction values, and the standard deviation of these prediction values is The safety factor is a multiplier set according to risk preference, usually taking a value of 2 or 3. is the hydrological condition increment; the current simulated flow rate is the groundwater pore or Darcy flow rate at the current time provided by the digital twin model; the historical average flow rate is the average flow rate at the same period or long term at the location; the correlation coefficient is a coefficient obtained by regression analysis of historical data, which represents the sensitivity of the change of the flow rate to the concentration of the pollutant, and is usually positive.

[0078] It needs to be explained that the prediction result output by the digital twin model after data assimilation and multi-step prediction is not a single number, but a set of multi-dimensional, spatio-temporal and uncertainty information containing data products. Mainly including: pollutant concentration field, probability distribution or confidence interval of each prediction value, water flow field and pollution plume migration path. Among them, the probability distribution or confidence interval of each prediction value is given in the form of standard deviation in this embodiment. The water flow field is the prediction of future water level and flow rate change, which helps to understand the driving force and direction of pollutant migration. The pollution plume migration path refers to the three-dimensional space trajectory and impact range of the pollutant in the groundwater from the pollution source to the downstream movement, diffusion and migration. In this embodiment, the pollution plume migration path can be visualized, and the moving track, diffusion range and arrival time of the pollution group from the source to the discharge point can be clearly seen.

[0079] The prediction result is compared with the dynamic threshold library, and the specific steps are as follows: for each target position (such as spring outlet, water intake of water source) that needs to be monitored, at each future prediction time point, the system will dynamically calculate a corresponding threshold value according to the formula introduced earlier, then compare the prediction result with the corresponding dynamic threshold value point by point in time sequence, and finally perform uncertainty analysis, which means not only comparing the "best prediction value", but also comparing the entire prediction interval. For example, even if the best prediction value is slightly lower than the threshold value, but if the upper limit of its 90% confidence interval has far exceeded the threshold value, the system will consider it as high risk.

[0080] Based on the comparison result, a graded early warning is realized. Specifically, when the predicted concentration exceeds 70% of the threshold value and the concentration shows an upward trend, a concern warning is triggered, which prompts on the management platform interface (such as yellow flashing), notifies the management personnel to pay more attention, and can arrange manual verification; when the predicted concentration continuously approaches the threshold value, or the uncertainty analysis shows high risk, the automatic start of the traceability analysis is started, the preliminary control scheme is generated, the warning information is pushed to the management personnel through the short message / App (Application, application program), and the emergency preparation is started; when the predicted concentration is confirmed to exceed the dynamic threshold value or the certain evidence (capsule triggering) from the field is received, the automatic trigger of the collaborative control mechanism is triggered, and the control instruction is issued to the executing agency (such as starting the dam). At the same time, all relevant persons are notified through the highest priority mode (telephone, short message).

[0081] It can be understood that the dynamic threshold library comparison method changes the dynamic threshold with environmental conditions and model confidence, so that the comparison is scientific; the comparison based on the prediction result can provide early warning before pollution occurs, providing valuable response time for management personnel; starting different response processes according to the risk level can avoid "making a big deal out of a small matter" or "missing a big event", and optimize resource allocation.

[0082] In step S104, in response to the hierarchical early warning signal, the position of the pollution source is reversely located by using the adjoint equation method combined with the particle backtracking algorithm, the trigger signals of the intelligent tracer capsule array are comprehensively cross-verified, and a pollution source probability distribution map is generated.

[0083] Among them, the adjoint equation method obtains the sensitivity field by constructing and solving the adjoint equation of the observation point concentration field to the pollution source release intensity, and obtains the pollution source contribution probability distribution map by normalizing the sensitivity field; the particle backtracking algorithm simulates the reverse migration of particles by releasing a large number of virtual particles at the early warning point and using the reverse flow field provided by the digital twin model, and counts the spatial distribution of all particles at the simulation termination time, and generates a probability distribution map; cross verification, by comparing the pollution source contribution probability distribution map, the probability distribution map, and the actual trigger position of the intelligent tracer capsule for consistency, a final high-confidence pollution source probability distribution map is generated. Specifically, the algorithm process of the adjoint equation method is as follows:

[0084] Step one: define a new variable (x, y, z, t), called the adjoint state variable. The physical meaning is the contribution of releasing a unit mass of pollutants at (x, y, z) point to the final concentration at the downstream observation point , ) in unit time.

