Sewage dosing automatic regulation and control method and system based on wireless sensor network

Abstract

The invention relates to the technical field of water treatment, in particular to a sewage dosing automatic regulation and control method and system based on a wireless sensor network, and the method comprises the steps: collecting water quality, flow and medicament state data through multiple types of sensors deployed at detection points; processing measured value drift caused by sensor probe pollution through an anti-drift model, constructing a feedforward model, and combining the feedforward model with the anti-drift model to predict water quality; based on the water quality prediction result, the dosage of the medicament is controlled, and the change information of the water quality is fed back in real time.

Description

Technical Field

[0001] This invention relates to the field of water treatment technology, and in particular to an automatic control method and system for wastewater chemical dosing based on wireless sensor networks. Background Technology

[0002] Wastewater chemical dosing is a crucial and common step in wastewater treatment processes. It refers to the process of adding chemical or biological agents to wastewater at a specific stage of treatment, utilizing physical, chemical, or biochemical reactions to remove pollutants, improve treatment efficiency, and ensure stable system operation.

[0003] The main purposes of wastewater chemical dosing are: 1. Removal of suspended solids and colloidal substances. Chemicals used include coagulants (such as polyaluminum chloride PAC, polyferric sulfate PFS, and aluminum sulfate AS) and flocculants (such as polyacrylamide PAM). Coagulants neutralize the surface charge of colloidal particles, destabilizing them; flocculants, through adsorption bridging or netting, aggregate destabilized fine particles into larger flocs, facilitating subsequent sedimentation, flotation, or filtration. Commonly used before primary sedimentation tanks (for enhanced sedimentation), before secondary sedimentation tanks (for enhanced biological sludge settling), and in advanced treatment (such as before high-efficiency sedimentation tanks and filters). 2. Phosphorus removal. Chemicals used include metal salts (such as aluminum salts PAC / AS and iron salts PFS / FeCl3) and lime Ca(OH)2. These react with phosphates in wastewater to form insoluble precipitates (such as aluminum phosphate, ferric phosphate, and hydroxyapatite). Dosing points include primary sedimentation tanks and biological treatment tanks. 3. Denitrification: The chemicals used are the carbon source required for denitrification. When the organic carbon source in the wastewater is insufficient, an external carbon source (such as methanol, ethanol, sodium acetate, glucose, etc.) needs to be added. 4. pH adjustment: The chemicals used are acids (such as sulfuric acid H₂SO₄, hydrochloric acid HCl) and alkalis (such as sodium hydroxide NaOH, lime Ca(OH)₂). This creates suitable conditions for biological treatment (such as nitrification requiring an alkaline environment) or chemical treatment (such as coagulation and chemical phosphorus removal having an optimal pH range). It also meets the pH requirements of discharge standards. Dosing points: inlet, biological treatment tank, outlet, etc.

[0004] Measuring water quality parameters using sensors connected to wireless networks is a common method for collecting data in existing wastewater treatment processes. However, in wastewater treatment, data drift caused by sensor probe contamination can lead to inaccurate measurements, affecting the prediction of sudden changes in water quality and the adjustment of chemical dosage. Summary of the Invention

[0005] This invention compensates for the drift of sensor measurement data in a sensor network to ensure the accuracy of the measurement data. Based on the acquired measurement data, it predicts sudden changes in water quality and determines the feedforward dosage in advance through a feedforward model. Based on the feedforward dosage, it obtains the control increment and solves the problem of lag in dosage adjustment when water quality changes suddenly.

[0006] The technical solution proposed in this invention is: an automatic control method for wastewater chemical dosing based on a wireless sensor network, the method comprising: An automatic control method for wastewater chemical dosing based on wireless sensor networks, the method comprising: Water quality, flow rate, and reagent status data are collected using various types of sensors deployed at monitoring points; The drift-proof model addresses the measurement drift caused by sensor probe contamination, and a feedforward model is constructed to combine with the drift-proof model for water quality prediction. Based on water quality prediction results, the dosage of chemicals is controlled, and water quality changes are reported in real time.

[0007] Preferably, the step of collecting water quality, flow rate, and reagent status data through multiple types of sensors deployed at detection points includes: Multiple types of sensors are deployed at the inlet, dosing point, mixing reaction zone, sedimentation tank, and outlet. The various types of sensors include: flow meters, pH meters, turbidity meters, online COD analyzers, online ammonia nitrogen analyzers, online total phosphorus analyzers, and Zeta level meters.

