Karst groundwater pollution treatment method based on near-source interception and drainage and clean-up diversion

Abstract

The invention relates to the technical field of environmental pollution abatement, in particular to a karst groundwater pollution abatement method based on near-source interception and drainage and sewage disposal diversion. The method comprises the following steps: acquiring comprehensive survey data of a target treatment area through basic geology, hydrogeology and environmental geology survey; based on the comprehensive survey data, a scheme type judgment model is used to obtain a treatment scheme type analysis result, and the treatment scheme type analysis result comprises a near-source interception and drainage scheme and a sewage disposal and diversion scheme; judging according to the treatment scheme type analysis result, and if the treatment scheme type is a near-source interception scheme, calling a near-source interception scheme decision AI model and optimizing to form a final near-source interception treatment scheme; otherwise, calling a sewage disposal and diversion scheme decision AI model, and generating a sewage disposal and diversion treatment scheme. According to the karst groundwater pollution treatment method based on near-source interception and drainage and clean-up diversion, the problems that an existing method is unstable in treatment effect, high in cost, complex in operation and the like are solved.

Description

Technical Field

[0001] The invention relates to the technical field of environmental pollution control, and in particular to a karst groundwater pollution control method based on near-source interception and drainage and clean and polluted water diversion. Background Art

[0002] Due to its unique hydrogeological structure, the groundwater system in karst areas is complex and vulnerable to pollution. Pollutants such as industrial wastewater, agricultural pollution and domestic sewage can easily enter the groundwater system through cracks and caves, leading to serious groundwater pollution problems.

[0003] Traditional groundwater pollution control methods have problems such as unstable control effects, high costs, and complex operations, and are difficult to meet the special needs of karst areas.

[0004] At present, the main problems in the control of karst groundwater pollution are as follows: the design of control plans relies more on expert experience, the formulation of control plans lacks scientific basis, the plan design cycle is long, and it is difficult to make targeted adjustments and optimizations to the target control areas; the implementation and operation of control plans lacks effective real-time monitoring and control methods, making it difficult to cope with the impact of environmental changes on control results. Summary of the invention

[0005] The technical problem solved by the present invention is to provide a karst groundwater pollution control method based on near-source interception and clean and polluted water diversion, which can solve the problems existing in the design process of existing control schemes, such as reliance on personal experience, long scheme design cycle, and insufficient implementation supervision and regulation.

[0006] The basic solution provided by the present invention is:

[0007] Karst groundwater pollution control methods based on near-source interception and clean-up and pollution diversion include:

[0008] Obtain comprehensive survey data of the target governance area through geological and hydrological surveys, wherein the geological and hydrological surveys include basic geological surveys, hydrogeological surveys and environmental geological surveys; the comprehensive survey data include geological data, hydrological data and environmental pollution source data;

[0009] Based on the comprehensive survey data, the scheme type discrimination model is used to obtain the treatment scheme type analysis results, which include: near-source interception and drainage scheme and pollution diversion scheme;

[0010] Make a judgment based on the analysis results of the governance plan type. If the governance plan type is a near-source interception and drainage plan, call up the near-source interception and drainage plan decision AI model and combine it with the optimization algorithm to form the final near-source interception and drainage governance plan;

[0011] If the type of treatment plan is a clean-up and pollution diversion plan, the clean-up and pollution diversion plan decision AI model is called to generate a clean-up and pollution diversion treatment plan.

[0012] Furthermore, the near-source interception and drainage scheme decision AI model includes a first near-source interception and drainage scheme decision AI model and a second near-source interception and drainage scheme decision AI model;

[0013] The AI ​​model for decision-making on near-source interception and drainage is retrieved and combined with the optimization algorithm to form the final near-source interception and drainage management plan, including:

[0014] According to the analysis results of the treatment plan type, if the treatment plan type is a near-source interception and drainage plan, the first near-source interception and drainage plan decision AI model is retrieved to generate the first near-source interception and drainage treatment plan;

[0015] Based on the comprehensive survey data, the governance plan data with a matching degree exceeding the preset value is retrieved from the historical governance plan database to generate a temporary training set, and the first near-source interception and drainage plan decision-making AI model is strengthened to form a second near-source interception and drainage plan decision-making AI model. Through the second near-source interception and drainage plan decision-making AI model, a second near-source interception and drainage governance plan is generated;

[0016] Determine the difference between the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme. If the difference is less than or equal to a preset value, the second near-source interception and drainage control scheme is adopted as the final near-source interception and drainage control scheme. If the difference is greater than the preset value, the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme are used as initial schemes and iteratively optimized based on the optimization algorithm to form the final near-source interception and drainage control scheme.