[0085] Step two: construct and solve the adjoint equation. Since the adjoint equation is the dual form of the original solute transport equation (advection-dispersion equation), the original solute transport equation is:

[0086]

[0087] Among them, is the dispersion coefficient tensor, is the concentration, is the flow velocity vector, is the reaction term coefficient, is the source-sink term (pollution source).

[0088] Therefore, the adjoint equation is:

[0089]

[0090] Among them, D is the dispersion coefficient tensor, D is the new variable, D is the flow velocity vector, D is the reaction term coefficient, D is the Dirac function, representing the pulse input, which occurs at the downstream observation point and observation time , D is the downstream observation point, D is the observation time.

[0091] Step three: solve the above adjoint equation to obtain the adjoint state field D at the final time t D, D, D, D, D, which is the required sensitivity field. The higher the value, the greater the contribution of the pollutant released at that point to the concentration at the downstream detection point.

[0092] Step four: normalize the adjoint state field D, D, D, D, D, and the pollutant source contribution probability distribution map can be obtained.

[0093] It can be understood that the adjoint equation method only needs to solve a partial differential equation once in the forward and backward directions, and the sensitivity of all positions in the entire field can be obtained, which is extremely efficient, while the traditional method needs to simulate the pollution source at each position, and the calculation amount increases linearly with the number of source points.

[0094] Particle backtracking algorithm is a stochastic simulation method based on Lagrangian framework. Its intuitive understanding is: release a large number of virtual particles at the position where the pollution is detected (downstream), and let them perform random walk in reverse time and against the direction of water flow. The area where these particles eventually stagnate or gather is the most likely location of the pollution source.

[0095] Specifically, the algorithm process is as follows:

[0096] Step one: use the calibrated digital twin model to simulate the stable flow field of the study area during the pollution event (or the current time), i.e. the flow velocity vector of each grid element.

[0097] Step two: release a large number of (e.g. 10000) virtual particles at the downstream point where the pollution is detected (e.g. spring mouth), simulate time in the negative direction, calculate the displacement of each virtual particle p at each time step, and the calculation formula is:

[0098]

[0099] where, is the displacement of each virtual particle p at each time step, is the inverse advection term, is the random dispersion term, is the time step.

[0100] It is noted that the random dispersion term is used to simulate the mechanical dispersion and molecular diffusion effects on the pollutant during migration. It is usually achieved through a stochastic model, for example:

[0101]

[0102] where, is a random vector of a standard normal distribution, is the hydrodynamic dispersion coefficient, is the time step.

[0103] Step three: stop the simulation when the particles reach the set simulation time origin (such as 30 days backtracking), reach the model boundary, or are stationary. Record the final coordinates of each particle when it stops.

[0104] Step four: divide the entire model area into a three-dimensional grid, count the number of particles that finally stop in each grid cell, calculate the probability value of each grid cell, which is the ratio of the number of particles that finally stop to the total number of particles, and visualize the probability value calculated above, i.e. get the pollution source probability distribution map. The darker the color (the denser the particles), the higher the likelihood of the pollution source.

[0105] Cross-validation is to superimpose and compare the "pollution source contribution probability distribution map" obtained by the adjoint equation method with the "probability distribution map" obtained by the particle backtracking method. If both indicate that the same area (A area) has the highest probability, the credibility of the A area is greatly improved. Compare the above overlapping high-probability area (A area) with the actual trigger location of the intelligent tracer capsule. If the capsule trigger point is located in the A area. This forms a complete evidence chain of "theoretical inference + numerical simulation + field evidence", and the confidence of the conclusion reaches the highest. That is, a final high-confidence pollution source probability distribution map can be generated, and the most likely source area can be clearly marked; if the capsule triggers in another B area, while the theoretical simulation points to the A area. The algorithm will identify this inconsistency and start uncertainty analysis. It may judge that the B area is the current active pollution source, while the A area is a potential old pollution source or a secondary pollution source. The final probability map will reflect this judgment comprehensively, and provide multiple possibility analysis for management personnel.