[0008] Preferably, the method of handling measurement drift caused by sensor probe contamination using an anti-drift model includes: The anti-drift model includes: Progressive drift model ; Step drift model ; in, , Indicates the sensor measurement value, This represents the actual measured value of water quality. Indicates the drift rate (calibrated in the laboratory using historical data); Random noise; Represents the unit step function. Indicates the time when the drift occurred. Indicates the step amplitude; Indicates the measurement compensation item; Solving for the parameters using Recursive Least Squares (RLS) Obtain measurement compensation value ; Output the actual water quality measurement value after drift compensation ; Set the drift monitoring metric, namely the average residual of the sliding window. ,when When a step shift occurs, the step drift model is triggered, and the measurement drift is processed using the step drift model; whereby, Indicates prior error. Sliding window length; Indicates the switching threshold. = .

[0009] Preferably, the construction of the feedforward model, combined with the anti-drift model, for predicting water quality includes: Determining whether water quality has undergone a sudden change includes: Water quality characteristics were extracted and normalized. Perform mutation detection: ;in, This represents the cumulative sum of water quality characteristics; among which, Represents the residuals corresponding to the water quality characteristics. This indicates the allowable deviation (e.g., 0.5). ; This represents the reference value under normal operating conditions. Indicates the standard deviation of normal fluctuation; when ;in, Indicates the mutation threshold; The water quality characteristics include one or more of the following: influent flow rate, pH value, turbidity, COD, ammonia nitrogen content, and total phosphorus value; By incorporating drift compensation, a feedforward model is constructed: ;in, Represents the autoregressive coefficient. Indicates the order of difference; Represents the moving average coefficient. Indicates the reference offset; Indicates the lag operator, ; This indicates a predicted water quality value at a future point in time.

[0010] Preferably, the step of controlling the dosage of the agent based on water quality prediction results and providing real-time feedback on water quality changes includes: Calculate feedforward dosing based on water quality predictions. ;in, Indicates the inflow rate; Indicates the feedforward gain coefficient; Calculate control increment ;in This represents the control quantity conversion coefficient; the control increment includes the actuator's action range (valve opening, pump speed, etc.) and system energy changes (changes in output power). Collect water quality data from monitoring points and send it back to the host computer.

[0011] Preferably, the step of processing the measurement drift caused by sensor probe contamination using an anti-drift model, constructing a feedforward model, and combining it with the anti-drift model to predict water quality further includes: Considering the nonlinear characteristics of sensor probe measurement drift and the correlation effects between multiple sensor probes, the anti-drift model is optimized, including: The true water quality values ​​are obtained by mapping nonlinear features using the radial basis function kernel (RBF). ;in, Indicates the first Real water quality measurements from individual probes; joint input vector from multiple probes ; Represents the RBF center point vector (determined through k-means clustering); Indicates kernel bandwidth; Indicates the weighting coefficient; The nonlinear characteristics of measurement drift were considered, and the feedforward prediction model was optimized, including: Constructing a state-space model: ;in, This represents the water quality state vector at the current moment. Indicates the control input at the current moment. This represents the data vector from multiple probes at the current moment. ; Indicates the number of probes; Represents the observation correction vector; , Represents the regression coefficient matrix. Represents the observation matrix; A feedforward prediction model is constructed using the LSTM-UKF algorithm. The input is a water quality state vector, and the output is the predicted water quality value. .

[0012] Preferably, the step of controlling the dosage of the agent based on water quality prediction results and providing real-time feedback on water quality changes further includes: Construct a multi-agent coordinated control model and establish the relationship between control inputs and water quality changes; A nonlinear multi-objective optimization function that integrates nonlinear drift compensation and multi-probe relationships is used to solve the multi-objective optimization function and obtain the optimal control input vector to coordinate the control of the dosage of various agents.

