Aeration system and aeration method for sewage treatment

Abstract

The application provides an aeration system and method for sewage treatment. The aeration system comprises a detection module for detecting water quality parameters in an oxidation ditch; a processing module for obtaining device parameters of an aeration execution mechanism, fine-tuning the device parameters according to a plurality of preset constraint conditions, and generating a plurality of candidate device parameter sets; the processing module is further configured to generate predicted effluent water quality parameters corresponding to each candidate device parameter set according to the water quality parameters and the plurality of candidate device parameter sets through a pre-trained prediction model; the processing module is further configured to input each candidate device parameter set and the corresponding predicted effluent water quality parameters into a preset objective function to obtain a corresponding objective function value; all candidate device parameter sets are screened based on the objective function values, and the screened candidate device parameter sets are used as control sequences to control the operation of the aeration execution mechanism. The application can realize fine operation management of the aeration process.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment technology, and in particular to an aeration system and aeration method for wastewater treatment. Background Technology

[0002] In the process of wastewater treatment, the aeration actuators (blowers, aerators and their control circuits) are the core to ensure the stable operation of the aerobic biochemical unit.

[0003] Traditional aeration control is mostly feedback control, which adjusts the aeration system based on currently detected water quality parameters. This approach is essentially a reactive measure; increasing aeration only after water quality deterioration is detected often already negatively impacts treatment effectiveness. Due to the large time delays and nonlinear characteristics of wastewater treatment processes, this lagging control strategy struggles to cope with rapid fluctuations in influent water quality, easily leading to unstable effluent quality. Therefore, improvements are needed. Summary of the Invention

[0004] This invention provides an aeration system and aeration method for wastewater treatment to solve the technical problem that existing control strategies with lag cannot cope with rapid fluctuations in influent water quality.

[0005] The present invention also proposes an aeration system for wastewater treatment, comprising:

[0006] The aeration actuator is installed inside the oxidation ditch of the sewage treatment plant;

[0007] The detection module is used to detect the influent water quality parameters of wastewater at multiple locations within the oxidation ditch;

[0008] The processing module is used to acquire the equipment parameters of the aeration actuator, fine-tune the equipment parameters according to multiple preset constraints, and generate multiple candidate equipment parameter sets; wherein, each candidate equipment parameter set includes multiple equipment parameters generated by multiple fine-tuning under the same constraints.

[0009] The processing module is further configured to generate predicted effluent water quality parameters corresponding to each candidate device parameter set based on the influent water quality parameters and multiple candidate device parameter sets using a pre-trained prediction model; wherein, the prediction model is an LSTM prediction model;

[0010] The processing module is further configured to input each candidate device parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value; filter all candidate device parameter sets based on the objective function value, and use the filtered candidate device parameter sets as a control sequence to control the operation of the aeration actuator according to the control sequence.

[0011] In one embodiment of the present invention, the processing module is further configured to:

[0012] After obtaining the influent water quality parameters and equipment parameters, the moving mean and standard deviation of the influent water quality parameters and equipment parameters are calculated respectively. Based on the moving mean and standard deviation, abnormal values ​​in the influent water quality parameters and equipment parameters are screened out and corrected.

[0013] The processing module is also used to fine-tune the corrected equipment parameters according to multiple preset constraints, generate multiple candidate equipment parameter sets, and generate predicted effluent water quality parameters corresponding to each candidate equipment parameter set through a pre-trained prediction model based on the corrected influent water quality parameters and multiple candidate equipment parameter sets.

[0014] In one embodiment of the present invention, the processing module includes:

[0015] The mutual information selection unit is used to calculate the mutual information value between each sub-parameter of the equipment parameters and the preset target effluent water quality parameter, select the sub-parameters of the equipment whose mutual information value is greater than the preset mutual information threshold, and retain the unselected sub-parameters of the equipment.

[0016] The correlation selection unit is used to select several equipment sub-parameters from the selected equipment sub-parameters based on the correlation degree, to obtain the equipment sub-parameters that need to be fine-tuned, and to retain the equipment sub-parameters that were not selected.

[0017] The generation unit is used to perform multi-step fine-tuning of the equipment sub-parameters that need to be fine-tuned for each of the constraints, and to form a corresponding candidate equipment parameter set by combining the fine-tuned equipment sub-parameters and the retained equipment sub-parameters; wherein the constraints include at least equipment parameter boundary constraints, equipment start-stop frequency constraints, and process stability constraints.

[0018] In one embodiment of the present invention, the correlation selection unit is further configured to:

[0019] The device sub-parameters selected by the mutual information selection unit are randomly grouped to obtain multiple device sub-parameter groups; wherein, different device sub-parameter groups include different device sub-parameters;

[0020] For each equipment sub-parameter group, calculate the correlation between each equipment sub-parameter and the target effluent water quality parameter, and calculate the average correlation of all obtained correlations. Calculate the similarity between any two equipment sub-parameters, and calculate the average similarity of all obtained similarities. Also, calculate the difference between the average correlation and the average similarity.

[0021] Select the equipment sub-parameter group with the largest difference between the average relevance and the average similarity, and confirm that all equipment sub-parameters in this equipment sub-parameter group are equipment sub-parameters that need to be fine-tuned.

[0022] In one embodiment of the present invention, the processing module includes:

[0023] The objective function calculation unit is used to input each candidate device parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value;

[0024] The filtering unit is used to sort all objective function values ​​and filter out a preset number of candidate device parameter sets in ascending order of value, and select one candidate device parameter set from the preset number of candidate device parameter sets as the control sequence.

[0025] A control unit is used to control the operation of the aeration actuator according to a control sequence.

[0026] In one embodiment of the present invention, the filtering unit includes:

[0027] The compliance value calculation subunit is used to calculate the corresponding compliance value based on the predicted effluent water quality parameters corresponding to each set of candidate equipment parameters and the preset emission standard limits.

[0028] The offset calculation subunit is used to calculate the corresponding offset value based on the set of parameters of each selected candidate device and the preset set of baseline control parameters.

[0029] The adjustment variable calculation subunit is used to calculate the corresponding number of adjustment variables based on the selected set of parameters for each candidate device and the preset set of baseline control parameters.

[0030] The energy consumption change calculation subunit is used to calculate the corresponding energy consumption change value based on the parameter set of each selected candidate device and the preset baseline control parameter set.

