Accurate control method and system for denitrification of sewage treatment plant and electronic equipment

Abstract

The invention relates to the technical field of accurate control scheme design for denitrification of a sewage treatment plant, in particular to an accurate control method and system for denitrification of a sewage treatment plant and electronic equipment. Data of carbon source concentration, dissolved oxygen concentration, pH value, temperature, alkalinity and nitrate concentration in sewage are collected in real time, parameters of a prediction model are dynamically corrected in combination with a machine learning algorithm and a mechanism model, and the denitrification rate is predicted. The carbon source adding amount and the dissolved oxygen target value are calculated according to the prediction result and the emission standard, the distribution of the carbon source and the dissolved oxygen in the reaction tank is monitored in real time through a distributed sensor network, and the operation states of a carbon source adding pump and aeration equipment are accurately adjusted through feedback control. According to the method, a pH value and alkalinity are taken as constraint conditions to cooperatively control a carbon source feeding rate and a dissolved oxygen concentration, and a control strategy is dynamically optimized through deep Q learning based on historical data and operation conditions, so that cross-cycle self-adaptive adjustment is realized, and accurate control on a denitrification process of a sewage treatment plant is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of precise control scheme design of wastewater treatment plant denitrification, and particularly relates to a precise control method and system for wastewater treatment plant denitrification and an electronic device. BACKGROUND

[0002] Wastewater treatment is an important link for environmental protection and water resource conservation, and the denitrification process plays a key role in removing nitrogen in wastewater. In the traditional wastewater treatment process, carbon source addition and dissolved oxygen control are mostly dependent on empirical parameters and fixed operation modes, which are difficult to adapt to the dynamic changes of influent load, water quality and environmental conditions, resulting in problems such as carbon source waste, low nitrate removal efficiency, pH and alkalinity fluctuations, etc. In the prior art, although automatic control methods based on physical models have appeared, due to the insufficient adaptability to complex dynamic processes, there are still defects such as low precision and slow response. Machine learning technology is gradually increasing in the application of industrial process optimization, which predicts system behavior by analyzing historical operation data, but a single machine learning method is difficult to cover the complex biochemical reaction mechanism. In addition, the existing system lacks real-time collaborative control of key parameters such as pH and alkalinity, and does not fully consider the integrated management of equipment state and control strategy. Therefore, there is an urgent need for a precise control scheme that integrates multi-source data, intelligent prediction models and multi-objective optimization algorithms to improve carbon source utilization efficiency, ensure water quality, and enhance the adaptive ability of the system.

[0003] Therefore, the prior art still needs further development. SUMMARY

[0004] The purpose of the present application is to overcome the above technical deficiencies and provide a precise control method and system for wastewater treatment plant denitrification and an electronic device to solve the problems existing in the prior art.

[0005] To achieve the above technical purpose, according to the first aspect of the present application, the present application provides a precise control method for wastewater treatment plant denitrification, comprising: S1: Real-time acquisition of carbon source concentration, dissolved oxygen concentration, pH value, temperature, alkalinity and nitrate concentration data in wastewater; S2: Constructing a dynamic denitrification rate prediction model based on multi-source data, and dynamically correcting model parameters by machine learning algorithm combined with process parameter changes; S3: According to the output of the prediction model, combined with the influent load of the wastewater treatment plant and the discharge standard, the carbon source addition amount and the dissolved oxygen control target value in the current stage are calculated; S4: Using a distributed sensor network to perform multi-point real-time monitoring of carbon source and dissolved oxygen distribution in the reaction tank, and adjusting the operation state of the carbon source addition pump and the aeration equipment through feedback control; S5: In the multi-parameter collaborative control stage, the carbon source dosing rate and the dissolved oxygen concentration are dynamically adjusted with pH value and alkalinity as the constraint conditions; S6: Based on historical data and current operating conditions, the control strategy is optimized through a reinforcement learning algorithm for cross-cycle adaptive adjustment; S7: Upload the control data to the cloud platform.

