Control system for sewage treatment plant

Abstract

The invention relates to the technical field of sewage treatment automatic control, and discloses a control system for a sewage treatment plant, which comprises a multi-modal biosensing module, the multi-modal biosensing module is deployed in an anaerobic zone and an aerobic zone of a biochemical reaction tank, and a cloud edge collaborative data processing module comprises an edge calculation module and a cloud optimization platform. The edge calculation module preprocesses data collected by the multi-mode biosensing module and executes emergency control, the cloud optimization platform operates the multi-target dynamic optimization module and the digital twin model, and the self-adaptive decoupling controller receives control parameters output by the multi-target dynamic optimization module. By deploying the multi-modal biosensing modules in the anaerobic zone and the aerobic zone, the real-time acquisition of biochemical reaction parameters of each zone along the way is realized, and the microbial metabolic activity change in the biological reaction process can be accurately captured in combination with the pretreatment of the edge calculation module and the analysis of the cloud optimization platform.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic control of sewage treatment, and particularly relates to a control system for a sewage treatment plant. BACKGROUND

[0002] Automatic control of sewage treatment refers to a technical system that uses sensors, controllers, actuators and computer systems to monitor water quality parameters (such as pH value, dissolved oxygen, turbidity, etc.), equipment operating conditions (such as start-stop and frequency of water pumps, fans and valves) in real time, analyze data and intelligently control them, so as to realize precise and efficient operation of the sewage treatment process, ensure that the effluent quality meets the standards, and reduce energy consumption, material consumption and labor intervention costs.

[0003] According to the search, the control system and control method of the water pump and aerator of the MC-MBBR process of the sewage treatment plant are disclosed in CN106336001B. The control system includes a PLC control system, an operation panel, multiple starters, multiple water pumps, multiple aerators, two online COD analyzers and two ammonia nitrogen detectors. The PLC control system includes a CPU module, a DC 24V direct current power module, an Ethernet communication module, an analog input module, multiple digital input modules and multiple digital output modules. The system can realize full-automatic control, better protect the water pump and aerator and improve the effluent quality.

[0004] However, the existing automatic control of sewage treatment mostly relies on water quality parameters at a single time point, lacks real-time capture of the kinetic state of the biological reaction process, and traditional sensors are usually installed at fixed positions such as the end of the reaction tank, which cannot reflect the changes in microbial metabolic activity in each sub-zone along the way. In addition, the key control variables in the biological treatment process, such as DO, internal reflux ratio, external reflux ratio and carbon source dosage, have strong coupling characteristics. The existing automatic control of sewage treatment mostly uses single-loop PID control, which will cause serious interference when adjusting DO and reflux ratio at the same time, resulting in large fluctuations in sludge concentration.

[0005] Therefore, it is necessary to provide a control system for a sewage treatment plant to solve the above problems. SUMMARY

[0006] The present application aims to provide a technical solution to solve the problems in the prior art mentioned in the background.

[0007] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0008] A control system for a sewage treatment plant, comprising a multi-modal biological sensing module, the multi-modal biological sensing module is deployed in the anaerobic zone and aerobic zone of a biochemical reaction tank for collecting microbial metabolic activity parameters, water quality spectral data and traditional electrochemical parameters of each sub-zone;

[0009] A cloud-edge collaborative data processing module, the cloud-edge collaborative data processing module comprises an edge computing module and a cloud optimization platform, the edge computing module pre-processes the data collected by the multi-modal biological sensing module and performs emergency control, and the cloud optimization platform runs a multi-objective dynamic optimization module and a digital twin model;

[0010] The multi-objective dynamic optimization module is integrated in the cloud optimization platform, and based on the pre-processed data, the objectives of water quality standard, energy consumption minimization and chemical reagent reduction are optimized synchronously;

[0011] The digital twin model comprises a digital twin fault prediction unit, which predicts the remaining life of the equipment and process abnormalities based on correlation analysis of equipment operation data and process parameters;

[0012] An adaptive decoupling controller receives the control parameters output by the multi-objective dynamic optimization module and processes the coupling relationship of control variables such as aeration quantity and reflux ratio through a dynamic decoupling matrix.

