Shallow lake water quality pollution treatment system and method

Abstract

The invention discloses a shallow lake water quality pollution treatment system and method, and relates to the technical field of water pollution, and the method comprises the steps: S1, building a disturbance intensity index DIX and ammonia nitrogen concentration change curve, achieving the quantitative recognition of bottom mud disturbance and pollution release behaviors, training a rebound risk prediction model according to the quantitative recognition, and outputting a rebound risk prediction coefficient Ft; and the capability of pre-judging the rebound trend of the ammonia nitrogen pollution is effectively improved. In the step S2, the adjusted aeration response frequency BAF is dynamically generated based on the rebound risk prediction coefficient Ft value, so that the aeration behavior is converted into a risk-driven system from a traditional timing system, the energy consumption and disturbance intensity are further reduced, and the treatment efficiency is improved. And S3, further introducing a governance residual coefficient GRC, constructing a complete feedback closed loop, and realizing accurate evaluation and secondary strategy optimization of the governance effect of each round, thereby avoiding the problem of effect back-off or excessive disturbance after governance.

Description

Technical Field

[0001] The present invention relates to the technical field of water pollution, and specifically to a water quality pollution control system and method for shallow lakes. Background Art

[0002] As an important part of urban wetlands, water source regulation, and regional ecological stability, shallow lakes have relatively shallow water depths, slow hydrodynamic forces, and strong pollutant deposition and retention capabilities. It is easy to have interactive effects of ammonia nitrogen accumulation and sediment release, resulting in prominent problems such as water quality deterioration, eutrophication, and treatment rebound. To improve the sustainability and stability of ammonia nitrogen pollution control in shallow lakes, it is urgent to construct an intelligent control mechanism that can dynamically respond to disturbance and pollution rebound trends.

[0003] Existing treatment methods for water quality pollution in shallow lakes mainly rely on artificial intervention, timed aeration, or fixed-point dosing. However, these methods generally rely on fixed regulation cycles and lack the ability to predict pollution rebound trends. Especially under the action of factors such as sediment disturbance and dissolved oxygen fluctuation after treatment, it often causes the "treatment rebound" phenomenon of a secondary increase in ammonia nitrogen concentration, which is not conducive to the effective treatment of water quality pollution in shallow lakes. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a water quality pollution control system and method for shallow lakes, which solves the problems in the above background art.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A method for treating water quality pollution in a shallow lake, including the following steps, S1: Use underwater monitoring equipment to continuously obtain the operation characteristics of the sediment in the shallow lake, and based on the operation characteristics, evaluate the disturbance intensity and its potential risk of pollutant release to obtain the disturbance intensity index DIX, and construct a curve of ammonia nitrogen concentration change before and after ammonia nitrogen treatment, extract the dynamic behavior of the rebound trend from it, and train a risk prediction model through the operation characteristics and the ammonia nitrogen concentration change curve to output the rebound risk prediction coefficient Ft; S2: Based on the value of the rebound risk prediction coefficient Ft, perform dynamic regulation to obtain the adjusted aeration response frequency BAF, and intervene in the aeration in the shallow lake according to the adjusted aeration response frequency BAF; S3: Continuously monitor the curve of ammonia nitrogen concentration change after intervening in the aeration in the shallow lake to provide feedback for the effect analysis of the dynamic regulation in S2, and execute the adjustment mechanism again to achieve the treatment of water quality pollution in the shallow lake.

[0006] Preferably, the specific steps of S1 include: S11: Install several sets of monitoring equipment in a shallow lake in advance to monitor and obtain the operation characteristics of the sediment in the shallow lake in real time. Among them, the operation characteristics include the number of disturbance events during the monitoring period in the sediment of the shallow lake , the ammonia nitrogen release intensity of the sediment and the minimum dissolved oxygen in the bottom water layer ; S12: Based on the operation characteristics, after dimensionless processing of the information in the operation characteristics, evaluate the disturbance intensity and its risk to potential pollutant release, so as to calculate and obtain the disturbance intensity index DIX. The disturbance intensity index DIX is specifically obtained through the following calculation method: ; In the formula, represents the number of disturbance events during the monitoring period, represents the ammonia nitrogen release intensity of the sediment, represents the minimum dissolved oxygen in the bottom water layer, represents a tiny constant.

[0007] Preferably, the specific steps of S1 further include: S13: Obtain the potential data in the water body in real time through an ammonia nitrogen ion-selective electrode sensor, and based on the Nernst equation, realize the conversion of the ammonia nitrogen concentration value, continuously sample and convert the ammonia nitrogen concentration sequence to form an ammonia nitrogen concentration change curve in the time dimension. Among them, the potential data refers to the electrode potential value E of the corresponding sampling period. The Nernst equation is specifically as follows: ; In the formula, is the standard electrode potential, is the gas constant, T is the temperature, is the ion charge number, F is the Faraday constant, is the ammonium ion concentration in the water body; S14: Extract the time point of each ammonia nitrogen treatment from the operation characteristics and mark it in the ammonia nitrogen concentration change curve to form an ammonia nitrogen concentration change curve before and after ammonia nitrogen treatment. Extract the dynamic behavior of the rebound trend from the ammonia nitrogen concentration change curve before and after ammonia nitrogen treatment. Among them, the dynamic behavior includes the start date of each treatment (reference benchmark) and the ammonium ion concentration in the water body every day after the corresponding start date of treatment ; ; S15: According to the dynamic behavior, analyze and calculate the daily growth rate Rn of the ammonium ion concentration in the water body after the corresponding treatment. Specifically: ; In the formula, is the daily growth rate of the ammonium ion concentration in the water body at the t-th moment after treatment, is the ammonium ion concentration in the water body at the t-th moment after the corresponding start date of treatment and is the ammonium ion concentration in the water body at the start date of the corresponding treatment and is the time number, is the start date of the corresponding treatment.