[0006] Specifically, the dynamic denitrification rate prediction model adopts a combination of LSTM neural network and mechanism model, the input layer includes water quality parameters, temperature and historical operation data, and the output layer is the carbon source consumption rate and nitrate degradation rate.

[0007] Specifically, the carbon source dosing pump is controlled by frequency conversion and multi-stage distribution valve to ensure uniform distribution of carbon source in the reaction tank.

[0008] Specifically, the dissolved oxygen control target value is set according to the partition of the reaction tank, the target value of the anoxic zone is 0.2-0.5mg / L, and the target value of the anaerobic zone is lower than 0.2mg / L.

[0009] Specifically, in the multi-parameter collaborative control, the pH value is adjusted by the alkali dosing pump, and the alkali concentration and dosing rate are dynamically matched according to the nitrate degradation rate.

[0010] Specifically, the reinforcement learning algorithm adopts deep Q learning, and the reward function is defined as the weighted sum of carbon source saving rate and nitrate removal rate.

[0011] Specifically, the reward function includes: wherein, is the carbon source saving rate, is the nitrate removal rate, DO is the current reaction tank dissolved oxygen concentration, is the target reaction tank dissolved oxygen concentration.

[0012] Specifically, the distributed sensor network includes optical sensors, electrochemical sensors and ultrasonic sensors, which are used to detect carbon source concentration, dissolved oxygen and sludge concentration respectively.

[0013] According to the second aspect of the present application, a precise control system for denitrification of a sewage treatment plant is provided, comprising: An acquisition module for real-time acquisition of carbon source concentration, dissolved oxygen concentration, pH value, temperature, alkalinity and nitrate concentration data in sewage; The control module is used for constructing a dynamic denitrification rate prediction model based on multi-source data, dynamically correcting model parameters by combining process parameter changes through a machine learning algorithm, calculating current stage carbon source dosage and dissolved oxygen control target values according to prediction model output, combining influent load and discharge standards of the sewage plant, adopting a distributed sensor network to perform multi-point real-time monitoring on carbon source and dissolved oxygen distribution in the reaction tank, adjusting the operation state of the carbon source dosing pump and the aeration equipment through feedback control, dynamically adjusting the carbon source dosing rate and the dissolved oxygen concentration by taking pH value and alkalinity as constraint conditions in the multi-parameter collaborative control stage, optimizing the control strategy through a reinforcement learning algorithm based on historical data and current operation conditions, performing cross-cycle adaptive adjustment, and uploading control data to a cloud platform.

[0014] According to a third aspect of the present application, an electronic device is provided, comprising a memory and a processor, wherein computer readable instructions are stored on the memory, and the computer readable instructions are executed by the processor to implement the precise control method for denitrification of a sewage treatment plant.

[0015] Advantages: The precise control method for denitrification of a sewage treatment plant provided by the present application has significant technical advantages and practical value. 1. Multi-source data fusion enhances prediction accuracy: By real-time acquisition of multi-element water quality parameter data such as carbon source concentration, dissolved oxygen, pH value, temperature, alkalinity and nitrate, a dynamic denitrification rate prediction model is constructed by combining LSTM neural network and mechanism model, which effectively improves the adaptability and prediction accuracy of the model to complex working conditions.

[0016] 2. Dynamic optimization of control strategy reduces resource waste: The carbon source dosage and dissolved oxygen control target values are dynamically adjusted according to real-time prediction results and influent load, and feedback control is realized through multi-point monitoring by a distributed sensor network. The carbon source dosing pump is controlled by variable frequency speed regulation and multi-stage distribution valve, and the aeration equipment is precisely adjusted by ORP, to ensure uniform distribution and efficient utilization of carbon source and dissolved oxygen, significantly reducing carbon source consumption and aeration energy consumption.