[0013] Preferably, the multi-modal biological sensing module comprises a multi-parameter fusion probe, the multi-parameter fusion probe is installed in each sub-zone of the anaerobic zone and the aerobic zone respectively, and the multi-parameter fusion probe comprises a microbial metabolic activity sensor, a miniature spectral sensor and an electrochemical sensor array;

[0014] The microbial metabolic activity sensor detects the concentration of NADH based on three-wavelength fluorescence technology, and reflects the metabolic state of microorganisms in real time;

[0015] The miniature spectral sensor continuously scans the sub-zone in the ultraviolet-visible light band, and monitors the nitrification and denitrification process in the sub-zone through absorbance change;

[0016] The electrochemical sensor array is used to detect pH, dissolved oxygen and oxidation-reduction potential in the sub-zone.

[0017] Preferably, the objective function of the multi-objective dynamic optimization module is:

[0018] minF(x)=[EC,EQ,CC];

[0019] Wherein, EC is the total energy consumed in the sewage treatment process; EQ is the standard degree of pollution indicators of discharged water after treatment; and CC is the consumption of additional carbon source added in the denitrification reaction or denitrification reaction;

[0020] Constraint: NH4 + -N≤2mg / L, TN≤15mg / L, 0.5≤DO≤3.0mg / L, sludge age is 10-20d.

[0021] Preferably, the multi-objective dynamic optimization module adopts NSGA-III algorithm, including population stratification mechanism, dynamic constraint processing unit and real-time optimization engine.

[0022] The population stratification mechanism divides the solution set into an elite layer accounting for 20% and an exploration layer accounting for 80%, and the elite layer adopts local search to accelerate convergence.

[0023] The dynamic constraint processing unit monitors key operating parameters in real time, including, for example, influent COD concentration, flow rate or temperature in wastewater treatment.

[0024] The real-time optimization engine uses GPU to accelerate calculation, shortens the optimization period from 20 minutes to 4 minutes, and realizes quasi-real-time control.

[0025] Preferably, the adaptive decoupling controller includes a feedforward compensator, a coupling analysis module and a decoupling matrix.

[0026] The feedforward compensator receives control instructions of aeration quantity, internal reflux ratio and external reflux ratio and performs preliminary decoupling, the coupling analysis module online identifies time-varying coupling coefficients τ between control variables, and the decoupling matrix constructs a dynamic decoupling model based on the coupling coefficients τ, and the formula is:

[0027]

[0028] u air : actual aeration quantity control quantity, such as valve opening degree or fan frequency;

[0029] u ref : actual reflux ratio control quantity, such as internal / external reflux pump frequency;

[0030] v air : target aeration quantity, ideal value output by cloud optimization;

[0031] v ref : target reflux ratio, ideal value output by cloud optimization;

[0032] τ 11 and τ 22 : dynamic coefficients of control quantities, such as response delay of aeration system or inertia of reflux pump;

[0033] τ 12 and τ 21 : cross-coupling coefficients between control quantities, τ 12τ 21 τ

[0034] s: Laplace operator, convert time-domain dynamic characteristics to frequency-domain transfer function;

[0035] u: actual control variable vector, determines process DO, sludge return efficiency.

[0036] Preferably, the edge computing module includes an edge layer, which deploys an industrial Internet gateway in each process section and performs the following operations:

[0037] Data preprocessing, including wavelet denoising and outlier rejection;

[0038] Emergency control, when DO is less than 1 mg / L or COD surge exceeds 500 mg / L, directly start PID control to adjust aeration quantity or carbon source dosage;

[0039] Data transmission, upload key characteristic parameters to the cloud, including activated sludge respiration rate OUR and nitrification rate NUR.

