Highly intelligent sewage treatment control system and method based on gas change

Abstract

The invention discloses a highly intelligent sewage treatment control system and method based on gas change, and relates to the field of sewage treatment, and the system comprises a data collection module which collects sensor data and sewage treatment system operation parameters; the data processing module preprocesses the sensor data and calculates the gas yield of the gas product; the core control module determines the target aeration rate of each aeration point, the DO concentration in the future one hour, the aeration effect, the sewage quality parameter and the activated sludge quality parameter according to the preprocessed sensor data; and the instruction generation module converts the target aeration rate of each aeration point into the current fan frequency, adjusts the current fan frequency according to the DO concentration, the aeration effect, the sewage quality parameter and the activated sludge quality parameter in the next hour, and adjusts the fan rotating speed and the valve to control the aeration rate. The problems of nonlinearity and time lag in the sewage treatment process can be solved, collaborative optimization of multiple aeration points is achieved, and meanwhile it is ensured that the effluent quality stably reaches the standard.

Description

Technical Field

[0001] This application relates to the field of wastewater treatment, and in particular to a highly intelligent wastewater treatment control system and method based on gas changes. Background Technology

[0002] With development, the pressure of urban water scarcity is increasing, and people are increasingly aware that the social cycle of water exceeds the carrying capacity of the natural water cycle. Wastewater treatment and reuse is a strategic choice of great significance. Urban wastewater treatment and reuse reduces wastewater discharge and lowers the city's sewage load while replacing clean water sources. It has the characteristics of stable water supply, short water transmission distance, and low water production cost, and can provide a safe and reliable alternative water source, making it a strategic choice for solving urban water shortage problems.

[0003] Traditional aeration control relies on proportional-integral-derivative (PID) control algorithms, which cannot adapt to the strong nonlinearity and time lag (treatment cycle delay of several hours) of wastewater treatment. This results in low dissolved oxygen (DO) control accuracy (error > 0.5 mg / L), leading to over- or under-aeration, and indirect carbon emissions accounting for up to 53.8%. Traditional aeration control causes large fluctuations in effluent quality. When the influent load changes (such as chemical oxygen demand (COD) fluctuations of ±30% during heavy rain), a fixed hydraulic retention time (HRT) strategy (typically 10¹² hours) increases the risk of ammonia nitrogen and total nitrogen (TN) concentrations exceeding standards. Existing control models based on DO or respiration rate do not integrate multi-parameter coupling relationships (such as pH and gas concentration), making it difficult to achieve global optimization.

[0004] Therefore, given the shortcomings of existing technologies, there is an urgent need to provide a new wastewater treatment control system that can solve the nonlinearity and time delay problems in the wastewater treatment process, achieve collaborative optimization of multiple aeration points, and ensure that the effluent water quality consistently meets the standards. Summary of the Invention

[0005] The purpose of this application is to provide a highly intelligent wastewater treatment control system and method based on gas change, which can solve nonlinear and time delay problems in the wastewater treatment process, achieve collaborative optimization of multiple aeration points, and ensure that the effluent water quality consistently meets the standards.

[0006] To achieve the above objectives, this application provides the following solution:

[0007] In a first aspect, this application provides a highly intelligent wastewater treatment control system based on gas change, including: a data acquisition module, a data processing module, a core control module, an instruction generation module, and an actuator;

[0008] The data acquisition module is used to collect sensor data and operating parameters of the wastewater treatment system. The sensor data includes: pH value in the anaerobic tank, pH value in the anoxic tank, pH value in the aerobic tank, DO concentration in the aerobic tank, and concentration of gaseous products. The concentration of gaseous products includes: concentrations of CH4, NH3, CO2, CO, O2, N2, and N2O. The operating parameters of the wastewater treatment system include: wastewater temperature, wastewater hydraulic retention time, activated sludge retention time, and activated sludge concentration.

[0009] The data processing module is connected to the data acquisition module; the data processing module is used to receive sensor data, preprocess the sensor data, and obtain the gas yield of the gas products based on the concentration of gas products in the preprocessed sensor data.