[0031] The first selection subunit is used to calculate the comprehensive score based on the compliance value, offset value, number of adjustment variables and energy consumption change value of each selected candidate device parameter set, and select the candidate device parameter set corresponding to the maximum comprehensive score as the control sequence.

[0032] In one embodiment of the present invention, the filtering unit includes:

[0033] The contribution calculation unit is used to compare the change of each sub-parameter of the selected candidate equipment parameters with the change of the predicted effluent water quality parameters, based on the selected candidate equipment parameter set and its corresponding predicted effluent water quality parameters, to obtain the contribution of each sub-parameter of the equipment to the predicted effluent water quality parameters.

[0034] The second selection sub-unit is used to identify the device sub-parameter with the greatest contribution, and select the set of candidate device parameters with the smallest change in the device sub-parameter from all the selected candidate device parameter sets as the control sequence.

[0035] In one embodiment of the present invention, the processing module adjusts the equipment parameters of the aeration actuator according to a preset adjustment step size based on a control sequence, so as to control the operation of the aeration actuator.

[0036] In one embodiment of the present invention, the training method of the prediction model includes:

[0037] Multiple training samples are obtained, including influent water quality parameters, equipment parameters, and effluent water quality parameters, wherein the effluent water quality parameters are sample labels corresponding to the influent water quality parameters and equipment parameters;

[0038] All training samples are input into the prediction model to be trained to obtain the corresponding prediction results;

[0039] Based on the sample labels of all training samples and the corresponding prediction results, calculate the loss value of the loss function of the prediction model;

[0040] Based on the loss value, the parameters in the prediction model to be trained are adjusted to obtain the pre-trained prediction model.

[0041] This invention proposes an aeration method for wastewater treatment, comprising:

[0042] The influent water quality parameters of the wastewater at multiple locations within the oxidation ditch were detected.

[0043] The equipment parameters of the aeration actuator are obtained, and the equipment parameters are fine-tuned according to multiple preset constraints to generate multiple candidate equipment parameter sets; wherein, each candidate equipment parameter set includes multiple equipment parameters generated by multiple fine-tuning under the same constraints.

[0044] Based on the influent water quality parameters and multiple candidate equipment parameter sets, the predicted effluent water quality parameters corresponding to each candidate equipment parameter set are generated through a pre-trained prediction model; wherein, the prediction model is an LSTM prediction model;

[0045] Each candidate device parameter set and its corresponding predicted effluent water quality parameters are input into a preset objective function to obtain the corresponding objective function value; based on the objective function value, all candidate device parameter sets are filtered, and the filtered candidate device parameter sets are used as control sequences to control the operation of the aeration actuator according to the control sequences.

[0046] The beneficial effects of this invention are as follows: This invention proposes an aeration system and method for wastewater treatment. A detection module detects influent water quality parameters at multiple locations within the oxidation ditch. A processing module fine-tunes the equipment parameters based on multiple preset constraints, generating multiple candidate equipment parameter sets for proactive intervention in the operation of the aeration actuator during wastewater treatment. Each candidate equipment parameter set includes multiple equipment parameters generated through multi-step fine-tuning under the same constraints. The processing module inputs the influent water quality parameters and multiple candidate equipment parameter sets into a pre-trained prediction model to generate predicted effluent water quality parameters corresponding to each candidate equipment parameter set. The processing module further inputs each candidate equipment parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value. Based on the objective function value, all candidate equipment parameter sets are filtered, and the filtered candidate equipment parameter sets are used as control sequences. The operation of the aeration actuator is controlled according to the control sequences, thereby achieving refined operation management of the aeration process. Attached Figure Description

[0047] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0048] In the attached diagram:

[0049] Figure 1 This is a structural block diagram of an oxidation ditch in a wastewater treatment plant provided in an embodiment of the present invention.

[0050] Figure 2 This is a structural block diagram of a sludge pumping station in a wastewater treatment plant, provided according to an embodiment of the present invention.

[0051] Figure 3 This is a structural block diagram of an aeration system for wastewater treatment provided in an embodiment of the present invention.

[0052] Figure 4 This is a structural block diagram of a control unit in an aeration system provided in an embodiment of the present invention.

[0053] Figure 5 This is another structural block diagram of the control unit in the aeration system provided in an embodiment of the present invention.

[0054] Figure 6 This is a schematic flowchart of an aeration method for wastewater treatment provided in an embodiment of the present invention.

[0055] The attached figures are labeled as follows:

[0056] 10. Detection module; 20. Processing module; 30. Aeration actuator;

[0057] 210. Mutual information selection unit; 220. Correlation selection unit; 230. Generation unit; 240. Objective function calculation unit; 250. Filtering unit; 260. Control unit;

[0058] 251. Sub-unit for calculating target value; 252. Sub-unit for calculating offset; 253. Sub-unit for calculating number of adjustment variables; 254. Sub-unit for calculating energy consumption change; 255. First selection sub-unit; 256. Sub-unit for calculating contribution; 257. Second selection sub-unit. Detailed Implementation

[0059] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0060] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0061] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0062] Please see Figures 1 to 6This invention proposes an aeration system and method for wastewater treatment, applicable to wastewater treatment fields such as municipal wastewater treatment plants, industrial wastewater treatment facilities, and aerobic biochemical treatment units in oxidation ditch processes. This enables precise control of the aeration process, energy consumption optimization, and stable effluent compliance. Detailed descriptions are provided below using specific embodiments.

[0063] Please see Figure 1 , Figure 2 and Figure 3 The present invention proposes an aeration system for wastewater treatment, which may include a detection module 10, a processing module 20 and an aeration actuator 30.

[0064] like Figure 1 As shown, the self-distribution well is the starting point of the wastewater treatment process. Its function is to evenly distribute the influent and ensure the stability of subsequent treatment processes. After distribution, the wastewater enters the oxidation ditch, which is a common aerobic biological treatment process in wastewater treatment plants. It typically uses plug-flow aeration equipment to drive the mixed liquor to circulate in a closed channel, providing sufficient dissolved oxygen for aerobic microorganisms to degrade organic matter in the wastewater and carry out nitrification. The effluent from the oxidation ditch then enters the secondary sedimentation tank for sludge-water separation.