[0017] Furthermore, the first near-source interception and drainage treatment scheme and the second near-source interception and drainage treatment scheme are used as initial schemes and iteratively optimized based on the optimization algorithm to form a final near-source interception and drainage treatment scheme, including:

[0018] An initial scheme generating step, generating an initial scheme group as a current scheme group according to the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme;

[0019] The fitness calculation step is to calculate the fitness of each individual solution in the current solution group;

[0020] Iterative optimization step: sort the individual solutions according to their fitness, select the top N individual solutions for solution cross-mutation, form an iterative solution group, and use the iterative solution group as the current solution group to repeatedly execute the fitness calculation step and iterative optimization step until the end condition is reached;

[0021] The individual scheme with the highest fitness is selected to correct the individual parameters of the scheme to the second near-source interception and drainage control scheme, and the corrected second near-source interception and drainage control scheme is used as the final near-source interception and drainage control scheme.

[0022] Furthermore, generating an initial solution group according to the first near-source interception and drainage treatment solution and the second near-source interception and drainage treatment solution specifically includes:

[0023] Obtain the difference between the first near-source interception and drainage treatment plan and the second near-source interception and drainage treatment plan;

[0024] Arrange the values ​​corresponding to the difference items of the first near-source interception and drainage treatment scheme as the first vector V1;

[0025] Arrange the values ​​corresponding to the difference items of the second near-source interception and drainage treatment scheme as the second vector V2;

[0026] Calculate the difference vector V between the second vector and the first vector d =V2-V1;

[0027] Calculate the intermediate vector V c =V1+a c *V d , where 0-r c <1+r, c is the serial number of the intermediate vector, r is the expansion radius;

[0028] Randomly adjust the first vector, the second vector, and all intermediate vectors:

[0029] V o =V i +b i , b i is a random vector with a modulus less than r; V i The first vector, the second vector or the intermediate vector is input; V o is the randomly adjusted vector;

[0030] All randomly adjusted vectors are taken as individual solutions, and the initial solution group is obtained by eliminating the individual solutions that do not meet the boundary conditions.

[0031] Furthermore, the geological data include basic geological characteristics, geological structural characteristics, and karst development characteristics;

[0032] The hydrological data include groundwater system characteristics, groundwater main runoff belt characteristics, groundwater dynamic condition characteristics and groundwater dynamic characteristics;

[0033] The environmental pollution source data includes soil pollution status data, groundwater pollution status data, pollution channel data and pollutant migration pattern data.

[0034] Furthermore, the geological and hydrological survey includes basic geological survey, which includes:

[0035] ​The basic features of the address are obtained through map data, and the map data and image data collected by drones are input into the joint and fissure identification model to identify the joints and fissures and extract their trend features;

[0036] The karst morphology and elevation distribution of the target treatment area are recorded and counted to obtain the karst morphological characteristics, karst plane characteristics and elevation distribution characteristics; through drilling detection, the karst development conditions in each hole depth range are obtained to obtain the karst vertical characteristics.

[0037] Furthermore, the scheme type discrimination model adopts a support vector machine model, which uses groundwater flow direction, pollution source location, karst development characteristics, groundwater main runoff zone characteristics, groundwater dynamic condition characteristics, groundwater dynamic characteristics, soil pollution status, groundwater pollution status, pollution channels and pollutant migration patterns as input feature dimensions of the support vector machine.

[0038] Furthermore, it also includes:

[0039] Implement the pollution diversion and treatment plan or the final near-source interception and treatment plan, collect the operating data of each device and regularly update the environmental pollution source data;

[0040] The water pollution control situation is judged based on the environmental pollution source data, and the distribution status of pollution levels in different areas is indicated by color.

[0041] Furthermore, it also includes:

[0042] The equipment operating parameters are automatically adjusted according to weather and rain forecast data as well as environmental pollution source data.