[0010] The core control module is connected to the data processing module. The core control module includes: a multi-input multi-output model based on DBN and MLP networks; a dynamic prediction model for a single aeration point based on LSTM and Bi-LSTM models; a global aeration optimization model based on SENet and CNN networks; and a gas yield simulation model. The multi-input multi-output model is used to determine the target aeration rate for each aeration point based on preprocessed sensor data. The dynamic prediction model for a single aeration point is used to predict the DO concentration for the next hour based on preprocessed sensor data. The global aeration optimization model is used to determine the aeration effect based on the time-series image obtained from the preprocessed sensor data. The gas yield simulation model is used to determine the wastewater quality parameters and activated sludge sludge parameters based on the gas yield of the gas products and the operating parameters of the wastewater treatment system. The aeration effect is a prediction of the aeration volume.

[0011] The instruction generation module is connected to the core control module; the instruction generation module is used to convert the target aeration rate of each aeration point into the current blower frequency, and then adjust the current blower frequency according to the DO concentration, aeration effect, wastewater quality parameters and activated sludge sludge parameters for the next hour, and generate control instructions.

[0012] The actuator is connected to the instruction generation module; the actuator is used to adjust the fan speed and valve according to the control instructions to control the aeration volume.

[0013] Optionally, the data acquisition module includes: a sensor module and a wastewater quality real-time monitoring module;

[0014] The sensor module is connected to the data processing module; the sensor module is used to collect sensor data.

[0015] The wastewater quality real-time monitoring module is used to collect operating parameters of the wastewater treatment system.

[0016] Optionally, the sensor module includes: a pH sensor, a DO sensor, and a gas concentration sensor;

[0017] The pH sensors are placed in the anaerobic tank, the anoxic tank, and the aerobic tank, respectively; the pH sensors collect the pH values ​​in the anaerobic tank, the anoxic tank, and the aerobic tank, respectively;

[0018] The DO sensor is placed in the aerobic tank; the DO sensor is used to collect the DO concentration in the aerobic tank.

[0019] The gas concentration sensors are placed in the anaerobic tank, the anoxic tank, and the aerobic tank, respectively; the gas concentration sensors are used to collect the concentration of gaseous products.

[0020] Optionally, the pH sensor has a range of 6.0-9.0 and an accuracy of ±0.1; the DO sensor has a range of 0 mg / L-10 mg / L and an accuracy of ±0.1 mg / L; the gas concentration sensor has a range of 0 ppm-1000 ppm; and the sampling frequency of the pH sensor, DO sensor, and gas concentration sensor is 1 time / minute.

[0021] Optionally, the data processing module includes: a standardization unit and a normalization unit;

[0022] The standardization unit is connected to the pH sensor; the standardization unit is used to standardize the pH values ​​collected in the anaerobic, anoxic, and aerobic tanks.

[0023] The normalization unit is connected to the DO sensor and the gas concentration sensor; the normalization unit is used to normalize the DO concentration and the concentration of gaseous products in the aerobic tank; and uses the formula Determine the gas yield of the gaseous products;

[0024] Where W(i) represents the gas yield of the i-th gaseous product in the reactor; wsensor(i) represents the concentration of the i-th gaseous product in the reactor; V1 represents the volume of the upper space of the reactor; wT(i) represents the solubility of the i-th gaseous product in water at temperature T; V2 represents the volume of the mixture in the reactor; Δt represents the time interval between two consecutive measurements; MLSS represents the concentration of suspended solids in the mixture; and f represents the ratio of the concentration of volatile suspended solids to the concentration of suspended solids in the mixture.

[0025] Optionally, the multiple-input multiple-output model specifically includes:

[0026] The target aeration rate for each aeration point is determined using the formulas Hidden1=ReLU(W1*X+b1), Hidden2=ReLU(W2*Hidden1+b2) and Output=W3*Hidden2+b3.

[0027] Where W1, W2, and W3 are weight matrices, b1, b2, and b3 are bias terms, ReLU is the activation function, X is the preprocessed sensor data, Hidden1 represents the underlying dynamics of the aeration system, Hidden2 represents the high-order coupling law of the aeration system, and Output represents the target aeration rate of each aeration point.

[0028] Optionally, the instruction generation module specifically includes:

[0029] Using formula f 风机 =20+0.4*v converts the target aeration rate at each aeration point into the fan frequency f. 风机 Where v is the target aeration rate at each aeration point.