[0065] like Figure 2 As shown, the mixed liquor from the oxidation ditch enters secondary settling tanks A and B respectively. The function of the secondary settling tanks is to separate the activated sludge from the treated water through gravity sedimentation. The supernatant is discharged as the final effluent, while the activated sludge settled at the bottom is discharged into the sludge pumping station. The sludge pumping station discharges part of the concentrated sludge as excess sludge from the system, and the other part is returned as sludge to the front end of the oxidation ditch to maintain a sufficient concentration of microorganisms in the oxidation ditch and ensure the effectiveness of the biological treatment.

[0066] The aeration actuator 30 can be installed in the oxidation ditch of the wastewater treatment plant. Specifically, it can be a surface aerator or a blower aeration device to supply oxygen to the aerobic biochemical unit to maintain the metabolic activities of microorganisms. The aeration actuator 30 can have multiple output ports, which are respectively installed at multiple locations within the oxidation ditch.

[0067] Please see Figure 3 The detection module 10 is deployed at multiple representative locations within the oxidation ditch to monitor influent water quality parameters in real time, such as dissolved oxygen concentration, ammonia nitrogen concentration, nitrate nitrogen concentration, chemical oxygen demand, water temperature, and pH value. These parameters are collected by online sensors at a fixed sampling frequency and transmitted to the processing module 20.

[0068] Please see Figure 3The processing module 20 acquires the equipment parameters of the aeration actuator 30. These parameters may include at least the blower speed, aeration valve opening, air supply setpoint, and dissolved oxygen setpoint. The processing module 20 fine-tunes the equipment parameters according to multiple preset constraints, generating multiple candidate equipment parameter sets. Each candidate equipment parameter set includes multiple equipment parameters generated through multiple fine-tuning steps under the same constraints. The constraints may include at least equipment parameter boundary constraints, equipment start-up and shutdown frequency constraints, and process stability constraints. Under the preset constraints, each fine-tuning generates one equipment parameter, and multiple fine-tuning steps generate a candidate equipment parameter set. That is, each candidate equipment parameter set may contain the adjusted values ​​of multiple equipment parameters at different time steps.

[0069] The processing module 20 is also used to generate predicted effluent water quality parameters for each candidate device parameter set based on the influent water quality parameters and multiple candidate device parameter sets using a pre-trained prediction model. The prediction model is an LSTM (Long Short-Term Memory) prediction model.

[0070] For example, after generating a candidate equipment parameter set each time, the processing module 20 inputs the current influent water quality parameters and the candidate equipment parameter set into the pre-trained prediction model. The prediction model uses historical data to learn the temporal influence of influent water quality and equipment parameters on effluent water quality, and outputs the predicted effluent water quality parameters corresponding to the candidate equipment parameter set. The predicted effluent water quality parameters include indicators such as ammonia nitrogen concentration, total nitrogen concentration, chemical oxygen demand, and suspended solids concentration at multiple future times.

[0071] Specifically, for each candidate device parameter set, the processing module 20 inputs the influent water quality parameter and one device parameter into the pre-trained prediction model in a preset order of all device parameters to generate the predicted effluent water quality parameter corresponding to each device parameter, and obtains the corresponding predicted effluent water quality parameter based on all predicted effluent water quality parameters.

[0072] In one embodiment of the present invention, the training method for the prediction model includes:

[0073] Multiple training samples were obtained, including influent water quality parameters, equipment parameters, and effluent water quality parameters. The effluent water quality parameters were the sample labels corresponding to the influent water quality parameters and equipment parameters.

[0074] All training samples are input into the prediction model to be trained to obtain the corresponding prediction results;

[0075] Based on the sample labels of all training samples and the corresponding prediction results, calculate the loss value of the loss function of the prediction model;

[0076] Based on the loss value, the parameters in the prediction model to be trained are adjusted to obtain the pre-trained prediction model.

[0077] The processing module 20 inputs each candidate equipment parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to calculate the objective function value. The objective function can comprehensively consider factors such as effluent compliance, equipment parameter adjustment range, and aeration energy consumption. The processing module 20 sorts and filters based on the objective function values ​​of all candidate equipment parameter sets, and selects the candidate equipment parameter set with the optimal objective function value as the control sequence. The control sequence contains equipment parameter adjustment instructions for multiple future time periods.

[0078] The processing module 20 can adjust the equipment parameters of the aeration actuator 30 step by step according to the control sequence. Specifically, it executes the control commands step by step according to the preset adjustment step size, and monitors the real-time feedback of key indicators such as dissolved oxygen and ammonia nitrogen in the oxidation ditch after each adjustment step. If an abnormal trend occurs, it triggers the rollback mechanism to restore the baseline control state.

[0079] Please see Figure 3 In one embodiment of the present invention, the processing module 20 is further configured to perform the following operations.

[0080] After obtaining the influent water quality parameters and equipment parameters, the processing module 20 calculates the moving average and standard deviation of the influent water quality parameters and equipment parameters respectively, and based on the moving average and standard deviation, filters out abnormal values ​​in the influent water quality parameters and equipment parameters and corrects them.

[0081] Specifically, for any sub-parameter among the influent water quality parameters and equipment parameters, if the value of the sub-parameter... satisfy: , Take 3. This is the moving average of the sub-parameter. The value of the sub-parameter is the standard deviation, and it is marked when the above inequality is satisfied. This is considered an anomaly and can be corrected or removed.

[0082] For example, for abnormal values, the corresponding normal values ​​can be calculated based on the linear / spline interpolation method, and the calculated normal values ​​can be corrected to obtain the corrected influent water quality parameters and equipment parameters.

[0083] In addition, the processing module 20 can also normalize the influent water quality parameters and equipment parameters after the outlier correction is completed, so that all sub-parameters are converted to a unified dimensional scale, eliminating the impact of different water quality indicators or equipment parameters on subsequent model training due to different units.

[0084] The processing module 20 is also used to fine-tune the corrected equipment parameters according to multiple preset constraints, generate multiple candidate equipment parameter sets, and generate the predicted effluent water quality parameters corresponding to each candidate equipment parameter set through a pre-trained prediction model based on the corrected influent water quality parameters and multiple candidate equipment parameter sets.

[0085] Please see Figure 3 In one embodiment of the present invention, the processing module 20 may include a mutual information selection unit 210, a correlation selection unit 220, and a generation unit 230.