[0040] Preferably, the digital twin fault prediction unit includes an electrical characteristic analysis channel, a process correlation analysis channel, and a D-S evidence theory fusion module. The electrical characteristic analysis channel is used to monitor the device current harmonic distortion rate and vibration spectrum. The process correlation analysis channel establishes a correlation matrix between device state and process parameters. The D-S evidence theory fusion module performs decision fusion on the results of the electrical characteristic analysis channel and the process correlation analysis channel, and reduces the false alarm rate to below 3%.

[0041] Preferably, the microbial metabolic activity sensor and the miniature spectrum sensor are arranged and installed at least 2 per partition, and are provided with an automatic cleaning device, which performs ultrasonic cleaning every 30 minutes to avoid sludge adhesion.

[0042] Preferably, the dynamic constraint processing unit automatically relaxes the effluent TN constraint to 20 mg / L when the influent COD mutation exceeds 30%, and gradually recovers to 15 mg / L within 24 hours according to a linear gradient.

[0043] Preferably, the adaptive decoupling controller further includes a robust control mode, which is automatically switched when the sludge concentration change rate exceeds 5% / h, suppresses coupled oscillation by enhancing feedback correction, and controls the sludge concentration fluctuation amplitude within ±120 mg / L.

[0044] The technical effects and advantages of the present application are as follows:

[0045] 1、The present application realizes real-time acquisition of biochemical reaction parameters in each sub-zone along the way by deploying multi-modal biological sensing modules in anaerobic and aerobic zones, combined with preprocessing by edge computing modules and analysis by cloud optimization platform, which can accurately capture the changes in microbial metabolic activity of the biological reaction process, breaking through the limitations of traditional sewage treatment automatic control relying on single time point water quality parameters and fixed position monitoring, providing comprehensive and real-time basis for subsequent precise regulation;

[0046] 2、The present application effectively solves the serious interference problem of traditional single-loop PID control when adjusting multiple parameters at the same time by performing multi-objective optimization calculation on the cloud optimization platform and through a double-channel adaptive decoupling controller, significantly reducing the fluctuation amplitude of sludge concentration and improving the stability of the operation of the sewage treatment system. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 The control system block diagram for the present application used in a sewage treatment plant. DETAILED DESCRIPTION

[0048] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and can include changes, modifications, additions or omissions of the functions and arrangements of the elements discussed without departing from the scope of the content of this specification. Various processes or components can be omitted, substituted or added according to the needs of the various examples. In addition, features described for some examples can be combined in other examples.

[0049] The present application provides a control system for a sewage treatment plant as shown in Figure 1 The present application provides a control system for a sewage treatment plant as shown in

[0050] The multi-modal biological sensing module is deployed in the anaerobic and aerobic zones of the biochemical reaction tank, and is used to collect microbial metabolic activity parameters, water quality spectral data and traditional electrochemical parameters in each sub-zone;

[0051] Specifically, the multi-modal biological sensing module includes a multi-parameter fusion probe, which is installed in each sub-zone of the anaerobic and aerobic zones, and the multi-parameter fusion probe includes a microbial metabolic activity sensor, a miniature spectral sensor and an electrochemical sensor array;

[0052] The microbial metabolic activity sensor detects the concentration of NADH based on a three-wavelength fluorescence technology, and reflects the metabolic state of microorganisms in real time. NADH is the concentration of reduced nicotinamide adenine dinucleotide, which is a key coenzyme for microbial metabolism. The fluorescence intensity of NADH is positively correlated with microbial activity. The three wavelengths are 370 / 440 / 550 nm for detecting the concentration of NADH. The background interference is eliminated by the ratio fluorescence method to improve the detection accuracy. When the activity decreases (such as insufficient DO or toxic shock), the NADH fluorescence weakens, and the system can early warn the activity inhibition of nitrifying bacteria or denitrifying bacteria.

[0053] The miniature spectral sensor continuously scans the sub-area in the ultraviolet-visible light band, and monitors the nitrification and denitrification process in the sub-area through the absorbance change. The band scans the sewage in the 190-800 nm band, and monitors the nitrification or denitrification process through the absorbance change. For example, nitrate has a characteristic absorption peak at 220 nm, and nitrite has a significant absorption at 210 nm. The absorbance difference can be used for quantitative analysis to realize real-time data collection of nitrification rate (NUR) and denitrification rate (DNUR).