[0030] Optionally, the actuator is a variable frequency aeration fan.

[0031] Optionally, the sensor data is wirelessly transmitted to the data processing module via RS-485 or LoRa.

[0032] Secondly, this application provides a highly intelligent wastewater treatment control method based on gas change, wherein the highly intelligent wastewater treatment control method based on gas change is applied to any of the highly intelligent wastewater treatment control systems based on gas change described in the previous application, and the highly intelligent wastewater treatment control method based on gas change includes:

[0033] Acquire sensor data and wastewater treatment system operating parameters; the sensor data includes: pH value in the anaerobic tank, pH value in the anoxic tank, pH value in the aerobic tank, DO concentration in the aerobic tank, and concentration of gaseous products; the concentration of gaseous products includes: concentrations of CH4, NH3, CO2, CO, O2, N2, and N2O; the wastewater treatment system operating parameters include: wastewater temperature, hydraulic retention time of wastewater, activated sludge retention time, and activated sludge concentration;

[0034] The sensor data is preprocessed, and the gas yield of the gas products is obtained based on the concentration of the gas products in the preprocessed sensor data.

[0035] Based on the preprocessed sensor data, the target aeration rate of each aeration point is determined by a multi-input multi-output model constructed using DBN and MLP networks.

[0036] Based on the preprocessed sensor data, a dynamic prediction model for a single aeration point, constructed using LSTM and Bi-LSTM models, predicts the DO concentration for the next hour.

[0037] Based on the time-series images obtained from the preprocessed sensor data, a global aeration optimization model constructed using SENet and CNN networks is used to determine the aeration effect.

[0038] Based on the gas yield of the gas products and the operating parameters of the wastewater treatment system, wastewater quality parameters and activated sludge sludge parameters are determined using a gas yield simulation model.

[0039] The target aeration rate of each aeration point is converted into the current blower frequency. Then, the current blower frequency is adjusted according to the DO concentration, aeration effect, wastewater quality parameters and activated sludge sludge parameters for the next hour, and control commands are generated.

[0040] Adjust the fan speed and valves according to control commands to control the aeration volume.

[0041] According to the specific embodiments provided in this application, this application has the following technical effects:

[0042] This application provides a highly intelligent wastewater treatment control system and method based on gas change. The system collects sensor data and wastewater treatment system operating parameters through a data acquisition module; standardizes and normalizes the sensor data through a data processing module, and calculates the gas yield of gaseous products based on the pre-processed sensor data; calculates the target aeration rate for each aeration point using a multi-input multi-output model, and then converts the target aeration rate of each aeration point into the current blower frequency. The aeration volume is controlled by controlling the blower speed and valves; a dynamic prediction model for a single aeration point predicts the DO concentration for the next hour, and the blower frequency is adjusted based on the DO concentration for the next hour to control the blower speed, thereby shortening the response time and dynamically balancing energy consumption and treatment effect; a global aeration optimization model obtains the aeration effect, enabling prediction of the aeration volume and coordinating the operation of multiple aeration points to avoid energy waste caused by local over-aeration or under-aeration; a gas yield simulation model determines wastewater quality parameters and activated sludge sludge parameters; and the blower is adjusted based on the aeration effect, wastewater quality parameters, and activated sludge sludge parameters to control the aeration volume. This application can solve the nonlinear and time-delay problems in the wastewater treatment process, achieve synergistic optimization of multiple aeration points, and ensure that the effluent quality consistently meets the standards. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, 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.

[0044] Figure 1 This is a schematic diagram of a highly intelligent wastewater treatment control system based on gas change in one embodiment of this application;

[0045] Figure 2 This is a schematic diagram of the process of a highly intelligent wastewater treatment control system method based on gas changes according to this application. Detailed Implementation

[0046] 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.

[0047] 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.

[0048] In one exemplary embodiment, such as Figure 1 As shown, a highly intelligent wastewater treatment control system based on gas change is provided. The system includes: a data acquisition module, a data processing module, a core control module, an instruction generation module, and an actuator.