[0086] The mutual information selection unit 210 is used to calculate the mutual information value between each equipment sub-parameter in the equipment parameters and the preset target effluent water quality parameters, select the equipment sub-parameters whose mutual information value is greater than the preset mutual information threshold, and retain the unselected equipment sub-parameters.

[0087] Specifically, the mutual information selection unit 210 is used to identify key control variables from the equipment parameters that are closely related to the preset target effluent water quality.

[0088] ,in, This is represented as each device sub-parameter in the device parameters. This represents the preset target effluent water quality parameters. Represented as a probability function, Represented as a logarithmic function, This is represented as a mutual information function.

[0089] Mutual information selection unit 210 selects the preset target effluent water quality parameters As the target variable, obtain each device sub-parameter from the device parameters. As candidate features, equipment sub-parameters may include blower speed, aeration valve opening, air supply setpoint, and dissolved oxygen setpoint. The mutual information selection unit 210 calculates the mutual information value between each equipment sub-parameter and the target effluent water quality parameter. The mutual information value measures the amount of information shared between two variables; the larger the value, the more significant the impact of the equipment sub-parameter on the effluent water quality.

[0090] The mutual information selection unit 210 compares the mutual information value of each device sub-parameter with the preset mutual information threshold. The mutual information threshold is pre-calibrated based on historical data distribution and process experience. Device sub-parameters with mutual information values ​​greater than the preset threshold are retained as the key feature set after initial screening. Other device sub-parameters with lower mutual information values ​​are considered to have a weak correlation with the effluent water quality and are not adjusted in subsequent fine-tuning.

[0091] The correlation selection unit 220 is used to select several device sub-parameters from the selected device sub-parameters based on the correlation degree, to obtain the device sub-parameters that need to be fine-tuned, and to retain the device sub-parameters that were not selected.

[0092] Specifically, the correlation selection unit 220 also performs the following operations: First, it randomly groups the equipment sub-parameters selected by the mutual information selection unit 210 to obtain multiple equipment sub-parameter groups; wherein different equipment sub-parameter groups include different equipment sub-parameters. Second, for each equipment sub-parameter group, it calculates the correlation between each equipment sub-parameter and the target effluent water quality parameter, and calculates the average correlation of all obtained correlations. It also calculates the similarity between any two equipment sub-parameters, calculates the average similarity of all obtained similarities, and calculates the difference between the average correlation and the average similarity. Then, it selects the equipment sub-parameter group with the largest difference between the average correlation and the average similarity, confirming that all equipment sub-parameters in this group are the equipment sub-parameters that need fine-tuning.

[0093] Specifically, the difference between the average relevance and the average similarity satisfies:

[0094] ,in, It is represented as a group of device sub-parameters, and is composed of some device sub-parameters among all the device sub-parameters selected by the mutual information selection unit 210. Represented as a combination of device sub-parameters The device sub-parameters in , Represented as a combination of device sub-parameters Any two device sub-parameters in the data, This is represented as a mutual information function. This is expressed as the average correlation. This is expressed as the average similarity.

[0095] The generation unit 230 is used to perform multi-step fine-tuning of the equipment sub-parameters that need to be fine-tuned for each constraint condition, and to form a corresponding candidate equipment parameter set by combining the fine-tuned equipment sub-parameters and the retained equipment sub-parameters; wherein, the constraint conditions include at least the equipment parameter boundary constraint conditions, the equipment start-stop frequency constraint conditions, and the process stability constraint conditions.

[0096] Specifically, the generation unit 230 is used to generate multiple candidate device parameter sets based on the set of device sub-parameters to be fine-tuned determined by the correlation selection unit 220. The generation unit 230 performs multi-step fine-tuning of the device sub-parameters to be fine-tuned each time according to different preset constraints, without making any adjustments to other device sub-parameters during the fine-tuning process.

[0097] Equipment parameter boundary constraints, namely, the adjustment range of each equipment sub-parameter must not exceed its physically allowed minimum and maximum values, to ensure that the generated candidate parameters are within the executable range of the equipment.

[0098] For example, .

[0099] in, Represented as Device parameters corresponding to the time , , These represent the minimum and maximum allowed values ​​for the device parameters, respectively. It should be understood that when a device parameter has a minimum and maximum allowed value, all device sub-parameters within that parameter also have corresponding minimum and maximum allowed values.

[0100] Equipment start-up and shutdown frequency constraints, namely, the number of times the equipment sub-parameters change per unit time shall not exceed the allowable upper limit, to prevent equipment wear or system oscillation caused by frequent adjustments.

[0101] For example, .

[0102] in, This is indicated as the 30th aeration actuator. An independently adjustable device or control loop, available Represented as Equipment parameters corresponding to independently adjustable devices or control loops. This is represented as the control time domain (prediction steps) corresponding to the device parameters. Indicates the current time. Represented as an index for a future time step. Represented as The control parameters of the equipment, Indicated as adjacent Control parameters of the equipment at any given time. This represents the single-step adjustment range. This is represented as the action dead zone threshold. Represented as an indicator function, This represents the maximum number of actions allowed.

[0103] The above formula means: for the th Device number, from the current moment Start to During the time period up to this point, adjust its control values ​​as needed, but the total number of times those adjustments are too drastic or excessive should not exceed the equipment's maximum capacity. .

[0104] The process stability constraint means that the fine-tuned combination of equipment parameters must be able to make the effluent water quality parameters meet the discharge standards. This constraint is calibrated by correlating the historical compliance data of the detection module with the corresponding equipment parameter range.

[0105] For example, .in, Represented as The corresponding device parameters at that moment Represented as a non-zero constant, These represent the maximum allowed values ​​for the device parameters, when... This allows the aeration volume at any output port of the aeration actuator 30 to be greater than a constant value, thereby enabling the wastewater treatment process to have a stable lower limit value.

[0106] Under the premise of satisfying all the above constraints, the generation unit 230 can generate multiple candidate device parameter sets by changing the fine-tuning step size, fine-tuning direction or fine-tuning time point based on the Counterfactual Explanation Generation (CEG) algorithm. Each candidate set contains device sub-parameter settings for multiple future times.