[0013] Preferably, the construction of the multi-agent coordinated control model to establish the relationship between control input and water quality changes includes: Establish a drug interaction matrix: ;in, Indicates the first The water quality deviation of the first type of water is the first The amount of change in water quality characteristics Represents the drug interaction matrix. Indicates use for control The input for controlling the dosage of the seed drug; Represents the control input vector; The nonlinear multi-objective optimization function, which integrates nonlinear drift compensation and multi-probe relationships, is solved to obtain the optimal control input vector for coordinating the dosage of various agents, including: Establish a multi-objective optimization function: ;in, Indicates the weight of water quality characteristics. This represents the Huber loss function. Measurements that represent water quality characteristics Set values ​​for water quality characteristics This represents the control increment vector. ; This represents the cost vector corresponding to the control increment; ; , These represent the control increment weight and the control error weight, respectively. Represents the ideal control input vector; Obtain water quality prediction values Substitute into the multi-objective optimization function To obtain the nonlinear multi-objective optimization function ; Solving the problem using the Sequential Quadratic Programming (SQP) method iteratively. ,include: With preset control parameter vector As a reference point, that is ; Calculate the current water quality deviation and the gradient of the Huber loss function. ; ;in, Indicates the k-th iteration. Water quality deviation; Constructing the Hessian matrix of the quadratic programming subproblem and gradient vector ;in, ; ; Represents the identity matrix; The subproblem of quadratic programming is solving the equation ;in Indicates the search direction. ; Represents the constraint matrix; Solve the quadratic programming subproblem using the effective set method or interior point method to obtain the search direction. ; The step size is found using the Armijio line search criterion. ; so that at the corresponding number of iterations, ; where constant ;(For example =10 -4 ); The optimal control input vector is obtained when the maximum number of iterations N is reached. ;in, This represents the control input obtained after the (N-1)th iteration. This represents the search direction obtained after the (N-1)th iteration; since This refers to obtaining the control input for the dosage of various drugs.

[0014] An automatic control system for wastewater dosing based on a wireless sensor network is provided, wherein the system is used to execute the aforementioned automatic control method for wastewater dosing based on a wireless sensor network.

[0015] A computer-readable storage medium storing a computer program that is executed by a processor to implement the aforementioned automatic control method for wastewater dosing based on a wireless sensor network.

[0016] The beneficial effects of this invention are: 1. This invention considers the nonlinear characteristics of sensor measurement data drift, i.e. the influence between multiple probes, and uses radial basis function kernel mapping to obtain the true value of water quality. Based on the above true value, water quality is predicted to obtain a digital predicted value, which improves the accuracy of measurement data in multi-probe measurement scenarios.

[0017] 2. This invention obtains the optimal control input vector by constructing and solving a nonlinear multi-objective optimization function, coordinating the dosing of multiple agents, and adapting to application scenarios requiring joint control of multiple agents. Attached Figure Description

[0018] Figure 1 This is a flowchart of the automatic control method for wastewater dosing based on a wireless sensor network according to the present invention. Detailed Implementation

[0019] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious modifications will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0020] It is understood that the term "a" should be understood as "at least one" or "one or more," that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0021] Example 1: refer to Figure 1 The technical solution provided by this invention is: an automatic control method for wastewater dosing based on a wireless sensor network, comprising the following steps: Step 1: Collect water quality, flow rate, and reagent status data using multiple types of sensors deployed at the detection points. Specifically, deploy multiple types of sensors at the inlet, dosing point, mixing reaction zone, sedimentation tank, and outlet to acquire the corresponding data. These multiple types of sensors include flow meters, pH meters, turbidity meters, online COD analyzers, online ammonia nitrogen analyzers, online total phosphorus analyzers, and Zeta level gauges.

[0022] Step 2: Address the measurement drift caused by sensor probe contamination using an anti-drift model, and construct a feedforward model to predict water quality in conjunction with the anti-drift model. This includes the following steps: The method for handling measurement drift caused by sensor probe contamination using an anti-drift model includes the following steps: The anti-drift model includes: Progressive drift model ; Step drift model ; in, , Indicates the sensor measurement value, This represents the actual measured value of water quality. Indicates the drift rate (calibrated in the laboratory using historical data); Random noise; Represents the unit step function. Indicates the time when the drift occurred. Indicates the step amplitude; Indicates the measurement compensation item; Solving for the parameters using Recursive Least Squares (RLS) Obtain measurement compensation value ; Output the actual water quality measurement value after drift compensation ; Set the drift monitoring metric, namely the average residual of the sliding window. ,when When a step shift occurs, the step drift model is triggered, and the measurement drift is processed using the step drift model; whereby, Indicates prior error. Sliding window length; Indicates the switching threshold. = .

