Sea-entering river pollution source targeted blocking and ecological restoration integrated prevention and control system

Abstract

The invention relates to the technical field of ecological restoration prevention and control, in particular to a sea-entering river pollution source targeted blocking and ecological restoration integrated prevention and control system which comprises a pollution source recognition unit, a targeted blocking unit, an ecological restoration unit and a cooperative control platform. The recognition unit combines a multispectral unmanned aerial vehicle with a water quality sensor and positions a pollution source through artificial intelligence traceability; the targeted blocking unit implements physical-chemical combined blocking according to pollution types; the ecological restoration unit purifies water and reconstructs ecology through a modular constructed wetland and the like; the cooperative control platform optimizes each unit strategy based on hydrological data; each unit correlation formula accurately calculates parameters, such as a recognition algorithm, a gate dam opening control formula and the like. According to the invention, multi-technology fusion and system collaboration are realized, pollution events can be quickly responded, measures are accurately applied, the sea-entering river pollution prevention and control efficiency is effectively improved, the treatment cost is reduced, and drainage basin water environment ecological restoration and sustainable development are promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of ecological restoration and prevention and control, and specifically to an integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in rivers flowing into the sea. Background Art

[0002] Traditional means of river pollution prevention and control have many limitations. In terms of pollution source identification, it mostly relies on manual inspections and fixed-point monitoring, with low efficiency and narrow coverage, making it difficult to quickly locate sudden pollution sources and track the pollution diffusion path; in the targeted blocking link, fixed sluice gates and conventional adsorption materials are mostly used, lacking the precise regulation ability for different pollution types, resulting in serious resource waste and limited blocking effect; in terms of ecological restoration, technologies such as constructed wetlands are often not closely combined with pollution characteristics, with a long restoration period and unstable effects. In addition, each link of the existing prevention and control system is independent and lacks a coordination mechanism, and it is unable to dynamically optimize the prevention and control strategy according to the real-time pollution situation and hydrological conditions, making it difficult to cope with complex and changeable pollution scenarios. Therefore, an integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in rivers flowing into the sea is proposed for the above problems. Summary of the Invention

[0003] The purpose of the present invention is to provide an integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in rivers flowing into the sea to solve the problems raised in the above background art.

[0004] To achieve the above purpose, the present invention provides the following technical solutions:

[0005] An integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in rivers flowing into the sea, including a pollution source identification unit, a targeted blocking unit, an ecological restoration unit, and a cooperative control platform:

[0006] The pollution source identification unit includes a multi-spectral drone scanning module, a water quality sensor array, and an artificial intelligence tracing and analysis module, and is used to identify the type, location, and diffusion path of pollution sources in real time;

[0007] The targeted blocking unit includes an intelligent sluice gate system, a dynamic adsorption device, and a degradable barrier film, and implements physical-chemical combined blocking in the target river section according to the pollution source type;

[0008] The ecological restoration unit includes a modular constructed wetland, a composite microbial agent releaser, and an aquatic vegetation restoration substrate bed, which are arranged downstream of the blocking unit;

[0009] The cooperative control platform is connected to each unit and dynamically optimizes the blocking and restoration strategies based on river hydrological data.

[0010] As a preferred solution, the artificial intelligence tracing and analysis module performs the following operations:

[0011] Receive the thermal map of the river channel scanned by the multispectral drone and the real-time parameters of the water quality sensor array;

[0012] Identify the spatial distribution of pollution plumes through an improved spatio-temporal convolutional neural network, and its calculation process is as follows:

[0013] Input the multispectral data and the water quality parameter fusion matrix into the dynamic convolution kernel, inject a gradient correction term in the convolution operation to enhance the identification of small-scale pollution sources, and output pollution feature values through the adaptive LeakyReLU activation function;

[0014] Combine the flow velocity-discharge model to predict the diffusion path, and output a classification report of pollution source types and a confidence evaluation report for source tracing.

[0015] As an optimal solution, the process of updating the weights of the dynamic convolution kernel is as follows:

[0016] Based on the gradient of the loss function of historical training samples, combine the time decay mechanism to dynamically adjust the weights of the convolution kernel, so that the weights decay as the time interval since the pollution event increases.

[0017] As an optimal solution, the targeted blocking unit satisfies:

[0018] The intelligent sluice and dam system automatically opens and closes according to the location of the pollution source, and its gate opening control logic is as follows:

[0019] Taking the maximum allowable opening as the benchmark, respond to the distance between the pollution source and the gate through the S-shaped function, and quickly reduce the opening when the distance is less than the critical response distance;

[0020] The dynamic adsorption device includes magnetic nano-adsorbents and ion exchange resins;

[0021] The degradable barrier film is made of poly(lactic-co-glycolic acid) and degrades within 48 hours after interception.

[0022] As an optimal solution, the calculation logic for the dosage of the dynamic adsorption device is as follows:

[0023] According to the cross-sectional flow rate of the river channel, the time-varying curve of pollutant concentration, and the adsorption reaction time constant, determine the dosage of the adsorbent through integral operation, and introduce a pollutant density correction factor for calibration.

[0024] As an optimal solution, in the ecological restoration unit:

[0025] The release rate regulation logic of the composite microbial agent releaser is as follows:

[0026] Taking the basic release rate as the benchmark, respond to the deviation between the real-time ammonia nitrogen concentration and the threshold through the hyperbolic tangent function, and the release rate increases non-linearly when the concentration exceeds the threshold;

[0027] The modular constructed wetland consists of an anaerobic sedimentation tank, an aerobic biological filter and an ecological gravel bed connected in series;

[0028] The aquatic vegetation restoration bed adopts a three-dimensional woven structure.

[0029] As a preferred solution, the dynamic decision-making engine of the collaborative control platform adopts a hydrological adaptability reinforcement learning model, and its decision-making mechanism includes:

[0030] Update the action value function through the experience replay mechanism;

[0031] The real-time reward function is calculated by weighting the pollution load reduction rate and the system energy consumption;

[0032] Introduce a policy gradient correction term to improve the adaptability to hydrological mutations.

[0033] As a preferred solution, the calculation logic of the policy gradient correction term is:

[0034] Based on the state-action trajectory data of the parallel environment, optimize the decision-making policy parameters by taking the derivative of the product of the policy network probability and the cumulative reward.

[0035] As a preferred solution, the prediction logic of the plant survival rate of the aquatic vegetation restoration bed is:

[0036] According to the weighted linear combination of dissolved oxygen concentration, water body pH value and turbidity, map the survival probability through the S-shaped function, where the weight of dissolved oxygen is the highest.