[0107] The core idea of ​​counterfactual interpretation generation is to use the current operating equipment parameters as a baseline state. By hypothetically changing the values ​​of certain equipment sub-parameters, a series of parameter sequences that differ from the baseline state but meet the constraints are generated, and the potential impact of these changes on the effluent water quality is observed. During each generation process, the generation unit 230 explores combinations of dimensions such as adjustment step size, adjustment direction, and adjustment time point for the set of equipment sub-parameters requiring fine-tuning. The adjustment step size determines the magnitude of each adjustment, the adjustment direction determines whether the parameter is increased or decreased, and the adjustment time point determines when the adjustment begins within the control time domain and how long the adjustment lasts. By traversing different combinations of these adjustment dimensions, the generation unit 230 generates multiple candidate equipment parameter sets. Each candidate set contains equipment sub-parameter set values ​​for multiple future time points, providing diverse control scheme options for subsequent evaluation and selection.

[0108] Please see Figure 3 In one embodiment of the present invention, the processing module 20 further includes an objective function calculation unit 240, a filtering unit 250, and a control unit 260.

[0109] The objective function calculation unit 240 is used to input each candidate device parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value.

[0110] Specifically, the objective function calculation unit 240 is used to quantitatively evaluate each candidate equipment parameter set. The objective function calculation unit 240 acquires multiple candidate equipment parameter sets output by the generation unit, and simultaneously acquires the predicted effluent water quality parameters corresponding to each candidate equipment parameter set. These predicted effluent water quality parameters may include predicted values ​​for indicators such as ammonia nitrogen concentration, total nitrogen concentration, and chemical oxygen demand at multiple future time points. The objective function calculation unit 240 substitutes each candidate equipment parameter set and its corresponding predicted effluent water quality parameters into a preset objective function for calculation, obtaining the corresponding objective function value.

[0111] Specifically, the objective function satisfies:

[0112] = .

[0113] in, Indicates the current time. This is represented as the control time domain (prediction steps) corresponding to the device parameters. Represented as from the current moment Start to The objective function value during this time period up to time. , These are weighting coefficients, all of which are constants.

[0114] objective function It consists of three items, the first of which is the loss due to failure to meet the standard. The first term measures the deviation between the predicted effluent water quality parameters and the discharge standard limits. If any water quality indicator exceeds the allowable range at any time, this term is assigned an infinite value to achieve a veto. The second term is the offset penalty term. This measure assesses the adjustment range between the candidate device parameter set and the current operating baseline device parameter set; a larger adjustment range results in a larger value for this item. The third item is the energy consumption item. It is used to estimate the aeration energy consumption corresponding to the set of candidate equipment parameters. The energy consumption is calculated based on the air supply or blower speed in the equipment parameters combined with the equipment energy consumption characteristic curve.

[0115] For example, .

[0116] .

[0117] in, Represented as the achievement loss function, Represented as from the current moment Start to The target values ​​achieved during this period up to the specified time.

[0118] This is represented as the predicted effluent water quality parameter. Representing the current time Start to Predicted water quality parameters for the period up to that point in time. Represented as Predicted water quality parameters at any time. This is expressed as the emission standard limit.

[0119] For example, .

[0120] in, Represented as an offset function, Represented as The corresponding device parameters at that moment Represented as The device parameters corresponding to the given time. Represented as a set of candidate device parameters, it is a set of parameters from the current time. To the future The vector sequence at time points, specifically in the form of: . The baseline device parameter set is a vector sequence with the same time domain as the candidate device parameter set. That is, from the current time... To the future The sequence of equipment parameter values ​​that the aeration actuator should maintain at each moment up to the specified time, provided that no optimization or adjustment is implemented and the current operating mode remains unchanged. Choose 1 or 2.

[0121] For example, .

[0122] in, Represented as an energy consumption function, Represented as from the current moment Start to The energy consumption value during this period up to the specified time. Represented as the first An independently adjustable device or control loop, available Represented as Equipment sub-parameters corresponding to independently controlled equipment or control loops. This indicates the total number of equipment sub-parameters, i.e., the number of equipment parameters involved in energy consumption calculations, such as blower speed and air supply. This is represented as the control time domain (prediction steps) corresponding to the device parameters. Indicates the current time. Represented as an index for a future time step. Represented as Control quantities of the device sub-parameters at any given time. The value range is between 1.5 and 3, and 2 can be used in this example. This is the equipment coefficient, obtained from on-site calibration. This is expressed as the sampling time interval or control step size.

[0123] The objective function calculation unit calculates the objective function value of each candidate device parameter set by weighted summation of the three terms. The smaller the objective function value, the better the overall performance of the candidate set in terms of compliance, stable operation and energy saving.

[0124] The filtering unit 250 is used to sort all objective function values ​​and filter out a preset number of candidate device parameter sets in ascending order of value, and select one candidate device parameter set from the preset number of candidate device parameter sets as the control sequence.

[0125] Specifically, the filtering unit 250 is used to select several sets of candidates with the best overall performance from a large number of candidate device parameter sets. The filtering unit 250 obtains all candidate device parameter sets and their corresponding objective function values ​​output by the objective function calculation unit, and sorts all objective function values ​​in ascending order of value. The smaller the value, the better the overall performance of the candidate set.

[0126] The screening unit 250 selects a number of candidate equipment parameter sets from the ranking results based on the preset number of candidate sets to be retained. These candidate sets have passed the veto test for compliance and have relative advantages in terms of adjustment range and energy consumption. The screening unit 250 outputs the selected candidate equipment parameter sets, their corresponding objective function values, and predicted effluent water quality parameters to the control unit 260.

[0127] The control unit 260 is used to control the operation of the aeration actuator according to the control sequence.

[0128] Specifically, the control unit 260 is used to convert the selected set of candidate equipment parameters into actual executable control instructions. The control unit 260 obtains the top-ranked set of candidate equipment parameters output by the screening unit 250, selects one of these candidate sets as the final control sequence, and the control sequence contains the equipment parameter set values ​​for multiple future times, specifically including the blower speed, aeration valve opening, air supply set value, and dissolved oxygen set point for each time.

[0129] The control unit 260 sends adjustment commands to the aeration actuator hourly according to the control sequence. The adjustment commands are executed step by step according to the preset step size to avoid shock to the biological system due to large changes. After each adjustment step, the control unit 260 continuously monitors the real-time influent water quality parameters fed back by the detection module. If the dissolved oxygen concentration is found to be below the safety lower limit or the ammonia nitrogen concentration shows an abnormal upward trend, the current control sequence is immediately stopped and the rollback mechanism is triggered to restore the equipment parameters to the baseline state to ensure process safety.

