Water quality adaptive control system and method based on multi-point detection

Abstract

The present invention discloses a water quality adaptive control system and method based on multi-point detection, which relates to the field of water quality adaptive control technology. A sensor network collects water quality data in each sub-area of ​​a target water area in real time, and screens out abnormal areas after determining the abnormal state of water quality in the sub-area by water quality abnormality score. An inspection drone is used to inspect each abnormal area, and a water area inspection data set is generated after obtaining corresponding inspection data; an intervention strategy is output for the target water area by a water quality intervention knowledge graph and then executed; an automatic control model is used to automatically control the operation of water treatment facilities and collect corresponding intervention feedback data, and the intervention strategy is adjusted and optimized according to the intervention feedback data; and a water quality monitoring report is sent to a data receiving end at a time node that meets the constraint conditions. When the water quality in the target water area is abnormal, automatic and adaptive adjustment control is realized to prevent the water quality in the target water area from deteriorating.

Description

Technical Field

[0001] The present invention relates to the technical field of water quality adaptive control, and in particular to a water quality adaptive control system and method based on multi-point detection. Background Art

[0002] Adaptive water quality control is an advanced water quality management method that integrates modern sensing technology, data analysis and artificial intelligence algorithms. By deploying a high-precision water quality monitoring sensor network, this technology can capture the changes in key parameters in the water body in real time and comprehensively, such as dissolved oxygen, pH value, turbidity, heavy metal content and microbial activity. Based on these real-time monitoring data, the system's built-in adaptive algorithm will quickly analyze the water quality status and intelligently adjust the operating parameters of the water treatment process, such as dosage, filtration speed or aeration intensity, according to the preset water quality standards or user-defined target values.

[0003] This process is dynamic and can automatically adapt to water quality fluctuations and seasonal changes, effectively respond to sudden pollution incidents, and ensure that water quality continues to meet standards. In addition, the water quality adaptive control system also has learning capabilities, which can continuously optimize algorithms in long-term operation, improve control efficiency and accuracy, and reduce unnecessary energy consumption and chemical use.

[0004] In the Chinese invention patent with application publication number CN115793585A, a method, system and medium for regulating water quality in shrimp farming are involved, which belongs to the technical field of shrimp farming water quality regulation. The present invention constructs a water quality change prediction model, and predicts the relevant hazardous substance data in the current shrimp farming area according to the data information of the planned feeding amount of bait and the water quality change prediction model to obtain the prediction result; formulates a water quality control plan for each shrimp growth stage according to the prediction result, and obtains the real-time colony structure information and algae data information in the current shrimp farming area; optimizes and adjusts the water quality control plan based on the real-time colony structure information and algae data information in the current shrimp farming area. This method can predict water quality data according to feeding data, and then effectively control ammonia nitrogen and nitrite nitrogen according to the ecological function data and synergistic effect of microalgae, heterotrophic bacteria and nitrifying bacteria in the farming area.

[0005] Combined with the above application and the contents of the prior art:

[0006] When aquaculture is carried out in large areas of water, the impact of water quality on aquaculture output is particularly important. Therefore, it is necessary to monitor various water quality parameters in the water area in real time. When water quality parameters are abnormal, the water quality in the water area can be intervened in time to avoid the negative impact of abnormal water quality on aquaculture.

[0007] In order to enable abnormal water quality in the water area to be discovered in time, an alarm is usually issued to the outside when water quality parameters are abnormal. After receiving the alarm command, the management personnel operate and control the water treatment facilities in the water area to adjust the abnormal water quality. However, considering that the efficiency of manual control after the alarm is low, the existing water quality intervention control usually introduces water quality adaptive regulation to deal with the abnormal water quality in the water area. However, when the water environment is abnormal, for example, there is a large amount of rainfall, the ambient temperature has a large change, or there is a large amount of fertilizer or similar nutrient input, the water quality in the water area may change greatly. However, when the water area is large, the water quality in different water areas will still be quite different. However, the existing adaptive regulation cannot effectively screen out some areas that need to be focused on, and the output adaptive regulation scheme is also less targeted.

[0008] To this end, the present invention provides a water quality adaptive control system and method based on multi-point detection. Summary of the invention

[0009] 1. Technical issues to be resolved

[0010] In view of the deficiencies of the prior art, the present invention provides a water quality adaptive control system and method based on multi-point detection. The abnormal water quality state in the sub-area is determined by the water quality abnormality score, and then the abnormal area is screened out. The inspection drone is used to inspect each abnormal area, and the water area inspection data set is generated after the corresponding inspection data is obtained; the intervention strategy is output for the target water area by the water quality intervention knowledge map, and the intervention strategy is executed after the intervention strategy is executed, and the operation of the water treatment facility is automatically controlled by the automatic control model and the corresponding intervention feedback data is collected, and the intervention strategy is adjusted and optimized according to the intervention feedback data; and the water quality monitoring report is sent to the data receiving end at the time node that meets the constraint conditions. When the water quality in the target water area is abnormal, automatic and adaptive adjustment control is realized to prevent the water quality in the target water area from deteriorating; thereby solving the technical problems recorded in the background technology.

[0011] (II) Technical solution

[0012] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0013] The water quality adaptive control method based on multi-point detection includes collecting scene condition data in the target water area, generating a scene abnormality degree in the target water area from the scene condition data , if the scene abnormality Exceeding expectations, the sensor network collects water quality data in real time in each sub-area of ​​the target water area, and aggregates them to generate a water quality data set;

[0014] Generate water quality anomaly scores from water quality data in sub-areas, determine water quality anomaly status in sub-areas based on water quality anomaly scores, and then screen out abnormal areas. Use inspection drones to inspect each abnormal area, and after obtaining corresponding inspection data, summarize and generate a water area inspection data set.

[0015] The water quality of the abnormal area is verified by the acquired inspection data to generate the water pollution degree If the water pollution Do not exceed expectations. Predict water quality data in abnormal waters. If the predicted water quality changes exceed expectations, issue pollution analysis instructions to the outside world.