[0049] The data acquisition module includes a sensor module and a wastewater quality real-time monitoring module. The sensor module is used to collect sensor data, and the wastewater quality real-time monitoring module is used to collect the operating parameters of the wastewater treatment system. The sensor data includes: pH value in the anaerobic tank, pH value in the anoxic tank, pH value in the aerobic tank, DO concentration in the aerobic tank, and concentration of gaseous products. The concentration of gaseous products includes: concentrations of CH4, NH3, CO2, CO, O2, N2, and N2O. The operating parameters of the wastewater treatment system include: wastewater temperature, hydraulic retention time of wastewater, activated sludge retention time (ASRT), and activated sludge concentration.

[0050] In an exemplary real-time example, the sensor module includes a pH sensor, a DO sensor, and a gas concentration sensor. The pH sensor is placed in the anaerobic, anoxic, and aerobic tanks respectively, to collect the pH values ​​in the anaerobic, anoxic, and aerobic tanks; the DO sensor is placed in the aerobic tank to collect the DO concentration in the aerobic tank; the gas concentration sensor is placed in the anaerobic, anoxic, and aerobic tanks respectively; the gas concentration sensor is used to collect the concentration of gaseous products.

[0051] Specifically, the pH sensor has a range of 6.0-9.0 and an accuracy of ±0.1; the DO sensor has a range of 0 mg / L-10 mg / L and an accuracy of ±0.1 mg / L; the gas concentration sensor has a range of 0 ppm-1000 ppm; and the sampling frequency of the pH sensor, DO sensor, and gas concentration sensor is 1 time / minute.

[0052] The data processing module is connected to the data acquisition module and is used to receive sensor data, preprocess the sensor data, and obtain the gas yield of the gas products based on the concentration of the preprocessed gas products. The sensor data is wirelessly transmitted to the data processing module via RS-485 or LoRa.

[0053] Specifically, the data processing module includes a standardization unit and a normalization unit. The standardization unit is connected to the pH sensor and is used to standardize the pH values ​​collected in the anaerobic, anoxic, and aerobic tanks. The normalization unit is connected to the DO sensor and the gas concentration sensor and is used to normalize the DO concentration and gaseous product concentration in the aerobic tank.

[0054] Specifically, the Z-score (standard score) standardization process is as follows: The pH value is standardized using the formula PH_std = (PH_raw - μ) / σ, where μ = 7.0 and σ = 0.5. The Min-Max normalization (deviation standardization) process is as follows: DO and gas concentrations are normalized using the formula DO_norm = (DO_raw - DO_min) / (DO_max - DO_min), where DO_min = 0 and DO_max = 8 mg / L. Here, PH_std represents the standardized pH value, PH_raw represents the pH value collected by the pH sensor, μ represents the average pH value, σ represents the standard deviation of the pH value, DO_norm represents the normalized DO concentration, DO_raw represents the DO concentration collected by the DO sensor, DO_min represents the minimum DO concentration, and DO_max represents the maximum DO concentration.

[0055] The data processing module uses formulas Determine the gas yield of the gaseous products; where W(i) represents the gas yield of the i-th gaseous product in the reactor; wsensor(i) represents the concentration of the i-th gaseous product in the reactor; V1 represents the volume of the upper space of the reactor; wT(i) represents the solubility of the i-th gaseous product in water at temperature T; V2 represents the volume of the mixture in the reactor; Δt represents the time interval between two consecutive measurements; MLSS represents the concentration of suspended solids in the mixture; and f represents the ratio of the concentration of volatile suspended solids to the concentration of suspended solids in the mixture.

[0056] In an exemplary embodiment, the concentration of gaseous products in the anaerobic and aerobic tanks is determined using the formula... The gas yield of gaseous products in the anaerobic and aerobic tanks is calculated. The gas yield of gaseous products in the anaerobic and aerobic tanks, along with the operating parameters of the wastewater treatment system, are input into the gas yield simulation model to obtain the wastewater quality parameters and activated sludge sludge quality parameters in the anaerobic and aerobic tanks. Based on these parameters, the required speed of the blower is determined and adjusted accordingly.