[0130] Please see Figure 4 In one embodiment of the present invention, the screening unit 250 may include a target value calculation subunit 251, an offset calculation subunit 252, an adjustment variable number calculation subunit 253, an energy consumption change calculation subunit 254, and a first selection subunit 255.

[0131] The compliance value calculation subunit 251 is used to calculate the corresponding compliance value based on the predicted effluent water quality parameters corresponding to each candidate equipment parameter set selected and the preset emission standard limit.

[0132] Specifically, the formula for calculating the target value satisfies:

[0133] .

[0134] in, This represents an index of the candidate scheme, used to distinguish different sets of candidate device parameters. Represented as the first The validity flag for the parameter set of each candidate device is set as follows: a value of 1 indicates that all effluent indicators meet the standards at all predicted times, and the device can proceed to the next evaluation. A value of 0 indicates that there is a risk of exceeding the standards, and the device will be rejected outright. Represented as a time index, with values ​​ranging from 1 to... , indicating the future That moment. This is expressed as the prediction time domain length, which is the total number of time steps for predicting the future. This is represented as an index of effluent water quality indicators, used to distinguish different types of pollutants or water quality parameters, such as ammonia nitrogen, total nitrogen, chemical oxygen demand, suspended solids, etc. Represented as in the first Under the candidate scheme, the predicted future number is... The moment of the first The values ​​of the effluent water quality parameters. Represented as the first The permissible range of the effluent water quality parameters, that is, the qualified range specified by the discharge standards. This is represented as an indicator function, which takes the value 1 when the condition within the parentheses is true, and 0 otherwise.

[0135] The offset calculation subunit 252 is used to calculate the corresponding offset value based on the set of parameters of each selected candidate device and the preset set of baseline control parameters.

[0136] Specifically, the formula for calculating the offset value satisfies:

[0137] .

[0138] in, Represented as the first The operational cost offset of a set of candidate device parameters is used to quantify the overall adjustment magnitude of the candidate scheme relative to the current baseline operating state. Represented as a time index, with values ​​ranging from 1 to... , indicating the future That moment. This is expressed as the prediction time domain length, which is the total number of time steps for predicting the future. Represented as the first In the set of candidate device parameters, the future... The device parameter vector at a given time moment contains the specific values ​​of all device sub-parameters at that time moment, such as blower speed, aeration valve opening, air supply setpoint, dissolved oxygen setpoint, etc. Represented as the baseline device parameter set, the future [number]th ... The device parameter vector at each moment represents the expected value of the device parameters at that moment if no adjustments are made. Represented as the L1 norm, it is used to calculate the difference between two vectors, specifically the sum of the absolute differences of their components. For a device parameter vector, the L1 norm calculates the absolute value of the adjustment magnitude of each device sub-parameter and then sums them up.

[0139] The adjustment variable calculation subunit 253 is used to calculate the corresponding number of adjustment variables based on the set of parameters of each candidate device selected and the preset set of baseline control parameters.

[0140] Specifically, the formula for calculating the number of adjustment variables satisfies:

[0141] .

[0142] in, Represented as the first The sparsity index of a candidate device parameter set indicates the number of device sub-parameters that undergo substantial adjustments relative to the baseline state in the candidate scheme. Represented as an index of a device sub-parameter, with values ​​ranging from 1 to... It is used to distinguish different controllable equipment parameters, such as blower speed, aeration valve opening, air supply setpoint, dissolved oxygen setpoint, etc. This represents the total number of device sub-parameters, i.e., the number of all control variables that may be involved in the adjustment. Represented as a time index, with values ​​ranging from 1 to... , indicating the future That moment. This is expressed as the prediction time domain length, which is the total number of time steps for predicting the future. Represented as the first In the set of candidate device parameters, the first... The device sub-parameter in the future The value taken at each moment. Represented as the first in the baseline device parameter set. The device sub-parameter in the future The value at each moment represents the expected value without adjustment. This is represented as the substantial change threshold, a preset constant used to determine whether an adjustment to a device sub-parameter is sufficiently significant. When the maximum adjustment exceeds this threshold, the parameter is considered to have undergone a substantial adjustment. This is represented as an indicator function, which takes the value 1 when the condition within the parentheses is true, and 0 otherwise.

[0143] The energy consumption change calculation subunit 254 is used to calculate the corresponding energy consumption change value based on the set of parameters of each selected candidate device and the preset set of baseline control parameters.

[0144] Specifically, the formula for calculating the value of energy consumption change satisfies:

[0145] .

[0146] in, Represented as the first The relative rate of change of energy consumption for a set of candidate device parameters represents the proportion of increase or decrease in energy consumption of the candidate scheme relative to the baseline operating state. Represented as the first The total energy consumption of the aeration system corresponding to the parameter set of each candidate device is the total electrical energy consumption estimated by combining the device parameter values ​​(such as air supply and blower speed) at each future time in the candidate scheme with the energy consumption model. The total energy consumption of the aeration system corresponding to the baseline set of equipment parameters is the total amount of electricity consumption predicted over a future period based on the current operating mode or while keeping the existing parameters unchanged. It serves as a benchmark value for energy consumption comparison.

[0147] The first selection subunit 255 is used to calculate the comprehensive score based on the compliance value, offset value, number of adjustment variables and energy consumption change value corresponding to each candidate device parameter set selected, and select the candidate device parameter set corresponding to the maximum comprehensive score as the control sequence.

[0148] Specifically, overall score The calculation formula satisfies:

[0149] .

[0150] in, Represented as the first The comprehensive score of the parameter set of each candidate device is used to quantify the overall performance of the solution in four dimensions: compliance, stable operation, energy saving and ease of operation. The higher the score, the better the solution. , , , They correspond to the first The set of parameters for each candidate device includes the target values, offset values, number of adjustment variables, and energy consumption changes. It is based on the original offset It is obtained after normalization and is used to eliminate the influence of different dimensions. , , These are weighting coefficients, which are non-zero constants. The function is an exponential function, which combines the weighted penalty terms within parentheses into a multiplicative form. The larger the sum of the penalty terms, the smaller the exponential function value, and the lower the overall score. The exponential function guarantees a positive score, and the penalty terms are multiplicative and independent of each other.