[0016] The water quality intervention knowledge graph outputs the intervention strategy for the target water area and then executes the intervention strategy. The automatic control model is used to automatically control the operation of the water treatment facilities and collect the corresponding intervention feedback data. The intervention strategy is adjusted and optimized based on the intervention feedback data.

[0017] Generate feedback coefficients from control feedback data of water quality regulation , if the feedback coefficient Do not exceed expectations, optimize the current water quality regulation process, and send the water quality monitoring report to the data receiving end at the time node that meets the constraints.

[0018] Furthermore, several monitoring points are set in the target area, and the environmental scene data in the target water area is collected at the monitoring points, and the scene condition data set is generated; the scene abnormality degree in the target water area is generated from the scene condition data ;

[0019] If the scene is abnormal When the abnormal threshold is exceeded, a water quality monitoring instruction is issued to the outside; a sensor network is deployed in the target water area, and the target area is divided into several sub-areas. After receiving the water quality monitoring instruction, the sensor network distributed in the sub-areas is started.

[0020] Furthermore, the ambient temperature , rainfall and the concentration of pollutants entering the target waters Perform linear normalization and map the corresponding data values ​​to the interval In the following way, the scene abnormality is generated :

[0021] ;

[0022] ;

[0023] ;

[0024] Weight coefficient, , , is the middle value of the scene, is the multidimensional data vector at time t, is the mean vector of multidimensional data, is the covariance matrix, is the Mahalanobis distance at time t, used to evaluate data anomalies, It is the water quality data monitoring cycle;

[0025] Furthermore, the water quality data in the sub-region is used as input, and the trained abnormal water quality model is used to perform abnormal assessment on the water quality in the sub-region to obtain the corresponding water quality abnormality score;

[0026] If the water quality abnormality score exceeds expectations, the corresponding time node is taken as the abnormal node, and the proportion of the water quality abnormality score exceeding expectations is taken as the water quality abnormality ratio, which is summarized to generate an abnormal state data set;

[0027] Generating water quality coefficients from abnormal state data sets If the water quality coefficient If the water quality exceeds the preset threshold, it will be regarded as an abnormal sub-area and the corresponding location information will be obtained. If the number of abnormal sub-areas exceeds expectations, a first-level alarm command will be issued to the outside.

[0028] Furthermore, based on the distribution of abnormal nodes and the corresponding water quality abnormality ratio Generated water quality coefficient , as follows:

[0029] ;

[0030] Weight coefficient, , ,and ; The value is 3.142; is the number of abnormal nodes, No. i The water quality anomaly ratio of abnormal nodes, It is i The abnormal node to j The time interval of abnormal nodes, is the water quality monitoring cycle, .

[0031] Furthermore, the abnormal sub-areas are marked on the electronic map. After receiving the first-level alarm command, the abnormal sub-areas are clustered using the pre-trained clustering algorithm to obtain the clustered abnormal areas; combined with the location information of several abnormal areas in the target waters, the pre-trained path planning algorithm is used to plan the path of the inspection drone;

[0032] Use inspection drones to inspect each abnormal area according to the obtained inspection path, obtain corresponding inspection data, and summarize and generate a water area inspection data set.

[0033] Furthermore, the water area inspection data is used as input, and the trained abnormal water quality model is used to perform abnormal analysis on the water quality in the abnormal area, obtain the corresponding water quality evaluation value, and calculate the current water pollution degree. If the water pollution If the water pollution threshold is exceeded, a secondary alarm command will be issued to the outside. The generation method is as follows:

[0034] ;

[0035] in, For the i The sum of the water quality anomaly score and the water quality evaluation value in each abnormal area. is the number of abnormal regions, is the mean of the sum of the two, is the weight coefficient, .

[0036] Furthermore, after receiving the secondary alarm command, the trained water quality prediction model is used to predict the water quality in the abnormal area and obtain the prediction data, and the water quality pollution degree is regenerated from the prediction data. ;

[0037] The water pollution degree at several consecutive time points Arrange along the time axis and analyze the water pollution degree The speed of change is calculated and used as the water quality change degree; if the obtained water quality change degree exceeds the change threshold, a pollution analysis instruction is issued to the outside.

[0038] Furthermore, after receiving the pollution analysis instruction, a sensitivity analysis of the water quality in the target area is performed based on the scene condition data, water area inspection data and water quality data to obtain sensitive factors affecting the water quality of the target water area;

[0039] A water quality intervention knowledge graph is pre-built with water quality intervention of aquaculture water bodies as the target word; the water quality intervention knowledge graph matches the corresponding intervention strategy for the target water area based on sensitive factors.

[0040] Further, the water quality intervention strategy for the target water area is implemented, and the corresponding feedback data is collected after a preset observation period, and the feedback data is aggregated to generate an intervention feedback data set;

[0041] Generate intervention coefficients from feedback data in the intervention feedback data set , several pre-coefficients Without exceeding the preset intervention threshold, the optimization goal is to improve the intervention coefficient, and the pre-trained genetic algorithm is used to adjust and iteratively optimize the intervention strategy in real time to obtain the optimized intervention strategy.

[0042] Furthermore, the water pollution degree before and after adjustment of each abnormal area is obtained. , the water pollution degree in the abnormal area Generate intervention coefficients , as follows:

[0043] ;

[0044] in, and Before and after intervention i The water quality coefficient on the abnormal area, and is the corresponding mean, i is the abnormal area number, , is the number of abnormal regions, For the i The intervention coefficient of the abnormal area, is the corresponding mean.

[0045] Furthermore, after executing the iteratively optimized water quality intervention strategy, the control feedback data of water quality regulation is collected periodically, and the feedback coefficient is generated from the control feedback data. , if the feedback coefficient Optimize the sensor network deployment and the frequency of inspections and water quality interventions without exceeding expectations; generate feedback coefficients The way is as follows:

[0046] ;

[0047] in, For the i Response speed during water quality intervention, is the corresponding response cost, , is the number of water quality interventions, weight coefficient: , ,and .