[0043] The beneficial effects of the present invention are:

[0044] 1. The support vector machine model is used to automatically identify the type of treatment plan by combining geological data, hydrological data and environmental pollution source data, thereby improving the scientificity and accuracy of the treatment plan selection.

[0045] 2. The pre-trained first near-source interception and drainage scheme decision-making AI model and the temporary enhanced training second near-source interception and drainage scheme decision-making AI model are used together. On the one hand, temporary enhanced training can be used to improve the pertinence of the scheme to the characteristics of the current target governance area; on the other hand, the mutual comparison of the results of the two models can also verify the generation results of the two models; the governance scheme is iteratively optimized using the optimization algorithm to ensure that the generated governance scheme has the best governance effect and the lowest cost. The decision-making AI model is strengthened by using historical governance scheme data to improve the accuracy and generalization ability of the model; during the optimization training process, the first near-source interception and drainage governance scheme and the second near-source interception and drainage governance scheme are used as the initial scheme to improve the quality of the initial scheme and the convergence speed of the optimization algorithm. The difference terms of the first near-source interception and drainage governance scheme and the second near-source interception and drainage governance scheme are used to expand the scheme based on the vector difference to form the initial scheme. A relatively comprehensive initial scheme individual can be quickly formed around the first near-source interception and drainage governance scheme and the second near-source interception and drainage governance scheme, while ensuring a certain degree of randomness. That is, the quality and convergence speed of the initial scheme are improved, and the globality of the optimal solution is guaranteed.

[0046] 3. Collect and analyze data during the treatment process in real time, and display the distribution status of pollution levels through color coding, so that personnel can adjust the treatment strategy in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 The present invention is a flowchart of a method for treating karst groundwater pollution based on near-source interception and drainage and clean and polluted water diversion according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following is further described in detail through specific implementation methods:

[0049] Embodiment 1

[0050] like Figure 1 As shown, the karst groundwater pollution control method based on near-source interception and drainage and clean and polluted water diversion of this embodiment includes:

[0051] Obtain comprehensive survey data of the target governance area through geological and hydrological surveys, wherein the geological and hydrological surveys include basic geological surveys, hydrogeological surveys and environmental geological surveys; the comprehensive survey data include geological data, hydrological data and environmental pollution source data;

[0052] In this embodiment, the geological data include basic geological characteristics, geological structure characteristics, and karst development characteristics; the hydrological data include groundwater system characteristics, groundwater main runoff belt characteristics, groundwater dynamic condition characteristics, and groundwater dynamic characteristics; the environmental pollution source data include soil pollution status data, groundwater pollution status data, pollution channel data, and pollutant migration pattern data.

[0053] Among them, the basic geological characteristics include geographical location, geological type, longitude and latitude coordinates, altitude range, etc.; the geological structure characteristics include joints, fissure direction characteristics, etc.; the karst development characteristics include karst morphological characteristics, karst plane characteristics, karst vertical characteristics, etc.

[0054] The characteristics of the groundwater system include the underground horizontal plane zoning characteristics, the groundwater vertical stratification characteristics, the spatial structure characteristics of the underground aquifer, etc.; the characteristics of the groundwater dynamic conditions include the groundwater flow velocity characteristics, the groundwater hydraulic gradient characteristics, the groundwater flow field characteristics, etc.; the dynamic characteristics of the groundwater include the seasonal precipitation characteristics, the upstream and downstream water storage and replenishment characteristics, etc.

[0055] Soil pollution status data and groundwater pollution status data both include pollution source location data, pollutant type data, pollutant range data, and pollution degree distribution data.

[0056] The geological and hydrological survey includes:

[0057] Basic geological survey, obtain the basic characteristics of the address through map data, input the map data and image data collected by drone into the joint and fissure recognition model to identify the joints and fissures and extract the trend characteristics; in this implementation, the yolov5 recognition model is used to extract image data for joint and fissure recognition, and the broken line texture map of the joints and fissures is obtained by performing graphic processing and edge extraction on the identified area. The broken line texture map is simplified by the Douglas-Peucker algorithm to obtain the direction of each joint and fissure, and then the joint and fissure direction data of each partition in the target treatment area are obtained through fitting analysis of the direction, and the joint and fissure direction rose diagram is drawn.

