A method and system for coordinated regulation of a double-pool aeration device based on LSTM

Abstract

The application discloses a kind of based on LSTM's double pool aeration equipment coordination regulation method and system, it is related to sewage intelligent processing field, method includes: obtaining historical operation data, based on LSTM model training double pool coupling aeration prediction model.Acquire fan frequency and valve opening, set two pool target aeration quantity;With online solution module, with the weighted minimization of aeration quantity tracking deviation and control quantity change amplitude as objective function, fan frequency and two valve openings are jointly optimized, in optimization process, the aeration quantity corresponding to each control strategy is dynamically predicted using the trained model, finally, the optimal control strategy is screened out to regulate and control equipment.The application accurately fits the relationship between double pool aeration equipment and aeration quantity based on LSTM's double pool coupling aeration prediction model, combined with the constraint with penalty term is jointly optimized, realizes smooth decoupling control, while meeting the process requirement, reduces the equipment regulation frequency and operating energy consumption.

Description

Technical Field

[0001] This application relates to the field of intelligent wastewater treatment, and in particular to a method and system for coordinated regulation of dual-tank aeration equipment based on LSTM. Background Technology

[0002] Aeration equipment is a core energy-consuming component of the biological treatment unit in wastewater treatment. In practical engineering, a configuration is often used where one blower simultaneously supplies air to two parallel biological treatment tanks. In this configuration, the blower frequency adjustment affects the total air supply to both tanks, while the branch valve opening adjustment affects the air distribution ratio between the two tanks, resulting in a strong coupling relationship. In existing technologies, manual adjustment by operators has a lag in response and limited accuracy; single-loop PID independent control ignores the cross-influence of branch valve adjustment on the other tank, leading to severe coupling interference between the two loops and easily causing repeated valve oscillations; while rule-based or fuzzy logic-based coordinated control considers the coupling relationship to some extent, rule formulation relies on expert experience and is difficult to adapt to changes in the system's dynamic characteristics. In existing technologies, a programmable logic controller (PLC) is typically used to control the start and stop of two aeration motors and corresponding solenoid valves based on liquid level and temperature signals, achieving batch processing of the two tanks. However, it is essentially still a sequential control of on / off quantities, designed to meet the phased aeration requirements of the SBR process. It does not involve the continuous adjustment of the fan frequency and valve opening, nor does it provide a technical solution for dealing with the coupling effect of air volume between the two pools. It is difficult to apply to continuous flow or circulating aeration scenarios where there is a need for continuous and precise control of aeration volume.

[0003] In summary, existing technologies cannot characterize the coupling relationship of aeration in dual-pool coupled aeration scenarios, making it difficult to achieve coordinated adjustment of equipment and resulting in high aeration energy consumption; it is also difficult to meet the independent aeration needs of the two pools while ensuring adjustment stability and long-term operational stability. Summary of the Invention

[0004] The purpose of this application is to provide a collaborative adjustment method and system for dual-pool aeration equipment based on LSTM. The relationship between dual-pool aeration equipment and aeration volume can be accurately fitted by a dual-pool coupled aeration prediction model based on LSTM. Combined with constraints with penalty terms, joint optimization is performed to achieve smooth decoupled control. While meeting process requirements, the frequency of equipment adjustment and operating energy consumption can be reduced.

[0005] To achieve the above objectives, this application provides the following solution.