[0048] Furthermore, water quality monitoring reports are generated from water quality data, inspection data, and intervention feedback data, and data analysis results are visualized to output water quality monitoring reports at intervals. Constraints are as follows:

[0049] ;

[0050] Where: is the number of water quality interventions in the target waters, For the i The scene abnormality during the water quality intervention, is the corresponding mean, For time node and The difference in water pollution between is the expected value of the water pollution difference, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0. The weight coefficient is , ,and .

[0051] Water quality adaptive control system based on multi-point detection, including:

[0052] The scene analysis unit collects scene condition data in the target water area and generates the scene abnormality degree in the target water area from the scene condition data. , if the scene abnormality Exceeding expectations, the sensor network collects water quality data in real time in each sub-area of ​​the target water area, and aggregates them to generate a water quality data set;

[0053] The water quality analysis unit generates a water quality anomaly score from the water quality data in the sub-area, determines the abnormal water quality status in the sub-area based on the water quality anomaly score, and then selects the abnormal area, uses the inspection drone to inspect each abnormal area, and after obtaining the corresponding inspection data, summarizes and generates a water area inspection data set;

[0054] The water quality prediction unit verifies the water quality of the abnormal area based on the acquired inspection data and generates the water pollution degree If the water pollution Do not exceed expectations. Predict water quality data in abnormal waters. If the predicted water quality changes exceed expectations, issue pollution analysis instructions to the outside world.

[0055] The water quality intervention adjustment unit outputs the intervention strategy for the target water area based on the water quality intervention knowledge graph and then executes the intervention strategy. It uses the automatic control model to automatically control the operation of the water treatment facilities and collects the corresponding intervention feedback data, and adjusts and optimizes the intervention strategy based on the intervention feedback data.

[0056] Feedback optimization unit, which generates feedback coefficients from control feedback data of water quality regulation , if the feedback coefficient Without exceeding expectations, the current water quality regulation process is optimized, and the water quality monitoring report is sent to the data receiving end at the time node that meets the constraints.

[0057] (III) Beneficial effects

[0058] The present invention provides a water quality adaptive control system and method based on multi-point detection, which has the following beneficial effects:

[0059] 1. Constructing scene abnormality from scene condition data , according to the abnormality of the scene The abnormal conditions of the scene in the target water area are evaluated and used as feedback when abnormal conditions in the scene in the target water area are detected. The water quality in the target water area can be adjusted in time to ensure the safety of the water quality in the target water area.

[0060] 2. Divide the target water area into several sub-areas, and use the sensor network to collect water quality data in the sub-areas in real time, so as to form real-time monitoring and collection of water quality in the target water area; The water quality status of the sub-areas is comprehensively evaluated, and the time scale is added based on the real-time water quality data to make the anomaly evaluation more reliable.

[0061] 3. By clustering the adjacent abnormal sub-areas together, the coverage of abnormal water quality treatment is wider. By inspecting each abnormal area and collecting image data of the water body, secondary detection of water quality can be achieved.

[0062] 4. Construct water pollution degree from two water quality data , realize the comprehensive evaluation of water quality, re-issue the alarm command when the water quality in the abnormal area is abnormal, and timely pay attention to the abnormal water quality in the target water area; by predicting the water quality data, the degree and speed of change of the abnormal water quality in the target water area can be judged. If the current change speed and speed exceed expectations, the urgency of water quality treatment can be determined.

[0063] 5. Match the corresponding intervention strategy according to the key factors or sensitivity factors, and the current abnormal water quality can be targetedly treated according to the intervention strategy; automatically control the water treatment facilities after automatically executing the intervention strategy, and realize automatic and adaptive adjustment control when the water quality in the target water area is abnormal based on the real-time monitoring of the water quality data in the target water area, so as to prevent the water quality in the target water area from deteriorating.

[0064] 6. Use the acquired water quality data as feedback to construct intervention coefficients , according to the intervention coefficient To judge the effectiveness of current water quality intervention, it is convenient to dynamically adjust and optimize the intervention strategy, and the automatic regulation of water quality is more targeted; Evaluate the effectiveness of current water quality intervention and regulation, adjust the current intervention methods and facilities, improve the efficiency of subsequent automated water quality regulation, and continuously revise the intervention costs.

[0065] 7. By periodically generating water quality monitoring reports and adapting the monitoring reports to the water quality status and current scene conditions, when environmental conditions and water quality conditions change, the water quality status in the target waters can be understood in a timely manner to facilitate intervention and adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a schematic diagram of the process of the water quality adaptive control method of the present invention;

[0067] Figure 2 It is a schematic diagram of the water quality adaptive control method system of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] See also Figure 1 The present invention provides a water quality adaptive control method based on multi-point detection, comprising:

[0070] Step 1: Collect scene condition data in the target waters, and generate scene abnormality in the target waters from the scene condition data , if the scene abnormality Exceeding expectations, the sensor network collects water quality data in real time in each sub-area of ​​the target water area, and aggregates them to generate a water quality data set;

[0071] The step 1 includes the following contents:

[0072] Step 101: After determining the coverage area of ​​the aquaculture water body, take it as the target water area, set up a number of monitoring points in the target area, and collect environmental scene data in the target water area at the monitoring points, including environmental temperature data, rainfall data, and light condition data in the current stage, as well as data on fertilizers and other pollutants entering the water body;

[0073] Among them, the abnormality of these environmental field data will cause the water quality in the target water area to change, and finally affect the dissolved oxygen content in the target water area. The collected scene condition data are aggregated to generate a scene condition data set;

[0074] Step 102: When the scene condition data in the target water area is dimensionless, the scene abnormality degree in the target water area is generated from the scene condition data. , where the ambient temperature , rainfall and the concentration of pollutants entering the target waters Perform linear normalization and map the corresponding data values ​​to the interval In the following way:

[0075] ;

[0076] ;

[0077] ;

[0078] Weight coefficient, , , is the middle value of the scene, is the multidimensional data vector at time t, is the mean vector of multidimensional data, is the covariance matrix, is the Mahalanobis distance at time t, used to evaluate data anomalies, It is the water quality data monitoring cycle;

[0079] Based on historical data and management expectations of water quality conditions in the target waters, abnormal thresholds are set in advance; if the scene abnormality is If the abnormal threshold is exceeded, it means that under the current environmental conditions, the water quality in the target waters may deteriorate, which will have a significant impact on the subsequent breeding status. At this time, a water quality monitoring instruction is issued to the outside world;

[0080] Ambient temperature data, rainfall data, lighting conditions data, and data on fertilizers and other pollutants entering the water body are all important factors affecting water quality.