[0008] Preferably, the specific steps of S1 further include: S16: By using the operation characteristics and the ammonia nitrogen concentration change curve as the input information of the convolutional neural network model, and dividing the input information into a training set and a validation set after dimensionless processing to train and validate the convolutional neural network model, and using the trained convolutional neural network model as the risk prediction model, the rebound risk prediction coefficient Ft is output from the output end of the risk prediction model.

[0009] Preferably, the specific steps of S2 include: S21: Preset a risk threshold, and judge whether it is necessary to dynamically regulate the current shallow lake by numerically comparing the risk threshold with the rebound risk prediction coefficient Ft. The specific content is as follows: If the rebound risk prediction coefficient Ft exceeds the risk threshold, it indicates that there is an abnormal rebound situation in the current shallow lake after ammonia nitrogen treatment, and at this time, the first adjustment mechanism will be triggered; If the rebound risk prediction coefficient Ft does not exceed the risk threshold, it indicates that there is no abnormal rebound situation in the current shallow lake after ammonia nitrogen treatment, and at this time, the first adjustment mechanism will not be triggered temporarily.

[0010] Preferably, the specific steps of S2 further include: S22: Start the first adjustment mechanism, and dynamically adjust the frequency during aeration based on the rebound risk prediction coefficient Ft to obtain the adjusted aeration response frequency BAF. The adjusted aeration response frequency BAF is specifically obtained by the following formula: ; In the formula, is the basic aeration frequency, is the response amplification coefficient, is the rebound risk prediction coefficient at the t-th moment after treatment, is the hyperbolic tangent function, and is used to ensure smooth regulation response; S23: Conduct ammonia nitrogen treatment on the current shallow lake according to the adjusted aeration response frequency BAF.

[0011] Preferably, the specific steps of S3 include: S31: On the basis of S23, monitor and obtain in real time the change curve of ammonia nitrogen concentration after the aeration in the shallow lake is intervened, so as to extract the ammonium ion concentration in the water body after the aeration in the shallow lake is initially intervened from the change curve of ammonia nitrogen concentration after the aeration in the shallow lake is intervened. And based on the change curve of ammonia nitrogen concentration after the aeration in the shallow lake is intervened, re-execute the content from S13 to S23 to determine again the adjusted aeration response frequency BAF in S22. And on the basis of determining again the adjusted aeration response frequency BAF in S22, update the change curve of ammonia nitrogen concentration again, and extract the ammonium ion concentration in the water body after the aeration in the shallow lake is intervened again from it. Mark the ammonium ion concentration in the water body after the aeration in the shallow lake is intervened again as the concentration after the second intervention. Mark the ammonium ion concentration in the water body after the aeration in the shallow lake is initially intervened as the concentration after the first intervention. .

[0012] Preferably, the specific steps of S3 further include: S32: According to the concentration after the first intervention obtained in S31 and the concentration after the second intervention, give feedback on the effect analysis of the dynamic regulation of the shallow lake to obtain the governance residual coefficient GRC for two times, and it is specifically obtained through the following calculation method: ; ; In the formula, and are respectively the governance residual coefficient after the aeration in the shallow lake is initially intervened and the governance residual coefficient after the aeration in the shallow lake is intervened again, is the target concentration threshold set for governance; S33: Repeat the content of S31 and S32 to obtain the governance residual coefficient GRC for several times, and sort the governance residual coefficient GRC for several times in ascending order to obtain a numerical sequence, and extract the change curves of ammonia nitrogen concentration after intervention corresponding to the first two groups of governance residual coefficient GRC from the numerical sequence.

[0013] Preferably, the specific steps of S3 further include: S34: Determine the intervention numbers of the first two groups of governance residual coefficient GRC according to the ammonia nitrogen concentration change curves after intervention corresponding to the first two groups of governance residual coefficient GRC, and determine the governance residual coefficient GRC and the difference in governance residual coefficient after intervention on the aeration in the shallow lake according to the two groups of intervention numbers. ; S35: Trigger the second adjustment mechanism for the ammonia nitrogen concentration change curves after intervention corresponding to the first two groups of governance residual coefficient GRC to achieve the treatment of water pollution in the shallow lake. Specifically: ; In the formula, is the aeration response intensity after the xth intervention, is the reference value of the aeration intensity, is the governance residual coefficient after the xth intervention, is the difference in governance residual coefficient after the xth intervention, x is the intervention number, and are respectively the governance residual coefficient after the xth intervention and the difference in governance residual coefficient after the xth intervention weights; S36: Based on the content of S35, determine the aeration response intensity after the first two groups of interventions, and according to the aeration response intensity after the first two groups of interventions, respectively perform intensity interventions on the current shallow lake. By re-monitoring and obtaining the ammonia nitrogen concentration change curves after intervention on the aeration in the shallow lake, determine the governance residual coefficient GRC after the first two groups of intensity interventions. By comparing the governance residual coefficient GRC after the first two groups of intensity interventions, select the aeration response intensity corresponding to the governance residual coefficient GRC with the smallest value as the target intervention to carry out the treatment of water pollution in the shallow lake.

[0014] A water pollution treatment system for shallow lakes includes a risk analysis module, a first intervention module and a second intervention module; The risk analysis module is used to use underwater monitoring equipment to obtain the operation characteristics of the bottom mud of the shallow lake in real time, and according to the operation characteristics, evaluate the disturbance intensity and its risk of potential pollutant release to obtain the disturbance intensity index DIX, and construct the ammonia nitrogen concentration change curves before and after ammonia nitrogen treatment, extract the dynamic behavior of the rebound trend from them, and train the risk prediction model through the operation characteristics and ammonia nitrogen concentration change curves to output the rebound risk prediction coefficient Ft; The first intervention module will perform dynamic regulation based on the value of the rebound risk prediction coefficient Ft to obtain the adjusted aeration response frequency BAF, and intervene in the aeration in the shallow lake according to the adjusted aeration response frequency BAF; The second intervention module is used to monitor in real time the change curve of ammonia nitrogen concentration after intervening in the aeration in the shallow lake, so as to provide feedback on the analysis of the effect of dynamic regulation, and execute the adjustment mechanism again to achieve the treatment of water pollution in the shallow lake.