[0017] 3. Multi-parameter collaborative control ensures process stability: Taking pH value and alkalinity as constraint conditions, the alkaline concentration and dosing rate are dynamically adjusted by an alkali dosing pump to match the nitrate degradation rate in real time, ensuring stable operation of the system within the optimal pH and alkalinity range, and avoiding the decline of denitrification efficiency caused by pH imbalance or insufficient alkalinity.

[0018] 4. Reinforcement learning algorithm improves intelligent decision-making ability: based on deep Q learning to construct control optimization strategy, through adaptive training of historical data and running conditions, realize cross-cycle parameter adjustment and strategy optimization. The reward function design takes into account the carbon source saving rate and nitrate removal rate, balances the economy and processing efficiency, and improves the overall optimization ability of the system.

[0019] In summary, the present application integrates multi-source data processing, intelligent prediction, dynamic optimization and cloud integration, realizes the precise control of the denitrification process of the sewage treatment plant, effectively solves the technical problems of carbon source waste, high energy consumption and poor process stability, and has significant economic and environmental benefits. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is a flowchart of the precise control method of the denitrification of the sewage treatment plant provided in the specific embodiments of the present application; Figure 2 is a system composition schematic diagram of the precise control system of the denitrification of the sewage treatment plant provided in the specific embodiments of the present application. DETAILED DESCRIPTION

[0021] In order for those skilled in the art to better understand the technical solutions of the present application, the technical solutions of the present application will be described clearly and completely below in combination with the drawings of the present application. Based on the embodiments in the present application, other similar embodiments obtained by those skilled in the art without creative labor shall belong to the scope of protection of the present application. In addition, the directional words mentioned in the following embodiments, such as "up", "down", "left", "right", etc. are only reference directions of the drawings, therefore, the directional words used are used to illustrate but not to limit the present application.

[0022] The present application will be further described below in combination with the drawings and preferred embodiments.

[0023] Please refer to Figure 1 , the present application provides a precise control method for denitrification of a sewage treatment plant, comprising: S1: Real-time acquisition of carbon source concentration, dissolved oxygen concentration, pH value, temperature, alkalinity and nitrate concentration data in sewage.

[0024] Specifically, the method comprises: 1. Sensor deployment: Sensor arrays are arranged in different areas (anoxic zone, anaerobic zone) of the sewage reaction tank, specifically including: Carbon source concentration sensor: infrared spectroscopy combined with ion chromatography detection can identify carbon source components such as acetic acid, methanol and glucose, and the detection accuracy is ±0.1 mg / L.

[0025] Dissolved oxygen sensor: based on fluorescence quenching principle, resolution 0.01 mg / L, response time <5 seconds, prevent signal drift caused by biofilm deposition through ultrasonic cleaning device.

[0026] pH / temperature / alkalinity sensor: composite electrode and pH compensation algorithm are used, temperature compensation range 4-35℃, alkalinity detection is based on potentiometric titration method, accuracy ±5%.

[0027] Distributed sensors upload data to edge computing gateway through LoRa / NB-IoT, sampling frequency is 1 time / minute.

[0028] 2. Data preprocessing: Outlier filtering: sliding window mean filtering combined with 3σ principle to remove noise points.

[0029] Data standardization: use Z-score method to map multi-source data to [0, 1] interval, eliminate dimension differences.

[0030] S2: Based on multi-source data, construct a dynamic denitrification rate prediction model, dynamically correct model parameters through machine learning algorithm combined with process parameter changes.

[0031] Specifically, the dynamic denitrification rate prediction model uses LSTM neural network combined with mechanism model, input layer includes water quality parameters, temperature and historical operation data, output layer is carbon source consumption rate and nitrate degradation rate.

[0032] It needs to be further explained that: 1. Model structure: Feature input: carbon source concentration (C), dissolved oxygen concentration (DO), pH, temperature (T), alkalinity (A), nitrate concentration (NO3 - ), influent flow (Q).