[0081] Temperature is one of the most important factors affecting aquatic ecosystems. Sustained high temperatures will change the physical and chemical properties of water bodies, thereby affecting the quality of the aquatic ecological environment. Specifically, it will reduce the self-purification capacity of water bodies: high temperatures will increase the evaporation of water bodies, and coupled with a decrease in precipitation, it will lead to insufficient water capacity and hydrodynamics, weakening the self-purification capacity of water bodies.

[0082] Reduce the concentration of dissolved oxygen in water: High temperature causes the solubility of oxygen in the air to decrease in water. At the same time, the oxygen consumption process of river sediment and water bodies is accelerated under high temperature conditions, resulting in a rapid decrease in the concentration of dissolved oxygen in water bodies. At the same time, high temperature is conducive to promoting the degradation of organic matter, resulting in a decrease in dissolved oxygen, an increase in the concentration of inorganic nutrients, and an acceleration of the release of nitrogen and phosphorus pollutants in sediments, thereby aggravating the degree of eutrophication of water bodies.

[0083] The amount and intensity of rainfall directly affect surface runoff and groundwater levels, which in turn change the concentration of pollutants in water bodies. Specifically, heavy rain may wash away the surface, bringing pollutants such as pesticides and fertilizer residues in the soil and domestic garbage into the water body, causing water quality to deteriorate, especially nitrogen, phosphorus and other nutrient salt indicators to rise significantly, causing eutrophication of water bodies.

[0084] Dilution effect: Excessive rainwater will dilute the water body, causing water instability, reduced total alkalinity and total hardness, and reduced water quality stability, which can easily cause algae problems and water quality deterioration.

[0085] Short-term heavy rainfall: Thunderstorms are frequent in summer. Short-term heavy rainfall will cause a large amount of non-point source pollutants to enter the river with runoff. In addition, the overflow of the sewage network during rainy days will cause a large amount of pollutants to be input into the river in a short period of time. Under the high temperature conditions in summer, it is very easy to cause a rapid decrease in dissolved oxygen in the water, which in turn causes water quality deterioration and the death of fish and shrimp.

[0086] Although light conditions do not directly affect water quality, they can indirectly affect it. Specifically:

[0087] Promote algae growth: Light is an important condition for algae to carry out photosynthesis. Excessive light may promote excessive reproduction of algae, consume a large amount of dissolved oxygen, and cause water quality deterioration.

[0088] Impact on the health of aquatic organisms: Some aquatic organisms do not like strong light, and long-term exposure to light may affect their health, which in turn affects water quality.

[0089] Data on the impact of fertilizers and other pollutants on water quality:

[0090] When fertilizers and other pollutants enter water bodies, they will have a serious impact on water quality. Specifically:

[0091] Eutrophication of water bodies: Nutrients such as nitrogen and phosphorus in fertilizers enter water bodies with precipitation and soil moisture, leading to eutrophication of water bodies, triggering massive reproduction of algae, consuming dissolved oxygen, and deteriorating water quality.

[0092] Heavy metal pollution: Some fertilizers and pesticides contain heavy metal elements. Once these elements enter the water body, they will have a toxic effect on aquatic organisms and enter the human body through the food chain, endangering human health.

[0093] Organic pollution: When organic matter contained in industrial wastewater, domestic sewage, etc. enters the water body, it will consume dissolved oxygen, produce odor, and affect water quality.

[0094] When in use, when the water area of ​​the aquaculture area is large, the water quality will be affected by more environmental conditions. At this time, the scene condition data in the target water area can be collected by setting the detection points. The scene anomaly degree can be constructed based on the scene condition data under the premise of having a reference standard. , can be based on the scene abnormality Evaluate abnormal scene conditions in the target waters, predict and verify the impact of scene conditions on the target waters, and use it as feedback when abnormal scene conditions in the target waters are detected to adjust the water quality in the target waters in a timely manner to ensure the safety of the water quality in the target waters.

[0095] Step 103: Arrange a sensor network in the target water area, including a pH sensor, a dissolved oxygen sensor, a conductivity sensor, a turbidity sensor, a chemical sensor, a biosensor, a flow rate and a water level sensor, etc.; divide the target area into a plurality of sub-areas, and after receiving a water quality monitoring instruction, start the sensor network distributed in the sub-areas, and use the sensor network to collect water quality data in real time in the target water area, such as temperature, turbidity, pH value, dissolved oxygen, conductivity, chemical oxygen demand, biochemical oxygen demand, heavy metals and pollutants, etc., and generate a water quality data set after aggregation;

[0096] When using, combine the contents in steps 101 to 103:

[0097] As a further content, considering that the water body in the target area is large, in order to improve the processing efficiency, the target water area is divided into several sub-areas for separate processing, so that the water quality data in the sub-areas can be collected in real time by the sensor network, forming real-time monitoring and collection of the water quality in the target water area.

[0098] In order to enable abnormal water quality in the water area to be discovered in time, an alarm is usually issued to the outside when water quality parameters are abnormal. After receiving the alarm command, the management personnel operate and control the water treatment facilities in the water area to adjust the abnormal water quality. However, considering that the efficiency of manual control after the alarm is low, the existing water quality intervention control usually introduces water quality adaptive regulation to deal with the abnormal water quality in the water area. However, when the water environment is abnormal, for example, there is a large amount of rainfall, the ambient temperature has a large change, or there is a large amount of fertilizer or similar nutrient input, the water quality in the water area may change greatly. However, when the water area is large, the water quality in different water areas will still be quite different. However, the existing adaptive regulation cannot effectively screen out some areas that need to be focused on, and the output adaptive regulation scheme is also less targeted.