[0037] It can be seen from the technical solution provided by the present invention above that an integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in estuarine rivers provided by the present invention has the following beneficial effects:

[0038] Efficient pollution prevention and control throughout the process: The pollution source identification unit uses multi-spectral drones and water quality sensor arrays, combined with intelligent algorithms, to quickly locate pollution sources and accurately predict the diffusion trend, gaining the initiative for prevention and control work; the targeted blocking unit, based on the identification results, quickly takes physical-chemical combined blocking measures through devices such as intelligent sluice gates and dynamic adsorption devices, effectively curbing the spread of pollution and greatly reducing the risk of pollutants spreading downstream, creating favorable conditions for subsequent ecological restoration;

[0039] Deep ecological restoration and system reconstruction: The ecological restoration unit constructs a multi-level ecological purification system including a modular constructed wetland, a composite microbial agent releaser and an aquatic vegetation restoration bed; through the synergistic effect of physics, chemistry and biology, deeply purify the residual pollutants and significantly improve water quality; at the same time, the reconstruction of aquatic vegetation and microbial communities promotes the balance of the water ecosystem, enhances the self-purification ability of the river, and realizes the transformation from pollution treatment to sustainable development of the ecosystem;

[0040] Intelligent Collaboration and Resource Optimization: As the core of the system, the collaborative control platform utilizes a hydrological adaptability reinforcement learning model to comprehensively consider factors such as pollution reduction and energy consumption, dynamically optimize the operation strategies of each unit, and achieve intelligent collaborative operation of the entire system. This intelligent decision-making method not only improves the accuracy and effectiveness of prevention and control measures but also reasonably allocates resources, reduces unnecessary inputs, and while ensuring the treatment effect, reduces the system operation cost and resource consumption.

[0041] Improvement of Comprehensive Environmental and Economic Benefits: It effectively reduces the total amount of pollutants in the rivers flowing into the sea, protects the marine ecological environment, maintains biodiversity, and plays a positive role in regional ecological balance. From an economic perspective, accurate pollution prevention and control and resource optimization strategies reduce the consumption of human, material, and financial resources in the treatment process. At the same time, a good ecological environment can drive the development of surrounding ecological tourism and other industries, creating more economic value, and realizing the organic unity of environmental benefits and economic benefits.

[0042] Technological Innovation and Wide Applicability: Integrating advanced technologies such as the Internet of Things, big data, and artificial intelligence, it realizes data sharing and function collaboration among multiple units, forming an innovative and technologically advanced integrated solution for pollution prevention and control and ecological restoration. The system adopts a modular design and can flexibly adjust the configuration of each unit according to the pollution characteristics, hydrological conditions, and treatment requirements of different rivers, with a wide range of application scenarios and strong expansion capabilities, providing reliable technical support for the treatment of river basin water environment. Description of the Drawings

[0043] Figure 1 It is a schematic diagram of the overall structure of an integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in rivers flowing into the sea according to the present invention. Detailed Embodiments

[0044] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0045] In order to better understand the above technical solutions, the above technical solutions will be described in detail below in conjunction with the drawings of the specification and specific embodiments.

[0046] As Figure 1 shown, an embodiment of the present invention provides an integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in rivers flowing into the sea, including a pollution source identification unit, a targeted blocking unit, an ecological restoration unit, and a collaborative control platform:

[0047] The pollution source identification unit includes a multispectral UAV scanning module, a water quality sensor array, and an artificial intelligence traceability analysis module, which are used to identify the type, location, and diffusion path of pollution sources in real time;

[0048] The targeted blocking unit includes an intelligent sluice and dam system, a dynamic adsorption device, and a degradable barrier film, which implement physical-chemical combined blocking in the target river section according to the type of pollution source;

[0049] The ecological restoration unit includes a modular artificial wetland, a composite microbial agent releaser, and an aquatic vegetation restoration substrate, which are arranged downstream of the blocking unit;

[0050] The collaborative control platform is connected to each unit and dynamically optimizes the blocking and restoration strategies based on river hydrological data.

[0051] In this embodiment, the artificial intelligence traceability analysis module performs the following operations:

[0052] Receive the river channel thermal map scanned by the multispectral UAV and the real-time parameters of the water quality sensor array;

[0053] Identify the spatial distribution of the pollution plume through an improved spatio-temporal convolutional neural network. The calculation process is as follows:

[0054] Input the multispectral data and water quality parameter fusion matrix into the dynamic convolutional kernel, inject a gradient correction term in the convolutional operation to enhance the identification of small-scale pollution sources, and output the pollution feature values through the adaptive LeakyReLU activation function;

[0055] Combine the flow velocity-discharge model to predict the diffusion path, and output a classification report of the pollution source type and a traceability confidence evaluation report;

[0056] The process of updating the weights of the dynamic convolutional kernel is as follows:

[0057] Based on the gradient of the loss function of historical training samples, dynamically adjust the weights of the convolutional kernel in combination with the time decay mechanism, so that the weights decay as the time interval of pollution events increases;

[0058] Furthermore, the pollution source identification unit serves as the "scout" of the integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in the river flowing into the sea. Through multi-source data collection and intelligent analysis, it accurately locates the pollution sources and provides key basis for subsequent blocking and restoration. The following elaborates in detail from aspects such as functions, technical principles, and work processes:

[0059] I. Overview of overall functions:

[0060] The pollution source identification unit integrates multispectral drone scanning, water quality sensor arrays, and artificial intelligence source tracing and analysis technology to achieve real-time monitoring, spatial positioning, and dynamic tracking of river pollution sources. Its core functions include: real-time collection of river spectral information and water quality parameters, rapid identification of pollution source types (such as industrial wastewater, domestic sewage, agricultural non-point source pollution, etc.) through algorithms, localization of pollution occurrence, and prediction of pollution diffusion paths in combination with hydrological data, providing data support for targeted blocking and ecological restoration.

[0061] 2. Submodule composition and functions:

[0062] (1) Multispectral data acquisition module:

[0063] Unmanned aerial vehicle system: Using drones equipped with high-resolution multispectral imagers covering visible light (400-700nm), near-infrared (700-1100nm), and short-wave infrared (1100-2500nm) bands, the system periodically scans the river at a cruising speed of 5-10 meters per second, generating river channel thermal maps with a spatial resolution of 0.5-1 meter, capturing areas of abnormal water surface color, temperature, and reflectivity.