[0015] The present invention provides a system and method for treating water pollution in a shallow lake, which has the following beneficial effects: (1) By constructing the disturbance intensity index DIX and the ammonia nitrogen concentration change curve in step S1, the quantification and identification of the sediment disturbance and pollution release behavior are realized, and the rebound risk prediction model is trained based on this, and the rebound risk prediction coefficient Ft is output, effectively improving the ability to predict the ammonia nitrogen pollution rebound trend. In step S2, based on the value of the rebound risk prediction coefficient Ft, the adjusted aeration response frequency BAF is dynamically generated, so that the aeration behavior is changed from the traditional timing system to the risk-driven system, further reducing the energy consumption and disturbance intensity, and improving the treatment efficiency. Step S3 further introduces the treatment residual coefficient GRC to construct a complete feedback closed loop, realizing the accurate evaluation of the treatment effect of each round and the secondary strategy optimization, so as to avoid the problems of the regression of the treatment effect or excessive disturbance after treatment. The overall method realizes the whole-process closed-loop control from disturbance identification, risk prediction, response intervention to effect evaluation, has the comprehensive treatment ability of feedforward prediction, adaptive regulation and multi-round optimization, and is applicable to the ammonia nitrogen pollution control of multiple types of shallow lakes.

[0016] (2) By using the sediment disturbance characteristics and the ammonia nitrogen concentration change curve as the model inputs in S16, and combining multi-dimensional parameters such as the historical ammonia nitrogen fluctuation coefficient, the disturbance intensity index DIX, and the ammonium ion concentration growth rate Rn, the trained neural network model can output the core index reflecting the pollution rebound intensity, and this prediction coefficient can discover the potential risks after ammonia nitrogen treatment in advance. In S21, by comparing the rebound risk prediction coefficient Ft with the set risk threshold in real time, it can automatically judge whether there is a rebound anomaly at present, and accurately trigger the first adjustment mechanism, realizing the transformation of pollution control from post-intervention to active prediction response.

[0017] (3) By marking the concentrations after the first and second interventions and continuously correcting the value of the aeration response frequency BAF based on the updated risk assessment results, the system can continuously iteratively optimize the treatment plan and gradually approach the optimal control path. This method effectively avoids the problems of over-aeration disturbance or response lag, improves the energy efficiency ratio and timeliness of the treatment behavior, and realizes the precise, dynamic and closed-loop feedback optimization control of pollution control.

[0018] (4) In S34 to S35, the system constructs an aeration response intensity function based on two sets of residual coefficients and their change differences, adaptively adjusts the aeration power by comprehensively considering the deviation degree and change trend, realizes the in-depth expansion from frequency regulation to intensity regulation. Finally, in S36, by comparing multiple sets of intervention intensities and response effects, the aeration intensity corresponding to the minimum governance residual is selected as the final control strategy to achieve the target intervention for shallow lakes. This method establishes a closed-loop self-learning control mechanism of "intensity - effect - feedback - optimization" in multiple rounds of interventions, which can effectively avoid the response deviation problem caused by a single control strategy and achieve stable control and precise governance input in high-frequency disturbance scenarios. Description of the Drawings

[0019] Figure 1 It is a schematic flow chart of a method for treating water pollution in shallow lakes according to the present invention; Figure 2 It is an overall logic diagram of a method for treating water pollution in shallow lakes according to the present invention; Figure 3 It is a partial logic diagram of a method for treating water pollution in shallow lakes according to the present invention; Figure 4 It is a block diagram of a system for treating water pollution in shallow lakes according to the present invention. Detailed Embodiments

[0020] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0021] Embodiment 1 Please refer to Figures 1 to 3 , the present invention provides a method for treating water pollution in shallow lakes, including the following steps. S1: Use underwater monitoring equipment to continuously obtain the operation characteristics of the bottom mud of the shallow lake, and based on the operation characteristics, evaluate the disturbance intensity and its potential risk of pollutant (such as ammonia nitrogen) release to obtain the disturbance intensity index DIX, and construct the ammonia nitrogen concentration change curve before and after ammonia nitrogen treatment, extract the dynamic behavior of the rebound trend from it, and train a risk prediction model through the operation characteristics and ammonia nitrogen concentration change curve to output the rebound risk prediction coefficient Ft; S2: Based on the value of the rebound risk prediction coefficient Ft, perform dynamic regulation to obtain the adjusted aeration response frequency BAF, and intervene in the aeration in the shallow lake according to the adjusted aeration response frequency BAF. S3: Real-time monitoring of the ammonia nitrogen concentration curve after intervention in aeration in shallow lakes to provide feedback on the effect analysis of dynamic regulation in S2, and to re-execute the regulation mechanism to achieve water pollution control in shallow lakes.

[0022] In this embodiment, by constructing a management process centered on disturbance behavior identification, ammonia nitrogen concentration trend extraction, risk prediction, and dynamic aeration regulation, the system realizes a fully closed-loop control mechanism of feedforward prediction, active response, and feedback optimization for ammonia nitrogen pollution in shallow lakes.

[0023] On the one hand, the disturbance intensity index DIX is used to quantify the potential risk of sediment disturbance to ammonia nitrogen release, which improves the ability to identify endogenous pollution release behavior; on the other hand, based on the real-time monitoring of ammonia nitrogen concentration change curve, the rebound trend before and after treatment is extracted, and the risk prediction coefficient Ft generated by training is used to guide the dynamic adjustment of aeration frequency BAF, thereby avoiding the governance rebound, energy waste and ecological disturbance risks caused by "delayed response" in traditional governance methods. For example, in the practice of a certain lake governance, the system monitored frequent sediment disturbances, low dissolved oxygen, and drastic fluctuations in ammonia nitrogen concentration, and the DIX value rose rapidly; at this time, the prediction model output Ft exceeded the threshold, and the system automatically increased the aeration frequency from 1 time / day to 2.4 times / day for emergency response. Subsequently, it was observed that the ammonia nitrogen concentration stabilized and decreased, and the governance residual coefficient GRC was significantly reduced, which effectively suppressed the secondary rebound of ammonia nitrogen. This method realizes "aeration on demand, quantitative control, and real-time correction", improves governance efficiency, reduces operating costs, and enhances the system's adaptability to complex pollution dynamics.