[0054] The microbial metabolic activity sensor and the miniature spectral sensor are arranged and installed at least two per sub-area, and an automatic cleaning device is built-in. Ultrasonic cleaning is performed every 30 minutes to avoid sludge adhesion. The single-point failure risk is eliminated through data fusion.

[0055] The electrochemical sensor array is used to detect pH, dissolved oxygen (DO) and oxidation-reduction potential (ORP) in the sub-area. DO reflects the aeration intensity in the aerobic zone, ORP assists in judging the anaerobic or anoxic state, and pH fluctuation can indicate the imbalance of nitrification reaction (acid production) or denitrification reaction (alkaline production).

[0056] The cloud-edge collaborative data processing module includes an edge computing module and a cloud optimization platform. The edge computing module pre-processes the data collected by the multi-modal biological sensor module and performs emergency control. The cloud optimization platform runs a multi-objective dynamic optimization module and a digital twin model.

[0057] Specifically, the edge computing module includes an edge layer, which deploys an industrial Internet gateway in each process section and performs the following operations:

[0058] Data preprocessing, including wavelet denoising and outlier removal. For high-frequency noise of the sensor, such as aeration bubble interference, db4 wavelet basis is used for 5-layer decomposition. After reconstruction, the signal-to-noise ratio is improved by 15-20 dB. Based on the IQR method, outliers deviating from the median by more than 3 times are identified and removed to avoid interference with subsequent analysis.

[0059] Emergency control corresponds, when DO is lower than 1mg / L or COD increases more than 500mg / L, directly start PID control to adjust aeration quantity or carbon source dosage;

[0060] Data transmission, upload key characteristic parameters to the cloud, key characteristic parameters include activated sludge respiratory rate OUR and nitrification rate NUR;

[0061] The digital twin model comprises a digital twin fault prediction unit, which predicts the remaining life of the equipment and process abnormalities based on correlation analysis of equipment operation data and process parameters.

[0062] The digital twin fault prediction unit comprises an electrical characteristic analysis channel, a process correlation analysis channel and a D-S evidence theory fusion module.

[0063] The process correlation analysis channel establishes a correlation matrix of equipment state (such as aeration machine speed) and process parameters (such as DO, NH4 + -N) and TN.

[0064] Taking the aeration machine speed as the core equipment state parameter, the correlation matrix is as follows:

[0065]

[0066] Row vector: S aer is the aeration machine speed, unit: r / min.

[0067] Column vector: P DO , P NH4 , and P TN represent DO, NH4 + -N, and TN, respectively.

[0068] Wherein, the matrix element M i,j : indicates the correlation strength of the i-th equipment state parameter to the j-th process parameter (value range: -1~1, the greater the absolute value, the stronger the correlation; positive sign indicates positive correlation, negative sign indicates negative correlation)

[0069] (Strong positive correlation): The aeration machine speed directly determines the aeration intensity (the higher the speed, the greater the aeration air volume or bubble density), the bubble and water contact area increases, the oxygen mass transfer efficiency is improved, which leads to the increase of the dissolved oxygen (DO) concentration in water.

[0070] (Medium strong negative correlation): NH4 +- N removal relies on nitrification (aerobic conditions, nitrifying bacteria convert NH4 + - N to NO3 - - N), while nitrifying bacteria activity strictly depends on DO (optimal DO is 2-3 mg / L).