[0064] Data preprocessing: Geometric correction, radiometric calibration, and atmospheric correction are performed on the collected multispectral images to remove interference factors such as clouds and light. The spectral characteristic curve of the water body is extracted and compared with the standard pollution spectrum library to preliminarily screen suspected polluted areas.

[0065] (2) Water quality sensor array module:

[0066] Sensor deployment: water quality sensor arrays are deployed along river sections and key nodes (such as sewage outlets and tributary junctions), including pH sensors, dissolved oxygen (DO) sensors, chemical oxygen demand (COD) sensors, ammonia nitrogen (NH4 + ) sensors and turbidity sensors to collect water quality parameters in real time at intervals of 1-5 minutes;

[0067] Data transmission and verification: Data is uploaded to the edge computing node via NB-IoT or LoRa wireless communication technology. The CRC check algorithm is used to verify data integrity, remove outliers, and smooth the data using the Kalman filter algorithm to improve data accuracy.

[0068] (3) Artificial Intelligence Source Tracing Analysis Module:

[0069] Data fusion: align multispectral image feature data with water quality sensor parameters in time and space to construct a multidimensional data matrix containing spatial coordinates, time, spectral reflectance and water quality indicators (in, for Time coordinates Multidimensional data matrix at is time coordinate Spectral reflectance data at is time coordinate Water quality index data at

[0070] Pollution identification algorithm: An improved spatio-temporal convolutional neural network (STCNN) is adopted, and its forward propagation formula is: (where is time coordinate Pollution feature output value at is the adaptive LeakyReLU activation function, which is used to enhance the nonlinear expression ability of the model; is the weight of the dynamic convolution kernel at position; is the value of the fusion matrix of the input multispectral data and water quality parameters at the coordinate position; is the gradient correction term, , is the loss function, which is used to measure the difference between the model prediction value and the true value, is the hidden layer parameter, and the model convergence is accelerated by introducing gradient correction);

[0071] Diffusion prediction model: Combining the river velocity-discharge model (based on the Manning formula) with the pollutant migration and diffusion equation, a three-dimensional dynamic prediction model is established to output the diffusion range and concentration change curve of the pollution plume in the next 6-24 hours, and generate a diagnostic report including the pollution source type, location coordinates and traceability confidence level (0-100%);

[0072] (IV) Model optimization and update module:

[0073] Dynamic weight adjustment: The formula is adopted (where is the weight of the dynamic convolution kernel at position at time; is the weight of the dynamic convolution kernel at position at time; is the learning rate of the convolution kernel, [[ID=S73]] is the loss function of the is the time decay coefficient, ; is the time interval between the current time and the time when the pollution event occurred, in hours; is the average value of the partial derivatives of the loss function of all training samples with respect to the weight to measure the impact of the weight on the diagnostic result; is the time decay factor, which weakens the influence of the training gradient of old pollution events on the current weight over time) Iteratively optimize the model parameters according to the data of newly occurred pollution events to improve the model adaptability;

[0074] Pollution spectral library update: Regularly collect the spectral data of actual pollution samples, classify and optimize the spectral library through clustering algorithms, add the spectral features of rare pollution sources, and ensure that the recognition algorithm keeps up with the times;

[0075] III. Key technical principles:

[0076] (I) Multi-source data fusion principle:

[0077] Based on the feature-level fusion technology, complementarily integrate the spatial information of multi-spectral images and the physical and chemical information of water quality parameters; Use principal component analysis (PCA) to reduce the dimension of high-dimensional data, retain key features, and map data with different sampling frequencies and different spatial resolutions to a unified coordinate system through a spatio-temporal alignment algorithm to construct a multi-dimensional feature space;

[0078] (II) Spatio-temporal convolutional neural network principle:

[0079] Extract the spatial features of the pollution area through a two-dimensional convolutional layer, and combine with a temporal convolutional layer to capture the evolution law of pollution over time; The dynamic convolutional kernel adaptively adjusts the weight according to the pollution type and the river channel environment, and the gradient correction term optimizes the model parameters through backpropagation to achieve accurate recognition of complex pollution scenarios;

[0080] (III) Pollution diffusion prediction principle:

[0081] Based on the law of conservation of mass and the theory of fluid mechanics, divide the river channel into multiple calculation units, combine the measured flow velocity, flow rate and pollutant degradation coefficient, and solve the convection-diffusion equation through the finite difference method to simulate the migration and diffusion process of pollutants under the action of water flow, providing a time window for formulating blocking plans;

[0082] IV. Working process of the module:

[0083] (I) Initialization stage:

[0084] The UAV system completes route planning, sensor calibration and communication link testing, and the water quality sensor array starts a self-check program to verify that the data acquisition module and the transmission module are functioning properly;

[0085] Load the initial parameters of the artificial intelligence traceability analysis model, the pollution spectral library, and the hydrological database, and establish a communication connection with the collaborative control platform;

[0086] (2) Data collection stage:

[0087] The drone executes the scanning task according to the preset route and transmits the multi-spectral image data after completing each flight segment; the water quality sensor array continuously collects real-time water quality parameters;

[0088] After receiving the data, the edge computing node preprocesses the image data and water quality data respectively to generate a standardized data set;

[0089] (3) Pollutant source identification stage:

[0090] Input the preprocessed multi-source data into the artificial intelligence traceability analysis module, calculate the pollution feature output value through the spatio-temporal convolutional neural network, and identify the pollution area;

[0091] Combine the diffusion prediction model, analyze the pollution diffusion trend, generate a diagnostic report including the pollutant source location, type, and diffusion path, and upload it to the collaborative control platform;

[0092] (4) Model optimization stage:

[0093] The collaborative control platform feeds back the actual pollution disposal results, compares and analyzes them with the identification report, and calculates the model error;

[0094] Trigger the model optimization program according to the error value, adjust the neural network parameters through the weight update formula, update the pollution spectral library, and complete the model iteration;

[0095] V. Application value of the module:

[0096] (1) Improve the efficiency of pollution prevention and control:

[0097] Achieve rapid discovery and accurate positioning of pollutant sources through a minute-level response speed and meter-level positioning accuracy, with an efficiency improvement of over 80% compared to traditional manual inspections, winning precious time for targeted blocking;

[0098] (2) Reduce treatment costs:

[0099] Accurately allocate blocking resources according to the pollution type and diffusion range, reduce waste of materials such as adsorbents and barrier membranes, and is expected to reduce treatment costs by 30%-50%;

[0100] (3) Support scientific decision-making:

[0101] Provide a quantitative analysis report including the confidence level of pollution traceability, provide data support for the environmental protection department to formulate river basin pollution treatment plans and evaluate treatment effects, and help build an intelligent water environment management system;

[0102] (IV) Enhance system synergy:

[0103] By real-time linkage with the targeted blocking unit and the ecological restoration unit, the pollution source identification results are directly converted into control instructions such as the opening and closing of dams and the deployment of adsorption devices, realizing the automation of the entire process of "monitoring-decision-making-execution" and improving the overall prevention and control efficiency of the system.