[0057] The core control module is connected to the data processing module; the core control module includes: a multi-input multi-output model based on DBN (Deep Belief Network) and MLP (Multilayer Perceptron Network), a dynamic prediction model for a single aeration point based on LSTM (Long Short-Term Memory Artificial Neural Network) and Bi-LSTM (Bidirectional Long Short-Term Memory Artificial Neural Network) models, a global aeration optimization model based on SENet (SE Network) and CNN (Convolutional Neural Network), and a gas yield simulation model;

[0058] The multi-input multi-output model determines the target aeration rate for each aeration point based on the preprocessed sensor data using the formulas Hidden1 = ReLU(W1*X+b1), Hidden2 = ReLU(W2*Hidden1+b2), and Output = W3*Hidden2+b3. Here, W1, W2, and W3 are weight matrices, b1, b2, and b3 are bias terms, ReLU is the activation function, X is the preprocessed sensor data, Hidden1 represents the underlying dynamics of the aeration system, Hidden2 represents the higher-order coupling rules of the aeration system, and Output represents the target aeration rate for each aeration point.

[0059] The single aeration point dynamic prediction model is used to predict the DO concentration for the next hour based on the preprocessed sensor data; the global aeration optimization model is used to determine the aeration effect based on the time-series image obtained from the preprocessed sensor data.

[0060] The gas yield simulation model is used to determine wastewater quality parameters and activated sludge sludge parameters based on the gas yield of the gas products and the operating parameters of the wastewater treatment system; the aeration effect is a prediction of the aeration volume.

[0061] In an exemplary embodiment, the network structure of the multi-input multi-output model contains an input layer with 12 nodes, 9 of which correspond to pH value, DO concentration, and gaseous product concentration, respectively. The three fully connected hidden layers in the network structure of the multi-input multi-output model have 512, 256, and 128 nodes, respectively, with the ReLU activation function and a dropout rate of 0.2. In the output layer with N nodes, N represents the number of aeration points. The network structure of the multi-input multi-output model also includes a linear activation function. The target aeration rate (0-100%) of each aeration point is then output according to the output layer of the multi-input multi-output model network structure.

[0062] The input layer of the network structure of the dynamic prediction model for a single aeration point is used to input 9-dimensional data (pH, DO, and concentration of gaseous products) within a 24-hour time window. The LSTM layer in the network structure includes 64 nodes with a time step of 24 and a tanh activation function. The output layer of the network structure has one node (used to predict the DO concentration for the next hour) and a linear activation function. The formulas h_t,c_t=LSTM(x_t,h_{t-1},c_{t-1}) and DO_pred=W*h_t+b are used to predict the DO concentration for the next hour. Here, h_t represents the hidden state, c_t represents the cell state, LSTM indicates training via a Long Short-Term Memory network, x_t represents the input data at time step t, t-1 represents the previous time step, DO_pred represents the predicted DO concentration for the next hour, W represents the inertial weight, and b represents the bias term.

[0063] The input layer of the global aeration optimization model's network structure is used to input the time-series image; the time-series image is determined based on the set 5 aeration points × 9 parameters × T time step; the convolutional layer of the global aeration optimization model's network structure includes: 3×3 convolutional kernels and output channel number C = 5 × 9 = 45; the SE module parameters of the global aeration optimization model's network structure are set as follows: compression ratio Ratio = 16, Cmid = 45 / 16 ≈ 3 (actually rounded to 3 or 4), the first FC (fully connected) layer is R45→R3 with ReLU activation function, the second FC layer is R3→R with Sigmoid activation function; thus realizing the generation of aeration effects; the fully connected layer of the global aeration optimization model's network structure is used to output each aeration effect.

[0064] The instruction generation module is connected to the core control module; the instruction generation module generates instructions according to formula f. 风机 =20+0.4*v converts the target aeration rate at each aeration point into the current fan frequency f. 风机 The frequency range is 20Hz-60Hz; where v is the target aeration rate at each aeration point. The current blower frequency is adjusted based on the DO concentration, aeration effect, wastewater quality parameters, and activated sludge parameters for the next hour, generating control commands. If the predicted DO concentration is lower than the set value (2.5mg / L), the frequency is increased proportionally (increment Δf = 5Hz / 0.5mg / L deviation). Real-time DO concentration data is transmitted back to the core control module, triggering model parameter updates, which are updated every 6 hours.