[0058] The karst development characteristics are collected and counted, including recording the karst morphology of the target treatment area, recording and counting the types of karst depressions, sinkholes, funnels, karst shafts, large caves, underground stream entrances, natural bridges, etc., to obtain the karst morphological characteristics and karst plane characteristics. The elevation distribution of karst points is counted to obtain the elevation distribution of karst in each elevation section; through drilling detection, the karst development in each hole depth range is obtained to obtain the vertical characteristics of karst.

[0059] Hydrological survey, based on GIS geographic information system and field investigation, obtains underground horizontal plane zoning characteristics, groundwater vertical stratification characteristics, underground aquifer spatial structure characteristics, etc.; adopts high-density, audio, transient and charging methods to comprehensively detect and track groundwater source channels, and obtains the characteristics of the main groundwater runoff zone through drilling verification.

[0060] Based on the comprehensive survey data, a scheme type discrimination model is used to obtain treatment scheme type analysis results, which include: near-source interception and drainage scheme and pollution diversion scheme. In this embodiment, the scheme type discrimination model adopts a support vector machine model.

[0061] Specifically, feature selection is first performed to select features that have an important influence on the discrimination of the type of treatment plan. In this embodiment, groundwater flow direction, pollution source location, karst development characteristics, groundwater main runoff zone characteristics, groundwater dynamic condition characteristics, groundwater dynamic characteristics, soil pollution status, groundwater pollution status, pollution channels and pollutant migration patterns are used as input feature dimensions of the support vector machine.

[0062] Then, the data set is constructed to obtain historical governance data. The geological data, hydrological data and environmental pollution source data in the historical governance data are preprocessed, including data cleaning, missing value processing, outlier processing and data standardization. Specifically, the noise and irrelevant information in the data are removed through data cleaning to ensure the accuracy of the data. The missing values ​​are filled by the mean filling, median filling and interpolation methods. The Z-score method is used to detect outliers, and the filling algorithm is used to correct and fill out the outliers. The Min-Max standardization is used to standardize the data to the same scale. The selected features are extracted to generate feature vectors for the input of the support vector machine model.

[0063] Then the SVM model is trained: the historical governance data is trained using the support vector machine algorithm to generate a solution type discrimination model. During the training process, the model parameters are optimized using the cross-validation method to ensure the accuracy and generalization ability of the model. The generated solution type discrimination model is verified using the cross-validation method. The verification indicators include accuracy, recall rate, F1 score, etc. The solution type discrimination model is obtained after the verification.

[0064] The current comprehensive survey data is input into the scheme type discrimination model, and the treatment scheme type analysis results can be obtained through model prediction. The treatment scheme type analysis results include near-source interception and drainage schemes and pollution diversion schemes.

[0065] Make a judgment based on the analysis results of the governance plan type. If the governance plan type is a near-source interception and drainage plan, call up the near-source interception and drainage plan decision AI model and combine it with the optimization algorithm to form the final near-source interception and drainage governance plan;

[0066] If the type of treatment plan is a clean-up and pollution diversion plan, the clean-up and pollution diversion plan decision AI model is called to generate a clean-up and pollution diversion treatment plan.

[0067] The near-source interception and drainage scheme decision-making AI model includes a first near-source interception and drainage scheme decision-making AI model and a second near-source interception and drainage scheme decision-making AI model;

[0068] The AI ​​model for decision-making on near-source interception and drainage is retrieved and combined with the optimization algorithm to form the final near-source interception and drainage management plan, including:

[0069] According to the analysis results of the treatment plan type, if the treatment plan type is a near-source interception and drainage plan, the first near-source interception and drainage plan decision AI model is retrieved to generate the first near-source interception and drainage treatment plan;

[0070] Based on the comprehensive survey data, the governance plan data with a matching degree exceeding the preset value is retrieved from the historical governance plan database to generate a temporary training set, and the first near-source interception and drainage plan decision-making AI model is strengthened to form a second near-source interception and drainage plan decision-making AI model. Through the second near-source interception and drainage plan decision-making AI model, a second near-source interception and drainage governance plan is generated;

[0071] Determine the difference between the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme. If the difference is less than or equal to a preset value, the second near-source interception and drainage control scheme is adopted as the final near-source interception and drainage control scheme. If the difference is greater than the preset value, the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme are used as initial schemes and iteratively optimized based on the optimization algorithm to form the final near-source interception and drainage control scheme.