[0106] It can be understood that by automatically comparing different results through an algorithm, the conclusion is no longer a black box output, but a verified and interpretable one. The finally generated "pollution source probability distribution map" is a visual heat map, so even non-professionals can immediately identify the key areas for investigation, greatly improving the efficiency and accuracy of the traceability work.

[0107] In step S105, according to the triggered warning level and the probability distribution map of the pollution source, a model predictive control optimization algorithm is used for multi-objective optimization solution, and an optimal control scheme with gate dam group collaborative scheduling as the core is calculated and specific execution instructions are generated.

[0108] It can be understood that the warning level determines the weight ratio of the optimization target. If the alarm level is reached, the water quality safety is prioritized at all costs, and the economic cost and device stability weights are reduced. If the warning level is reached, the treatment effect, economy and stability are balanced. If the attention level is reached, it may focus more on low-cost and small-scale monitoring scheduling. The pollution source probability distribution map provides crucial spatial information, which determines the main area of treatment. It tells the optimization algorithm which area is most likely to be polluted, and the algorithm will prioritize scheduling water conservancy facilities upstream or downstream of the area, rather than blindly operating throughout the entire watershed.

[0109] The model predictive control optimization algorithm is not a one-time static planning, but a rolling optimization strategy. Optimization is carried out around a mathematical function that quantifies the comprehensive cost of a "good" scheme:

[0110]

[0111] Among them, In order to minimize the comprehensive cost, is the water quality target, aiming to minimize the pollutant concentration at sensitive points such as spring outlets. is the stability target, aiming to avoid excessive or intense action of gates and pumps, protect equipment, and ensure system stability. is the economic target, aiming to minimize treatment cost. 、 、 is the weight coefficient, and the three coefficients are dynamically determined by the warning level. When the warning level is alarm, the weight is very large, 、 the weight is very small; when the warning level is warning, the weights of the three are comparable, seeking balance; when the warning level is attention, the weight is relatively large.

[0112] Specifically, the model predictive control optimization algorithm minimizes pollution concentration, gate operation amount and treatment economic cost as the target to construct a multi-objective cost function, and performs rolling optimization solution in a limited time domain, wherein the multi-objective cost function is:

[0113]

[0114] wherein, to minimize the comprehensive cost, to minimize the deviation of pollution concentration from the target value, to minimize the variation range of control variables, to minimize the economic cost of treatment, , , are weight coefficients.

[0115] It is easy to imagine that, is the above , is the above , is the above .

[0116] The optimization solver will find a control sequence that minimizes the total cost from thousands of possible scheduling combinations. The optimal control scheme usually includes a combination of the following strategies: upstream dilution: scheduling the reservoir or dam upstream of the pollution source to increase the discharge flow, flushing and diluting the pollutants with clean water, and occupying the water flow channel; downstream throttling: scheduling the dam downstream of the pollution group to close or reduce the discharge flow at the right time, artificially raising the local groundwater level, slowing down or even temporarily blocking the advance speed of the pollution group to the spring, and gaining time for treatment; lateral diversion: by scheduling, the polluted water body is guided to the preset artificial wetland or ecological pond for natural treatment, avoiding its direct entry into sensitive water areas.

[0117] The final execution instructions are specific and executable, for example: instruction 1: linearly adjust gate A from the current 30% opening to 55% opening within 600 seconds; instruction 2: start pump station B and set its speed to 1200 RPM, target flow rate 40 m³ / h; instruction 3: 24 hours later, restore gate A to 40% opening.

[0118] These instructions are directly issued to the PLC (Programmable Logic Controller) on site through the industrial control system (such as SCADA system) for automatic execution.

[0119] It can be understood that the generated decision instructions are based on model simulation and optimization algorithms rather than empiricism, making the generation of instructions more scientific; The entire river basin water conservancy facilities are considered as a whole for coordinated scheduling, realizing "one chess" management; From early warning to generating instructions to execution, the whole process is automated, realizing unprecedented response speed and processing accuracy.