[0071] Therefore, increasing the aeration machine speed -> DO increases -> nitrifying bacteria metabolic activity increases -> NH4 + - N degradation accelerates -> concentration decreases (negative correlation)

[0072] (Moderately negative correlation): N (total nitrogen) removal requires nitrification (aerobic) + denitrification (anoxic) synergy. If the aeration machine speed is too high -> the DO in the aerobic zone spills over into the anoxic zone -> inhibits denitrifying bacteria (denitrification requires anoxic environment, DO > 0.5 mg / L when inhibited) -> TN degradation efficiency decreases (positive correlation trend); but if the speed is too low -> nitrification is incomplete (NH4 + - N remains) -> TN increases (negative correlation trend). In practice, the matrix takes "moderately negative correlation" based on "normal operating conditions (DO controlled at 0.5-3 mg / L), moderately increasing the speed can promote complete nitrification, indirectly reducing TN, the correlation strength is weak (coefficient absolute value 0.5)

[0073] When the digital twin fault prediction unit detects an abnormal increase in NH4 + - N, the matrix can quickly locate whether the aeration machine speed is too low (DO deficiency leads to nitrification inhibition), narrowing down the scope of investigation, providing the D-S evidence theory fusion module with process dimension correlation weights, and cross- verifying the results with the electrical characteristic analysis channel (such as aeration machine current distortion rate), to reduce the false positive rate.

[0074] The D-S evidence theory fusion module fuses the results of the electrical characteristic analysis channel and the process correlation analysis channel, the fusion rule is to calculate the support degree of each channel based on the belief function (Belief Function), to merge the evidence using the Dempster combination rule, and to reduce the false positive rate to below 3%, which is better than single-channel monitoring.

[0075] The multi-objective dynamic optimization module is integrated in the cloud optimization platform, based on pre-processed data to simultaneously optimize the water quality compliance, energy consumption minimization, and chemical reagent reduction targets;

[0076] Specifically, the objective function of the multi-objective dynamic optimization module is:

[0077] minF(x) = [EC, EQ, CC];

[0078] wherein EC is the total energy consumed in the wastewater treatment process, mainly from the aeration blower and the reflux pump; EQ is the compliance degree of the pollution indicators of the discharged water after treatment, expressed as NH4 + N, TN, and COD are the core indicators; CC is the consumption of the additional carbon source (such as sodium acetate) added in the denitrification reaction or the denitrification reaction;

[0079] The constraint conditions are: NH4 + N≤2 mg / L (strictly control the nitrification effect), TN≤15 mg / L (conventional constraint), relax to 20 mg / L when the COD mutation is >30% (dynamic adjustment), 0.5≤DO≤3.0 mg / L (balance nitrification and denitrification), and the sludge age is 10-20 d (prevent sludge aging or loss).

[0080] The multi-objective dynamic optimization module adopts the NSGA-III algorithm, including a population stratification mechanism, a dynamic constraint processing unit, and a real-time optimization engine.

[0081] The population stratification mechanism divides the solution set into an elite layer accounting for 20% and an exploration layer accounting for 80%. The elite layer adopts local search to accelerate convergence, and the exploration layer adopts global search to avoid prematureness.

[0082] The dynamic constraint processing unit monitors the key working condition parameters of the system in real time. The key working condition parameters include, for example, the influent COD concentration, flow rate, or temperature in wastewater treatment. For example, when the influent COD mutation exceeds 30%, the dynamic constraint processing unit automatically relaxes the effluent TN constraint to 20 mg / L and gradually restores it to 15 mg / L within 24 hours according to a linear gradient, which not only responds to the impact load but also avoids long-term over-standard.

[0083] The real-time optimization engine uses GPU acceleration to calculate, shortens the optimization period from 20 minutes to 4 minutes, and realizes quasi-real-time control.

[0084] The adaptive decoupling controller receives the control parameters output by the multi-objective dynamic optimization module and processes the coupling relationship of the control variables such as the aeration amount and the reflux ratio through a dynamic decoupling matrix.

[0085] Specifically, the adaptive decoupling controller includes a feedforward compensator, a coupling analysis module, and a decoupling matrix.

[0086] The feedforward compensator receives the control instructions of the aeration amount, the internal reflux ratio, and the external reflux ratio and performs preliminary decoupling. The coupling analysis module online identifies the time-varying coupling coefficients τ between the control variables. The decoupling matrix constructs a dynamic decoupling model based on the coupling coefficients τ, and the formula is:

[0087]

[0088] u air : actual aeration amount control variable, such as valve opening degree or blower frequency.