[0065] The actuator is connected to the instruction generation module; the actuator is used to adjust the fan speed and valves according to control instructions to control the aeration volume. Specifically, by judging the positive or negative value of the aeration volume, the operation of the fan and valves is dynamically adjusted. When the aeration volume > 0, the total air volume is gradually reduced, and the valves of the aeration module are closed in sequence at the inlet end; when the aeration volume < 0, the closed valves are opened in sequence at the outlet end to optimize the airflow distribution. In an exemplary embodiment, the actuator is a variable frequency aeration fan.

[0066] In another exemplary embodiment, such as Figure 2 As shown, a highly intelligent wastewater treatment control method based on gas change is provided, including the following steps S101 to S108. Wherein:

[0067] S101: Acquire sensor data and wastewater treatment system operating parameters; the sensor data includes: pH value in the anaerobic tank, pH value in the anoxic tank, pH value in the aerobic tank, DO concentration in the aerobic tank, and concentration of gaseous products; the concentration of gaseous products includes: concentrations of CH4, NH3, CO2, CO, O2, N2, and N2O; the wastewater treatment system operating parameters include: wastewater temperature, wastewater hydraulic retention time, activated sludge retention time, and activated sludge concentration.

[0068] S102: Preprocess the sensor data; and obtain the gas yield of the gas products based on the concentration of the gas products in the preprocessed sensor data.

[0069] S103: Based on the preprocessed sensor data, the target aeration rate of each aeration point is determined using a multi-input multi-output model constructed using DBN and MLP networks.

[0070] S104: Based on the preprocessed sensor data, a dynamic prediction model for a single aeration point, constructed using LSTM and Bi-LSTM models, predicts the DO concentration for the next hour.

[0071] S105: Based on the time-series images obtained from the preprocessed sensor data, the global aeration optimization model constructed using SENet and CNN networks determines the aeration effect;

[0072] S106: Based on the gas yield of the gas products and the operating parameters of the wastewater treatment system, determine the wastewater quality parameters and activated sludge sludge quality parameters using a gas yield simulation model.

[0073] S107: Convert the target aeration rate of each aeration point into the current blower frequency, and then adjust the current blower frequency based on the DO concentration, aeration effect, wastewater quality parameters and activated sludge sludge parameters for the next hour, and generate control commands.

[0074] S108: Adjusts the fan speed and valves according to control commands to control the aeration volume. When the aeration volume > 0, gradually reduces the total air volume and closes the valves of the aeration module in sequence at the water inlet end; when the aeration volume < 0, opens the closed valves in sequence at the water outlet end to optimize the airflow distribution.

[0075] This application proposes intelligent algorithm fusion: combining PSO (Particle Swarm Optimization) + BPNN (Backpropagation Neural Network) adaptive control with LSTM time-series prediction to solve nonlinear and time-delay problems; multi-dimensional data integration: constructing a time-series image (SENet+CNN) and a multi-input-multi-output model based on DBN and MLP networks to achieve collaborative optimization of multiple aeration points; and a dynamic feedback mechanism: dynamically adjusting HRT and aeration strategies based on GBDT (Gradient Boosting Decision Tree) and SVR (Support Vector Regression) models to predict effluent quality (GBDT AUC = 0.92) and ensure stable compliance.

[0076] This application employs a PSO+BPNN adaptive control algorithm to dynamically adjust the aeration rate, overcoming the lag and nonlinearity problems of traditional PID control. A global aeration optimization model coordinates the operation of multiple aeration points, avoiding energy waste caused by local over-aeration or under-aeration. By optimizing dissolved oxygen control in the aeration system, the total greenhouse gas yield is significantly reduced by approximately 20%, and energy consumption is reduced by approximately 15%, achieving both greenhouse gas emission reduction and energy consumption reduction. A multi-input multi-output model predicts and adjusts the aeration strategy in real time to adapt to dynamic changes in influent water quality. Feedback control combined with a gas yield simulation model adjusts process parameters promptly. The fluctuation range of effluent COD, ammonia nitrogen, and TN concentrations is reduced by 30%, and the compliance rate is increased to over 98%, providing an indication of effluent water quality stability. An LSTM / BiLSTM model captures the time-series dependence of dissolved oxygen, predicting trends for the next 16 hours. Adaptive Dynamic Programming (ADP) enables intelligent decision-making, dynamically balancing energy consumption and treatment efficiency. This improves DO control accuracy (RMSE - root mean square error) to within ±0.2 mg / L and reduces response time to within 10 minutes. Control accuracy and response speed are optimized.