[0006] In one aspect, this application provides a coordinated adjustment method for a dual-pool aeration device based on LSTM. The dual-pool aeration device includes: a blower, a first valve, and a second valve. The blower is connected to an external first aeration pool and an external second aeration pool via a first pipeline and a second pipeline, respectively. The first valve and the second valve are respectively installed on the first pipeline and the second pipeline. Both the first valve and the second valve are configured to, under the condition that the total air intake of the blower is constant, adjust the air intake of their respective pipelines by controlling their opening degrees, and indirectly adjust the air intake of the other pipeline in the opposite direction. The method includes: acquiring historical operating data of the dual-pool aeration device; the historical operating data specifically includes: blower frequency, first valve opening degree, second valve opening degree, aeration volume of the first aeration pool, and aeration volume of the second aeration pool; selecting training samples based on the historical operating data, using the blower frequency, first valve opening degree, and second valve opening degree as inputs, and the aeration volume of the first aeration pool and the aeration volume of the second aeration pool as outputs, to train a dual-pool coupled aeration prediction model; the dual-pool coupled aeration prediction model is based on an LSTM model. The following steps are taken: The current fan frequency, first valve opening, and second valve opening are obtained; target aeration rates for the first and second aeration tanks are set; using an online solver module, the objective function for joint optimization is the weighted minimization of the aeration rate tracking deviation and the variation amplitude of the control quantity of the dual-tank aeration equipment. The control quantities of the fan, the first valve, and the second valve are used as optimization objects. Based on constraint terms, joint optimization is performed on the fan frequency, the first valve opening, and the second valve opening. During the joint optimization process, a dual-tank coupled aeration prediction model is used to... The predicted aeration volume of the first aeration tank and the predicted aeration volume of the second aeration tank corresponding to each control strategy are dynamically predicted. The optimal control strategy is obtained through screening, and finally, the dual-tank aeration equipment is adaptively adjusted in real time according to the optimal control strategy. The aeration volume tracking deviation includes the aeration volume deviation of the first aeration tank and the aeration volume deviation of the second aeration tank. The aeration volume deviation of the first aeration tank is the difference between the predicted aeration volume of the first aeration tank and the target aeration volume of the first aeration tank. The aeration volume deviation of the second aeration tank is the difference between the predicted aeration volume of the second aeration tank and the target aeration volume of the second aeration tank.

[0007] In a second aspect, this application also provides a computer system, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the LSTM-based coordinated adjustment method for dual-pool aeration equipment described in the first aspect.

[0008] Based on the specific embodiments provided in this application, the following technical effects are disclosed.

[0009] This application uses an LSTM-based dual-pool coupled aeration prediction model to accurately fit the relationship between dual-pool aeration equipment and aeration volume. Combined with constraints with penalty terms, it performs joint optimization to achieve stable decoupled control, thereby reducing equipment adjustment frequency and operating energy consumption while meeting process requirements. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic flowchart of the LSTM-based coordinated adjustment method for dual-pool aeration equipment provided in the embodiments of this application.

[0012] Figure 2 This is a schematic diagram of the structure of the dual-pool aeration device provided in the embodiments of this application.

[0013] Figure 3 A comparison chart of the target and actual values ​​of the aeration volume in the dual-pool system provided in this application embodiment.

[0014] Figure 4 This is an internal structure diagram of a computer system provided in an embodiment of this application. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] Example 1, such as Figures 1-2As shown, this embodiment provides a collaborative adjustment method for a dual-pool aeration device based on LSTM. The dual-pool aeration device includes: a blower, a first valve, and a second valve. The blower is connected to an external first aeration pool and an external second aeration pool through a first pipeline and a second pipeline, respectively. The first valve and the second valve are respectively installed on the first pipeline and the second pipeline. Both the first valve and the second valve are configured to: under the condition that the total air intake of the blower is constant, adjust the air intake of their respective pipelines by controlling their opening degree, and indirectly adjust the air intake of the other pipeline in the opposite direction. The method includes the following steps.

[0018] S1. Obtain historical operating data of the dual-tank aeration equipment; the historical operating data specifically includes: blower frequency, first valve opening degree, second valve opening degree, aeration volume of the first aeration tank and the aeration volume of the second aeration tank.

[0019] In practical applications, the adjustment methods for dual-pool aeration equipment are as follows: increasing the fan frequency can simultaneously increase the aeration volume of the first and second aeration pools, while decreasing the fan frequency can simultaneously decrease the aeration volume of the first and second aeration pools.

