Multi-source perception intelligent closed-loop water quality restoration system and control method thereof

Abstract

The present application relates to the technical field of water quality restoration, and more particularly to a multi-source sensing intelligent closed-loop water quality restoration system and a control method thereof, which is suitable for river management, landscape water maintenance, industrial wastewater treatment and the like, and is particularly suitable for small and medium-sized water body restoration projects that need to dynamically respond to water quality changes. By means of multi-channel online monitoring, physicochemical index data and model organism activity data are obtained, combined with artificial sampling data, an evaluation model is used to comprehensively evaluate the water quality and grade the abnormality level; and according to the abnormality level, a graded restoration program is automatically triggered, and automatic backflow, automatic sample reservation and remote alarm are simultaneously executed, so as to realize closed-loop, accurate and efficient restoration of water quality and whole-process traceable management, improve restoration efficiency and reliability, and reduce the frequency of manual intervention.

Description

Technical Field

[0001] This invention relates to the field of water quality remediation technology, and in particular to a multi-source sensing intelligent closed-loop water quality remediation system and its control method. It is applicable to scenarios such as river management, landscape water body maintenance, and industrial wastewater treatment, and is especially suitable for small and medium-sized water body remediation projects that require dynamic response to changes in water quality. Background Technology

[0002] Water pollution has become a global environmental problem. Especially with the acceleration of industrialization and rapid urbanization, water pollution poses a serious threat to the balance of the ecological environment and human health. Currently, although various water remediation technologies and equipment have been put into use for river management, landscape water body maintenance, and industrial wastewater treatment, the following significant shortcomings still exist in practical applications: 1. Most existing water remediation equipment uses fixed treatment processes (such as single physical filtration, biodegradation, or chemical dosing), which cannot dynamically adjust the remediation strategy according to real-time changes in water quality. When faced with complex and fluctuating polluted water bodies, existing equipment often exhibits poor adaptability and rigid processes, resulting in water quality that still fails to meet standards after remediation.

[0003] 2. Traditional water quality monitoring systems mainly rely on physicochemical sensors (such as pH, dissolved oxygen, turbidity, etc.) for data collection. However, this monitoring method has obvious limitations. For example, conventional physicochemical sensors cannot fully cover specific indicators such as heavy metal pollution, lack the ability to perceive the biological toxicity of water bodies in real time, and cannot truly reflect the ecological safety of water bodies.

[0004] 3. When the effluent quality is abnormal, the existing remediation system lacks the functions of automatic interception and recirculation treatment, which may lead to the direct discharge of unqualified water into the natural environment. At the same time, it also lacks the function of sample retention and traceability, making it difficult to trace the source of pollution and analyze the cause of the accident.

[0005] Therefore, this case is brought. Summary of the Invention

[0006] The purpose of this invention is to provide a multi-source sensing intelligent closed-loop water quality remediation system and its control method, so as to achieve precise and efficient water quality remediation and full-process traceability management.

[0007] To achieve the above objectives, the technical solution of the present invention is as follows: A multi-source sensing intelligent closed-loop water quality remediation system includes: The graded water quality remediation unit includes a return water inlet, a raw water inlet, a remediation water main outlet, and a built-in multi-stage remediation module, which is used to perform graded remediation of the incoming water according to control commands. The detection unit is used to receive the remediation water output from the graded water quality remediation unit, and to detect the remediation water. At the same time, it retains samples of the remediation water according to control instructions. The data collected by the detection includes at least physicochemical index data and model organism activity data. The drainage control unit is used to receive the remediation water output from the detection unit and control the remediation water to be discharged or returned to the graded water quality remediation unit according to the control command. The intelligent control unit receives data collected by the detection unit and data from manual sampling, processes and analyzes the received data using a built-in evaluation model to assess the quality of the remediated water, and issues control commands to the graded water quality remediation unit, detection unit, and drainage control unit based on the assessment results.