[0099] Step 2: Generate a water quality anomaly score based on the water quality data in the sub-area, determine the abnormal water quality status in the sub-area based on the water quality anomaly score, and then select the abnormal area, use the inspection drone to inspect each abnormal area, and after obtaining the corresponding inspection data, summarize and generate a water area inspection data set;

[0100] The step 2 includes the following contents:

[0101] Step 201: train a machine learning algorithm with the labeled sample data to obtain a trained abnormal water quality model; use the water quality data in the sub-region as input, use the trained abnormal water quality model to perform abnormal assessment on the water quality in the sub-region, and obtain a corresponding water quality abnormality score; if the water quality abnormality score exceeds expectations, take the corresponding time node as an abnormal node, and take the ratio of the water quality abnormality score exceeding expectations as the water quality abnormality ratio, and generate an abnormal state data set after aggregation;

[0102] When in use, the water quality in the sub-area is evaluated based on the completion of water quality data collection, and it is determined whether there is any abnormality in the water quality at the current time node. Moreover, based on the real-time update of water quality data, continuous abnormal monitoring of water quality in the target water area is completed.

[0103] Step 202: The distribution state of abnormal nodes in the abnormal state data set and the corresponding water quality abnormality ratio Generated water quality coefficient , as follows:

[0104] ;

[0105] Weight coefficient, , ,and ; The value is 3.142; is the number of abnormal nodes, No. i The water quality anomaly ratio of abnormal nodes, It is i The abnormal node to j The time interval of abnormal nodes, is the water quality monitoring cycle, ;

[0106] If the water quality coefficient If the water quality exceeds the preset threshold, it means that the water quality in the abnormal area is abnormal or has potential risks. If the water quality is abnormal, it will be regarded as an abnormal sub-area and the corresponding location information will be obtained. If the number of abnormal sub-areas exceeds expectations, a first-level alarm command will be issued to the outside.

[0107] When used, as a further content, based on the real-time abnormal monitoring of each abnormal water quality, the water quality coefficient is constructed from the abnormal monitoring data , based on the water quality coefficient Further comprehensive evaluation of the water quality status of the sub-areas is conducted, and the time scale is added based on the real-time water quality data to make the anomaly evaluation more reliable;

[0108] Step 203: construct an electronic map covering the target waters, mark the abnormal sub-areas on the electronic map, and after receiving the first-level alarm instruction, use the pre-trained clustering algorithm to cluster the abnormal sub-areas to obtain the clustered abnormal areas, so as to determine the areas that need to be monitored based on the water quality data; combine the location information of several abnormal areas in the target waters, and use the pre-trained path planning algorithm to plan the path of the inspection drone;

[0109] Use inspection drones to inspect each abnormal area according to the obtained inspection path, obtain corresponding inspection data, including infrared and multi-spectral images, water temperature anomaly data, pollution source data, plant and biodiversity data, etc., and summarize and generate a water area inspection data set;

[0110] When using, combine the contents in steps 201 to 203:

[0111] As a further content, after the abnormal sub-areas with persistent abnormal water quality have been screened out, considering that water quality is usually in a spreading state, the adjacent abnormal sub-areas are linked together through clustering processing, so that the coverage area when handling abnormal water quality is wider. By inspecting each abnormal area and collecting image data of the water body, secondary detection of water quality can be achieved to form a verification effect.

[0112] Step 3: Verify the water quality of the abnormal area with the inspection data obtained and generate the water pollution degree If the water pollution Do not exceed expectations. Predict water quality data in abnormal waters. If the predicted water quality changes exceed expectations, issue pollution analysis instructions to the outside world.

[0113] The step three includes the following contents:

[0114] Step 301: Using the water area inspection data as input, use the trained abnormal water quality model to perform abnormal analysis on the water quality in the abnormal area, obtain the corresponding water quality evaluation value, verify the water quality in the abnormal area with the obtained inspection data, and calculate the current water pollution degree by combining the water quality abnormality score and the water quality evaluation value. The generation method is as follows:

[0115] ;

[0116] in, For the i The sum of the water quality anomaly score and the water quality evaluation value in each abnormal area. is the number of abnormal regions, is the mean of the sum of the two, is the weight coefficient, ;

[0117] If the water pollution degree is obtained If the pollution threshold is exceeded, it means that the water quality in the abnormal area is highly abnormal and needs to be processed and responded to. At this time, a secondary alarm command is issued to the outside.

[0118] When used, as a further content, based on the secondary verification of water quality data, the water quality pollution degree is constructed from the two water quality data. , and once again realize the comprehensive evaluation of water quality. On the basis of obtaining the comprehensive evaluation, the alarm command is re-issued when the water quality in the abnormal area is abnormal. By repeatedly issuing the alarm command and even switching the alarm mode, the water quality in the target water area can be noticed in time when it is abnormal.

[0119] Step 302: Obtain a trained water quality prediction model from the labeled sample data convolutional neural network algorithm; after receiving the secondary alarm instruction, use the trained water quality prediction model to predict the water quality in the abnormal area and obtain the prediction data, and regenerate the water quality pollution degree from the prediction data. ; The water pollution degree at several consecutive time points Arrange along the time axis and analyze the water pollution degree The rate of change of water quality is used as the degree of change of water quality;

[0120] Based on historical data and the expected water quality management in the target water area, a water quality change threshold is pre-set; if the acquired water quality change exceeds the change threshold, it means that the water quality in the target water area has changed significantly, and the possible water quality deterioration rate exceeds the expected value, requiring immediate intervention. At this time, a pollution analysis instruction is issued to the outside;

[0121] When using, combine the contents in steps 301 and 302:

[0122] As a further content, before intervening in the water quality in the target water area, by predicting the water quality data and constructing the water quality change degree based on the historical data, the degree and speed of abnormal water quality changes in the target water area can be judged. If the current change rate and speed exceed expectations, it means that the water quality problems faced in the current stage are more serious, and the urgency of water quality treatment can be determined.

[0123] Step 4: Output the intervention strategy for the target water area from the water quality intervention knowledge graph and then execute the intervention strategy. Use the automatic control model to automatically control the operation of the water treatment facilities and collect the corresponding intervention feedback data. Adjust and optimize the intervention strategy based on the intervention feedback data.