[0024] Among them, aeration is an intervention method that suppresses the anaerobic environment of the sediment through oxygenation and reduces the release of ammonia nitrogen and phosphorus.

[0025] Example 2 Please refer to Figure 1 , specifically: S1 specific steps include: S11: Install several sets of monitoring equipment in shallow lakes in advance to monitor and obtain the operating characteristics of shallow lake sediments in real time, where the operating characteristics include the number of disturbance events in the shallow lake sediments during the monitoring period. , release intensity of ammonia nitrogen from sediment and the minimum dissolved oxygen in bottom water ; S12: Based on the operation characteristics, after dimensionless processing of the information in the operation characteristics, the disturbance intensity and its risk of potential release of pollutants (such as ammonia nitrogen) are evaluated to calculate the disturbance intensity index DIX. The disturbance intensity index DIX is specifically obtained by the following calculation method: ; In the formula, Indicates the number of disturbance events during the monitoring period, including wind waves, aquatic animals, and mechanical agitation, Indicates the release intensity of sediment ammonia nitrogen, Indicates the minimum dissolved oxygen in the bottom water layer, Indicates a tiny constant used to prevent division-by-zero errors; Is a logarithmic function; the log function can compress extreme values, giving the indicator a continuous, controllable, and smoothly varying response structure. When the pollution release intensity increases rapidly, not using the log function may result in overly large output values; The number of disturbance events during the monitoring period Is used to reflect the frequency of sediment disturbance. The higher its value, the more frequent the disturbance and the easier it is to promote pollutant release, which is obtained through monitoring with an acoustic bottom sediment disturbance probe; The release intensity of sediment ammonia nitrogen Is used to measure the release intensity of sediment pollutants. The larger its value, the higher the pollutant load in the sediment and the more likely it is to cause rebound. It uses ammonia nitrogen ion-selective electrode sensors deployed at two depth positions on the sediment surface and the overlying water body to obtain the ammonia nitrogen concentration difference between the two layers in real time, and combines the effective diffusion coefficient and the sampling point spacing, and uses an estimation model based on Fick's diffusion law for derivation and calculation to obtain the ammonia nitrogen release rate from the sediment to the overlying water body per unit time; The minimum dissolved oxygen in the bottom water layer Affects the sediment release behavior. The lower its value, the easier it is to release (therefore it appears in the denominator position), which is obtained through monitoring with an optical dissolved oxygen sensor; Through , if the release is strong and the oxygen content is low, it is easier to release ammonia nitrogen (high risk of rebound); Represents the activity level of the disturbance behavior, which is a combined logic similar to "trigger probability × risk level" (similar to the idea of constructing risk factors in environmental risk calculations); The disturbance intensity index DIX is used to represent the disturbance intensity and its potential risk to pollutant release. The higher the value of the disturbance intensity index DIX, the greater the risk of sediment disturbance and pollution release, and the higher the risk of endogenous pollution re-release.

[0026] The specific steps of S1 also include: S13: Real-time obtain the potential data in the water body through an ammonia nitrogen ion-selective electrode sensor, and based on the Nernst equation, achieve the conversion of ammonia nitrogen concentration values, continuously sample and convert the ammonia nitrogen concentration sequence to form an ammonia nitrogen concentration change curve in the time dimension. Among them, the potential data refers to the electrode potential value E of the corresponding sampling period. The Nernst equation is as follows: ; In the formula, Is the standard electrode potential, where \(R\) is the gas constant and \(T\) is the temperature, where \(z\) is the ionic charge number and \(F\) is the Faraday constant, and \([NH_4^+]\) is the concentration of ammonium ions in the water body; since the response of the ion-selective electrode to ions is "logarithmic-linear" (for every 10-fold change in concentration, the potential increases by approximately 59 mV), logarithmic conversion must be performed.

[0027] S14: Extract the time points of each ammonia nitrogen treatment from the operation characteristics and mark them in the ammonia nitrogen concentration change curve to form the ammonia nitrogen concentration change curve before and after the ammonia nitrogen treatment. Extract the dynamic behavior of the rebound trend from the ammonia nitrogen concentration change curve before and after the ammonia nitrogen treatment, where the dynamic behavior includes the start date of each treatment (reference benchmark) and the concentration of ammonium ions in the water body every day after the corresponding start date of the treatment ; S15: According to the dynamic behavior, analyze and calculate the daily growth rate \(R_n\) of the ammonium ion concentration in the water body after the corresponding treatment, specifically: ; In the formula, is the daily growth rate of the ammonium ion concentration in the water body at the \(t\)th moment after the treatment, is the concentration of ammonium ions in the water body at the \(t\)th moment after the corresponding start date of the treatment ; is the concentration of ammonium ions in the water body at the corresponding start date of the treatment ; is the time number, and \(t_0\) is the corresponding start date of the treatment.

[0028] In this embodiment, the present invention constructs a pre-treatment modeling mechanism for pollution control based on disturbance behavior recognition and concentration trend tracking by finely decomposing step S1. First, through step S12, underwater monitoring equipment is deployed to obtain the characteristics of sediment disturbance, and the disturbance intensity index DIX is calculated by introducing three factors: the number of disturbance events, the ammonia nitrogen release rate, and the minimum dissolved oxygen, so as to quantify the risk of endogenous pollution release caused by disturbance. Secondly, in S13, the ammonia nitrogen ion-selective electrode sensor is used to collect potential data, which is converted into a continuous concentration sequence through the Nernst equation to form a concentration change curve before and after the treatment. Then, the dynamic behavior of the rebound trend is extracted through treatment marking to accurately identify the pollution rebound cycle and response window. Finally, in S15, the daily growth rate \(R_n\) is calculated to form a dynamic rebound trend index chain, providing a direct quantitative basis for risk prediction and control decision-making.