[0072] In this embodiment, the pollution diversion scheme decision AI model and the first near-source interception and drainage scheme decision AI model both adopt CNN neural network models. The construction and training steps of the CNN neural network model include:

[0073] Construct a model, wherein the CNN neural network model includes a convolutional layer, a pooling layer, and a fully connected layer, and the fully connected layer includes an input layer, a hidden layer, and an output layer;

[0074] Obtain historical data on the treatment plans for near-source interception and drainage and clean-up and diversion, process and classify the treatment case data to form a data set, and divide the data set into a test set and a training set;

[0075] The initial decision model is trained based on the test set and the training set to obtain a trained governance solution decision model.

[0076] In this embodiment, the first near-source interception and drainage treatment scheme and the second near-source interception and drainage treatment scheme are used as initial schemes and iteratively optimized based on the optimization algorithm to form a final near-source interception and drainage treatment scheme, which specifically includes:

[0077] The initial scheme generation step generates an initial scheme group as the current scheme group according to the first near-source interception and drainage treatment scheme and the second near-source interception and drainage treatment scheme; generating the initial scheme group according to the first near-source interception and drainage treatment scheme and the second near-source interception and drainage treatment scheme specifically includes:

[0078] Obtain the difference between the first near-source interception and drainage treatment plan and the second near-source interception and drainage treatment plan. For example, the difference between the two plans lies in the different designs of the position coordinates of the interception wells, the different depth structure parameters of the interception wells, etc.

[0079] The values ​​corresponding to the difference items of the first near-source interception and drainage treatment scheme are arranged as the first vector V1=(X 11 ,…X 1m ), where X 1m The parameter value of the mth difference term in the first near-source interception and drainage treatment scheme;

[0080] The values ​​corresponding to the difference items of the second near-source interception and drainage treatment scheme are arranged as the second vector V2=(X 21 ,…X 2m ); where X 2m The parameter value of the mth difference term in the second near-source interception and drainage treatment scheme;

[0081] Calculate the difference vector V between the second vector and the first vector d =V2-V1;

[0082] Calculate the intermediate vector V c =V1+a c *V d , where 0-r c <1+r, c is the serial number of the intermediate vector, r is the extension radius, and in this implementation, r=0.2;

[0083] Randomly adjust the first vector, the second vector, and all intermediate vectors:

[0084] V o =V i +b i , b i is a random vector with a modulus less than r; V i The first vector, the second vector or the intermediate vector is input; V o is the randomly adjusted vector;

[0085] All randomly adjusted vectors are taken as individual solutions, and the individual solutions that do not meet the boundary conditions are eliminated to obtain the initial solution group; in this embodiment, the boundary conditions include cost boundary conditions, interception well quantity boundary conditions, interception well range boundary conditions, interception well parameter boundary conditions, etc.; after each individual solution is formed, it is determined whether the cost, interception well quantity, interception well area coordinates, interception well depth structure parameters, etc. are within the above-mentioned boundary range requirements. If not, it is a solution individual that does not meet the boundary conditions. Such an individual cannot meet the project implementation requirements and should be eliminated.

[0086] ​The fitness calculation step is to calculate the fitness of each individual solution in the current solution group. In this implementation, the fitness is calculated according to the following function:

[0087]

[0088] Where: C i is the total cost, T i is the total time, E i is the governance effect, and α, β, and γ are weights.

[0089] In this embodiment, E i The results of the simulation model are used to evaluate. Specifically, the current treatment plan is loaded into the mechanism simulation model, and the simulation is run to obtain the distribution of groundwater pollution concentration C after treatment. g ; Obtain the groundwater pollution concentration distribution C without treatment ung . Calculate the percentage reduction in pollution concentration after treatment: Among them, ∑(C g ) and ∑(C ung ) represent the total pollution concentration of each block in the target treatment area after treatment and before treatment, respectively.

[0090] Total cost C i =C con +C op , where C con is the construction cost of the interception well, including drilling, grouting and other costs. op The operation and maintenance costs of the interception wells, including the electricity costs and maintenance costs of the pumping equipment, are calculated through the cost budget system.