[0020] Increasing the opening of the first valve can increase the aeration volume of the first aeration tank and decrease the aeration volume of the second aeration tank.

[0021] Increasing the opening of the second valve can increase the aeration rate of the second aeration tank and decrease the aeration rate of the first aeration tank.

[0022] S2. Based on historical operating data, training samples are selected, and the fan frequency, the opening degree of the first valve and the opening degree of the second valve are used as inputs, and the predicted aeration volume of the first aeration tank and the predicted aeration volume of the second aeration tank are used as outputs to train the dual-tank coupled aeration prediction model; the dual-tank coupled aeration prediction model is constructed based on the LSTM model.

[0023] Furthermore, in addition to the LSTM model, the dual-pool coupled aeration prediction model can also be constructed based on gradient boosting tree model, random forest model, support vector regression model, and neural network model.

[0024] In practical applications, the inputs of the dual-pool coupled aeration prediction model include: the fan frequency at the current time and one or more historical times, the opening degree of the first valve, and the opening degree of the second valve. The output is the aeration volume of the first aeration pool and the aeration volume of the second aeration pool at future times, or the change in the aeration volume of the first aeration pool and the change in the aeration volume of the second aeration pool.

[0025] S3. Obtain the current fan frequency, the opening degree of the first valve and the opening degree of the second valve, and set the target aeration rate of the first aeration tank and the target aeration rate of the second aeration tank.

[0026] S4. Using an online solver module, the objective function for joint optimization is the weighted minimization of the aeration volume tracking deviation and the variation amplitude of the control quantity of the dual-pool aeration equipment. The control quantities of the blower, the first valve, and the second valve are used as optimization objects. Based on constraint terms, joint optimization is performed on the blower frequency, the opening degree of the first valve, and the opening degree of the second valve. During the joint optimization process, the predicted aeration volume of the first and second aeration pools corresponding to each control strategy is dynamically predicted using a dual-pool coupled aeration prediction model. The optimal control strategy is obtained through screening, and finally, the dual-pool aeration equipment is adaptively adjusted in real time according to the optimal control strategy. The aeration volume tracking deviation includes the aeration volume deviation of the first and second aeration pools. The aeration volume deviation of the first aeration pool is the difference between the predicted aeration volume and the target aeration volume of the first aeration pool; the aeration volume deviation of the second aeration pool is the difference between the predicted aeration volume and the target aeration volume of the second aeration pool.

[0027] Furthermore, the objective function expression for the joint optimization is as follows.

[0028] .

[0029] In the formula, To jointly optimize the objective function; , , , and All are weighting coefficients; and The predicted aeration volume for the first aeration tank and the predicted aeration volume for the second aeration tank are respectively. and These are the target aeration rates for the first aeration tank and the second aeration tank, respectively. The frequency of the fan to be solved; The current frequency of the wind turbine; Let the opening degree of the first valve be the one to be solved. This represents the first valve opening at the current moment; Let the opening degree of the second valve be the one to be solved. This represents the current valve opening degree.

[0030] Furthermore, and The range of values ​​is ; , and The range of values ​​is .

[0031] Furthermore, the constraints include at least one of the following: upper and lower limits of fan frequency, upper and lower limits of first valve opening, and upper and lower limits of second valve opening.

[0032] The expressions for the upper and lower limits of the wind turbine frequency constraints are as follows.

[0033] .

[0034] In the formula, For the frequency of the fan, The lower limit of the fan frequency is set according to the surge threshold or minimum aeration requirement of the fan. This is the upper limit of the fan frequency, set according to the rated frequency of the fan.

[0035] The expressions for the upper and lower limits of the opening of the first valve are as follows.

[0036] .

[0037] In the formula, This represents the opening degree of the first valve.

[0038] The expressions for the upper and lower limits of the second valve opening are as follows.