[0120] In step S106, the executing mechanism receives the execution instructions and accurately adjusts the gate opening or pump station speed to achieve the management goal of hydraulic coordination and control.

[0121] Specifically, the executing mechanism receives the accurate instructions described above, and accurately adjusts the gate opening or pump station speed according to the instructions, for example, gate A is linearly adjusted from the current 30% opening to 55% opening within 600 seconds, and pump station B is started, and its speed is set to 1200 RPM, and the target flow is 40 m³ / h.

[0122] It should be noted that after the above step S106, there is also an effect feedback and adaptive optimization step: the execution state of the executing mechanism and the environmental response data after management are fed back to the digital twin model in real time, which is used to evaluate the management performance and optimize the model.

[0123] It should be noted that the data content of the execution state feedback includes the actual opening, actual speed, actual flow, current, voltage and other state data of the executing mechanism (gate, pump station). The purpose is to verify whether the instructions are accurately executed. For example, the output instruction is to open the gate to 50%, but the actual feedback is only 48%, which indicates that there may be mechanical jamming or error, so this deviation will be recorded. The data content of the environmental response feedback is the change data of the environmental parameters such as groundwater level, water quality (pollutant concentration), temperature after the management starts. The purpose is to evaluate the actual effect of the management action on the real environment. This is the final standard for testing whether the decision is correct.

[0124] After receiving the feedback data, the corresponding module will automatically perform quantitative evaluation of the management performance: according to the evaluation indexes such as concentration reduction rate, pollution plume control efficiency, cost-benefit ratio and instruction execution coincidence degree, the performance report of this management action is automatically generated, for example: "This time, the resistance action took 12 hours, the nitrate concentration at the spring mouth decreased from 25 mg / L to 12 mg / L, the reduction rate was 52%, a total of 3XXX degrees of electric energy was consumed, and the average coincidence degree of instruction execution was 99.5%."

[0125] At the same time, the actual water level and water quality response data during the treatment are taken as new and high-quality inputs, and are input into the digital twin model again for data assimilation. The model parameters (such as the permeability coefficient K and the dispersivity) are adjusted in reverse optimization to make the model simulate the system response during the treatment more accurately. This means that the model has a better understanding of the behavior of groundwater through this practice.

[0126] During the analysis of the "instruction-effect" data, it may be found that the weight coefficients in the cost function are not reasonable. For example, it may be found that a little more electricity bill ( The weight can be appropriately reduced) can be exchanged for a substantial increase in treatment effect ( weight). The system will accordingly fine-tune the weights to make the next decision better.

[0127] It can be understood that through the feedback optimization closed loop, the prediction and decision making can be more and more in line with the characteristics of the actual basin, and the accuracy continues to improve; it can adapt to changes in the environment (such as new engineering buildings changing hydrogeological conditions, new pollution sources appearing), and through self-learning to adjust itself, without the need for manual reprogramming or large-scale adjustment of the model, so that the model can automatically and continuously evolve.

[0128] According to the karst groundwater intelligent early warning and coordination method proposed in the embodiments of the present application, by deploying pollution trigger type intelligent tracer capsule array, distributed optical fiber sensing network and multi-parameter sensor node group, 24-hour real-time monitoring and minute-level data acquisition are realized, and the early warning mode is innovated from "alarm after exceeding the standard" to "predictive early warning before pollution occurs or in the early stage of diffusion", which gains valuable time for response; through the periodic active hydraulic disturbance diagnosis mechanism, the aquifer permeability coefficient field is inverted and updated, the channel patency is evaluated, and early diagnosis of system health is realized; by fusing multiple tracing technologies such as concomitant equation method and particle backtracking algorithm, and combining with the field evidence of intelligent tracer, fast, accurate and reliable positioning of the pollution source is realized, and the great difficulty of pollution source tracing in karst channels is overcome; through digital twin and model predictive control optimization algorithm for multi-objective optimization solution, the treatment effect and economic cost of different schemes can be automatically simulated, and a scientific optimal solution is given, which fundamentally improves the scientificity and accuracy of decision making; the extensive treatment mode is abandoned, and the precise treatment strategy of "hydraulic coordination control" is proposed, the intelligent scheduling of the gate dam and pump station in the spring basin is realized, the clean water flow is used to occupy the channel, block the pollution group, and guide it to the preset treatment area, and the low-cost, high-efficiency and environment-friendly treatment goal is realized.