[0104] In this embodiment, the targeted blocking unit satisfies:

[0105] The intelligent dam system automatically opens and closes according to the location of the pollution source. The gate opening control logic is as follows:

[0106] Based on the maximum allowable opening, the distance between the pollution source and the gate is responded to through an S-shaped function. When the distance is less than the critical response distance, the opening is quickly reduced.

[0107] The dynamic adsorption device includes a magnetic nano-adsorbent and an ion exchange resin;

[0108] The degradable barrier film is made of polylactic acid-glycolic acid copolymer and degrades within 48 hours after interception;

[0109] Furthermore, the targeted blocking unit serves as the "interception guard" of the integrated prevention and control system for targeted blocking of pollution sources and ecological restoration in rivers entering the sea. It accurately intercepts pollution sources through a combination of physical and chemical means, creating conditions for downstream ecological restoration. The following is a detailed explanation of its functional positioning, core components, working principles, and application value:

[0110] 1. Overview of overall functions:

[0111] Based on the diagnostic reports generated by the pollution source identification unit, the targeted blocking unit rapidly activates the intelligent sluice gate system, dynamic adsorption device, and biodegradable barrier membrane to implement a graded and classified blocking strategy for different types and scales of pollution. Its core functions include: automatically adjusting the sluice gate opening according to the location and spread rate of pollution to intercept the spread of pollution; precisely deploying adsorbents based on the concentration and characteristics of pollutants to reduce pollutant loads; and using biodegradable barrier membranes to create a physical barrier to prevent pollution from spreading downstream, buying time for the ecological restoration unit to process the pollution.

[0112] 2. Submodule composition and functions:

[0113] (1) Intelligent sluice and dam system:

[0114] Hardware composition: It consists of a hydraulically driven dam, a water level and flow monitor, a high-precision displacement sensor, and a remote control terminal. The dam is made of high-strength, corrosion-resistant materials, has a maximum load-bearing capacity of 500 tons, and is capable of rapid opening and closing (full stroke time ≤ 2 minutes).

[0115] Control Logic: According to the location information provided by the pollution source identification unit, the opening degree of the sluice dam is automatically calculated through the formula ( is the actual opening degree of the gate, in degrees; is the maximum allowable opening degree, in degrees; is the response sensitivity coefficient, ; is the distance between the pollution source and the gate, in meters; is the critical response distance, m). When the pollution source approaches the critical distance, the sluice dam quickly contracts to slow down the water flow speed and reduce the pollution diffusion range;

[0116] Emergency Mode: When a sudden high-concentration pollution event is detected, the sluice dam can be switched to the fully enclosed mode to form a double barrier of "sluice dam interception - adsorption and purification" in cooperation with the dynamic adsorption device;

[0117] (2) Dynamic Adsorption Device:

[0118] Device Design: Adopts a modular adsorption bin structure, including an activated carbon adsorption module, a nano-scale ion exchange resin module, and a biological ceramsite adsorption module, which can be assembled according to the types of pollutants (organic matter, heavy metals, ammonia nitrogen, etc.); the effective adsorption area of a single adsorption bin reaches 20 ㎡, and the maximum load capacity is 500 kg of adsorbent;

[0119] Feeding Control: According to the formula (where, is the feeding amount of the adsorbent, in kilograms; is the pollutant density correction factor, ; is the cross-sectional flow rate of the river channel, in cubic meters per second; is the pollutant concentration at time is the adsorption reaction time constant, s; is the start time of pollution, [[ID=?]] is the calculation end time) to calculate the feeding amount of the adsorbent; the adsorption bin is fixed-point deployed in the polluted area through an automatic feeding system to achieve rapid adsorption and degradation of pollutants;

[0120] Regeneration and Replacement: The adsorption bin is equipped with a saturation monitoring sensor. When the adsorption efficiency drops to 30% of the initial value, a recovery instruction is triggered, and replacement or regeneration treatment is carried out through an offshore operation platform;

[0121] (3) Degradable Barrier Film:

[0122] It should be noted that there seems to be an incomplete or incorrect tag reference in the original text at line where it says " [[ID=?]] is the calculation end time", which might be a formatting or content error. This has been translated as best as possible while maintaining the integrity of the provided text.Material characteristics: It is made by compounding poly (lactic-co-glycolic acid) (PLGA) with natural fibers. The film thickness is 0.5 - 1 mm, the tensile strength is ≥15 MPa, and it can be completely degraded in 3 - 6 months in the water environment. The film surface has superhydrophobic characteristics, which can effectively block pollutants such as oil stains and suspended solids.

[0123] Deployment method: A roll-type automatic laying device is adopted, which can complete the laying of a 500-meter river blocking film within 20 minutes. Both ends of the film body are fixed to the bank anchor piles, and the middle part is kept in a tensioned state through a floating ball array to form a flexible physical barrier.

[0124] Adaptive design: The height of the blocking film can be automatically adjusted according to the water level change. When the water level fluctuates by more than 1 meter, the hydraulic lifting mechanism is triggered to adjust the position of the film body to ensure effective interception.

[0125] III. Key technical principles:

[0126] (I) Intelligent sluice-dam linkage control principle:

[0127] Based on Bernoulli's equation and the sluice-dam discharge formula in fluid mechanics, a dynamic model of sluice-dam opening - flow rate - pollution diffusion speed is established. By real-time monitoring of the river water level, flow velocity, and pollution source location, a nonlinear response mechanism is constructed using the Sigmoid function to achieve the adaptive adjustment of the sluice-dam opening, while intercepting pollution and avoiding the upstream water level exceeding the limit.

[0128] (II) Dynamic adsorption optimization principle:

[0129] Combined with adsorption kinetics and the law of conservation of mass, an integral algorithm is used to calculate the pollutant accumulation. The exponential decay term simulates the saturation process of the adsorbent. By dynamically adjusting the dosage, the adsorption efficiency is always maintained above 70%. At the same time, the adsorption isotherm model (such as Langmuir, Freundlich model) is used to optimize the selection and ratio of the adsorbent.