[0089] u ref : actual internal / external reflux pump frequency;

[0090] v air : target aeration rate, ideal value output by cloud optimization;

[0091] v ref : target reflux ratio, ideal value output by cloud optimization;

[0092] τ 11 and τ 22 : dynamic coefficient of control quantity itself, such as response delay of aeration system, inertia of reflux pump;

[0093] τ 12 and τ 21 : cross-coupling coefficient between control quantities, τ 12 is the interference intensity of reflux ratio on aeration, τ 21 is the interference intensity of aeration on reflux;

[0094] s: Laplace operator, converts time-domain dynamic characteristics into frequency-domain transfer function;

[0095] u: actual control quantity vector, determines process DO and sludge reflux efficiency.

[0096] When the actual reflux ratio u ref suddenly increases, that is, more nitrate nitrogen enters the anoxic zone, if there is no decoupling: the change of reflux ratio will indirectly affect the nitrification demand in the aerobic zone, causing the aeration rate u air to deviate from v air , and the system appears coupled oscillation;

[0097] There is a decoupling matrix: This item will compensate the interference of reflux ratio on aeration in advance, so that the actual aeration rate u air still closely tracks the target v air , and is not affected by the fluctuation of reflux ratio.

[0098] Further, the adaptive decoupling controller also includes a robust control mode, which is automatically switched when the sludge concentration change rate exceeds 5% / h, and suppresses coupled oscillation by enhancing feedback correction, so that the sludge concentration fluctuation amplitude is controlled within ±120mg / L.

[0099] The above describes embodiments of the present application, but the present application is not limited to the above specific embodiments, and the above specific embodiments are only illustrative and not limiting, and those skilled in the art can make many forms under the inspiration of the present application, which are all within the protection of the present application.

Claims

1. A control system for a sewage treatment plant, characterized in that: The multimodal biosensor module is deployed in the anaerobic and aerobic zones of the biochemical reaction tank to collect microbial metabolic activity parameters, water quality spectral data, and traditional electrochemical parameters of each zone; A cloud-edge collaborative data processing module, comprising an edge computing module and a cloud optimization platform. The edge computing module pre-processes the data collected by the multimodal biosensor module and performs emergency control. The cloud optimization platform runs a multi-objective dynamic optimization module and a digital twin model. The multi-objective dynamic optimization module is integrated into the cloud optimization platform, and optimizes the goals of achieving water quality standards, minimizing energy consumption, and reducing chemical dosage based on pre-processed data; The digital twin model includes a digital twin fault prediction unit, which predicts the remaining life of the equipment and process anomalies based on the correlation analysis of equipment operation data and process parameters; An adaptive decoupling controller receives the control parameters output by the multi-objective dynamic optimization module and processes the coupling relationship of control variables such as aeration volume and reflux ratio through a dynamic decoupling matrix.

2. A control system for a sewage treatment plant according to claim 1, characterized in that: The multimodal biosensing module includes a multi-parameter fusion probe, which is installed in each partition of the anaerobic zone and the aerobic zone respectively. The multi-parameter fusion probe includes a microbial metabolic activity sensor, a micro-spectral sensor and an electrochemical sensor array; The microbial metabolic activity sensor detects NADH concentration based on three-wavelength fluorescence technology and reflects the metabolic status of microorganisms in real time; The micro-spectral sensor uses the ultraviolet-visible light band to continuously scan the partition and monitor the nitrification and denitrification process in the partition through the change of absorbance; The electrochemical sensor array is used to detect pH, dissolved oxygen and redox potential in the partition.

3. A control system for a sewage treatment plant according to claim 1, characterized in that: The objective function of the multi-objective dynamic optimization module is: minF(x)=[EC,EQ,CC]; Among them, EC is the total energy consumed in the sewage treatment process; EQ is the degree of compliance of the pollution index of the treated effluent; CC is the consumption of the external carbon source added for denitrification reaction or denitrification reaction; The constraints are: NH4 + -N≤2mg / L, TN≤15mg / L, 0.5≤DO≤3.0mg / L, sludge age is 10-20d.