[0029] For example, on a certain day, the first treatment was initiated in a certain lake area. Through the monitoring from S13 to S15, it was found that since the start date, the ammonia nitrogen concentration has been rising at a rate of 0.22 mg / L per day, and the DIX value is as high as 2.91. The system immediately outputs a high-risk warning, indicating the initiation of aeration strengthening intervention to further avoid the possible event of excessive ammonia nitrogen concentration in the later stage. Thus, it can be seen that this method has the ability to identify and respond to the rebound trend early, significantly improving the foresight and initiative of the treatment.

[0030] Example 3 Please refer to Figure 1 , specifically: The specific steps of S1 also include: S16: By taking the operation characteristics and the ammonia nitrogen concentration change curve as the input information of the convolutional neural network model, and dividing the input information into a training set and a validation set after dimensionless processing to train and validate the convolutional neural network model, and taking the trained convolutional neural network model as the risk prediction model, the rebound risk prediction coefficient Ft is output from the output end of the risk prediction model. The rebound risk prediction coefficient Ft is obtained through the following calculation method: ; In the formula, is the rebound risk prediction coefficient at the t-th moment after treatment, is the historical ammonia nitrogen fluctuation coefficient (standard deviation, indicating the system stability), , and respectively represent the disturbance intensity index DIX, the daily growth rate of ammonium ion concentration in the water body at the t-th moment after treatment and the historical ammonia nitrogen fluctuation coefficient weighting coefficients, where the weighting coefficients can be calculated and obtained through the risk prediction model; The way to obtain the rebound risk prediction coefficient Ft is to judge whether there is a rebound phenomenon of ammonia nitrogen at a future moment through multi-dimensional fusion. The output rebound risk prediction coefficient Ft is the core control index of the whole system.

[0031] Through the aforementioned S1 steps, a prediction basic model for the pollution "rebound trend" is constructed.

[0032] The specific steps of S2 include: S21: Preset a risk threshold, and judge whether it is necessary to dynamically regulate the current shallow lake by numerically comparing the risk threshold with the rebound risk prediction coefficient Ft. The specific content is as follows: If the rebound risk prediction coefficient Ft exceeds the risk threshold, it indicates that there is an abnormal rebound in the current shallow lake after ammonia nitrogen treatment. At this time, the first adjustment mechanism will be triggered; When the rebound risk prediction coefficient Ft does not exceed the risk threshold, it indicates that there is no abnormal rebound in the current shallow lake after ammonia nitrogen treatment. At this time, the first adjustment mechanism will not be triggered temporarily.

[0033] In this embodiment, in step S16, the system takes the bottom mud disturbance operation characteristics and the ammonia nitrogen concentration change curve as inputs, constructs multi-dimensional training data, and through dimensionless processing and sample division, realizes the effective training and verification of the CNN model. The rebound risk prediction coefficient Ft output by the model comprehensively integrates key indicators such as the disturbance intensity DIX, the ammonia nitrogen concentration growth rate Rn, and the historical fluctuation coefficient, and can dynamically evaluate the pollution rebound risk.

[0034] Subsequently, by comparing Ft with the risk threshold set by the system, when the rebound risk prediction coefficient Ft exceeds the threshold, the first adjustment mechanism is automatically triggered to perform precise aeration control intervention to avoid pollution diffusion. For example, if the ammonia nitrogen concentration in a certain lake area does not increase significantly on the 3rd day after treatment, but due to the CNN model identifying high disturbance frequency, low oxygen level and large volatility, the calculated rebound risk prediction coefficient Ft = 0.82 is higher than the threshold of 0.75, that is, an early aeration response is triggered, thus avoiding the possible rebound peak on the 5th day. This mechanism has the advantages of "predicting first and regulating actively", significantly improving the stability and forward-looking of the treatment effect.

[0035] By constructing a dynamic prediction and response control process for ammonia nitrogen rebound in shallow lakes, the trend risk of "ammonia nitrogen rising again" is identified in advance in the later stage of treatment, and the intensity of the linkage regulation mechanism is corrected to prevent the failure of the treatment rebound.

[0036] Example 4 Please refer to Figure 1 , specifically: The specific steps of S2 also include: S22: Start the first adjustment mechanism, and based on the rebound risk prediction coefficient Ft, dynamically adjust the frequency during aeration to obtain the adjusted aeration response frequency BAF. The adjusted aeration response frequency BAF is specifically obtained through the following formula: ; In the formula, is the basic aeration frequency (system default initial value), is the response amplification coefficient, used to control the adjustment amplitude, is the rebound risk prediction coefficient at the t-th moment after treatment, is the hyperbolic tangent function, and using is to ensure smooth regulation response and suppress extreme amplification; S23: Carry out ammonia nitrogen treatment on the current shallow lake according to the adjusted aeration response frequency BAF.

[0037] The specific steps of S3 include: S31: Based on S23, continuously monitor and obtain the change curve of ammonia nitrogen concentration after intervening in the aeration in the shallow lake, so as to extract the ammonium ion concentration in the water body after initially intervening in the aeration in the shallow lake from the change curve of ammonia nitrogen concentration after intervening in the aeration in the shallow lake , and based on the change curve of ammonia nitrogen concentration after intervening in the aeration in the shallow lake, re - execute the content from S13 to S23 to determine the adjusted aeration response frequency BAF in S22 again. On the basis of re - determining the adjusted aeration response frequency BAF in S22, update the ammonia nitrogen concentration change curve again, and extract the ammonium ion concentration in the water body after intervening in the aeration in the shallow lake again , and mark the ammonium ion concentration in the water body after intervening in the aeration in the shallow lake again as the concentration after the second intervention , and mark the ammonium ion concentration in the water body after initially intervening in the aeration in the shallow lake as the concentration after the first intervention .