[0129] Thus, the technical problems of pollution discovery lag, tracing difficulty and treatment difficulty are solved.

[0130] In some embodiments, a famous karst spring basin in northern China is an important source of drinking water and a tourist attraction. Within its recharge area, there are farmlands and scattered towns. In the 2024 rainy season, the system successfully warned and disposed of an agricultural non-point source nitrate pollution event.

[0131] In the karst fissures downstream of the main farmland area in the recharge area and near the sewage outlets of the towns, dozens of nitrate-triggered capsules (shell sensitive to nitrate concentration, rhodamine WT inside) were pre-embedded. And along the known main groundwater runoff channels and around the spring mouth, a few kilometers of temperature / vibration optical fibers were laid; In the monitoring wells at the spring mouth, the water intake of the water source, and the main runoff path, sensor nodes that can monitor water level, pH, conductivity, and nitrate concentration in real time were laid. All equipment transmits data back to the central platform every 5 minutes through Internet of Things technology.

[0132] The digital twin model assimilated rainfall forecast data and predicted that the nitrate concentration at the spring mouth would rise in the next 72 hours. The system carried out the planned quarterly "active hydraulic disturbance" and pumped water at the spring mouth for 6 hours. By analyzing the response curve, it was found that the response speed of monitoring point A in the northern agricultural area was significantly slower than the historical "health fingerprint", indicating that the patency of the channel had decreased and there was a possibility of accumulation, making it easier for pollutants to accumulate. At 10 am on a certain day, the model predicted that the concentration would continue to rise, but it had not yet exceeded the standard. Because of the identification of abnormal point A, the system automatically triggered a "attention" level warning (yellow signal), prompting the management personnel to pay attention to the northern area.

[0133] At 2 pm on the same day, an intelligent capsule downstream of the northern farmland area was triggered, releasing a rhodamine WT tracer. At the same time, the sensor showed that the nitrate concentration in this area had risen sharply. The system immediately upgraded to the "warning" level (orange signal).

[0134] Through the adjoint equation method, it is quickly calculated that the sensitivity field of the spring mouth concentration has the highest contribution rate in area B downstream of the northern farmland area. Through particle backtracking, the simulation shows that a large number of particles end in area B. Through cross-validation, it is known that area B is highly coincident with the location of the intelligent capsule trigger. Automatically generate a "high-confidence pollution source probability distribution map" to accurately locate the pollution source within an area of about 0.5 square kilometers in the northern farmland area.

[0135] The model predicts that the pollution plume will affect the water quality of the spring mouth in 36 hours. The system automatically triggers an "alarm" level (red signal).

[0136] Through the model predictive control optimization algorithm, the multi-objective optimization solution is finally calculated as follows:

[0137] Instruction 1: Immediately open the gate of the ecological reservoir G1 in the northern upstream area and increase the discharge by 20% to dilute and flush the pollutants with clean water.

[0138] Instruction two: close the regulating gate Z2 downstream of the pollution plume by 30%, artificially raise the water level, form a "hydraulic barrier", and slow down the advance of the pollution plume towards the spring outlet.

[0139] Instruction three: divert a portion of the water body to the artificial wetland in the abandoned mine pit on the east side for natural purification.

[0140] The instructions are issued to the PLC of G1 reservoir and Z2 gate through industrial Internet of Things, and the actuator automatically completes precise adjustment.

[0141] The actual opening of G1 gate coincides with the instruction with a degree of 99%. The monitoring network shows that the concentration in the northern region begins to decrease, and the migration speed of the pollution plume slows down significantly. After 36 hours, the concentration at the spring outlet only fluctuates slightly, which is far below the warning line. The system evaluation report shows: "This control is successful, avoiding a major pollution event, with a power consumption cost of only XXX yuan."

[0142] The data of the whole process of this event (such as the actual effect of Z2 gate closing by 30% on water level rise) is used for incremental learning, optimizing the parameters of the digital twin model and the cost function of MPC, and the system becomes more "intelligent".