4. A control system for a sewage treatment plant according to claim 1, characterized in that: The multi-objective dynamic optimization module adopts the NSGA-III algorithm, including a population stratification mechanism, a dynamic constraint processing unit and a real-time optimization engine; The population stratification mechanism divides the solution set into an elite layer (20%) and an exploration layer (80%). The elite layer uses local search to accelerate convergence. The dynamic constraint processing unit monitors key operating parameters of the system in real time, and the key operating parameters include influent COD concentration, flow rate or temperature in sewage treatment; The real-time optimization engine uses GPU accelerated computing to shorten the optimization cycle from 20 minutes to 4 minutes, achieving quasi-real-time control.

5. The control system for a sewage treatment plant according to claim 1, characterized in that: The adaptive decoupling controller includes a feedforward compensator, a coupling analysis module and a decoupling matrix; The feedforward compensator receives control instructions for aeration volume, internal recirculation ratio, and external recirculation ratio and performs preliminary decoupling. The coupling analysis module online identifies the time-varying coupling coefficient τ between the control variables. The decoupling matrix constructs a dynamic decoupling model based on the coupling coefficient τ. The formula is: u air : Actual aeration volume control amount; u ref : Actual reflux ratio control value; v air : Target aeration volume, the ideal value output by cloud optimization; v ref : Target reflux ratio, the ideal value output by cloud optimization; τ 11 and τ 22 : dynamic coefficient of the controlled variable itself; τ 12 and τ 21 : Cross-coupling coefficient between control variables, τ 12 is the interference intensity of the reflux ratio on aeration, τ 21 is the interference intensity of aeration on backflow; s: Laplace operator, which converts the time domain dynamic characteristics into the frequency domain transfer function; u: Actual control quantity vector, which determines the process DO and sludge return efficiency.

6. A control system for a sewage treatment plant according to claim 1, characterized in that: The edge computing module includes an edge layer that deploys an industrial IoT gateway in each process segment and performs the following operations: Data preprocessing, including wavelet noise reduction and outlier removal; Emergency control: when DO is lower than 1mg / L or COD suddenly increases to more than 500mg / L, PID control is directly activated to adjust the aeration volume or carbon source dosage; Data transmission, uploading key characteristic parameters to the cloud, the key characteristic parameters including activated sludge respiration rate OUR and nitrification rate NUR.

7. A control system for a sewage treatment plant according to claim 1, characterized in that: The digital twin fault prediction unit includes an electrical feature analysis channel, a process correlation analysis channel and a DS evidence theory fusion module. The electrical feature analysis channel is used to monitor the equipment current harmonic distortion rate and vibration spectrum. The process correlation analysis channel establishes a correlation matrix between equipment status and process parameters. The DS evidence theory fusion module makes decision-making fusion on the predicted results of the electrical feature analysis channel and the process correlation analysis channel, and reduces the false alarm rate to below 3%.

8. The control system for a sewage treatment plant according to claim 2, characterized in that: The microbial metabolic activity sensors and micro-spectral sensors are arranged and installed with at least two sensors in each partition, and a built-in automatic cleaning device is installed to perform ultrasonic cleaning every 30 minutes to avoid sludge adhesion.

9. A control system for a sewage treatment plant according to claim 4, characterized in that: The dynamic constraint treatment unit automatically relaxes the effluent TN constraint to 20 mg / L when the influent COD suddenly changes by more than 30%, and gradually recovers to 15 mg / L according to a linear gradient within 24 hours.

10. The control system for a sewage treatment plant according to claim 1, characterized in that: The adaptive decoupling controller also includes a robust control mode, which automatically switches when the sludge concentration change rate exceeds 5% / h, and suppresses coupled oscillation by enhancing feedback correction to control the sludge concentration fluctuation range within ±120mg / L.

Images