[0039] .

[0040] In the formula, This refers to the opening degree of the second valve.

[0041] Furthermore, during the joint optimization process, when the opening degrees of the first valve and the second valve reach the preset boundary conditions, but still cannot reach the target aeration volume of the first aeration tank and the target aeration volume of the second aeration tank, the blower frequency compensation adjustment is triggered; when the blower frequency meets the total aeration volume requirement of the target aeration volume of the first aeration tank and the target aeration volume of the second aeration tank, but there is still a deviation in the aeration distribution between the first aeration tank and the second aeration tank, the difference is corrected first through the first valve and the second valve.

[0042] Furthermore, the online solution module employs a particle swarm optimization algorithm.

[0043] Optionally, after step S4, the following steps are also included.

[0044] S5. Update the difference between the predicted and actual aeration volume to the historical operating data, and perform incremental updates or periodic retraining on the trained dual-pool coupled aeration prediction model.

[0045] Furthermore, the dual-tank aeration equipment further includes: a first aeration volume monitoring device and a second aeration volume monitoring device; the first aeration volume monitoring device is disposed between the first valve in the first pipeline and the external first aeration tank; the second aeration volume monitoring device is disposed between the second valve in the second pipeline and the external second aeration tank; the first aeration volume monitoring device is used to obtain the aeration volume of the first aeration tank; the second aeration volume monitoring device is used to obtain the aeration volume of the second aeration tank.

[0046] In practical applications, each detection instrument is connected to the PLC, which collects and stores the operating data in real time. The PLC is also connected to the online solution module.

[0047] As an optional implementation method, the specific execution process of the LSTM-based coordinated adjustment method for dual-pool aeration equipment is described in detail below.

[0048] Step 1: Collect the operating data of the dual-pool aeration equipment using a PLC at a fixed sampling period of once per minute and store it in the historical database. The collected raw data needs to undergo data cleaning and preprocessing, including outlier removal, missing value filling, and data normalization, to ensure the data quality for subsequent model training.

[0049] Step 2: Construct training samples based on historical operating data, and establish a dual-pool coupled aeration prediction model using machine learning algorithms. The dual-pool coupled aeration prediction model is used to characterize the mapping relationship between the aeration volume of the first and second aeration pools and the blower frequency, the opening degree of the first valve, and the opening degree of the second valve.

[0050] The training sample construction process involves organizing preprocessed historical operational data into sample pairs required for supervised learning in chronological order. The input feature vector includes the fan frequency, first valve opening, second valve opening, actual aeration volume of the first aeration tank, and actual aeration volume of the second aeration tank at the current time and at least one historical time. By introducing historical time features to capture the time lag effect of the aeration process, the model can learn the dynamic response process of aeration volume after the control action is issued. The output label is configured as the aeration volume of the first and second aeration tanks within a preset future prediction time domain, or as the change in aeration volume of the first and second aeration tanks. The prediction time domain is determined based on the system response characteristics, with a typical range from several seconds to several minutes.

[0051] The dual-pool coupled aeration prediction model is constructed using a Long Short-Term Memory (LSTM) network. The LSTM network utilizes its gating mechanism to capture long-distance dependencies in time-series data, adapting to the dynamic temporal characteristics of the aeration process. The LSTM network structure includes an input layer, several LSTM hidden layers, a fully connected layer, and an output layer. The number of nodes in the input layer equals the dimension of the input feature vector. The number of LSTM hidden layers and the number of neurons per layer are determined experimentally based on the data scale and system complexity; a typical configuration is 1 to 3 hidden layers, each containing 32 to 128 neurons. The output layer contains two output nodes, corresponding to the predicted aeration volume of the first and second aeration pools, respectively. During model training, mean squared error is used as the loss function, the Adam optimizer is used for parameter updates, and early stopping and L2 regularization are employed to suppress overfitting, thereby improving the model's generalization performance under unknown conditions.