[0091] Iterative optimization step, sorting the individual schemes according to fitness, selecting the first N individual schemes for scheme crossover mutation, forming an iterative scheme group, in this embodiment, N is preferably 10, and in other embodiments of the present application, N can also select other values ​​according to the actual situation. The fitness calculation step and the iterative optimization step are repeated with the iterative scheme group as the current scheme group until the end condition is reached; in this embodiment, the optimization algorithm preferably adopts a genetic optimization algorithm, and the process of crossover mutation is the same as the process of generating an initial scheme group according to the first near-source interception and drainage governance scheme and the second near-source interception and drainage governance scheme, and the first N schemes are selected in pairs to expand to form an iterative scheme group, wherein r decreases as the number of iterations increases. In other embodiments of the present application, other optimization algorithms such as the ant lion optimization algorithm and the particle swarm optimization algorithm can also be used.

[0092] The individual scheme with the highest fitness is selected to correct the individual parameters of the scheme to the second near-source interception and drainage control scheme, and the corrected second near-source interception and drainage control scheme is used as the final near-source interception and drainage control scheme.

[0093] Embodiment 2

[0094] The difference between this embodiment and the first embodiment is that this embodiment further includes:

[0095] Implement the pollution diversion and treatment plan or the final near-source interception and treatment plan, collect the operating data of each device and regularly update the environmental pollution source data;

[0096] The water pollution control situation is judged based on the environmental pollution source data, and the distribution status of pollution levels in different areas is indicated by color.

[0097] According to the weather and rain forecast data and the environmental pollution source data, the equipment operating parameters are automatically adjusted. In this embodiment, an existing neural network model is adopted, with the weather and rain forecast data and the environmental pollution source data as input, and the operating parameters of the equipment involved, such as water pumps, gates, etc., as output, to train the corresponding control model, and output the equipment operating parameters according to the real-time acquired weather and rain forecast data and the environmental pollution source data, and perform automatic adjustment.

[0098] The above are only embodiments of the present invention. Common knowledge such as the known specific structures and characteristics in the scheme is not described in detail here. Ordinary technicians in the relevant field know all the common technical knowledge in the technical field to which the invention belongs before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A method for treating karst groundwater pollution based on near-source interception and drainage and clean and polluted water diversion, characterized by: include: Obtain comprehensive survey data of the target governance area through geological and hydrological surveys, wherein the geological and hydrological surveys include basic geological surveys, hydrogeological surveys and environmental geological surveys; the comprehensive survey data include geological data, hydrological data and environmental pollution source data; Based on the comprehensive survey data, the scheme type discrimination model is used to obtain the treatment scheme type analysis results, which include: near-source interception and drainage scheme and pollution diversion scheme; Make a judgment based on the analysis results of the governance plan type. If the governance plan type is a near-source interception and drainage plan, call up the near-source interception and drainage plan decision AI model and combine it with the optimization algorithm to form the final near-source interception and drainage governance plan; If the type of treatment plan is a clean-up and pollution diversion plan, the clean-up and pollution diversion plan decision AI model is called to generate a clean-up and pollution diversion treatment plan.

2. The karst groundwater pollution control method based on near-source interception and drainage and clean and polluted water diversion according to claim 1 is characterized by: The near-source interception and drainage scheme decision-making AI model includes a first near-source interception and drainage scheme decision-making AI model and a second near-source interception and drainage scheme decision-making AI model; The AI ​​model for decision-making on near-source interception and drainage is retrieved and combined with the optimization algorithm to form the final near-source interception and drainage management plan, including: Retrieve the first near-source interception and drainage plan decision-making AI model to generate the first near-source interception and drainage management plan; Based on the comprehensive survey data, the governance plan data with a matching degree exceeding the preset value is retrieved from the historical governance plan database to generate a temporary training set, and the first near-source interception and drainage plan decision-making AI model is strengthened to form a second near-source interception and drainage plan decision-making AI model. Through the second near-source interception and drainage plan decision-making AI model, a second near-source interception and drainage governance plan is generated; Determine the difference between the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme. If the difference is less than or equal to a preset value, the second near-source interception and drainage control scheme is adopted as the final near-source interception and drainage control scheme. If the difference is greater than the preset value, the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme are used as initial schemes and iteratively optimized based on the optimization algorithm to form the final near-source interception and drainage control scheme.