[0038] Through the aforementioned S2 step, it is used to quantify the "potential rebound risk" faced by the current shallow lake and serve as the input basis for subsequent regulation.

[0039] In this embodiment, the present invention realizes the adaptive adjustment and response optimization of the aeration frequency by introducing a dynamic regulation mechanism based on the rebound risk prediction coefficient Ft, enhancing the control flexibility and treatment closed - loop for ammonia nitrogen pollution rebound.

[0040] In S22, the system calls the hyperbolic tangent function for smooth amplification processing according to the Ft value to avoid sudden changes in the aeration frequency caused by abnormal disturbances, thereby obtaining the adjusted aeration response frequency BAF, and applying it to the actual aeration treatment process through S23, effectively improving the treatment accuracy.

[0041] S31 further constructs a cyclic feedback mechanism to continuously obtain the change curve of ammonia nitrogen concentration after aeration intervention, recalculate the aeration response frequency BAF and update the regulation strategy in real - time, forming a closed - loop structure of "monitoring - adjustment - feedback - readjustment". This mechanism significantly improves the timeliness and intelligence of pollution treatment, avoids over - aeration disturbance and resource waste, and is applicable to multi - round ammonia nitrogen treatment in complex dynamic lake environments.

[0042] Embodiment 5 Please refer to Figure 1 and Figure 3 , specifically: The specific steps of S3 also include: S32: According to the concentration after the first intervention obtained in S31 And the concentration after the second intervention , feedback on the effect analysis of the dynamic regulation of shallow lakes is carried out to obtain the governance residual coefficients GRC for the two times, which are specifically obtained through the following calculation methods: ; ; In the formula, and are the governance residual coefficients after the initial intervention of aeration in the shallow lake and the governance residual coefficients after the re - intervention of aeration in the shallow lake respectively, is the target concentration threshold set for governance (such as the national surface water standard Class I ≤ 1.0 mg / L); The larger the governance residual coefficient GRC, the more serious the deviation of the current governance effect from the target; S33: Repeat the content of S31 and S32 to obtain several governance residual coefficients GRC, and arrange the several governance residual coefficients GRC in ascending order to obtain a numerical sequence. Extract the ammonia nitrogen concentration change curves after intervention corresponding to the first two groups of governance residual coefficients GRC from the numerical sequence.

[0043] The specific steps of S3 also include: S34: According to the ammonia nitrogen concentration change curves after intervention corresponding to the first two groups of governance residual coefficients GRC, determine the intervention numbers (i.e., the number of times of intervention) of the first two groups of governance residual coefficients GRC, and determine the governance residual coefficients GRC and the difference in governance residual coefficients after the intervention of aeration in the shallow lake under the two groups of intervention numbers ; S35: Trigger the second adjustment mechanism for the ammonia nitrogen concentration change curves after intervention corresponding to the first two groups of governance residual coefficients GRC to achieve the water pollution control of the shallow lake. Specifically: ; In the formula, is the aeration response intensity after the x - th intervention, is the aeration intensity reference value, is the governance residual coefficient after the x - th intervention, is the difference in governance residual coefficients after the x - th intervention, x is the intervention number, and are the governance residual coefficients after the x - th intervention and the difference in governance residual coefficients after the x - th intervention respectively, and are used to adjust the response sensitivity (empirically set or learned and optimized); Among them, the difference in governance residual coefficients Refers to the difference between the governance residual coefficient after the x-th intervention and the governance residual coefficient after the previous intervention. The difference in governance residual coefficients Indicates the trend of deviation change, that is, whether the deviation is increasing or decreasing, and is used to adjust the response acceleration trend. When only considering the difference in governance residual coefficients If the current deviation is already large but the change is very small ( ≈0), it may cause misjudgment of system stability and stop governance. If the difference in governance residual coefficients itself has strong volatility and is easily affected by measurement interference, resulting in unstable or oscillating control; The governance residual coefficient GRC is used to adjust the basic amplitude of the response. When only considering the governance residual coefficient GRC, the system can only see the magnitude of the deviation and cannot perceive whether the deviation is improving, which will cause over-control (energy consumption waste), and the system cannot distinguish the potential out-of-control situation where the deviation is increasing, resulting in delayed response and missing the best governance window.

[0044] S36: Based on the content of S35, determine the aeration response intensity of the first two groups after intervention and, according to the aeration response intensity of the first two groups after intervention respectively perform intensity intervention on the current shallow lake. By re-monitoring and obtaining the ammonia nitrogen concentration change curve after intervening in the aeration in the shallow lake, determine the governance residual coefficient GRC of the first two groups after intensity intervention. By comparing the governance residual coefficient GRC of the first two groups after intensity intervention, select the aeration response intensity corresponding to the smallest governance residual coefficient GRC as the target intervention for water pollution treatment of the shallow lake.

[0045] Through the aforementioned S3 steps, it is used to determine whether each round of control behavior is effective or lagging, provide a basis for closed-loop feedback, and construct a "result identifier" for the system governance performance, promoting the continuous optimization of the governance path and the intelligent evolution of strategies.

[0046] In this embodiment, the present invention constructs an iterative intensity intervention system for ammonia nitrogen pollution treatment in shallow lakes by introducing a multi-round feedback analysis and response intensity optimization mechanism based on the governance residual coefficient GRC, greatly improving the governance accuracy and decision-making optimization efficiency.

[0047] In steps S32 - S36, the system extracts the ammonia nitrogen concentration change curves under different intervention rounds respectively, calculates the corresponding governance residual coefficient GRC, and constructs an ascending sequence based on the deviation degree to determine the intervention numbers with relatively better response effects in history.