[0151] Please see Figure 4 In one embodiment of the present invention, the screening unit 250 may include a contribution calculation subunit 256 and a second selection subunit 257.

[0152] The contribution calculation unit 256 is used to compare the change of each sub-parameter of the selected candidate equipment parameters with the change of the predicted effluent water quality parameters based on the selected candidate equipment parameter set and its corresponding predicted effluent water quality parameters, and obtain the contribution of each sub-parameter of the equipment to the predicted effluent water quality parameters.

[0153] Specifically, the contribution calculation unit 256 is used to quantify the impact of each equipment sub-parameter on the effluent water quality. The contribution calculation unit 256 acquires multiple candidate equipment parameter sets and their corresponding predicted effluent water quality parameters, as well as the baseline equipment parameter set and its corresponding baseline predicted effluent water quality parameters. For each candidate set, the contribution calculation unit 256 calculates the average change of each equipment sub-parameter relative to the baseline value, and simultaneously calculates the average change of each predicted effluent water quality parameter relative to the baseline predicted value.

[0154] The contribution calculation unit 256 compares the average change of each equipment sub-parameter with the average change of the effluent water quality parameters, and determines the contribution of the equipment sub-parameter by the ratio of the two. The larger the ratio, the more significant the impact of the parameter on the effluent water quality. The contribution calculation unit 256 integrates the calculation results of all candidate sets to obtain the final contribution of each equipment sub-parameter.

[0155] The second selection subunit 257 is used to identify the device sub-parameter with the largest contribution, and select the candidate device parameter set with the smallest change in the device sub-parameter from all the selected candidate device parameter sets as the control sequence.

[0156] Specifically, the second selection sub-unit 257 is used to determine the final control sequence based on the contribution. The second selection sub-unit 257 obtains the contribution of all device sub-parameters and identifies the device sub-parameter with the largest contribution as the target device sub-parameter. The second selection sub-unit 257 traverses all candidate device parameter sets, extracting the adjustment range of the target device sub-parameter in each candidate set. The adjustment range is measured by the degree of deviation of the parameter from the baseline value throughout the entire control time domain. The second selection sub-unit 257 compares the adjustment ranges of the target device sub-parameters in all candidate sets and selects the candidate device parameter set with the smallest adjustment range as the control sequence, which is then output to the control unit 260 for execution.

[0157] Please see Figure 6 In one embodiment of the present invention, an aeration method for wastewater treatment is also proposed, comprising the following steps.

[0158] Step S10: Detect the influent water quality parameters of sewage at multiple locations within the oxidation ditch.

[0159] Step S20: Obtain the equipment parameters of the aeration actuator, fine-tune the equipment parameters according to multiple preset constraints, and generate multiple candidate equipment parameter sets; wherein, each candidate equipment parameter set includes multiple equipment parameters generated by multiple fine-tuning under the same constraints.

[0160] Step S30: Based on the influent water quality parameters and multiple candidate equipment parameter sets, generate the predicted effluent water quality parameters corresponding to each candidate equipment parameter set through a pre-trained prediction model; wherein, the prediction model is an LSTM prediction model.

[0161] Step S40: Input each candidate device parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value; filter all candidate device parameter sets based on the objective function value, and use the filtered candidate device parameter sets as a control sequence to control the operation of the aeration actuator according to the control sequence.

[0162] In summary, this invention proposes an aeration system and method for wastewater treatment. The system and method detect influent water quality parameters at multiple locations within an oxidation ditch using a detection module. A processing module fine-tunes the equipment parameters based on multiple preset constraints, generating multiple candidate equipment parameter sets for proactive intervention in the operation of the aeration actuator during wastewater treatment. Each candidate equipment parameter set includes multiple equipment parameters generated through multi-step fine-tuning under the same constraints. The processing module inputs the influent water quality parameters and multiple candidate equipment parameter sets into a pre-trained prediction model to generate predicted effluent water quality parameters for each candidate equipment parameter set. The processing module further inputs each candidate equipment parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value. Based on the objective function value, all candidate equipment parameter sets are filtered, and the filtered sets are used as control sequences to control the operation of the aeration actuator, thereby achieving refined operation management of the aeration process.

[0163] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. An aeration system for wastewater treatment, characterized in that, include: The aeration actuator is installed inside the oxidation ditch of the sewage treatment plant; The detection module is used to detect the influent water quality parameters of wastewater at multiple locations within the oxidation ditch; The processing module is used to acquire the equipment parameters of the aeration actuator, fine-tune the equipment parameters according to multiple preset constraints, and generate multiple candidate equipment parameter sets; wherein, each candidate equipment parameter set includes multiple equipment parameters generated by multiple fine-tuning under the same constraints. The processing module is further configured to generate predicted effluent water quality parameters corresponding to each candidate device parameter set based on the influent water quality parameters and multiple candidate device parameter sets using a pre-trained prediction model; wherein, the prediction model is an LSTM prediction model; The processing module is further configured to input each candidate device parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value; filter all candidate device parameter sets based on the objective function value, and use the filtered candidate device parameter sets as control sequences; The processing module adjusts the equipment parameters of the aeration actuator according to a preset adjustment step size based on the control sequence, and controls the operation of the aeration actuator. After each adjustment step is completed, the real-time influent water quality parameters fed back by the detection module are continuously detected. When the real-time influent water quality parameters are lower than the safety lower limit, the current control sequence is stopped and the equipment parameters are restored to the preset baseline state value. The baseline state value represents the equipment parameter value corresponding to maintaining the current operating mode unchanged. The preset objective function satisfies: = ； in, Indicates the current time. This is represented as the control time domain corresponding to the device parameters. Represented as from the current moment Start to The objective function value during this time period up to time. , These are weighting coefficients, all of which are non-zero constants; The objective function consists of three terms, the first of which is the achievement loss term. The first item measures the deviation between the predicted effluent water quality parameters and the discharge standard limits. If any water quality indicator exceeds the allowable range at any time, this item is assigned an infinite value to achieve a veto. The second item is the deviation penalty item. The first item measures the adjustment range between the candidate device parameter set and the current operating baseline device parameter set; the larger the adjustment range, the larger the value of this item. The third item is the energy consumption item. It is used to estimate the aeration energy consumption corresponding to the set of candidate equipment parameters. The energy consumption is calculated based on the air supply or blower speed in the equipment parameters combined with the equipment energy consumption characteristic curve.