[0008] Furthermore, the built-in multi-level repair module includes a first-level repair module, a second-level repair module, and a third-level repair module connected in series. The bottom of the primary repair module is equipped with an aeration device, and the side wall is equipped with an activated carbon inlet, which is connected to the activated carbon silo through a metering device. The secondary repair module is equipped with a microbial inlet, which is connected to the bacterial solution storage tank via a metering device. The three-level repair module is equipped with a chemical agent inlet, which is connected to a chemical agent storage tank through a metering and dispensing device; The metering and dispensing device controls the amount of material dispensed into each level of the repair module according to control commands; Both the return water inlet and the raw water inlet are connected to the inlet of the first-stage repair module, and the outlet of the third-stage repair module is connected to the main outlet of the repair water.

[0009] Furthermore, the detection unit includes a physicochemical sensing unit, a model organism sensing unit, and a sample retention unit. The main outlet of the tiered water quality remediation unit is divided into three branches via pipelines. The three pipelines are respectively connected to the inlets of the physicochemical sensing unit, the model organism sensing unit, and the sample retention unit. The outlets of the physicochemical sensing unit and the model organism sensing unit are connected to the inlet of the drainage control unit via pipelines.

[0010] Furthermore, the physicochemical sensing unit includes a physicochemical sensor group for performing physicochemical detection on the remediation water. The physicochemical sensor group includes at least temperature, turbidity, TDS, pH, dissolved oxygen, oxidation-reduction potential, permanganate index, five-day biochemical oxygen demand, ammonia nitrogen, total nitrogen, and total phosphorus physicochemical sensors.

[0011] Furthermore, the model organism sensing unit includes a transparent container and a camera. The inlet and outlet of the transparent container are equipped with flow regulating valves. The model organism is placed inside the transparent container, and the camera is used to collect video of the model organism's activities.

[0012] Furthermore, the sample retention unit includes a sample retention water pump, a multi-channel valve module, and sample retention containers. The input end of the sample retention water pump is connected to the inlet of the sample retention unit, and the output end of the sample retention water pump is connected to the input end of the multi-channel valve module. The multi-channel valve module includes multiple output channels and a sample retention opening and closing valve set on each output channel. After the repair water output from the sample retention water pump enters the multi-channel valve module, it enters each output channel. The number of sample retention containers corresponds to the number of output channels of the multi-channel valve module and is used to receive the repair water output from the output channels. The sample retention pump and sample retention valve are opened and closed according to control commands.

[0013] Furthermore, the drainage control unit is equipped with a diversion pipe, a return water pump, and a drain valve. The inlet of the diversion pipe is connected to the inlet of the drainage control unit. The first outlet of the diversion pipe is connected to the input end of the return water pump. The output end of the return water pump is connected to the return water inlet of the graded water quality restoration unit. The second outlet of the diversion pipe is used as a drain outlet and is equipped with a drain valve. The return water pump and drain valve are opened and closed according to control commands.

[0014] Furthermore, it includes a communication module, through which the intelligent control unit communicates with at least one of a cloud server, a mobile control terminal, or a local control terminal.

[0015] Furthermore, it includes an information display interface electrically connected to the intelligent control unit, used to display detection data, equipment operating status and processing status, and to provide operation command input and data recording functions.

[0016] A control method for the multi-source sensing intelligent closed-loop water quality remediation system includes the following steps: S1. Collect physicochemical index data, model organism activity data, and manually input sampling and detection data; S2. Perform outlier removal, missing value imputation, and Min-Max standardization on the collected data to shrink all data to the 0~1 range; S3. The physicochemical index data and model organism activity data at the same time point are spliced ​​together point by point to obtain the fused feature time series; S4. Input the fused feature time series into the trained water quality anomaly comprehensive assessment model based on bidirectional gated recurrent units; the model analyzes the trend of pollution impact on biological behavior through the forward GRU layer, analyzes the previous pollution changes corresponding to behavioral anomalies through the reverse GRU layer, outputs the probability value of the current remediated water belonging to each level of anomaly, and takes the category with the highest probability and exceeding the preset threshold as the final water quality anomaly level determination result. S5. If the water quality is determined to be normal, the drain valve is opened and the return water pump is turned off to discharge the repair water; If the water quality is determined to be abnormal, the abnormal water quality sampling is initiated. At the same time, the drain valve is closed and the return water pump is turned on to return the repair water to the graded water quality repair unit. Based on the water quality abnormality level and the preset repair logic, graded repair is carried out.