[0077] The HRT (Hydraulic Retention Time) adjustment range for the aeration system is 12-15 hours (optimal value 15 hours, COD (Chemical Oxygen Demand) removal rate >90%, minimum greenhouse gas yield 804.9 kg CO2 / d). The PSO algorithm parameters are set as follows: particle number 50, iteration count 100, learning factors c1 = 2.0, c2 = 2.0, and inertia weight w = linearly decreasing from 0.9 to 0.4. The BPNN (Backpropagation Neural Network) structure includes: a 9-node input layer, 3 hidden layers (128 / 64 / 32 nodes) containing Sigmoid activation functions, and a 1-node output layer (DO setting).

[0078] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0079] 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 highly intelligent wastewater treatment control system based on gas change, characterized in that, The highly intelligent wastewater treatment control system based on gas change includes: a data acquisition module, a data processing module, a core control module, an instruction generation module, and an actuator; The data acquisition module is used to collect sensor data and operating parameters of the wastewater treatment system. The sensor data includes: pH value in the anaerobic tank, pH value in the anoxic tank, pH value in the aerobic tank, DO concentration in the aerobic tank, and concentration of gaseous products. The concentration of gaseous products includes: concentrations of CH4, NH3, CO2, CO, O2, N2, and N2O. The operating parameters of the wastewater treatment system include: wastewater temperature, wastewater hydraulic retention time, activated sludge retention time, and activated sludge concentration. The data processing module is connected to the data acquisition module; the data processing module is used to receive sensor data, preprocess the sensor data, and obtain the gas yield of the gas products based on the concentration of gas products in the preprocessed sensor data. The core control module is connected to the data processing module. The core control module includes: a multi-input multi-output model based on DBN and MLP networks; a dynamic prediction model for a single aeration point based on LSTM and Bi-LSTM models; a global aeration optimization model based on SENet and CNN networks; and a gas yield simulation model. The multi-input multi-output model is used to determine the target aeration rate for each aeration point based on preprocessed sensor data. The dynamic prediction model for a single aeration point is used to predict the DO concentration for the next hour based on preprocessed sensor data. The global aeration optimization model is used to determine the aeration effect based on the time-series image obtained from the preprocessed sensor data. The gas yield simulation model is used to determine the wastewater quality parameters and activated sludge sludge parameters based on the gas yield of the gas products and the operating parameters of the wastewater treatment system. The aeration effect is a prediction of the aeration volume. The instruction generation module is connected to the core control module; the instruction generation module is used to convert the target aeration rate of each aeration point into the current blower frequency, and then adjust the current blower frequency according to the DO concentration, aeration effect, wastewater quality parameters and activated sludge sludge parameters for the next hour, and generate control instructions. The actuator is connected to the instruction generation module; the actuator is used to adjust the fan speed and valve according to the control instructions to control the aeration volume.

2. The highly intelligent wastewater treatment control system based on gas change according to claim 1, characterized in that, The data acquisition module includes: a sensor module and a wastewater quality real-time monitoring module; The sensor module is connected to the data processing module; the sensor module is used to collect sensor data. The wastewater quality real-time monitoring module is used to collect operating parameters of the wastewater treatment system.

3. The highly intelligent wastewater treatment control system based on gas change according to claim 2, characterized in that, The sensor module includes: a pH sensor, a DO sensor, and a gas concentration sensor; The pH sensors are placed in the anaerobic tank, the anoxic tank, and the aerobic tank, respectively; the pH sensors collect the pH values ​​in the anaerobic tank, the anoxic tank, and the aerobic tank, respectively; The DO sensor is placed in the aerobic tank; the DO sensor is used to collect the DO concentration in the aerobic tank. The gas concentration sensors are placed in the anaerobic tank, the anoxic tank, and the aerobic tank, respectively; the gas concentration sensors are used to collect the concentration of gaseous products.

4. The highly intelligent wastewater treatment control system based on gas change according to claim 3, characterized in that, The pH sensor has a range of 6.0-9.0 and an accuracy of ±0.1; the DO sensor has a range of 0 mg / L-10 mg / L and an accuracy of ±0.1 mg / L; the gas concentration sensor has a range of 0 ppm-1000 ppm; and the sampling frequency of the pH sensor, DO sensor, and gas concentration sensor is 1 time / minute.