2. The aeration system for wastewater treatment according to claim 1, characterized in that, The processing module is also used for: After obtaining the influent water quality parameters and equipment parameters, the moving mean and standard deviation of the influent water quality parameters and equipment parameters are calculated respectively. Based on the moving mean and standard deviation, abnormal values ​​in the influent water quality parameters and equipment parameters are screened out and corrected. The processing module is also used to fine-tune the corrected equipment parameters according to multiple preset constraints, generate multiple candidate equipment parameter sets, and generate predicted effluent water quality parameters corresponding to each candidate equipment parameter set through a pre-trained prediction model based on the corrected influent water quality parameters and multiple candidate equipment parameter sets.

3. The aeration system for wastewater treatment according to claim 1, characterized in that, The processing module includes: The mutual information selection unit is used to calculate the mutual information value between each sub-parameter of the equipment parameters and the preset target effluent water quality parameter, select the sub-parameters of the equipment whose mutual information value is greater than the preset mutual information threshold, and retain the unselected sub-parameters of the equipment. The correlation selection unit is used to select several equipment sub-parameters from the selected equipment sub-parameters based on the correlation degree, to obtain the equipment sub-parameters that need to be fine-tuned, and to retain the equipment sub-parameters that were not selected. The generation unit is used to perform multi-step fine-tuning of the equipment sub-parameters that need to be fine-tuned for each of the constraints, and to form a corresponding candidate equipment parameter set by combining the fine-tuned equipment sub-parameters and the retained equipment sub-parameters; wherein the constraints include at least equipment parameter boundary constraints, equipment start-stop frequency constraints, and process stability constraints.

4. The aeration system for wastewater treatment according to claim 3, characterized in that, The correlation selection unit is also used for: The device sub-parameters selected by the mutual information selection unit are randomly grouped to obtain multiple device sub-parameter groups; wherein, different device sub-parameter groups include different device sub-parameters; For each equipment sub-parameter group, calculate the correlation between each equipment sub-parameter and the target effluent water quality parameter, and calculate the average correlation of all obtained correlations. Calculate the similarity between any two equipment sub-parameters, and calculate the average similarity of all obtained similarities. Also, calculate the difference between the average correlation and the average similarity. Select the equipment sub-parameter group with the largest difference between the average relevance and the average similarity, and confirm that all equipment sub-parameters in this equipment sub-parameter group are equipment sub-parameters that need to be fine-tuned.

5. The aeration system for wastewater treatment according to claim 1, characterized in that, The processing module includes: The objective function calculation unit is used to input each candidate device parameter set and its corresponding predicted effluent water quality parameters into a preset objective function to obtain the corresponding objective function value; The filtering unit is used to sort all objective function values ​​and filter out a preset number of candidate device parameter sets in ascending order of value, and select one candidate device parameter set from the preset number of candidate device parameter sets as the control sequence. A control unit is used to control the operation of the aeration actuator according to a control sequence.

6. The aeration system for wastewater treatment according to claim 5, characterized in that, The filtering unit includes: The compliance value calculation subunit is used to calculate the corresponding compliance value based on the predicted effluent water quality parameters corresponding to each set of candidate equipment parameters and the preset emission standard limits. The offset calculation subunit is used to calculate the corresponding offset value based on the set of parameters of each selected candidate device and the preset set of baseline control parameters. The adjustment variable calculation subunit is used to calculate the corresponding number of adjustment variables based on the selected set of parameters for each candidate device and the preset set of baseline control parameters. The energy consumption change calculation subunit is used to calculate the corresponding energy consumption change value based on the parameter set of each selected candidate device and the preset baseline control parameter set. The first selection subunit is used to calculate the comprehensive score based on the compliance value, offset value, number of adjustment variables and energy consumption change value of each selected candidate device parameter set, and select the candidate device parameter set corresponding to the maximum comprehensive score as the control sequence.

7. The aeration system for wastewater treatment according to claim 5, characterized in that, The filtering unit includes: The contribution calculation unit is used to compare the change of each sub-parameter of the selected candidate equipment parameters with the change of the predicted effluent water quality parameters, based on the selected candidate equipment parameter set and its corresponding predicted effluent water quality parameters, to obtain the contribution of each sub-parameter of the equipment to the predicted effluent water quality parameters. The second selection sub-unit is used to identify the device sub-parameter with the greatest contribution, and select the set of candidate device parameters with the smallest change in the device sub-parameter from all the selected candidate device parameter sets as the control sequence.

8. The aeration system for wastewater treatment according to claim 1, characterized in that, The training methods for the prediction model include: Multiple training samples are obtained, including influent water quality parameters, equipment parameters, and effluent water quality parameters, wherein the effluent water quality parameters are sample labels corresponding to the influent water quality parameters and equipment parameters; All training samples are input into the prediction model to be trained to obtain the corresponding prediction results; Based on the sample labels of all training samples and the corresponding prediction results, calculate the loss value of the loss function of the prediction model; Based on the loss value, the parameters in the prediction model to be trained are adjusted to obtain the pre-trained prediction model.

9. An aeration method for wastewater treatment, using the wastewater treatment aeration system as described in any one of claims 1 to 8, characterized in that, include: Detect the influent water quality parameters at multiple locations within the oxidation ditch; The equipment parameters of the aeration actuator are obtained, and the equipment parameters are fine-tuned according to multiple preset constraints to generate multiple candidate equipment parameter sets; wherein, each candidate equipment parameter set includes multiple equipment parameters generated by multiple fine-tuning under the same constraints. Based on the influent water quality parameters and multiple candidate equipment parameter sets, the predicted effluent water quality parameters corresponding to each candidate equipment parameter set are generated through a pre-trained prediction model; wherein, the prediction model is an LSTM prediction model; Each candidate device parameter set and its corresponding predicted effluent water quality parameters are input into a preset objective function to obtain the corresponding objective function value; based on the objective function value, all candidate device parameter sets are filtered, and the filtered candidate device parameter sets are used as control sequences to control the operation of the aeration actuator according to the control sequences.

Images