[0130] (III) Degradable barrier film action mechanism:

[0131] Using the principles of surface tension and pore size screening, the barrier film physically intercepts pollutants with a particle size > 10 μm. The PLGA material is gradually decomposed into carbon dioxide and water under the action of microorganisms and hydrolysis, and no secondary pollution is produced during the degradation process, meeting the requirements of ecological restoration.

[0132] IV. Working process of the module:

[0133] (I) Early warning response stage:

[0134] Receive the diagnostic report from the pollution source identification unit, and analyze key information such as pollution type, location, and diffusion speed.

[0135] The intelligent sluice dam system starts the pre-adjustment program, gradually contracts the sluice dam opening according to the distance of the pollution source, and at the same time, the dynamic adsorption device completes the module assembly and deployment preparation;

[0136] (II) Blocking implementation stage:

[0137] Hierarchical blocking:

[0138] Mild pollution: Only the degradable barrier film is enabled to form a primary interception;

[0139] Moderate pollution: Start the intelligent sluice dam and the dynamic adsorption device to reduce the pollution concentration;

[0140] Severe pollution: The three work together to implement the full-process blocking;

[0141] The dynamic adsorption device puts the adsorbent into the adsorption bin according to the calculated amount, and monitors the adsorption saturation in real time, and replenishes or replaces it in time;

[0142] (III) Effect evaluation and adjustment stage:

[0143] Through the water quality monitoring points deployed downstream of the blocking area, the change of pollutant concentration is detected in real time to evaluate the blocking effect;

[0144] According to the detection data, the collaborative control platform adjusts parameters such as the sluice dam opening and the adsorbent dosage to optimize the blocking strategy;

[0145] (IV) Post-treatment stage:

[0146] After the pollution incident ends, the residual fragments of the degradable barrier film and the saturated adsorption bin are recovered for harmless treatment;

[0147] Carry out maintenance on the intelligent sluice dam system, update the equipment operation parameters, and prepare for the next blocking task;

[0148] V. Application value of the module:

[0149] (I) High efficiency of emergency response:

[0150] For sudden pollution incidents, the blocking deployment can be completed within 30 minutes, reducing the pollution diffusion range by more than 60%, and effectively reducing the impact of pollution on the downstream ecological environment;

[0151] (II) Resource conservation and environmental protection:

[0152] The precise placement of the dynamic adsorption device and the application of the degradable barrier film reduce material waste and the risk of secondary pollution. Compared with the traditional interception method, the adsorbent dosage is reduced by 40% and the disposal cost is reduced by 35%;

[0153] (III) Ecologically friendly design:

[0154] The complete degradation characteristics of the degradable barrier film avoid the long-term interference of traditional rigid interception facilities on the river ecosystem, ensure the smooth passage of aquatic organisms, and maintain the integrity of the river ecosystem;

[0155] (4) System synergy enhancement:

[0156] It is seamlessly connected with the pollution source identification unit and the ecological restoration unit to form a "monitoring-blocking-restoration" closed loop, improve the overall efficiency of the integrated prevention and control system, and provide reliable technical support for the water environment governance of the basin.

[0157] The calculation logic for the dosage of the dynamic adsorption device is as follows:

[0158] According to the cross-sectional flow of the river, the time-varying curve of pollutant concentration, and the adsorption reaction time constant, the dosage of the adsorbent is determined through integral operation, and a pollutant density correction factor is introduced for calibration.

[0159] In this embodiment, in the ecological restoration unit:

[0160] The release rate regulation logic of the composite microbial agent releaser is as follows:

[0161] Based on the basic release rate, the deviation between the real-time ammonia nitrogen concentration and the threshold is responded through the hyperbolic tangent function, and the release rate increases non-linearly when the concentration exceeds the threshold;

[0162] The modular artificial wetland is composed of a series-connected anaerobic sedimentation tank, an aerobic biological filter, and an ecological gravel bed;

[0163] The aquatic vegetation restoration bed adopts a three-dimensional weaving structure;

[0164] Further, as the "green guard" of the integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in the estuary river, the ecological restoration unit undertakes the key tasks of purifying water quality and reconstructing the ecosystem; through the construction of a multi-level ecological purification system, after the pollution source is blocked, pollutants are further reduced and the aquatic ecological balance is restored; the following elaborates in detail from aspects such as functional positioning, core components, technical principles, and application value:

[0165] I. Overall function overview:

[0166] The ecological restoration unit is located downstream of the targeted blocking unit. Through the synergistic effect of the modular artificial wetland, the composite microbial agent releaser, and the aquatic vegetation restoration bed, it realizes the deep purification of residual pollutants and the reconstruction of ecological functions; its core functions include: using the combined physical-chemical-biological effects of the artificial wetland to remove pollutants such as nitrogen, phosphorus, and organic matter; accelerating the degradation of pollutants through the composite microbial agent; relying on the aquatic vegetation restoration bed to construct a stable water ecosystem, improving the self-purification ability and biodiversity of the river, and ultimately achieving the dual goals of water quality compliance and ecological restoration;

[0167] II. Composition and Functions of Sub-modules:

[0168] (I) Modular Constructed Wetland:

[0169] Structural Design: Adopt a four-stage series structure of "pretreatment area - vertical flow wetland - horizontal subsurface flow wetland - surface flow wetland"; set a grille and a grit chamber in the pretreatment area to remove large particulate suspended matters; fill the vertical flow wetland with filter materials such as volcanic rock and ceramsite to strengthen the nitrification of ammonia nitrogen; lay activated carbon and zeolite in the horizontal subsurface flow wetland to adsorb heavy metals and organic matters; plant emergent and floating plants in the surface flow wetland to form a landscape ecological belt; the area of a single module is 200 - 500 square meters, and it can be flexibly combined according to the width of the river channel;

[0170] Water Flow Regulation: Adjust the water level difference of each unit through an intelligent water level control system to achieve uniform water flow distribution; adopt an intermittent water inlet mode, with 1 hour of water inlet every 6 hours, and the residence time is controlled within 24 - 48 hours to ensure the full degradation of pollutants;

[0171] Plant Configuration: Select plant species according to the characteristics of water quality pollution. For example, plant reeds and calamus for ammonia nitrogen pollution; plant water hyacinths and Pteris vittata for heavy metal pollution; the plant density is 16 - 25 plants per square meter, and the above-ground part is regularly harvested to prevent the secondary release of pollutants;

[0172] (II) Composite Microbial Inoculant Releaser:

[0173] Inoculant Formula: A composite microbial community composed of nitrifying bacteria, denitrifying bacteria, photosynthetic bacteria and polyphosphate-accumulating bacteria is used to effectively degrade ammonia nitrogen, total phosphorus and COD; the inoculant adopts an embedding and immobilization technology to wrap the microorganisms in sodium alginate - polyvinyl alcohol gel particles with a particle size of 2 - 3 mm to improve the stability of the inoculant under water flow impact;

[0174] Release Control: According to the formula ( is the actual release rate of the microbial inoculant, with the unit of g / h; is the basic release rate, with the unit of g / h; is the ammonia nitrogen concentration monitored in real time, with the unit of mg / L; is the concentration threshold, mg / L; is the concentration adjustment amplitude, mg / L), dynamically regulate the release rate; when the ammonia nitrogen concentration is higher than the threshold, the release amount of the inoculant increases; after the concentration decreases, the release rate automatically adjusts back;

[0175] Intelligent Supply: The releaser is equipped with a sensor for the stock of the inoculant. When the remaining amount of the inoculant is lower than 30%, an automatic supply instruction is triggered, and the inoculant is supplemented through the pipeline transportation system;

[0176] (3) Aquatic vegetation restoration bed:

[0177] Bed structure: Adopt a three-layer structure of "ecological concrete framework + porous bio-ceramsite + slow-release fertilizer layer"; the porosity of the ecological concrete framework reaches 30%, providing habitat space for aquatic organisms; the particle size of the porous bio-ceramsite is 5 - 10 mm, adsorbing suspended solids in the water body; the slow-release fertilizer layer contains nitrogen, phosphorus, potassium and trace elements, continuously supplying fertilizers for plant growth; the bed size is 2m × 2m × 0.5m, and is spliced into a continuous belt through connectors;

[0178] Prediction of plant survival rate: Use the formula (where is the vegetation survival probability, unit: %; DO is the dissolved oxygen concentration, unit: mg / L; pH is the water body acidity and alkalinity; Turb is the turbidity, unit: NTU; , , are weight coefficients, , , ) to evaluate the plant growth conditions in real time; according to the prediction results, automatically adjust parameters such as the bed water level and light intensity;

[0179] Ecological function: After vegetation restoration, a three-dimensional community of "submerged plants (Vallisneria natans, Hydrilla verticillata) - floating-leaved plants (Nymphaea tetragona, Nymphoides peltata) - emergent plants (Typha orientalis, Scirpus validus)" is formed, providing habitats for fish and benthic organisms, promoting the reconstruction of the food chain, and enhancing the ecological stability of the water body;

[0180] III. Key technical principles:

[0181] (1) Principle of artificial wetland purification:

[0182] Based on the synergistic effect of physical filtration, chemical adsorption and biological degradation; the filter media intercepts suspended solids, activated carbon adsorbs heavy metals, microorganisms remove nitrogen and phosphorus through nitrification-denitrification and phosphorus accumulation, and plant roots absorb nutrients and release oxygen to promote microbial metabolism, forming a complete ecological purification chain;

[0183] (2) Mechanism of action of composite microbial inoculant:

[0184] Utilize the high activity and stress resistance of immobilized microorganisms. Nitrifying bacteria convert ammonia nitrogen into nitrate, denitrifying bacteria reduce nitrate to nitrogen, and phosphorus-accumulating bacteria excessively absorb phosphorus and discharge it with the excess sludge, achieving efficient degradation of pollutants; dynamic release control ensures that the concentration of the inoculant matches the pollution load;

[0185] (3) Theory of aquatic vegetation restoration:

[0186] Based on the niche principle and the law of community succession, a stable water ecosystem is constructed by optimizing the plant configuration and the substrate environment; the plant survival rate prediction model comprehensively considers key indicators such as dissolved oxygen, pH, and turbidity, quantifies the impact of ecological factors on plant growth, and guides precise ecological restoration;

[0187] IV. Workflow of the module:

[0188] (I) Initialization stage:

[0189] Complete the assembly of the modular constructed wetland and the commissioning of the water flow system, fill the filter media and plant pioneer plants;

[0190] Install the composite microbial agent releaser and inject the initial agent; lay the substrate for aquatic vegetation restoration and sow the seeds of submerged plants;

[0191] Start the water quality monitoring system and establish the initial operating parameters of each unit;

[0192] (II) Operation stage:

[0193] Water quality purification: The river water treated by the targeted blocking unit enters the constructed wetland and passes through the four-stage treatment unit in sequence to gradually reduce pollutants;

[0194] Microbial regulation: Monitor the ammonia nitrogen concentration in real time and dynamically adjust the release rate of the composite microbial agent according to the formula to accelerate pollutant degradation;

[0195] Vegetation management: Evaluate the growth status of aquatic plants through the survival rate prediction model, regularly prune and replant to maintain the stability of the community;

[0196] (III) Optimization stage:

[0197] Collect the water quality data of the inlet and outlet of the constructed wetland, microbial activity indicators and vegetation growth parameters, and analyze the restoration effect;

[0198] According to the data analysis results, adjust the water flow residence time, agent release strategy and substrate environment parameters of the constructed wetland to optimize the ecological restoration efficiency;

[0199] (IV) Maintenance stage:

[0200] Regularly clean the sediment in the sedimentation tank and on the surface of the filter media of the constructed wetland, and replace the saturated adsorption materials;

[0201] Supplement the composite microbial agent, repair the releaser equipment; repair or reconstruct the damaged substrate for aquatic vegetation restoration;

[0202] V. Application value of the module:

[0203] (I) Deep purification of water quality:

[0204] For the river water after the blocking treatment, ammonia nitrogen (80%-90%), total phosphorus (70%-85%) and COD (60%-75%) can be further removed to ensure stable water quality compliance and reduce the pollution pressure on the marine ecosystem;

[0205] (II) Ecosystem reconstruction:

[0206] By restoring aquatic vegetation and constructing microbial communities, the biodiversity of the river can be significantly enhanced, the habitats of fish and birds can be restored, a healthy water ecosystem can be rebuilt, and the self-purification ability of the river can be strengthened;

[0207] (III) Environmental landscape improvement:

[0208] The modular artificial wetland and the aquatic vegetation belt form an ecological landscape, improve the environment around the river channel, provide a leisure space for residents, and achieve the unity of ecological and social benefits;

[0209] (IV) System synergy and efficiency improvement:

[0210] Closely cooperate with the pollution source identification unit and the targeted blocking unit to form a complete closed-loop of pollution prevention and control - ecological restoration, improve the comprehensive efficiency of the integrated prevention and control system, and provide a long-term solution for the water environment governance of the basin.