[0143] Second, with reference to the accompanying drawings, the karst groundwater intelligent early warning and coordination system according to the embodiments of the present application is described.

[0144] In this embodiment, the anomaly is perceived through a three-dimensional monitoring network, early warning is achieved through digital twin prediction and active diagnosis, the pollution source is quickly locked through multi-technology fusion tracing, precise positioning of the pollution source is achieved; water conservancy facilities are intelligently dispatched to achieve hydraulic control; feedback is continuously optimized through a closed loop, and the system becomes more intelligent with use.

[0145] Figure 2 is a structural schematic diagram of the karst groundwater intelligent early warning and coordination system according to the embodiments of the present application.

[0146] As shown in Figure 2 , the karst groundwater intelligent early warning and coordination system 10 includes an intelligent perception module 100, a data fusion and communication module 200, a digital twin and early warning module 300, an intelligent tracing module 400, a collaborative decision engine 500, an execution control module 600, and an adaptive module 700.

[0147] The intelligent perception module 100 includes an intelligent tracer capsule array, a distributed optical fiber sensing network and a multi-parameter sensor node group, is used for executing an active water disturbance mechanism and collecting multi-source heterogeneous data; the data fusion and communication module 200 is used for cleaning, aligning and transmitting the multi-source heterogeneous data; the digital twin and early warning module 300 internally has a digital twin model, is used for completing simulation, prediction and triggering a hierarchical early warning; the intelligent tracing module 400 is used for running a concomitant equation method and a particle backtracking algorithm, and generates a pollution source probability distribution map; the collaborative decision engine 500 is used for running a model predictive control optimization algorithm, and generates an optimal blocking control scheme and an execution instruction; the execution control module 600 is used for receiving and executing the instruction, and controls an execution mechanism to complete an action; and the adaptive module 700 is used for completing governance performance evaluation and model incremental learning.

[0148] It should be noted that the foregoing explanation and description of the embodiment of the karst groundwater intelligent early warning and collaboration method also applies to the karst groundwater intelligent early warning and collaboration system of the embodiment, which will not be described here again.

[0149] The karst groundwater intelligent early warning and collaboration system according to the embodiment of the present application deploys a pollution trigger type intelligent tracer capsule array, a distributed optical fiber sensing network and a multi-parameter sensor node group, realizes 24-hour real-time monitoring and minute-level data collection, and innovates the early warning mode from “alarm after exceeding the standard” to “predictive early warning before pollution occurs or in the early stage of diffusion”, so as to gain valuable time; through a periodic active water disturbance diagnosis mechanism, the aquifer permeability coefficient field is inverted and updated, the channel patency is evaluated, and early diagnosis of system health is realized; through fusion of various tracing technologies such as the concomitant equation method and the particle backtracking algorithm, and in combination with the field evidence of the intelligent tracer, fast, accurate and reliable positioning of the pollution source is realized, and the great difficulty of pollution tracing in the karst channel is overcome; through digital twin and model predictive control optimization algorithm, multi-objective optimization solution is realized, the governance effect and economic cost of different schemes can be simulated automatically, and a scientific optimal solution is given, so as to fundamentally improve the scientificity and accuracy of decision-making; the extensive governance mode is abandoned, and a “hydraulic collaborative blocking” precise governance strategy is proposed, the intelligent scheduling of the gate dam and the pump station in the spring area is realized, the channel is occupied by clean water flow, the pollution group is blocked, and the pollution group is guided to the preset treatment area, so as to realize the governance goal of low cost, high efficiency and environmental friendliness.

[0150] Thus, the technical problems of pollution discovery lag, tracing difficulty and large governance difficulty are solved.

[0151] Figure 3 The structure schematic diagram of the electronic device provided by the embodiment of the present application is shown. The electronic device can include:

[0152] The memory 301, the processor 302 and the computer program stored in the memory 301 and executable on the processor 302.

[0153] The processor 302 implements the karst groundwater intelligent early warning and coordination method provided in the above embodiments when executing a program.