[0048] Furthermore, by comparing the residual value with the residual difference, a second adjustment mechanism is introduced to calculate and apply a differential aeration response intensity, achieving targeted intensity adjustment. After multiple rounds of intervention, by comparing the new round of treatment residual coefficient GRC, the response intensity corresponding to the minimum deviation is finally selected as the target value to ensure the self-optimization and convergence of the system. In summary, this method further improves the consistency of the treatment effect and the target convergence ability, effectively avoiding problems such as the out-of-control response amplitude and rigid strategies in traditional control, and demonstrating the evolution and self-adjustment ability of intelligent pollution treatment.

[0049] Example 6 Please refer to Figure 4 , specifically: A water pollution treatment system for a shallow lake, including a risk analysis module, a first intervention module, and a second intervention module; The risk analysis module is used to use underwater monitoring equipment to obtain the operation characteristics of the bottom mud of the shallow lake in real time, and based on the operation characteristics, evaluate the disturbance intensity and its potential risk of pollutant release to obtain the disturbance intensity index DIX, and construct a curve of the ammonia nitrogen concentration change before and after ammonia nitrogen treatment, extract the dynamic behavior of the rebound trend from it, and train a risk prediction model through the operation characteristics and the ammonia nitrogen concentration change curve to output a rebound risk prediction coefficient Ft; The first intervention module will perform dynamic regulation based on the value of the rebound risk prediction coefficient Ft to obtain the adjusted aeration response frequency BAF, and intervene in the aeration in the shallow lake according to the adjusted aeration response frequency BAF; The second intervention module is used to monitor the curve of the ammonia nitrogen concentration change in real time after intervening in the aeration in the shallow lake to provide feedback on the analysis of the effect of dynamic regulation, and execute the adjustment mechanism again to achieve the water pollution treatment of the shallow lake.

[0050] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for treating water pollution in shallow lakes, characterized in that: It includes the following steps: S1: Use underwater monitoring equipment to obtain the operation characteristics of the sediment in a shallow lake in real time. Based on the operation characteristics, evaluate the disturbance intensity and its potential risk of pollutant release to obtain the disturbance intensity index DIX. Construct the ammonia nitrogen concentration change curve before and after ammonia nitrogen treatment, extract the dynamic behavior of the rebound trend from it, and train a risk prediction model through the operation characteristics and the ammonia nitrogen concentration change curve to output the rebound risk prediction coefficient Ft; S2: Based on the value of the rebound risk prediction coefficient Ft, perform dynamic regulation to obtain the adjusted aeration response frequency BAF, and intervene in the aeration in the shallow lake according to the adjusted aeration response frequency BAF; S3: Real-time monitor the ammonia nitrogen concentration change curve after intervening in the aeration in the shallow lake to provide feedback for the effect analysis of the dynamic regulation in S2, and execute the adjustment mechanism again to achieve the treatment of water pollution in the shallow lake.

2. A method for treating water pollution in a shallow lake according to claim 1, characterized in that: The specific steps of S1 include: S11: Install several groups of monitoring devices in a shallow lake in advance to monitor and obtain the operation characteristics of the sediment in the shallow lake in real time. Among them, the operation characteristics include the number of disturbance events during the monitoring period in the sediment of the shallow lake, the ammonia nitrogen release intensity of the sediment, and the minimum dissolved oxygen in the bottom water body. ; S12: Based on the operation characteristics, after dimensionless processing of the information in the operation characteristics, evaluate the disturbance intensity and its potential risk of pollutant release to calculate and obtain the disturbance intensity index DIX. The disturbance intensity index DIX is specifically obtained through the following calculation method: ； In the formula, represents the number of disturbance events during the monitoring period, represents the release intensity of sediment ammonia nitrogen, represents the minimum dissolved oxygen in the bottom water layer, represents a tiny constant.

3. A method for treating water pollution in a shallow lake according to claim 2, characterized in that: The specific steps of S1 also include: S13: Use an ammonia nitrogen ion selective electrode sensor to obtain the potential data in the water body in real time, and based on the Nernst equation, realize the conversion of the ammonia nitrogen concentration value. Continuously sample and convert the ammonia nitrogen concentration sequence to form an ammonia nitrogen concentration change curve in the time dimension. Among them, the potential data refers to the electrode potential value E in the corresponding sampling period. The Nernst equation is specifically as follows: ； In the formula, is the standard electrode potential, is the gas constant, T is the temperature, is the ionic charge number, F is the Faraday constant, is the ammonium ion concentration in water; S14: Extract the time points of each ammonia nitrogen treatment from the operation characteristics and mark them on the ammonia nitrogen concentration change curve to form the ammonia nitrogen concentration change curve before and after the ammonia nitrogen treatment. Extract the dynamic behavior of the rebound trend from the ammonia nitrogen concentration change curve before and after the ammonia nitrogen treatment, where the dynamic behavior includes the start date of each treatment and the concentration of ammonium ions in the water body every day after the corresponding start date of the treatment ; ; S15: According to the dynamic behavior, analyze and calculate the daily growth rate Rn of the ammonium ion concentration in the water body after treatment, specifically: ； In the formula, is the daily growth rate of the ammonium ion concentration in the water body at the t-th moment after treatment, is the ammonium ion concentration in the water body at the t-th moment after the corresponding start date of treatment , is the ammonium ion concentration in the water body at the corresponding start date of treatment , is the time number, is the corresponding start date of treatment.

4. A method for treating water pollution in a shallow lake according to claim 3, characterized in that: The specific steps of S1 also include: S16: Use the operation characteristics and the ammonia nitrogen concentration change curve as the input information of the convolutional neural network model. After dimensionless processing of the input information, divide it into a training set and a validation set to train and validate the convolutional neural network model, and use the trained convolutional neural network model as the risk prediction model to output the rebound risk prediction coefficient Ft from the output end of the risk prediction model.