5. The highly intelligent wastewater treatment control system based on gas change according to claim 3, characterized in that, The data processing module includes: a standardization unit and a normalization unit; The standardization unit is connected to the pH sensor; the standardization unit is used to standardize the pH values ​​collected in the anaerobic, anoxic, and aerobic tanks. The normalization unit is connected to the DO sensor and the gas concentration sensor; the normalization unit is used to normalize the DO concentration and the concentration of gaseous products in the aerobic tank; and uses the formula Determine the gas yield of the gaseous products; Where W(i) represents the gas yield of the i-th gaseous product in the reactor; wsensor(i) represents the concentration of the i-th gaseous product in the reactor; V1 represents the volume of the upper space of the reactor; wT(i) represents the solubility of the i-th gaseous product in water at temperature T; V2 represents the volume of the mixture in the reactor; Δt represents the time interval between two consecutive measurements; MLSS represents the concentration of suspended solids in the mixture; and f represents the ratio of the concentration of volatile suspended solids to the concentration of suspended solids in the mixture.

6. The highly intelligent wastewater treatment control system based on gas change according to claim 1, characterized in that, The multi-input multi-output model specifically includes: The target aeration rate for each aeration point is determined using the formulas Hidden1=ReLU(W1*X+b1), Hidden2=ReLU(W2*Hidden1+b2) and Output=W3*Hidden2+b3. Where W1, W2, and W3 are weight matrices, b1, b2, and b3 are bias terms, ReLU is the activation function, X is the preprocessed sensor data, Hidden1 represents the underlying dynamics of the aeration system, Hidden2 represents the high-order coupling law of the aeration system, and Output represents the target aeration rate of each aeration point.

7. The highly intelligent wastewater treatment control system based on gas change according to claim 1, characterized in that, The instruction generation module specifically includes: Using formula f 风机 =20+0.4*v converts the target aeration rate at each aeration point into the fan frequency f. 风机 Where v is the target aeration rate at each aeration point.

8. The highly intelligent wastewater treatment control system based on gas change according to claim 1, characterized in that, The actuator is a variable frequency aeration blower.

9. The highly intelligent wastewater treatment control system based on gas change according to claim 1, characterized in that, The sensor data is wirelessly transmitted to the data processing module via RS-485 or LoRa.

10. A highly intelligent wastewater treatment control method based on gas change, characterized in that, The highly intelligent wastewater treatment control method based on gas change is applied to the highly intelligent wastewater treatment control system based on gas change according to any one of claims 1-9, wherein the highly intelligent wastewater treatment control method based on gas change includes: Acquire sensor data and wastewater treatment system operating parameters; the sensor data includes: pH value in the anaerobic tank, pH value in the anoxic tank, pH value in the aerobic tank, DO concentration in the aerobic tank, and concentration of gaseous products; the concentration of gaseous products includes: concentrations of CH4, NH3, CO2, CO, O2, N2, and N2O; the wastewater treatment system operating parameters include: wastewater temperature, hydraulic retention time of wastewater, activated sludge retention time, and activated sludge concentration; The sensor data is preprocessed; and the gas yield of the gas products is obtained based on the concentration of the gas products in the preprocessed sensor data. Based on the preprocessed sensor data, the target aeration rate of each aeration point is determined by a multi-input multi-output model constructed using DBN and MLP networks. Based on the preprocessed sensor data, a dynamic prediction model for a single aeration point, constructed using LSTM and Bi-LSTM models, predicts the DO concentration for the next hour. Based on the time-series images obtained from the preprocessed sensor data, a global aeration optimization model constructed using SENet and CNN networks is used to determine the aeration effect. Based on the gas yield of the gas products and the operating parameters of the wastewater treatment system, wastewater quality parameters and activated sludge sludge parameters are determined using a gas yield simulation model. The target aeration rate of each aeration point is converted into the current blower frequency. Then, the current blower frequency is adjusted according to the DO concentration, aeration effect, wastewater quality parameters and activated sludge sludge parameters for the next hour, and control commands are generated. Adjust the fan speed and valves according to control commands to control the aeration volume.

Images