[0211] In this embodiment, the dynamic decision-making engine of the collaborative control platform adopts a hydrological adaptive reinforcement learning model, and its decision-making mechanism includes:

[0212] Update the action value function through the experience replay mechanism;

[0213] The real-time reward function is calculated by weighting the pollution load reduction rate and the system energy consumption;

[0214] Introduce a policy gradient correction term to improve the adaptability to hydrological mutations;

[0215] The calculation logic of the policy gradient correction term is:

[0216] Based on the state-action trajectory data of the parallel environment, optimize the decision-making strategy parameters by taking the derivative of the product of the policy network probability and the cumulative reward;

[0217] Further, as the "intelligent brain" of the integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in the river flowing into the sea, the collaborative control platform undertakes the core tasks of integrating system resources, optimizing decision-making strategies, and realizing the collaborative operation of each unit; the following will elaborate in detail from aspects such as platform architecture, core functions, key technologies, work processes, and application values:

[0218] I. Overall function overview:

[0219] The collaborative control platform, based on Internet of Things, big data, and artificial intelligence technologies, collects the operation data of the pollution source identification unit, the targeted blocking unit, and the ecological restoration unit in real time, and dynamically optimizes the parameters and collaborative strategies of each unit through a hydrological adaptability reinforcement learning model. Its core functions include: data integration and analysis, intelligent decision-making generation, device remote control, system status monitoring and early warning, ensuring the efficient and stable operation of the entire prevention and control system in complex hydrological environments, and realizing precise and intelligent management of pollution prevention and control and ecological restoration;

[0220] II. Platform Architecture and Sub-module Functions:

[0221] (I) Data Integration and Processing Center:

[0222] Multi-source data collection: Through communication technologies such as 5G and LoRa, the pollution monitoring data (multi-spectral images, water quality parameters) of the pollution source identification unit, the device operation status (gate opening, adsorbent dosage) of the targeted blocking unit, and the environmental indicators (dissolved oxygen, microbial activity) of the ecological restoration unit are obtained in real time, and the collection frequency reaches the second level;

[0223] Data cleaning and storage: The original data is preprocessed using median filtering and outlier detection algorithms to remove noise and invalid data; A spatio-temporal database is constructed using distributed storage technology (Hadoop HDFS), and the storage period is not less than 5 years, supporting historical data backtracking and trend analysis;

[0224] Data visualization: Through 3D GIS maps and dynamic charts, the pollution distribution, device locations, and system operation status are intuitively displayed, providing a visual decision-making interface for management personnel;

[0225] (II) Intelligent Decision-making Engine:

[0226] Hydrological adaptability reinforcement learning model: Using the formula (where is the value evaluation of performing action in state ; is the experience replay weight factor, , used to balance the influence of new and old experiences on decision-making; is the real-time reward function, The system energy consumption comprehensively considers the treatment effect and operation cost; is the future revenue discount rate, , reflecting the emphasis on long-term revenue; is the optimal value of the next state output by the target network; is the policy gradient correction term, , is the policy function, used to optimize the decision-making strategy);

[0227] Policy gradient correction: According to the formula (where represents the gradient operation on the policy network parameters , is the probability that the policy network selects the action in the state ; is at time the cumulative reward of the -th trajectory; is the number of environments for parallel training;

[0228] Decision generation: According to the output of the learning model, generate control instructions for each unit, such as adjusting the opening of the intelligent sluice gate, optimizing the adsorbent dosing plan, adjusting the release rate of microbial agents, etc.;

[0229] (3) Equipment collaborative control module:

[0230] Remote control: Through industrial Ethernet or wireless network, remotely start and stop, and adjust parameters of equipment such as intelligent sluice gates, dynamic adsorption devices, and microbial agent release devices, with a control delay not exceeding 1 second;

[0231] Linkage mechanism: Establish a three-level linkage rule of "pollution source identification - targeted blocking - ecological restoration". For example, when high-concentration pollution is detected, trigger collaborative actions such as automatically closing the sluice gate, running the adsorption device at full load, and doubling the release of microbial agents;

[0232] Fault emergency response: Real-time monitor the operation status of equipment. When a fault is detected, automatically switch to standby equipment or start an emergency plan, and at the same time generate a fault work order and push it to maintenance personnel;

[0233] (4) System status monitoring and early warning module:

[0234] Health assessment: Build a system health assessment index system, and comprehensively quantify and score the overall operation status of the system based on parameters such as equipment failure rate, data transmission stability, and treatment effect compliance rate;

[0235] Early warning classification: Set a three-level early warning mechanism (blue - potential risk, yellow - local anomaly, red - serious fault). When the monitored indicators exceed the threshold, send an early warning through methods such as text messages and APP push, and automatically generate disposal suggestions;

[0236] Simulation and deduction: Based on historical data and real-time monitoring information, use digital twin technology to simulate scenarios such as pollution diffusion and equipment failures, and formulate coping strategies in advance;

[0237] III. Key Technical Principles:

[0238] (I) Principle of Hydrological Adaptation Reinforcement Learning:

[0239] Taking the hydrological conditions (flow rate, flow velocity, water level) and pollution status (type, concentration, diffusion range) of the river as the state space, and the operation instructions (gate opening, adsorption capacity, repair parameters) of each unit as the action space, optimize the decision-making strategy by maximizing the long-term cumulative reward (the balance between pollution reduction rate and energy consumption), so that the system can adapt to the pollution prevention and control requirements under different hydrological environments;

[0240] (II) Principle of Multi-Unit Cooperative Control:

[0241] Based on event-driven and rule engine technologies, establish the logical association between each unit; by defining the trigger conditions (such as pollution concentration, equipment status) and response actions, realize the full-process automatic cooperation from pollution monitoring to disposal, and avoid decision-making delays and operation errors caused by manual intervention;

[0242] (III) Principle of Digital Twin and Simulation Deduction:

[0243] Utilize 3D modeling and data mapping technologies to construct a virtual model that is synchronized with the physical system in real time; by inputting different pollution scenarios and control strategies, simulate the operation effect of the system, evaluate the feasibility of the decision-making scheme, and provide data support for optimizing the prevention and control strategy;

[0244] IV. Platform Workflow:

[0245] (I) Data Access and Initialization Phase:

[0246] Establish a data communication link with each unit, and complete device registration and parameter configuration;

[0247] Load historical hydrological data, pollution cases and initial model parameters, and start the data collection and preprocessing program;