[0124] The step 4 includes the following contents:

[0125] Step 401: after receiving the pollution analysis instruction, based on the scene condition data, water area inspection data and water quality data, a sensitivity analysis is performed on the water quality in the target area to obtain sensitive factors affecting the water quality of the target water area;

[0126] Taking water quality intervention of aquaculture water bodies as the target word, a pre-built water quality intervention knowledge graph is constructed after deep retrieval and entity relationship construction; according to the correspondence between sensitive factors and water quality intervention strategies, the water quality intervention knowledge graph matches the corresponding intervention strategies for the water quality changes in the target waters;

[0127] When used, as a further content, by performing sensitivity analysis or similar principal component analysis on abnormal water quality and related data, the corresponding key factors can be determined, and the corresponding intervention strategy can be matched according to the key factors or sensitivity factors, and the current abnormal water quality can be targeted according to the intervention strategy;

[0128] Step 402: After training the machine learning algorithm with the labeled sample data, obtain the trained automatic control model; execute the water quality intervention strategy for the target water area, use the trained automatic control model to automatically control the operation status of the water treatment facilities, such as aeration facilities, oxygenation facilities and water quality filtration facilities, etc., and perform dosing, aeration and oxygenation, and water quality filtration in the target water area; collect corresponding feedback data after a preset observation period, such as various water quality change data, and generate an intervention feedback data set after aggregating the feedback data;

[0129] As a further content, when it is necessary to carry out targeted treatment of the water quality in the target water area, after determining the time node for water quality treatment and the area to be treated, the water treatment facilities are automatically controlled after the intervention strategy is automatically executed. On the basis of real-time monitoring of the water quality data in the target water area, when the water quality in the target water area is abnormal, automatic and adaptive adjustment control is realized to prevent the water quality in the target water area from deteriorating;

[0130] Step 403: Generate intervention coefficients from feedback data in the intervention feedback data set , where the water pollution degree before and after adjustment of each abnormal area is obtained , the water pollution degree in the abnormal area Generate intervention coefficients , as follows:

[0131] ;

[0132] in, and Before and after intervention i The water quality coefficient on the abnormal area, and is the corresponding mean, i is the abnormal area number, , is the number of abnormal regions, For the i The intervention coefficient of the abnormal area, is the corresponding mean;

[0133] If the intervention coefficient is obtained If the pre-set intervention threshold is not exceeded, it means that the current intervention strategy has failed to achieve the expected intervention effect and needs to be adjusted and optimized. With the improvement of the intervention coefficient as the optimization goal, the pre-trained genetic algorithm is used to adjust and iteratively optimize the intervention strategy in real time to obtain the optimized intervention strategy.

[0134] When used, combine the contents in steps 401 to 403:

[0135] As a further step, based on the automatic control of water quality in the target water area, the water quality data changes are tracked in real time, and the obtained water quality data is used as feedback to construct the intervention coefficient. , according to the intervention coefficient Judging the effectiveness of current water quality interventions will facilitate dynamic adjustment and optimization of intervention strategies and better targeted automated regulation of water quality.

[0136] Step 5: Generate feedback coefficients from water quality control feedback data , if the feedback coefficient The current water quality regulation process is optimized without exceeding expectations, and the water quality monitoring report is sent to the data receiving end at the time node that meets the constraint conditions;

[0137] The step five includes the following contents:

[0138] Step 501: After executing the iteratively optimized water quality intervention strategy, periodically collect the control feedback data of water quality regulation, including response speed, reliability and cost-effectiveness, and generate a control response data set after aggregation; under dimensionless conditions, generate feedback coefficients from the control feedback data in the control response data set , as follows:

[0139] ;

[0140] in, For the i Response speed during water quality intervention, is the corresponding response cost, , is the number of water quality interventions, weight coefficient: , ,and ;

[0141] If the feedback coefficient If it does not exceed expectations, it means that the comprehensive control feedback effect of the current water quality regulation process has not met expectations. For example, the response speed may be slow and the cost is high, and it is necessary to continuously optimize the sensor network deployment and the frequency of inspection and water quality intervention;

[0142] When in use, after completing the automatic control and adjustment of water quality in the target area, the feedback data of the control and adjustment is collected and the feedback coefficient is constructed. , with the feedback coefficient Evaluate the effectiveness of current water quality interventions and regulation. If the water quality is relatively low, the current intervention methods and facilities can be adjusted to improve the efficiency of subsequent automatic water quality regulation, and the intervention costs can also be continuously revised.

[0143] Step 502: Generate a water quality monitoring report based on water quality data, inspection data, and intervention feedback data, and visualize the data analysis results. Constraints are imposed, and water quality monitoring reports are sent to the data receiving end at time nodes that meet the constraints. The constraints are as follows:

[0144] ;

[0145] Where: is the number of water quality interventions in the target waters, For the i The abnormality of the scene during the water quality intervention, is the corresponding mean, For time node and The difference in water pollution between is the expected value of the water pollution difference, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0. The weight coefficient is , ,and , the value is consistent with the previous value;

[0146] When using, combine the contents in steps 501 and 502:

[0147] As a further aspect, by periodically generating water quality monitoring reports and adapting the monitoring reports to the water quality status and current scenario conditions, managers can timely understand the water quality status in the target water area when environmental conditions and water quality conditions change, so as to facilitate intervention and adjustment.

[0148] The construction method of the knowledge graph of aquaculture water quality intervention is a complex but systematic process, which involves multiple steps and techniques. The following is a detailed construction method:

[0149] First, we need to clarify the goal of building a knowledge graph, which is to provide a structured, easy-to-query and understand knowledge base for understanding the relevant knowledge of water quality intervention in aquaculture. Structured data: including formatted data such as water quality monitoring reports, aquaculture logs, and expert advice. Semi-structured data: such as relevant research reports, abstracts and conclusions of academic papers, these data contain certain structures but also need further processing.