[0033] Network architecture: LSTM layer (neuron number 64, time step 5) connected to fully connected layer (ReLU activation function), fully connected layer (ReLU activation function connected to output layer.

[0034] Mechanism model fusion: introduce denitrification kinetics equation where, is the reaction rate, C is the carbon source concentration, is the rate constant, predicted by LSTM in real time, is a piecewise function (activated when pH is 6-8, range 0.8-1.0), is the DO concentration when the inhibition rate reaches 50%, DO is the current reaction tank dissolved oxygen concentration.

[0035] Specifically: If , 0; If is otherwise, .

[0036] 2. Dynamic updating mechanism: Update the weight every hour using the online learning algorithm (incremental SGD), and set the forgetting factor to 0.95 to ensure that the model remembers the short-term trend.

[0037] S3: According to the prediction model output, combined with the influent load and discharge standard of the sewage plant, calculate the carbon source dosage and dissolved oxygen control target value in the current stage.

[0038] Specifically, the method comprises: 1. Nitrogen removal target setting: According to the discharge standard (such as GB18918-2002), the upper limit of nitrate concentration in effluent is ≤10mg / L, and the total nitrogen removal rate required is calculated (for example, 70%) combined with the influent load.

[0039] 2. Carbon source demand Calculation: Where: Carbon source consumption rate (mg / L·min).

[0040] Reaction tank volume ( .

[0041] Carbon source utilization rate (0.8~0.95), indicating the proportion of carbon source actually utilized by microorganisms.

[0042] Carbon source solution concentration (mg / L). Solution density (kg / , default 1000kg / .

[0043] 3. Dissolved oxygen control logic: The DO set point of the anoxic zone is determined by fuzzy rules, and the fuzzy rules are shown in Table 1: Table 1 Fuzzy rules The aeration blower adopts feedforward plus feedback PI control, the feedforward is determined by the basic opening of the influent flow, and the feedback is corrected by the online DO sensor error.

[0044] S4: Real-time monitoring of carbon source and dissolved oxygen distribution in the reaction tank by distributed sensor network, and adjusting the operation state of carbon source dosing pump and aeration equipment through feedback control.

[0045] Specifically, the distributed sensor network includes optical sensors, electrochemical sensors, and ultrasonic sensors for detecting carbon source concentration, dissolved oxygen, and sludge concentration, respectively.

[0046] Specifically, the carbon source dosing pump is controlled by frequency conversion and multi-stage distribution valve to ensure uniform distribution of carbon source in the reaction tank.

[0047] It should be further pointed out that the method comprises: 1. Subarea monitoring network: The reaction tank is divided into 8 monitoring areas, each area is equipped with a group of sensors (carbon source + DO + pH), and a total of 64 data points are uploaded to the edge end every minute.

[0048] Abnormal area determination: if the carbon source concentration in a certain area is >10 mg / L and Last >20 mg / L, trigger local dosing pump power increase by 10%.

[0049] 2. Dynamic compensation algorithm: When the feedback data shows that the nitrate degradation is lagging (actual rate <80% of predicted value), additional carbon source 15%.

[0050] S5: In the multi-parameter collaborative control stage, the pH value and alkalinity are used as constraint conditions to dynamically adjust the carbon source dosing rate and dissolved oxygen concentration.

[0051] Specifically, the target value of dissolved oxygen control is set according to the partition of the reaction tank, and the target value of the anoxic zone is 0.2-0.5 mg / L, and the target value of the anaerobic zone is lower than 0.2 mg / L.

[0052] Specifically, in the multi-parameter collaborative control, the pH value is adjusted by the alkali dosing pump, and the alkali concentration and dosing rate are dynamically matched according to the nitrate degradation rate.

[0053] Specifically, regarding alkali dosing control: Alkali concentration ( : Dynamically matched according to the nitrate degradation rate (empirical formula): Wherein, Empirical coefficient, calibrated by experiment, preferably k=0.5 in the present application.