[0052] After training, the model accuracy is evaluated using an independent test set.

[0053] In this embodiment, the mean squared error (MSE) of the model prediction results on the test set is 12450.12, the mean absolute error (MAE) is 62.35, the root mean square error (RMSE) is 111.58, and the coefficient of determination (R²) is... 2 The value is 0.9764, and the mean absolute percentage error (MAPE) is 5.21%. 2 A value close to 1 and a MAPE value below 10% indicate that the model fits well and the predicted values ​​match the measured values ​​well.

[0054] Step 3: The target aeration rates of the first aeration tank and the second aeration tank are manually set according to the effluent quality requirements and the influent load.

[0055] Step 4: Input the current fan frequency, first valve opening, second valve opening, target aeration volume of the first aeration tank, and target aeration volume of the second aeration tank into the online solution module. Based on the dual-tank coupled aeration prediction model, jointly optimize the fan frequency, first valve opening, and second valve opening to obtain the target control quantity combination.

[0056] The particle swarm optimization algorithm is used to solve the above optimization problem. First, several particles are randomly initialized within the constraints, and the position vector of each particle is represented as follows: The velocity vector is randomly initialized. Algorithm parameters such as the maximum number of iterations, inertia weight, individual learning factor, and social learning factor are set. Then, the iterative optimization process begins. For each particle in each iteration: 1) Convert the particle's current position vector Input the dual-pool coupled aeration prediction model, and the model outputs the predicted aeration rate of the first aeration pool under this control combination. Predicted aeration volume for the second aeration tank .

[0057] 2) Calculate the objective function value corresponding to this particle. .

[0058] 3) If the current objective function value is less than the historical best value of the particle, then update the individual's optimal position; if the current objective function value is less than the group's historical best value, then update the group's optimal position.

[0059] 4) Adjust the velocity of each particle according to the velocity update formula of the particle swarm algorithm, thereby updating the position of the particle and constraining the new position to the feasible region.

[0060] The iterative process continues until the maximum number of iterations is reached or the change in the population's optimal objective function value is less than a preset convergence threshold. After the iterations conclude, the population's optimal position vector... This refers to the combination of target control variables, used for issuing and executing commands.

[0061] During the joint optimization process, when the opening degrees of the first and second valves have reached the preset boundaries but still cannot meet the target aeration volume of the two tanks, the blower frequency compensation adjustment is triggered. That is, while keeping the valve openings unchanged, the blower frequency is increased or decreased individually until the predicted aeration volume approaches the target value. When the blower frequency meets the total aeration demand but there is still a deviation in the aeration distribution between the two tanks, the difference is corrected first through the first and second valves. That is, while keeping the blower frequency unchanged, only the opening degrees of the two valves are adjusted to redistribute the air volume. The boundary condition handling logic is executed in the online solution module as a verification and correction step after optimization.

[0062] Step 5: Combine the target control variables (Optimal control strategy) The PLC sends the control to the fan frequency converter, the first valve actuator, and the second valve actuator to complete the adjustment action according to the target value. After the adjustment is completed, the system enters the next control cycle.

[0063] Step 6: After each control execution, continuously collect the actual aeration volume of the first and second aeration tanks, compare it with the predicted value of the dual-tank coupled aeration prediction model under the corresponding control quantity, and calculate the prediction error. When the accumulated number of new samples reaches a preset threshold (e.g., 500 sets), trigger the model incremental update process, that is, perform several rounds of fine-tuning (e.g., 10 rounds) using new samples based on the existing model parameters. After the model update is evaluated and deemed qualified, it can replace the original online model, thereby ensuring that the prediction model is continuously optimized as the system runs and maintains stable prediction performance even when operating conditions fluctuate.