[0150] Unstructured data: including text data such as farming experience sharing, forum discussions, and expert interview records. These data need to be converted into structured data through steps such as information extraction and entity recognition. Data cleaning: remove duplicate, erroneous, and irrelevant data to ensure data accuracy and consistency. Entity recognition: use named entity recognition (NER) technology to extract key entities such as aquaculture, water quality indicators, and intervention measures from text data. Relationship extraction: extract the relationship between entities through semantic analysis, such as "pH value affects the growth of farmed fish" and "dissolved oxygen is too low and oxygen needs to be increased".

[0151] Define the model layer: Based on the collected data, define the model layer of the knowledge graph, including entities, attributes, and relationships. For example, you can define "water quality indicators" as an entity category, whose attributes include "name", "normal value range", "influencing factors", etc., and relationships include "influence", "regulatory measures", etc. Fill the data layer: Based on the preprocessed data, fill the data layer of the knowledge graph. Store entities, attributes, and relationships in the form of triples, such as "(pH value, normal value range, 6.5~8.5)" "(dissolved oxygen, influencing factors, high temperature, excessive breeding density)" and so on.

[0152] Knowledge fusion: Integrate data from different sources, deal with information redundancy and inconsistency, and form a unified and coherent knowledge system. Ontology construction: Further improve the ontology library of the knowledge graph, define more detailed entity categories, attributes and relationships, as well as the constraints and rules between them. Quality assessment: Conduct quality assessment on the constructed knowledge graph, including data accuracy, completeness, consistency and availability.

[0153] Iterative optimization: Based on the evaluation results, the knowledge graph is iteratively optimized to continuously improve its quality and usability. Visual display: Using visualization tools, the knowledge graph is displayed in a graphical way, which is convenient for users to intuitively understand and query. Interactive function: Add interactive functions such as query, search, and recommendation, so that users can easily obtain the required knowledge and information.

[0154] See also Figure 2 The present invention provides a water quality adaptive control system based on multi-point detection, comprising:

[0155] The scene analysis unit collects scene condition data in the target water area and generates the scene abnormality degree in the target water area from the scene condition data. , if the scene abnormality Exceeding expectations, the sensor network collects water quality data in real time in each sub-area of ​​the target water area, and aggregates them to generate a water quality data set;

[0156] The water quality analysis unit generates a water quality anomaly score from the water quality data in the sub-area, determines the abnormal water quality status in the sub-area based on the water quality anomaly score, and then selects the abnormal area, uses the inspection drone to inspect each abnormal area, and after obtaining the corresponding inspection data, summarizes and generates a water area inspection data set;

[0157] The water quality prediction unit verifies the water quality of the abnormal area based on the acquired inspection data and generates the water pollution degree If the water pollution Do not exceed expectations. Predict water quality data in abnormal waters. If the predicted water quality changes exceed expectations, issue pollution analysis instructions to the outside world.

[0158] The water quality intervention adjustment unit outputs the intervention strategy for the target water area based on the water quality intervention knowledge graph and then executes the intervention strategy. It uses the automatic control model to automatically control the operation of the water treatment facilities and collects the corresponding intervention feedback data, and adjusts and optimizes the intervention strategy based on the intervention feedback data.

[0159] Feedback optimization unit, which generates feedback coefficients from control feedback data of water quality regulation , if the feedback coefficient Without exceeding expectations, the current water quality regulation process is optimized, and the water quality monitoring report is sent to the data receiving end at the time node that meets the constraints.

[0160] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0161] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0162] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0163] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0164] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A water quality adaptive control method based on multi-point detection, characterized in that: include, Collect scene condition data in the target water area, and generate scene abnormality degree in the target water area from the scene condition data , if the scene abnormality Exceeding expectations, the sensor network collects water quality data in real time in each sub-area of ​​the target water area, and aggregates them to generate a water quality data set; Generate water quality anomaly scores from water quality data in sub-areas, determine water quality anomaly status in sub-areas based on water quality anomaly scores, and then screen out abnormal areas. Use inspection drones to inspect each abnormal area, and after obtaining corresponding inspection data, summarize and generate a water area inspection data set. The water quality of the abnormal area is verified by the acquired inspection data to generate the water pollution degree If the water pollution Do not exceed expectations. Predict water quality data in abnormal waters. If the predicted water quality changes exceed expectations, issue pollution analysis instructions to the outside world. The water quality intervention knowledge graph outputs the intervention strategy for the target water area and then executes the intervention strategy. The automatic control model is used to automatically control the operation of the water treatment facilities and collect the corresponding intervention feedback data. The intervention strategy is adjusted and optimized based on the intervention feedback data. Generate feedback coefficients from control feedback data for water quality regulation , if the feedback coefficient Without exceeding expectations, the current water quality regulation process is optimized, and the generated water quality monitoring report is sent to the data receiving end at the time node that meets the constraints.

2. The method for adaptively controlling water quality based on multi-point detection according to claim 1 is characterized in that: Several monitoring points are set up in the target area, and the environmental scene data in the target water area is collected at the monitoring points. The scene condition data set is aggregated to generate the scene anomaly degree in the target water area. ; If the scene is abnormal When the abnormal threshold is exceeded, a water quality monitoring instruction is issued to the outside, a sensor network is deployed in the target water area, and the target area is divided into several sub-areas. After receiving the water quality monitoring instruction, the sensor network distributed in the sub-areas is started.

3. The method for adaptively controlling water quality based on multi-point detection according to claim 2 is characterized in that: Taking the water quality data in the sub-region as input, the trained abnormal water quality model is used to evaluate the water quality in the sub-region abnormally and obtain the corresponding water quality abnormality score; If the water quality abnormality score exceeds expectations, the corresponding time node is taken as the abnormal node, and the proportion of the water quality abnormality score exceeding expectations is taken as the water quality abnormality ratio, which is summarized to generate an abnormal state data set; Generating water quality coefficients from abnormal state data sets If the water quality coefficient If the water quality exceeds the preset threshold, it will be regarded as an abnormal sub-area and the corresponding location information will be obtained. If the number of abnormal sub-areas exceeds expectations, a first-level alarm command will be issued to the outside.