5. A method for treating water pollution in a shallow lake according to claim 4, characterized in that: The specific steps of S2 include: S21: Preset a risk threshold, and compare the risk threshold with the rebound risk prediction coefficient Ft numerically to judge whether it is necessary to perform dynamic regulation on the current shallow lake. The specific content is as follows: If the rebound risk prediction coefficient Ft exceeds the risk threshold, it indicates that there is an abnormal rebound situation in the current shallow lake after ammonia nitrogen treatment. At this time, the first adjustment mechanism will be triggered; When the rebound risk prediction coefficient Ft does not exceed the risk threshold, it indicates that there is no abnormal rebound in the current shallow lake after ammonia nitrogen treatment. At this time, the first adjustment mechanism will not be triggered temporarily.

6. A method for treating water pollution in a shallow lake according to claim 5, characterized in that: The specific steps of S2 further include: S22: Start the first adjustment mechanism, and based on the rebound risk prediction coefficient Ft, dynamically adjust the frequency during aeration to obtain the adjusted aeration response frequency BAF. The adjusted aeration response frequency BAF is specifically obtained by the following formula: ； In the formula, is the basic aeration frequency, is the response amplification coefficient, is the rebound risk prediction coefficient at the t-th moment after treatment, is the hyperbolic tangent function, and is used to ensure the smoothness of the regulation response; S23: Treat the current shallow lake for ammonia nitrogen according to the adjusted aeration response frequency BAF.

7. A method for treating water pollution in a shallow lake according to claim 6, characterized in that: The specific steps of S3 include: S31: Based on S23, monitor and obtain in real time the change curve of ammonia nitrogen concentration after the aeration in the shallow lake is intervened, so as to extract the ammonium ion concentration in the water body after the aeration in the shallow lake is initially intervened from the change curve of ammonia nitrogen concentration after the aeration in the shallow lake is intervened , and based on the change curve of ammonia nitrogen concentration after the aeration in the shallow lake is intervened, re-execute the content from S13 to S23 to determine again the adjusted aeration response frequency BAF in S22. On the basis of determining again the adjusted aeration response frequency BAF in S22, update the change curve of ammonia nitrogen concentration again, and extract the ammonium ion concentration in the water body after the aeration in the shallow lake is intervened again from it , and take the ammonium ion concentration in the water body after the aeration in the shallow lake is intervened again and mark it as the concentration after the second intervention , and take the ammonium ion concentration in the water body after the aeration in the shallow lake is initially intervened and mark it as the concentration after the first intervention .

8. A method for treating water pollution in a shallow lake according to claim 7, characterized in that: The specific steps of S3 further include: S32: According to the concentration after the first intervention obtained in S31 and the concentration after the second intervention , feedback is carried out on the effect analysis of the dynamic regulation of the shallow lake to obtain the governance residual coefficients GRC for the two times, and the specific calculation method is as follows: ； ； In the formula, and are the treatment residual coefficients after the initial intervention on the aeration in the shallow lake and the treatment residual coefficients after the subsequent intervention on the aeration in the shallow lake, respectively, is the target concentration threshold set for the treatment; S33: Repeat the content of S31 and S32 to obtain several treatment residual coefficients GRC, and arrange the several treatment residual coefficients GRC in ascending order to obtain a numerical sequence. Extract the ammonia nitrogen concentration change curves after intervention corresponding to the first two groups of treatment residual coefficients GRC from the numerical sequence.

9. A method for treating water pollution in a shallow lake according to claim 8, characterized in that: The specific steps of S3 further include: S34: Determine the intervention numbers corresponding to the ammonia nitrogen concentration change curves after intervention for the first two groups of governance residual coefficient GRCs, and based on the two groups of intervention numbers, determine the governance residual coefficient GRC and the difference in governance residual coefficient after the aeration in the shallow lake is intervened under the two groups of intervention numbers ; S35: Trigger the second adjustment mechanism for the ammonia nitrogen concentration change curves after intervention corresponding to the first two groups of treatment residual coefficients GRC to achieve the treatment of water pollution in the shallow lake. Specifically: ； In the formula, is the aeration response intensity after the x-th intervention, is the reference value of the aeration intensity, is the treatment residual coefficient after the x-th intervention, is the difference in the treatment residual coefficient after the x-th intervention, where x is the intervention number, and are respectively the treatment residual coefficient after the x-th intervention and the weight of the difference in the treatment residual coefficient after the x-th intervention; S36: Based on the content of S35, determine the aeration response intensity of the first two groups after the intervention , and based on the aeration response intensity of the first two groups after the intervention , perform intensity interventions on the current shallow lake respectively. By monitoring again and obtaining the ammonia nitrogen concentration change curve after the aeration in the shallow lake is intervened, determine the governance residual coefficient GRC of the first two groups after the intensity intervention. By comparing the governance residual coefficient GRC of the first two groups after the intensity intervention, select the aeration response intensity corresponding to the governance residual coefficient GRC with the smallest value as the target intervention to carry out the water pollution treatment of the shallow lake.

10. A water pollution treatment system for shallow lakes, which is used to implement the water pollution treatment method for shallow lakes described in any one of the above claims 1 to 9, and is characterized in that: It includes a risk analysis module, a first intervention module, and a second intervention module; The risk analysis module is used to use underwater monitoring equipment to continuously obtain the operation characteristics of the bottom mud of the shallow lake, and according to the operation characteristics, evaluate the disturbance intensity and its risk of potential pollutant release to obtain the disturbance intensity index DIX, and construct the ammonia nitrogen concentration change curves before and after ammonia nitrogen treatment, extract the dynamic behavior of the rebound trend from them, and train a risk prediction model through the operation characteristics and ammonia nitrogen concentration change curves to output the rebound risk prediction coefficient Ft; The first intervention module will perform dynamic regulation based on the value of the rebound risk prediction coefficient Ft to obtain the adjusted aeration response frequency BAF, and intervene in the aeration in the shallow lake according to the adjusted aeration response frequency BAF; The second intervention module is used to continuously monitor the ammonia nitrogen concentration change curve after intervening in the aeration in the shallow lake to feedback the analysis of the effect of dynamic regulation, and execute the adjustment mechanism again to achieve the treatment of water pollution in the shallow lake.

Images