[0054] Dosing rate (Q_base): wherein, alkaline liquor storage concentration (mg / L), is the volume of the reaction tank.

[0055] It needs to be further explained that the present application introduces an improved NSGA-II to solve the carbon source dosage optimization problem under the constraints of pH and alkalinity: 1. Individual representation: [carbon source flow (m³ / h), DO set point (mg / L)], chromosome length 16 bits.

[0056] 2. Fitness function: wherein, is the weight coefficient (relative importance of carbon source cost and energy consumption), which dynamically changes over time (0.7 at the initial stage to 0.3 at the later stage), wherein the constraint condition is: mg / L, , represents the removal efficiency of nitrate, which is the core index for measuring denitrification effect, represents the influent nitrate concentration, represents the effluent nitrate concentration, F2 takes into account the carbon source cost and energy consumption, wherein, represents the carbon source cost, represents the energy loss.

[0057] Specifically, regarding the dynamic adjustment of the weight: Initial debugging stage (w1=0.7): preferentially control the cost to avoid waste of carbon source.

[0058] Stable operation stage (w1=0.3): pay more attention to energy consumption optimization to reduce operating costs.

[0059] It needs to be further explained that regarding the weight function design: The parameter explanations of the above formula are shown in Table 2: Table 2 Parameter Explanation It can be understood that: Initial stage (t=0): w1=0.7, preferentially ensure the nitrate removal rate to avoid effluent exceeding standard due to energy saving.

[0060] Stable stage (t=24h): w1=0.5, balance the cost and energy consumption; t=72h after gradually transition to w1=0.3, strengthen economic optimization.

[0061] The following is illustrated by a specific example: Assume that the operating parameters at a certain time are as follows: (mid-term operating state).

[0062] Calculation results: ; ; If it is detected at this time (exceeding the standard), the penalty mechanism is triggered: ; (where is the penalty coefficient).

[0063] Specifically, regarding the coupling of the constraint condition: When F2 optimization results in ΔpH>0.3 or ΔA<2, the solution needs to be inhibited through the penalty function mechanism.

[0064] 3. Constraint processing: (1) pH value fluctuation limit: Reason: A sharp change in pH value (>0.5) will cause inhibition of denitrifying bacteria activity, and the removal rate of nitrate will decrease by more than 30%.

[0065] Countermeasures: When , immediately reduce the carbon source dosage by 10% and start alkali compensation.

[0066] (2) Alkalinity maintenance requirement: Where, is the actual alkalinity, is the initial alkalinity.

[0067] Reason: Insufficient alkalinity (<2mg / L) will cause system acidification, affecting the survival environment of microorganisms.

[0068] Correlation control: Alkalinity and pH value are coupled through the following empirical formula: (3) Other process constraints Dissolved oxygen threshold: DO≤0.5mg / L in the anoxic zone and DO≤0.2mg / L in the anaerobic zone.

[0069] Carbon source upper limit: single addition amount is not more than 120% of the theoretical demand (to prevent excessive inhibition).

[0070] S6: Based on historical data and current operating conditions, the control strategy is optimized through a reinforcement learning algorithm for cross-cycle adaptive adjustment.

[0071] Specifically, the reinforcement learning algorithm uses deep Q learning, and the reward function is defined as the weighted sum of carbon source saving rate and nitrate removal rate.

[0072] It should be further explained that regarding reinforcement learning optimization, the design scheme of the present application includes: Deep Q-Learning architecture State space: integrate 7 types of sensor data + current control action history.

[0073] Action space: carbon source addition adjustment amount (-15% to +15%), DO set point fine tuning ( .

[0074] Reward function: wherein, is the carbon source saving rate, is the nitrate removal rate, DO is the current reaction tank dissolved oxygen concentration, is the target reaction tank dissolved oxygen concentration.