[0017] The advantages of this invention are: 1. The system can match physical, biological or chemical graded remediation strategies based on the real-time assessment of water quality anomalies; once the effluent fails to meet the standards, it will immediately and automatically cut off the discharge path and start the return water pump to send the water back to the front end for reprocessing, while triggering the automatic sampling function; thus achieving precise and efficient water quality remediation and full-process traceability management.

[0018] 2. The proposed comprehensive assessment and grading model for water quality anomalies breaks through the traditional limitations of "separate judgment of physicochemical indicators" and "qualitative description of biological indicators". It uses water body physicochemical parameters (reflecting the components and intensity of pollutants), zebrafish movement behavior parameters (reflecting comprehensive ecotoxicity), and artificial sampling and detection data as input sources to achieve complementary coupling of multiple parameters, complete the accurate quantitative comprehensive assessment of water quality anomaly levels, realize the early identification and warning of water body compound pollution and sudden toxic events, and grade water quality anomaly levels, significantly improving ecological safety.

[0019] 3. With the support of remote communication and information display interface, it supports seamless switching between remote monitoring and manual intervention, taking into account both automation and operational flexibility. It significantly reduces the cost of manual operation and maintenance and the risk of misjudgment, and is particularly suitable for scenarios such as river management, landscape water body maintenance, and industrial wastewater treatment. It has significant environmental benefits and intelligent application value. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the multi-source sensing intelligent closed-loop water quality remediation system in the embodiment; Figure 2 This is a schematic diagram of the intelligent control part of the multi-source sensing intelligent closed-loop water quality remediation system in the embodiment. Figure 3 This is a schematic diagram of the control process of the multi-source sensing intelligent closed-loop water quality remediation system in the embodiment; Figure 4 This is a detailed schematic diagram of the sampling unit in the embodiment. Detailed Implementation

[0021] The present invention will be further described in detail below with reference to embodiments. It should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., used in this document indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the present invention.

[0022] like Figure 1 As shown in the figure, this embodiment proposes a multi-source sensing intelligent closed-loop water quality remediation system. This system is an integrated system of "sensing-control-remediation-sample retention", mainly including a graded water quality remediation unit, a physicochemical sensing unit, a model organism sensing unit, a sample retention unit, a drainage control unit, an intelligent control unit, and an information display and interactive interface.

[0023] The graded water quality remediation unit includes three built-in remediation modules: a primary module, a secondary module, and a tertiary module, which are connected in series via guide pipes. The primary module has an aeration device at its bottom and an activated carbon inlet on its side wall, connected to an activated carbon silo via a metering device. The secondary module has a microbial inlet connected to a bacterial solution storage tank via a metering device. The tertiary module has a chemical reagent inlet connected to a reagent storage tank via a metering device. The metering device controls the amount of material added to each remediation module based on received control commands.

[0024] The graded water quality remediation unit includes a raw water inlet, a return water inlet, and a remediation water main outlet. Both the raw water inlet and the return water inlet are connected to the inlet of the first-level remediation module, and the outlet of the third-level remediation module is connected to the remediation water main outlet.

[0025] The main outlet of the graded water quality remediation unit is divided into three pipelines, which are respectively connected to the inlets of the physicochemical sensing unit, the model organism sensing unit, and the sample retention unit. The outlets of the physicochemical sensing unit and the model organism sensing unit are connected to the inlet of the drainage control unit through pipelines.

[0026] In this embodiment, the physicochemical sensing unit device integrates a group of physicochemical sensors, including temperature, turbidity, TDS (total dissolved solids), pH value, dissolved oxygen (DO), oxidation-reduction potential (ORP), and permanganate index (COD). mn Physicochemical sensors for five-day biochemical oxygen demand (BOD5), ammonia nitrogen (NH3-N), total nitrogen (TN), and total phosphorus (TP).

[0027] The model organism sensing unit includes a transparent container and a camera. Both the inlet and outlet of the transparent container are equipped with flow regulating valves to adjust the water flow within the container. The model organism (a zebrafish in this embodiment) is placed inside the transparent container, and the camera is fixed to the outer wall to capture video of the model organism's activities.