[0248] (II) Real-Time Monitoring and Analysis Phase:

[0249] Continuously collect the operation data of each unit, and perform cleaning, storage and visualization display;

[0250] Use machine learning algorithms to analyze the data and identify pollution trends and system anomalies;

[0251] (III) Intelligent Decision-Making and Instruction Issuance Phase:

[0252] Input the real-time status data into the hydrological adaptation reinforcement learning model to calculate the optimal decision-making scheme;

[0253] Generate device control instructions and send them via the network to the targeted blocking unit and the ecological restoration unit for execution;

[0254] (4) Execution feedback and optimization stage:

[0255] Receive the feedback on the execution results of the device, compare the actual effect with the expected goal, and calculate the decision error;

[0256] Adjust the parameters of the learning model according to the error, optimize the decision-making strategy, and form a closed-loop control of "monitoring - decision-making - execution - optimization";

[0257] (5) System maintenance and upgrade stage:

[0258] Regularly repair vulnerabilities and upgrade functions of the platform software, and update the pollution prevention and control knowledge base;

[0259] Conduct inspections and maintenance on the hardware devices to ensure the stable operation of the communication network and the control terminal;

[0260] V. Application value of the platform:

[0261] (1) Improve prevention and control efficiency and accuracy:

[0262] Through intelligent decision-making and collaborative control, shorten the pollution response time to the minute level, increase the decision-making accuracy by more than 30%, and achieve precise and efficient pollution prevention and control;

[0263] (2) Reduce operation and maintenance costs:

[0264] Based on data-driven fault warning and intelligent maintenance, reduce the equipment failure rate by 40% and lower the manual inspection and repair costs; optimize energy consumption management to reduce the system operation energy consumption by 25%;

[0265] (3) Support scientific decision-making:

[0266] Provide analysis reports such as pollution trend prediction and prevention and control effect evaluation, and provide data basis for the environmental protection department to formulate river basin treatment plans, policies and regulations;

[0267] (4) Enhance system scalability:

[0268] Adopt a modular architecture design, support the rapid access and integration of new devices and new algorithms, facilitate system function upgrade and application scenario expansion, and adapt to the diverse needs of future water environment governance.

[0269] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An integrated prevention and control system for targeted blocking of pollution sources and ecological restoration of an estuarine river, characterized in that: It includes a pollution source identification unit, a targeted blocking unit, an ecological restoration unit and a collaborative control platform: The pollution source identification unit includes a multispectral UAV scanning module, a water quality sensor array and an artificial intelligence tracing and analysis module, which are used to identify the type, location and diffusion path of pollution sources in real time; The targeted blocking unit includes an intelligent sluice system, a dynamic adsorption device and a degradable barrier film, which implement physical-chemical combined blocking in the target river section according to the type of pollution source; The ecological restoration unit includes a modular artificial wetland, a composite microbial agent releaser and an aquatic vegetation restoration bed, which are arranged downstream of the blocking unit; The collaborative control platform is connected to each unit and dynamically optimizes the blocking and restoration strategies based on river hydrological data.

2. The integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in an estuarine river according to claim 1, characterized in that: The artificial intelligence tracing and analysis module performs the following operations: Receiving the river channel thermal map scanned by the multispectral UAV and the real-time parameters of the water quality sensor array; Identifying the spatial distribution of pollution plumes through an improved spatio-temporal convolutional neural network. The calculation process is as follows: Inputting the multispectral data and water quality parameter fusion matrix into a dynamic convolutional kernel, injecting a gradient correction term into the convolutional operation to enhance the identification of small-scale pollution sources, and outputting pollution characteristic values through an adaptive LeakyReLU activation function; Combining with the flow velocity-discharge model to predict the diffusion path, and outputting a classification report of pollution source types and a tracing confidence evaluation report.

3. The integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in an estuarine river according to claim 2, characterized in that: The weight update process of the dynamic convolutional kernel is as follows: Based on the gradient of the loss function of historical training samples, dynamically adjusting the convolutional kernel weights in combination with a time decay mechanism, so that the weights decay as the time interval of pollution events increases.

4. An integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in an estuarine river according to claim 1, characterized in that: The targeted blocking unit satisfies: The intelligent sluice system automatically opens and closes according to the location of the pollution source. Its gate opening control logic is: Taking the maximum allowable opening as the benchmark, responding to the distance between the pollution source and the gate through an S-shaped function, and quickly reducing the opening when the distance is less than the critical response distance; The dynamic adsorption device contains magnetic nano-adsorbents and ion exchange resins; The degradable barrier film is made of poly(lactic-co-glycolic acid) and degrades within 48 hours after interception.

5. The integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in an estuarine river according to claim 4, characterized in that: The calculation logic for the dosage of the dynamic adsorption device is: According to the river channel cross-section flow rate, the time-varying curve of pollutant concentration and the adsorption reaction time constant, determining the adsorbent dosage through integral operation and introducing a pollutant density correction factor for calibration.

6. The integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in an estuary river according to claim 1, wherein: In the ecological restoration unit: The release rate regulation logic of the composite microbial agent releaser is: Taking the basic release rate as the benchmark, responding to the deviation between the real-time ammonia nitrogen concentration and the threshold through a hyperbolic tangent function, and the release rate increases non-linearly when the concentration exceeds the threshold; The modular artificial wetland consists of a series of anaerobic sedimentation tanks, aerobic biological filters and ecological gravel beds; The aquatic vegetation restoration bed adopts a three-dimensional braided structure.

7. An integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in an estuarine river, as claimed in claim 1, wherein: The dynamic decision-making engine of the collaborative control platform adopts a hydrological adaptive reinforcement learning model. Its decision-making mechanism includes: Updating the action value function through an experience replay mechanism; The real-time reward function is calculated by weighting the pollution load reduction rate and the system energy consumption; Introducing a policy gradient correction term to improve the adaptability to hydrological mutations.

8. An integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in an estuarine river, according to claim 7, characterized in that: The calculation logic of the policy gradient correction term is: Based on the state-action trajectory data in a parallel environment, optimize the decision-making policy parameters by taking the derivative of the product of the policy network probability and the cumulative reward.

9. An integrated prevention and control system for targeted blocking and ecological restoration of pollution sources in rivers flowing into the sea according to claim 6, characterized in that: The prediction logic for the plant survival rate of the aquatic vegetation restoration bed is as follows: According to the weighted linear combination of dissolved oxygen concentration, water body pH value, and turbidity, map the survival probability through the S-shaped function, where the weight of dissolved oxygen is the highest.

Images