[0064] The method described in this embodiment was applied and verified in a wastewater treatment plant. Figure 3The diagram shows a comparison curve of the target aeration rate and the actual aeration rate in the dual-tank system during a typical time period. The target aeration rate for the first aeration tank is set to [value missing] in the first half of the diagram. The latter half was adjusted upwards. The target aeration rate for the second aeration tank is set at [value] in the first half. The latter half was adjusted upwards. As can be seen from the figure, the actual aeration rates of both pools closely track their respective target setpoints, with the tracking deviation remaining within a small range. When the target value undergoes a step change, the system can respond quickly and smoothly transition to the new target value, without significant overshoot or oscillation during the transition process. This indicates that the proposed method has good target tracking capability and regulation stability under dual-pool coupling conditions.

[0065] The coupling relationship between the two pools is explicitly represented using an LSTM machine learning model, and the coordinated adjustment of the blower frequency and the opening of the two valves is achieved through joint optimization using a particle swarm optimization method with a penalty term. This reduces operating energy consumption and equipment wear while meeting the independent aeration requirements of both pools, and demonstrates good adaptability during long-term operation.

[0066] The technical effects of this application are as follows: This application achieves stable decoupled control by accurately fitting the relationship between dual-pool aeration equipment and aeration volume using an LSTM dual-pool coupled aeration prediction model, combined with constraints with penalty terms for joint optimization. This reduces equipment adjustment frequency and operating energy consumption while meeting process requirements.

[0067] Example 2: This example provides a computer system, which can be a server or a terminal, and its internal structure diagram can be as follows. Figure 4 As shown, the computer system includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and databases. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media to run. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements the aforementioned methods.

[0068] Those skilled in the art will understand that Figure 4The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer system to which the present application is applied. A specific computer system may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0069] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0070] The processors involved in the various embodiments provided in this application may be general-purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited thereto.

[0071] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0072] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A coordinated adjustment method for a dual-pool aeration device based on LSTM, characterized in that, The dual-pool aeration equipment includes: a blower, a first valve, and a second valve; the blower is connected to an external first aeration pool and an external second aeration pool respectively through a first pipeline and a second pipeline; the first valve and the second valve are respectively installed on the first pipeline and the second pipeline; both the first valve and the second valve are configured to: under the condition that the total air intake of the blower is constant, adjust the air intake of their respective pipelines by controlling their opening degree, and indirectly adjust the air intake of the other pipeline in the opposite direction, the method including: Obtain historical operating data of the dual-tank aeration equipment; the historical operating data specifically includes: blower frequency, first valve opening degree, second valve opening degree, aeration volume of the first aeration tank and the aeration volume of the second aeration tank; Training samples were selected based on historical operating data. The fan frequency, the opening degree of the first valve and the opening degree of the second valve were used as inputs, and the aeration volume of the first aeration tank and the aeration volume of the second aeration tank were used as outputs to train a dual-tank coupled aeration prediction model. The dual-tank coupled aeration prediction model was constructed based on the LSTM model. Obtain the current fan frequency, the opening degree of the first valve and the opening degree of the second valve, and set the target aeration rate of the first aeration tank and the target aeration rate of the second aeration tank. Using an online solver module, the objective function for joint optimization is the weighted minimization of the aeration volume tracking deviation and the variation amplitude of the control quantity of the dual-pool aeration equipment. The control quantities of the blower, the first valve, and the second valve are used as optimization objects. Based on constraint terms, joint optimization is performed on the blower frequency, the opening degree of the first valve, and the opening degree of the second valve. During the joint optimization process, a dual-pool coupled aeration prediction model dynamically predicts the predicted aeration volume of the first and second aeration pools corresponding to each control strategy. The optimal control strategy is obtained through screening, and finally, the dual-pool aeration equipment is adaptively adjusted in real time according to the optimal control strategy. The aeration volume tracking deviation includes the aeration volume deviation of the first and second aeration pools. The aeration volume deviation of the first aeration pool is the difference between the predicted aeration volume and the target aeration volume of the first aeration pool; the aeration volume deviation of the second aeration pool is the difference between the predicted aeration volume and the target aeration volume of the second aeration pool.

2. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 1, characterized in that, Using an online solution module, the objective function for joint optimization is the weighted minimization of the aeration volume tracking deviation and the variation amplitude of the control quantity of the dual-pool aeration equipment. The control quantities of the blower, the first valve, and the second valve are used as optimization objects. Based on constraint terms, joint optimization is performed on the blower frequency, the opening degree of the first valve, and the opening degree of the second valve. During the joint optimization process, a dual-pool coupled aeration prediction model is used to dynamically predict the predicted aeration volume of the first and second aeration pools corresponding to each control strategy. The optimal control strategy is obtained through screening. Finally, after real-time adaptive adjustment of the dual-pool aeration equipment according to the optimal control strategy, the following steps are also included: The difference between the predicted and actual aeration volume is updated into the historical operating data, and the trained dual-pool coupled aeration prediction model is incrementally updated or periodically retrained.

3. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 1, characterized in that, The objective function expression for the joint optimization is as follows: ； In the formula, To jointly optimize the objective function; , , , and All are weighting coefficients; and The predicted aeration volume for the first aeration tank and the predicted aeration volume for the second aeration tank are respectively. and These are the target aeration rates for the first aeration tank and the second aeration tank, respectively. The frequency of the fan to be solved; The current frequency of the wind turbine; Let the opening degree of the first valve be the one to be solved. This represents the first valve opening at the current moment; Let the opening degree of the second valve be the one to be solved. This represents the current valve opening degree.

4. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 3, characterized in that, and The values ​​range from 1.0 to 10.0; , and The values ​​range from 0.1 to 1.

0.

5. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 1, characterized in that, The constraints include at least one of the following: upper and lower limits of fan frequency, upper and lower limits of first valve opening, and upper and lower limits of second valve opening. The expressions for the upper and lower limit constraints of the wind turbine frequency are as follows: ； In the formula, For the frequency of the fan, The lower limit of the fan frequency is set according to the surge threshold or minimum aeration requirement of the fan. The upper limit of the fan frequency is set according to the rated frequency of the fan. The expressions for the upper and lower limits of the opening of the first valve are as follows: ； In the formula, This represents the opening degree of the first valve. The expressions for the upper and lower limits of the second valve opening are as follows: ； In the formula, This refers to the opening degree of the second valve.

6. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 1, characterized in that, During the joint optimization process, when the opening of the first valve and the second valve reaches the preset boundary condition, but still cannot reach the target aeration volume of the first aeration tank and the target aeration volume of the second aeration tank, the blower frequency compensation adjustment is triggered; when the blower frequency meets the total aeration volume requirement of the target aeration volume of the first aeration tank and the target aeration volume of the second aeration tank, but there is still a deviation in the aeration distribution between the first aeration tank and the second aeration tank, the difference is corrected first through the first valve and the second valve.

7. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 1, characterized in that, The online solution module uses the particle swarm optimization algorithm.

8. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 1, characterized in that, In addition to the LSTM model, the dual-pool coupled aeration prediction model can also be constructed based on gradient boosting tree model, random forest model, support vector regression model, and neural network model.

9. The LSTM-based coordinated adjustment method for dual-pool aeration equipment according to claim 1, characterized in that, The dual-pool aeration equipment also includes: a first aeration volume monitoring device and a second aeration volume monitoring device; The first aeration volume monitoring device is installed between the first valve in the first pipeline and the external first aeration tank; the second aeration volume monitoring device is installed between the second valve in the second pipeline and the external second aeration tank; the first aeration volume monitoring device is used to obtain the aeration volume of the first aeration tank; the second aeration volume monitoring device is used to obtain the aeration volume of the second aeration tank.

10. A computer system, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the LSTM-based coordinated adjustment method for a dual-pool aeration device according to any one of claims 1-9.

Images