[0028] like Figure 4 As shown, the sample retention unit includes a sample retention water pump, a multi-channel valve module, and sample retention containers. The input end of the sample retention water pump is connected to the inlet of the sample retention unit, and the output end of the sample retention water pump is connected to the input end of the multi-channel valve module. The multi-channel valve module includes multiple output channels and a sample retention on / off valve installed on each output channel. The remedial water output from the sample retention water pump enters the multi-channel valve module and then flows into each output channel. The number of sample retention containers corresponds to the number of output channels of the multi-channel valve module and is used to receive the remedial water output from the output channels. The sample retention water pump and the sample retention on / off valves are opened and closed according to received control commands.

[0029] The drainage control unit includes a diversion pipeline, a return water pump, and a drain valve. The inlet of the diversion pipeline is connected to the inlet of the drainage control unit. The first outlet of the diversion pipeline is connected to the input of the return water pump, and the output of the return water pump is connected to the return water inlet of the graded water quality restoration unit. The second outlet of the diversion pipeline serves as a drain outlet and is equipped with a drain valve. The return water pump and drain valve are opened and closed according to received control commands.

[0030] like Figure 2 As shown, the intelligent control unit is used to receive data from online monitoring physicochemical sensors, video stream data from the video acquisition module (camera), and manual sampling and detection data. It processes and analyzes the received data using a built-in evaluation model to complete the evaluation of the remediation water quality. Based on the evaluation results, it issues control commands to the metering and dispensing device of the graded water quality remediation unit, the sample retention pump and sample retention valve of the sample retention unit, and the return water pump and drainage valve of the drainage control unit.

[0031] The system also includes a communication module, through which the intelligent control unit communicates with at least one of a cloud server, a mobile control terminal, or a local control terminal, supporting local and remote push notifications of alarm information and switching between "automatic" and "manual" modes. The information display interface is electrically connected to the intelligent control unit, used to display detection data, equipment operating status, and processing status, and provides operation command input and data recording functions.

[0032] The following example illustrates a landscape wastewater remediation scenario, along with related information. Figure 3 This explains the workflow of the aforementioned multi-source sensing intelligent closed-loop water quality remediation system.

[0033] I. Routine Operation Raw water flows into the graded water quality remediation unit through the raw water inlet, and then flows through the primary remediation module, the secondary remediation module, and the tertiary remediation module in sequence before flowing out through the main remediation water outlet.

[0034] II. Multi-source detection The physicochemical sensing unit performs bypass sampling of the effluent and monitors key physicochemical indicators in real time, including: Physical properties: temperature, turbidity, TDS (total dissolved solids); Chemical indicators: pH value, dissolved oxygen (DO), oxidation-reduction potential (ORP), permanganate index (COD) mn ), five-day biochemical oxygen demand (BOD5), ammonia nitrogen (NH3-N), total nitrogen (TN), and total phosphorus (TP).

[0035] In the model organism sensing unit, zebrafish move in the water flow inside a transparent container. A camera continuously collects behavioral videos, and the intelligent control unit identifies the activity status data through image analysis, specifically including data on individual zebrafish behavior changes and fish school behavior changes.

[0036] Indicators of individual behavioral changes in zebrafish include: (1) Speed ​​refers to the distance that the geometric center point of the zebrafish moves per unit time; (2) Acceleration refers to the acceleration of the geometric center point of the zebrafish body; (3) Turning corner, refers to the change in the zebrafish's direction of travel; (4) Swerveness, which refers to the ratio of the change in the direction of movement of the zebrafish to the distance it moves; (5) Activity refers to the activity level of the zebrafish, that is, the percentage change in pixels between the current zebrafish sample and the previous sample in the video monitoring image.

[0037] Indicators of changes in zebrafish group behavior include: (1) Group average speed refers to the average distance swam by all zebrafish in the group per unit time, reflecting the overall activity intensity of the group; (2) Group speed standard deviation refers to the dispersion of the instantaneous speed of all zebrafish in the group, reflecting the consistency of behavior and the degree of stress disorder; (3) Group average turning rate, the average number of times an individual in the group undergoes a significant change in direction of movement (≥60°); (4) Group aggregation degree, the average distance of all individuals in the group to the geometric center of the group. The smaller the distance, the higher the aggregation degree. (5) The proportion of time the group spends near the edge: The proportion of the total time that zebrafish spends in the edge area of ​​the tank to the total observation time reflects anxiety and stress behavior; (6) The proportion of time zebrafish spend in the middle layer of the water body is the proportion of time they spend in the safe area (middle layer), which reflects their environmental adaptability; (7) The proportion of time the group is stationary, which is the proportion of time spent by individuals with a speed below the threshold (usually <2 mm / s), reflects the degree of poisoning inhibition; (8) Group up-and-down shuttle frequency: the average number of times a group shuttles between the upper and lower layers of the water per unit time, reflecting hypoxia or toxic stress.