[0075] S7: Upload control data to the cloud platform.

[0076] It can be understood that the cloud platform is used to store historical data for subsequent analysis.

[0077] Please refer to Figure 2 , the present application provides another embodiment, which provides a precise control system for denitrification of a sewage treatment plant, comprising: An acquisition module 100 is configured to collect carbon source concentration, dissolved oxygen concentration, pH value, temperature, alkalinity and nitrate concentration data in sewage in real time. The control module 200 is configured to construct a dynamic denitrification rate prediction model based on multi-source data, dynamically correct model parameters by combining process parameter changes through a machine learning algorithm, output the prediction model, combine influent load and discharge standards of the sewage plant, calculate carbon source dosage and dissolved oxygen control target values in the current stage, monitor carbon source and dissolved oxygen distribution in the reaction tank through a distributed sensor network, adjust the operation state of the carbon source dosing pump and the aeration equipment through feedback control, dynamically adjust the carbon source dosage rate and the dissolved oxygen concentration by taking pH value and alkalinity as constraint conditions in the multi-parameter collaborative control stage, optimize the control strategy through a reinforcement learning algorithm based on historical data and current operating conditions, perform cross-cycle adaptive adjustment, and upload control data to a cloud platform.

[0078] In a preferred embodiment, the present application further provides an electronic device, which comprises: a memory and a processor, wherein the memory has computer readable instructions stored thereon, and the computer readable instructions, when executed by the processor, implement the precise control method for denitrification of a sewage treatment plant. The computer device can be a server, a terminal, or any other electronic device with necessary computing and / or processing capabilities. In an embodiment, the computer device can include a processor, a memory, a network interface, a communication interface, and the like connected by a system bus. The processor of the computer device can be used to provide necessary computing, processing, and / or control capabilities. The memory of the computer device can include a non-volatile storage medium and an internal memory. The non-volatile storage medium or the operating system and computer programs stored therein or thereon can include an operating system, a computer program, and the like. The internal memory can provide an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface and the communication interface of the computer device can be used to connect and communicate with external devices through a network. The computer program, when executed by the processor, performs the steps of the method of the present application.

[0079] The application can be implemented as a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, causes the steps of the method of the embodiments of the application to be performed. In one embodiment, the computer program is distributed over a network of coupled computer devices or processors such that the computer program is stored, accessed and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, can be performed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations can be performed by one or more computer devices or processors and one or more other method steps / operations can be performed by one or more other computer devices or processors. One or more computer devices or processors can perform a single method step / operation, or perform two or more method steps / operations.

[0080] It will be appreciated by the person of ordinary skill in the art that the method steps of the application can be instructed by a computer program to relevant hardware such as a computer device or processor, the computer program being stored in a non-transitory computer-readable storage medium which, when executed, causes the steps of the application to be performed. Any reference herein to a memory, storage, database or other medium can include non-volatile and / or volatile memory as appropriate. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid-state disk, and so on. Examples of volatile memory include random access memory (RAM), external cache memory, and so on.

[0081] The various technical features described above can be combined in any manner. Although not all possible combinations are described, any combination of the technical features should be considered to be within the scope of the present description, as long as such a combination does not result in a contradiction.

[0082] The specific embodiments described above do not constitute an exhaustive description of the scope of the application. Any other corresponding changes and modifications according to the technical concept of the application should be included within the scope of the claims of the application.