[0154] Further, the electronic device further comprises:

[0155] The communication interface 303 is configured to communicate between the memory 301 and the processor 302.

[0156] The memory 301 is configured to store a computer program executable on the processor 302.

[0157] The memory 301 can include a high-speed RAM memory, and can also include a non-volatile memory, for example, at least one disk memory.

[0158] If the memory 301, the processor 302 and the communication interface 303 are implemented independently, the communication interface 303, the memory 301 and the processor 302 can be connected to each other through a bus and complete communication between each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 3 In the figure, only one thick line is used to represent, but it does not mean that there is only one bus or one type of bus.

[0159] Optionally, in a specific implementation, if the memory 301, the processor 302 and the communication interface 303 are integrated on a chip, the memory 301, the processor 302 and the communication interface 303 can complete communication between each other through an internal interface.

[0160] The processor 302 can be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0161] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the karst groundwater intelligent early warning and coordination method as above.

[0162] The embodiment of the present application further provides a computer program product, which stores a computer program, and the program is executed by a processor to realize the karst groundwater intelligent early warning and cooperation method.

[0163] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the specification and the features of the different embodiments or examples without contradiction.

[0164] In addition, the terms "first", "second" are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise explicitly specified.

[0165] Any process or method descriptions in flow charts or otherwise described herein can be understood as representing code modules, segments, or portions of code that include one or more executable instructions for performing specific logic functions (or steps) or portions of the application, and the various embodiments of the application include additional or fewer functions performed in the same order or in a different order, combined with previously described functions, performed with previously described functions in a different manner, or performed using a different structure of code from those described or illustrated. It is understood that many modifications can be made by one skilled in the art, and all modifications that fall within the scope of the underlying principles of the application are intended to be part of this application.

[0166] It should be understood that various portions of the application can be implemented in hardware, software, firmware or a combination thereof. In the above embodiments, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. As in another embodiment, if implemented in hardware, any of the following technologies known in the art or their combinations can be used: discrete logic circuit with logic gate circuit for implementing logic functions on data signals, application specific integrated circuit with suitable combination of logic gate circuits, programmable gate array (PGA), field programmable gate array (FPGA), etc.

[0167] Those skilled in the art can understand that all or part of the steps of the foregoing method embodiments can be completed by programs instructing relevant hardware, and the programs can be stored in a computer readable storage medium. When the programs are executed, the steps of the method embodiments or a combination thereof are included.

Claims

1. A method for intelligent early warning and collaborative control of karst groundwater, characterized in that, Includes the following steps: S1. Deploy a pollution-triggered intelligent tracer capsule array, a distributed optical fiber sensor network, and a multi-parameter sensor node group to collaboratively monitor target pollutants and multiple physicochemical parameters in karst groundwater, collecting multi-source heterogeneous data. The pollution-triggered intelligent tracer capsule array has capsule shells made of environmentally responsive materials, containing fluorescent tracers, and the degradation rate of the capsule shell is positively correlated with the concentration of the target pollutant, thus triggering release when pollution levels exceed standards. The distributed optical fiber sensor network is used to collect distributed physical field data. The multi-parameter sensor node group is used to collect water quality and hydrological data at specific locations. The multi-source heterogeneous data includes the distributed physical field data and the water quality and hydrological data. The distributed physical field data includes temperature field data and vibration field data. The water quality and hydrological data includes water level, pH, conductivity, dissolved oxygen, turbidity, and target ion concentration data. S2. Construct a digital twin model, integrate the multi-source heterogeneous data, and introduce a periodic active hydraulic disturbance diagnosis mechanism. Dynamically update the state of the digital twin model using data assimilation technology to achieve real-time simulation and multi-step advance prediction of the flow field and solute concentration field. The periodic active hydraulic disturbance diagnosis mechanism includes: artificially creating a hydraulic excitation by conducting short-term, quantitative pumping or injection experiments at a preset downstream location through periodic setting or emergency triggering; the digital twin model analyzes the water level response, water quality response, and tracer response of each upstream monitoring point to this excitation, inverting and updating the permeability coefficient field and porosity field parameters of the karst aquifer to achieve model calibration; storing the standard response curve measured during the disturbance period in a database, and comparing the newly measured response curve with the standard response curve in subsequent periodic disturbances to identify anomalies and achieve early anomaly location; S3. Compare the prediction results of the digital twin model with the dynamic threshold library to automatically trigger graded early warning signals with three levels: attention, warning, and alarm. S4. In response to the graded early warning signal, the location of the pollution source is located in reverse by using the adjoint equation method and the particle backtracking algorithm, and the trigger signals of the intelligent tracer capsule array are cross-validated to generate a pollution source probability distribution map. S5. Based on the triggered warning level and the pollution source probability distribution map, a model predictive control optimization algorithm is used to perform multi-objective optimization, calculate the optimal control scheme with the coordinated scheduling of the dam group as the core, and generate specific execution instructions. The multi-objective optimization of the model predictive control optimization algorithm includes: constructing a multi-objective cost function with the objectives of simultaneously minimizing pollution concentration, gate operation volume, and treatment economic cost, and performing rolling optimization within a finite time domain. The multi-objective cost function is: ； in, To minimize overall cost, To minimize the deviation of pollution concentration from the target value, To minimize the magnitude of change in the control variables, To minimize the economic cost of governance, , , These are the weighting coefficients; S6. The actuator receives the execution command and precisely adjusts the gate opening or pump station speed to achieve the governance goal of hydraulic coordinated resistance control.