3. The karst groundwater pollution control method based on near-source interception and drainage and clean and polluted water diversion according to claim 2 is characterized by: The first near-source interception and drainage treatment scheme and the second near-source interception and drainage treatment scheme are used as the initial schemes and iteratively optimized based on the optimization algorithm to form the final near-source interception and drainage treatment scheme, including: An initial scheme generating step, generating an initial scheme group as a current scheme group according to the first near-source interception and drainage control scheme and the second near-source interception and drainage control scheme; The fitness calculation step is to calculate the fitness of each individual solution in the current solution group; Iterative optimization step: sort the individual solutions according to their fitness, select the top N individual solutions for solution cross-mutation, form an iterative solution group, and use the iterative solution group as the current solution group to repeatedly execute the fitness calculation step and iterative optimization step until the end condition is reached; The individual scheme with the highest fitness is selected to correct the individual parameters of the scheme to the second near-source interception and drainage control scheme, and the corrected second near-source interception and drainage control scheme is used as the final near-source interception and drainage control scheme.

4. The karst groundwater pollution control method based on near-source interception and drainage and clean and polluted water diversion according to claim 3 is characterized by: The generating of the initial solution group according to the first near-source interception and drainage treatment solution and the second near-source interception and drainage treatment solution specifically includes: Obtain the difference between the first near-source interception and drainage treatment plan and the second near-source interception and drainage treatment plan; Arrange the values ​​corresponding to the difference items of the first near-source interception and drainage treatment scheme as the first vector V1; Arrange the values ​​corresponding to the difference items of the second near-source interception and drainage treatment scheme as the second vector V2; Calculate the difference vector V between the second vector and the first vector d =V2-V1; Calculate the intermediate vector V c =V1+a c *V d , where 0-r c <1+r, c is the serial number of the intermediate vector, r is the expansion radius;​ Randomly adjust the first vector, the second vector, and all intermediate vectors: V o =V i +b i , b i is a random vector with a modulus less than r; V i The first vector, the second vector or the intermediate vector is input; V o is the randomly adjusted vector; All randomly adjusted vectors are taken as individual solutions, and the initial solution group is obtained by eliminating the individual solutions that do not meet the boundary conditions.

5. The method for treating karst groundwater pollution based on near-source interception and drainage and clean and polluted water diversion according to claim 1 is characterized by: The geological data include basic geological characteristics, geological structural characteristics, and karst development characteristics; The hydrological data include groundwater system characteristics, groundwater main runoff belt characteristics, groundwater dynamic condition characteristics and groundwater dynamic characteristics; The environmental pollution source data includes soil pollution status data, groundwater pollution status data, pollution channel data and pollutant migration pattern data.

6. The method for treating karst groundwater pollution based on near-source interception and drainage and clean and polluted water diversion according to claim 5 is characterized by: The geological and hydrological survey includes basic geological survey, which includes: The basic features of the address are obtained through map data, and the map data and image data collected by drones are input into the joint and fissure identification model to identify the joints and fissures and extract their trend features; The karst morphology and elevation distribution of the target treatment area are recorded and counted to obtain the karst morphological characteristics, karst plane characteristics and elevation distribution characteristics; through drilling detection, the karst development conditions in each hole depth range are obtained to obtain the karst vertical characteristics.

7. The method for treating karst groundwater pollution based on near-source interception and drainage and clean and polluted water diversion according to claim 1 is characterized by: The scheme type discrimination model adopts the support vector machine model, and uses groundwater flow direction, pollution source location, karst development characteristics, groundwater main runoff zone characteristics, groundwater dynamic conditions characteristics, groundwater dynamic characteristics, soil pollution status, groundwater pollution status, pollution channels and pollutant migration patterns as the input feature dimensions of the support vector machine.

8. The method for treating karst groundwater pollution based on near-source interception and drainage and clean and polluted water diversion according to claim 1 is characterized by: Also includes: Implement the pollution diversion and treatment plan or the final near-source interception and treatment plan, collect the operating data of each device and regularly update the environmental pollution source data; The water pollution control situation is judged based on the environmental pollution source data, and the distribution status of pollution levels in different areas is indicated by color.

9. The method for treating karst groundwater pollution based on near-source interception and drainage and clean and polluted water diversion according to claim 8 is characterized by: Also includes: The equipment operating parameters are automatically adjusted according to weather and rain forecast data as well as environmental pollution source data.

Images