[0038] (9) Synchronicity of group movement, the degree of correlation between the direction and speed of individual movement within the group, reflecting the group's coordination; (10) Total frequency of abnormal behavior in the group: the total number of times the group exhibits abnormal behaviors such as frantic swimming, sudden stopping, rolling over, imbalance, floating and convulsions during the observation period.

[0039] Manual sampling and testing involves routine and irregular inspections, supplementing and enhancing online monitoring (physicochemical sensors + model organisms). It also serves as an input parameter path for the intelligent control unit, triggering subsequent water quality anomaly assessments along with physicochemical index data and zebrafish movement behavior data. Manual sampling and testing primarily focuses on water pollution and ecological-related detection, and the data includes: (1) Conventional water quality parameters: water temperature, pH, dissolved oxygen (DO), COD, BOD5, ammonia nitrogen, total phosphorus, total nitrogen, etc.; (2) Heavy metal content parameters: copper, zinc, lead, cadmium, mercury, arsenic, chromium, etc.; (3) Other relevant parameters: volatile phenol concentration, anionic surfactant concentration, petroleum-related substance concentration, etc.

[0040] Manually sampled data can be input through the water quality monitoring interface. Online monitoring data from physicochemical sensors and model organisms, as well as manually sampled data, can all be displayed on the water quality display interface.

[0041] III. Comprehensive Assessment of Water Quality Anomalies This invention employs a bidirectional gated recurrent unit (BiGRU) to construct a comprehensive water quality anomaly assessment model. As an improved model of recurrent neural networks (RNN), BiGRU effectively captures the sequential dependencies of time series through bidirectional coupling of forward and reverse GRUs. It excels at mining deep correlation features between multidimensional parameters. Compared with traditional machine learning models (SVM, BP neural network) and single GRU models, it has stronger feature extraction and generalization capabilities, and is suitable for the fusion analysis of multidimensional time series data such as "physicochemical parameters + biological behavior parameters + artificial sampling and detection parameters".

[0042] This invention's comprehensive water quality anomaly assessment model adopts a "multi-input, single-output" structure. The inputs are two types of parameters monitored online: water quality physicochemical parameters and zebrafish movement behavior parameters, as well as water quality parameters detected through manual sampling. The output is the water quality anomaly level (Level 1, Level 2, Level 3). (Reference) Figure 3 The specific assessment includes the following process: S1. Collect physicochemical index data, zebrafish movement behavior data, and manually input sampling and detection data; S2. Preprocess the collected data, including outlier removal, missing value imputation, and Min-Max standardization, reducing all data to the 0~1 range; S3. The preprocessed physical parameters (3 items), chemical parameters (8 items) + zebrafish individual behavior parameters (6 items) + zebrafish group behavior parameters (10 items) are spliced ​​together by time, that is, the physical and chemical index data at the same time point are spliced ​​together with the model organism activity data point by point to obtain the fusion feature time series sequence; S4. Input the fused feature time series into the trained water quality anomaly comprehensive assessment model (BiGRU model) based on bidirectional gated recurrent units. The BiGRU model is constructed based on PyTorch and includes a 4-layer core structure: Input layer: used to receive time series data fused from physicochemical parameters, zebrafish behavior parameters, and artificial detection parameters; BiGRU layer: bidirectionally extracts time series features, forward GRU: analyzes how pollution affects behavior from the past to the present, and backward GRU: analyzes the changes in pollution corresponding to previous pollution from the present to the past; Feature compression layer: compresses the time series information of all time points into a whole feature; Fully connected + Output layer: The fully connected layer maps the features extracted by BiGRU into 3 water quality levels, and Softmax turns the output into a probability of 0~1, with a sum of 1. The output layer outputs the probability of each sample corresponding to the first, second, and third level of abnormal water quality; the category with the highest probability and exceeding the preset threshold is taken as the final water quality anomaly level determination result. The preset threshold mentioned here means that when the probability is ≥0.7, the judgment result is considered valid, and when the probability is <0.7, the judgment result is marked as "pending review", and the final water quality anomaly level needs to be manually confirmed before proceeding to the next step. S5. If the water quality is determined to be normal, the drain valve is opened and the return water pump is turned off to discharge the repair water; If the water quality is determined to be abnormal, the abnormal water quality sampling is initiated. At the same time, the drain valve is closed and the return water pump is turned on to return the repair water to the graded water quality repair unit. Based on the water quality abnormality level and the preset repair logic, graded repair is carried out.