2. The intelligent early warning and collaborative control method for karst groundwater according to claim 1, characterized in that, The warning thresholds in the dynamic threshold library are dynamically calculated and generated by the digital twin model based on the current water flow state, historical background concentration, and the uncertainty of the prediction itself. The core calculation formula is: ； in, For the final dynamic threshold, As the background baseline value, For safety reasons, The standard deviation of the predicted value. The correlation coefficient, The current simulated flow rate, This represents the historical average flow velocity.

3. The intelligent early warning and collaborative control method for karst groundwater according to claim 1, characterized in that, The adjoint equation method constructs and solves the adjoint equation of the concentration field at the observation point to the release intensity of the pollution source, obtains the sensitivity field, and normalizes the sensitivity field to obtain the pollution source contribution probability distribution map. The particle backtracking algorithm releases a large number of virtual particles at the warning point and uses the reverse flow field provided by the digital twin model to simulate the reverse migration of the particles, statistically analyzes the spatial distribution of all particles at the end of the simulation, and generates a probability distribution map. The cross-validation compares the pollution source contribution probability distribution map and the probability distribution map with the actual triggering position of the smart tracer capsule to generate the final high-confidence pollution source probability distribution map.

4. The intelligent early warning and collaborative control method for karst groundwater according to claim 1, characterized in that, Step S6 is followed by an effect feedback and adaptive optimization step: the execution status of the implementing agency and the environmental response data after governance are fed back to the digital twin model in real time to evaluate governance performance and to perform incremental learning and optimization of the model.

5. A karst groundwater intelligent early warning and collaborative system, used to implement the method described in any one of claims 1-4, characterized in that, include: The system comprises an intelligent sensing module, a data fusion and communication module, a digital twin and early warning module, an intelligent traceability module, a collaborative decision-making engine, an execution control module, and an adaptive module. The intelligent sensing module includes an intelligent tracer capsule array, a distributed optical fiber sensor network, and a group of multi-parameter sensor nodes, which are used to execute an active hydraulic disturbance mechanism and collect multi-source heterogeneous data. The data fusion and communication module is used to clean, align, and transmit the multi-source heterogeneous data; The digital twin and early warning module has a built-in digital twin model, which is used to simulate, predict and trigger graded early warnings; The intelligent source tracing module is used to run the adjoint equation method and particle backtracking algorithm to generate a pollution source probability distribution map; The collaborative decision engine is used to run model predictive control optimization algorithms to generate optimal resistance and control schemes and execution instructions; The execution control module is used to receive and execute the instructions, and control the execution mechanism to complete the action; The adaptive module is used to complete governance performance evaluation and incremental model learning.

6. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the intelligent early warning and collaborative control method for karst groundwater as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the intelligent early warning and collaborative control method for karst groundwater as described in any one of claims 1-4.

Images