4. The method for adaptively controlling water quality based on multi-point detection according to claim 3 is characterized in that: The distribution state of abnormal nodes in the abnormal state data set and the corresponding water quality abnormality ratio Generated water quality coefficient , as follows: ； Weight coefficient, , ,and ; The value is 3.142; is the number of abnormal nodes, No. i The water quality anomaly ratio of abnormal nodes, It is i The abnormal node to the j The time interval of abnormal nodes, is the water quality monitoring cycle, .

5. The method for adaptively controlling water quality based on multi-point detection according to claim 4 is characterized in that: Mark the abnormal sub-area on the electronic map, and after receiving the first-level alarm instruction, use the pre-trained clustering algorithm to cluster the abnormal sub-area to obtain the clustered abnormal area; Combined with the location information of several abnormal areas in the target waters, the pre-trained path planning algorithm plans the path for the inspection drone; Use inspection drones to inspect each abnormal area according to the obtained inspection path, obtain corresponding inspection data, and summarize and generate a water area inspection data set.

6. The method for adaptively controlling water quality based on multi-point detection according to claim 5 is characterized in that: Taking water area inspection data as input, the trained abnormal water quality model is used to analyze the water quality in the abnormal area, obtain the corresponding water quality evaluation value, and calculate the current water pollution degree. ; If water pollution If the water pollution threshold is exceeded, a secondary alarm command will be issued to the outside. The generation method is as follows: ； in, For the i The sum of the water quality anomaly score and the water quality evaluation value in each abnormal area. is the number of abnormal regions, is the mean of the sum of the two, is the weight coefficient, .

7. The method for adaptively controlling water quality based on multi-point detection according to claim 6 is characterized in that: After receiving the secondary alarm command, the trained water quality prediction model is used to predict the water quality in the abnormal area and obtain the prediction data, and the water quality pollution degree is regenerated from the prediction data. ; The water pollution degree at several consecutive time points Arrange along the time axis and analyze the water pollution degree The speed of change is calculated and used as the water quality change degree; if the obtained water quality change degree exceeds the change threshold, a pollution analysis instruction is issued to the outside.

8. The method for adaptively controlling water quality based on multi-point detection according to claim 7 is characterized in that: After receiving the pollution analysis instruction, based on the scene condition data, water area inspection data and water quality data, a sensitivity analysis is performed on the water quality in the target area to obtain the sensitive factors that affect the water quality of the target water area; Taking water quality intervention of aquaculture water bodies as the target word, a water quality intervention knowledge graph is pre-built, and the water quality intervention knowledge graph matches the corresponding intervention strategy for the target water area based on sensitive factors.

9. The method for adaptively controlling water quality based on multi-point detection according to claim 8, characterized in that: Execute water quality intervention strategies for target waters, collect corresponding feedback data after a preset observation period, and aggregate the feedback data to generate an intervention feedback data set; Generate intervention coefficients from feedback data in the intervention feedback data set , several pre-coefficients Without exceeding the preset intervention threshold, the optimization goal is to improve the intervention coefficient, and the pre-trained genetic algorithm is used to adjust and iteratively optimize the intervention strategy in real time to obtain the optimized intervention strategy.

10. The method for adaptively controlling water quality based on multi-point detection according to claim 9, characterized in that: Obtain the water pollution degree before and after adjustment of each abnormal area , the water pollution degree in the abnormal area Generate intervention coefficients , as follows: ； in, and Before and after intervention i The water quality coefficient on the abnormal area, and is the corresponding mean, i is the abnormal area number, , is the number of abnormal regions, For the i The intervention coefficient of the abnormal area, is the corresponding mean.

11. The method for adaptively controlling water quality based on multi-point detection according to claim 10, characterized in that: After executing the iteratively optimized water quality intervention strategy, the control feedback data of water quality regulation is collected periodically, and the feedback coefficient is generated from the control feedback data. , if the feedback coefficient Optimize the sensor network deployment and the frequency of inspections and water quality interventions without exceeding expectations; generate feedback coefficients The way is as follows: ； in, For the i Response speed during water quality intervention, is the corresponding response cost, , is the number of water quality interventions, weight coefficient: , ,and .

12. The method for adaptively controlling water quality based on multi-point detection according to claim 11, characterized in that: Generate water quality monitoring reports from water quality data, inspection data, and intervention feedback data, and visualize data analysis results; Output interval of water quality monitoring report Constraints are as follows: ； Where: is the number of water quality interventions in the target waters, For the i The abnormality of the scene during the water quality intervention, is the corresponding mean, For time node and The difference in water pollution between is the expected value of the water pollution difference, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0. The weight coefficient is , ,and .

13. The water quality adaptive control system based on multi-point detection is characterized by: include, The scene analysis unit collects scene condition data in the target water area and generates the scene abnormality degree in the target water area from the scene condition data. , if the scene abnormality Exceeding expectations, the sensor network collects water quality data in real time in each sub-area of ​​the target water area, and aggregates them to generate a water quality data set; The water quality analysis unit generates a water quality anomaly score from the water quality data in the sub-area, determines the abnormal water quality status in the sub-area based on the water quality anomaly score, and then selects the abnormal area, uses the inspection drone to inspect each abnormal area, and after obtaining the corresponding inspection data, summarizes and generates a water area inspection data set; The water quality prediction unit verifies the water quality of the abnormal area based on the acquired inspection data and generates the water pollution degree If the water pollution Do not exceed expectations. Predict water quality data in abnormal waters. If the predicted water quality changes exceed expectations, issue pollution analysis instructions to the outside world. The water quality intervention adjustment unit outputs the intervention strategy for the target water area based on the water quality intervention knowledge graph and then executes the intervention strategy. It uses the automatic control model to automatically control the operation of the water treatment facilities and collects the corresponding intervention feedback data, and adjusts and optimizes the intervention strategy based on the intervention feedback data. Feedback optimization unit, which generates feedback coefficients from control feedback data of water quality regulation , if the feedback coefficient Without exceeding expectations, the current water quality regulation process is optimized, and the water quality monitoring report is sent to the data receiving end at the time node that meets the constraints.

Images