Claims

1. A method for precise control of denitrification in a wastewater treatment plant, characterized in that, The method comprises: S1: real-time acquisition of carbon source concentration, dissolved oxygen concentration, pH value, temperature, alkalinity and nitrate concentration data in sewage; S2: constructing a dynamic denitrification rate prediction model based on multi-source data, and dynamically correcting model parameters through machine learning algorithm combined with process parameter changes; S3: calculating the current stage carbon source dosage and dissolved oxygen control target value according to the prediction model output combined with the influent load of the sewage plant and the discharge standard; S4: using a distributed sensor network to perform multi-point real-time monitoring of the carbon source and dissolved oxygen distribution in the reaction tank, and adjusting the operation state of the carbon source dosing pump and the aeration equipment through feedback control; S5: in the multi-parameter collaborative control stage, dynamically adjusting the carbon source dosing rate and dissolved oxygen concentration with pH value and alkalinity as the constraint conditions; S6: optimizing the control strategy through reinforcement learning algorithm based on historical data and current operating conditions for cross-cycle adaptive adjustment; S7: uploading the control data to the cloud platform.

2. The method for precision control of denitrification in a wastewater treatment plant according to claim 1, characterized in that, The dynamic denitrification rate prediction model adopts the combination of LSTM neural network and mechanism model, the input layer includes water quality parameters, temperature and historical operation data, and the output layer is carbon source consumption rate and nitrate degradation rate.

3. The method for precision control of denitrification in a wastewater treatment plant according to claim 2, characterized in that, The carbon source dosing pump is controlled by frequency conversion and multi-stage distribution valve to ensure uniform distribution of carbon source in the reaction tank.

4. The method for precision control of denitrification in a wastewater treatment plant according to claim 1, characterized in that, The dissolved oxygen control target value is set according to the partition of the reaction tank, the target value of the anoxic zone is 0.2-0.5mg / L, and the target value of the anaerobic zone is lower than 0.2mg / L.

5. The method for precision control of denitrification in a wastewater treatment plant according to claim 2, characterized in that, In the multi-parameter collaborative control, the pH value is adjusted by the alkali dosing pump, and the alkali concentration and dosing rate are dynamically matched according to the nitrate degradation rate.

6. The method for precision control of denitrification in a wastewater treatment plant according to claim 1, characterized in that, The reinforcement learning algorithm adopts deep Q learning, and the reward function is defined as the weighted sum of carbon source saving rate and nitrate removal rate.

7. The method for precision control of denitrification in a wastewater treatment plant according to claim 6, characterized in that, The reward function comprises: wherein, is the carbon source saving rate, is the nitrate removal rate, DO is the current reactor dissolved oxygen concentration, is the target reactor dissolved oxygen concentration.

8. The method for precision control of denitrification in a wastewater treatment plant according to claim 1, characterized in that, The distributed sensor network includes optical sensors, electrochemical sensors and ultrasonic sensors, which are used to detect carbon source concentration, dissolved oxygen and sludge concentration respectively.

9. A precision control system for denitrification in a wastewater treatment plant, characterized in that, It comprises: An acquisition module is configured to acquire carbon source concentration, dissolved oxygen concentration, pH value, temperature, alkalinity and nitrate concentration data in sewage in real time; A control module is configured to construct a dynamic denitrification rate prediction model based on multi-source data, and dynamically correct model parameters through machine learning algorithm combined with process parameter changes; calculate the current stage carbon source dosage and dissolved oxygen control target value according to the prediction model output combined with the influent load of the sewage plant and the discharge standard; and use a distributed sensor network to perform multi-point real-time monitoring of the carbon source and dissolved oxygen distribution in the reaction tank, and adjust the operation state of the carbon source dosing pump and the aeration equipment through feedback control; In the multi-parameter collaborative control stage, the carbon source dosing rate and dissolved oxygen concentration are dynamically adjusted with pH value and alkalinity as the constraint conditions; the control strategy is optimized through reinforcement learning algorithm based on historical data and current operating conditions for cross-cycle adaptive adjustment; and the control data is uploaded to the cloud platform.

10. An electronic device, comprising: It comprises: A memory; and a processor, the memory having computer readable instructions stored thereon, the computer readable instructions, when executed by the processor, implementing the method for precision control of denitrification in a wastewater treatment plant according to any one of claims 1 to 8.

Images