[0043] IV. Graded Response Level 1 Anomaly: Automatically enhances the physical repair intensity of the Level 1 repair module, that is, activates the metering and dispensing device to add activated carbon for pollutant adsorption.

[0044] Level 2 Anomaly: Based on the Level 1 measures, the bioremediation measures of the Level 2 remediation module are initiated, namely, the metering and dispensing device is activated to introduce microbial liquid to decompose pollutants.

[0045] Level 3 anomaly: Based on the Level 2 measures, the chemical remediation measures of the Level 3 remediation module are initiated, that is, the metering and dosing device is activated to add chemical agents for water purification.

[0046] Synchronous execution: Close the drain valve, start the return water pump, and return the drain water to the graded water quality remediation unit for reprocessing; the sample retention unit starts automatically, retains the current water sample in the sample retention container, and records the abnormal time, level, and system operation log.

[0047] V. Human Interaction Operators can view test data, anomaly records, and sample retention information in real time through the local information display interface or mobile terminal, and can switch to manual mode to take over control at any time.

[0048] VI. Data Management All operating parameters, monitoring data, and operation logs are encrypted and stored on local storage devices, and can be exported as needed for process analysis and optimization.

[0049] This solution constructs a multi-source fusion sensing system that integrates online monitoring of "physicochemical indicators + zebrafish biological behavior" with manual sampling and detection. Combined with the BiGRU deep learning model, it accurately captures the latent toxicity and temporal evolution characteristics of water bodies, achieving early warning and graded assessment of water quality risks. Utilizing an intelligent closed-loop control mechanism of "monitoring-assessment-sample retention and traceability-automatic backflow-graded remediation," it can dynamically match physical, biological, and chemical three-level remediation strategies according to the real-time water quality level, ensuring "zero discharge" of abnormal water bodies and automatically retaining evidence. This effectively solves the pain points of traditional technologies, such as single monitoring dimensions, delayed response, and direct discharge of pollutants exceeding standards, significantly improving the accuracy of water quality remediation, ecological safety, and automated operation and maintenance level.

[0050] The above implementation is only used to explain the concept of the present invention, and is not intended to limit the protection of the present invention. Any non-substantial modifications made to the present invention using this concept should fall within the protection scope of the present invention.

Claims

1. A multi-source sensing intelligent closed-loop water quality remediation system, characterized in that, include: The graded water quality remediation unit includes a return water inlet, a raw water inlet, a remediation water main outlet, and a built-in multi-stage remediation module, which is used to perform graded remediation of the incoming water according to control commands. The detection unit is used to receive the remediation water output from the graded water quality remediation unit, and to detect the remediation water. At the same time, it retains samples of the remediation water according to control instructions. The data collected by the detection includes at least physicochemical index data and model organism activity data. The drainage control unit is used to receive the remediation water output from the detection unit and control the remediation water to be discharged or returned to the graded water quality remediation unit according to the control command. The intelligent control unit receives data collected by the detection unit and data from manual sampling, processes and analyzes the received data using a built-in evaluation model to assess the quality of the remediated water, and issues control commands to the graded water quality remediation unit, detection unit, and drainage control unit based on the assessment results.

2. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 1, characterized in that, The built-in multi-level repair module includes a first-level repair module, a second-level repair module, and a third-level repair module connected in series. The bottom of the primary repair module is equipped with an aeration device, and the side wall is equipped with an activated carbon inlet, which is connected to the activated carbon silo through a metering device. The secondary repair module is equipped with a microbial inlet, which is connected to the bacterial solution storage tank via a metering device. The three-level repair module is equipped with a chemical agent inlet, which is connected to a chemical agent storage tank through a metering and dispensing device; The metering and dispensing device controls the amount of material dispensed into each level of the repair module according to control commands; Both the return water inlet and the raw water inlet are connected to the inlet of the first-stage repair module, and the outlet of the third-stage repair module is connected to the main outlet of the repair water.

3. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 1, characterized in that, The detection unit includes a physicochemical sensing unit, a model organism sensing unit, and a sample retention unit. The main outlet of the tiered water quality remediation unit is divided into three pipelines, which are respectively connected to the inlets of the physicochemical sensing unit, the model organism sensing unit, and the sample retention unit. The outlets of the physicochemical sensing unit and the model organism sensing unit are connected to the inlet of the drainage control unit through pipelines.

4. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 3, characterized in that, The physicochemical sensing unit includes a physicochemical sensor group for physicochemical detection of the remediation water. The physicochemical sensor group includes at least temperature, turbidity, TDS, pH, dissolved oxygen, oxidation-reduction potential, permanganate index, five-day biochemical oxygen demand, ammonia nitrogen, total nitrogen, and total phosphorus physicochemical sensors.

5. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 3, characterized in that, The model organism sensing unit includes a transparent container and a camera. The inlet and outlet of the transparent container are equipped with flow regulating valves. The model organism is placed inside the transparent container, and the camera is used to collect video of the model organism's activities.

6. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 3, characterized in that, The sample retention unit includes a sample retention water pump, a multi-channel valve module, and sample retention containers. The input end of the sample retention water pump is connected to the inlet of the sample retention unit, and the output end of the sample retention water pump is connected to the input end of the multi-channel valve module. The multi-channel valve module includes multiple output channels and a sample retention on / off valve installed on each output channel. After the remediation water output from the sample retention water pump enters the multi-channel valve module, it enters each output channel. The number of sample retention containers corresponds to the number of output channels of the multi-channel valve module and is used to receive the remediation water output from the output channels. The sample retention pump and sample retention valve are opened and closed according to control commands.

7. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 1, characterized in that, The drainage control unit is equipped with a diversion pipeline, a return water pump and a drain valve. The inlet of the diversion pipeline is connected to the inlet of the drainage control unit. The first outlet of the diversion pipeline is connected to the input end of the return water pump. The output end of the return water pump is connected to the return water inlet of the graded water quality restoration unit. The second outlet of the diversion pipeline is used as a drain outlet and is equipped with a drain valve. The return water pump and drain valve are opened and closed according to control commands.

8. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 1, characterized in that, The system includes a communication module, through which the intelligent control unit communicates with at least one of a cloud server, a mobile control terminal, or a local control terminal.

9. The multi-source sensing intelligent closed-loop water quality remediation system as described in claim 1, characterized in that, It includes an information display interface that is electrically connected to the intelligent control unit, used to display detection data, equipment operating status and processing status, and provides operation command input and data recording functions.

10. A control method for the multi-source sensing intelligent closed-loop water quality remediation system according to any one of claims 1 to 9, characterized in that, Includes the following steps: S1. Collect physicochemical index data, model organism activity data, and manually input sampling and detection data; S2. Perform outlier removal, missing value imputation, and Min-Max standardization on the collected data to shrink all data to the 0~1 range; S3. The physicochemical index data and model organism activity data at the same time point are spliced ​​together point by point to obtain the fused feature time series; S4. Input the fused feature time series into the trained water quality anomaly comprehensive assessment model based on bidirectional gated recurrent units; the model analyzes the trend of pollution impact on biological behavior through the forward GRU layer, analyzes the previous pollution changes corresponding to behavioral anomalies through the reverse GRU layer, outputs the probability value of the current remediated water belonging to each level of anomaly, and takes the category with the highest probability and exceeding the preset threshold as the final water quality anomaly level determination result. S5. If the water quality is determined to be normal, the drain valve is opened and the return water pump is turned off to discharge the repair water; If the water quality is determined to be abnormal, the abnormal water quality sampling is initiated. At the same time, the drain valve is closed and the return water pump is turned on to return the repair water to the graded water quality repair unit. Based on the water quality abnormality level and the preset repair logic, graded repair is carried out.

Images