[0023] The process of constructing a feedforward model and combining it with an anti-drift model to predict water quality includes the following steps: Determining whether water quality has undergone a sudden change includes: Water quality characteristics were extracted and normalized. Perform mutation detection: ;in, This represents the cumulative sum of water quality characteristics; among which, Represents the residuals corresponding to the water quality characteristics. This indicates the allowable deviation (e.g., 0.5). ; This represents the reference value under normal operating conditions. Indicates the standard deviation of normal fluctuation; when ;in, Indicates the mutation threshold; Water quality characteristics include one or more of the following: influent flow rate, pH value, turbidity, COD, ammonia nitrogen content, and total phosphorus value; By incorporating drift compensation, a feedforward model is constructed: ;in, Represents the autoregressive coefficient. Indicates the order of difference; Represents the moving average coefficient. Indicates the reference offset; Indicates the lag operator, ; This represents the predicted water quality value at a future point in time. Step 3: Based on the water quality prediction results, control the dosage of the chemical and provide real-time feedback on water quality changes, including the following steps: Calculate feedforward dosing based on water quality predictions. ;in, Indicates the inflow rate; Indicates the feedforward gain coefficient; Calculate control increment ;in This represents the control quantity conversion coefficient; the control increment includes the actuator's action range (valve opening, pump speed, etc.), system energy change (output power change), and reagent dosage change; water quality data at the detection points are collected and fed back to the host computer.

[0024] Example 2: Sensor measurement data drift exhibits nonlinear characteristics. For example, multiple probes drifting simultaneously and influencing each other can cause linear correction to fail. A compensation model capable of capturing this cross-influence is needed. Therefore, based on Example 1, we propose the following technical solution: Considering the nonlinear characteristics of sensor probe measurement drift and the correlation effects between multiple sensor probes, the anti-drift model is optimized, including the following steps: The true water quality values ​​are obtained by mapping nonlinear features using the radial basis function kernel (RBF). ;in, Indicates the first Real water quality measurements from individual probes; joint input vector from multiple probes ; Represents the RBF center point vector (determined through k-means clustering); Indicates kernel bandwidth; Indicates the weighting coefficient; The nonlinear characteristics of measurement drift were considered, and the feedforward prediction model was optimized, including: Constructing a state-space model: ;in, This represents the water quality state vector at the current moment. Indicates the control input at the current moment. This represents the data vector from multiple probes at the current moment. ; Indicates the number of probes; Represents the observation correction vector; , Represents the regression coefficient matrix. Represents the observation matrix; A feedforward prediction model is constructed using the LSTM-UKF algorithm. The input is a water quality state vector, and the output is the predicted water quality value. .

[0025] To adapt to scenarios involving coordinated control of multiple agents, the following steps are also included: Constructing a multi-agent coordinated control model and establishing the relationship between control inputs and water quality changes includes the following steps: Establish a drug interaction matrix: ;in, Indicates the first The water quality deviation of the first type of water is the first The amount of change in water quality characteristics Represents the drug interaction matrix. Indicates use for control Input for controlling the dosage of the drug.

[0026] A nonlinear multi-objective optimization function integrating nonlinear drift compensation and multi-probe relationships is used to solve the nonlinear multi-objective optimization function and obtain the optimal control input vector to coordinate the dosing of various agents. The specific steps include: Establish a multi-objective optimization function: ;in, Indicates the weight of water quality characteristics. This represents the Huber loss function. Measurements that represent water quality characteristics Set values ​​for water quality characteristics This represents the control increment vector. ; This represents the cost vector corresponding to the control increment; ; , These represent the control increment weight and the control error weight, respectively. Represents the ideal control input vector; Obtain water quality prediction values Substitute into the multi-objective optimization function To obtain the nonlinear multi-objective optimization function ; Solving the problem using the Sequential Quadratic Programming (SQP) method iteratively. ,include: With preset control parameter vector As a reference point, that is ; Calculate the current water quality deviation and the gradient of the Huber loss function. ; ;in, Indicates the k-th iteration. Water quality deviation; Constructing the Hessian matrix of the quadratic programming subproblem and gradient vector ;in, ; ; Represents the identity matrix; The subproblem of quadratic programming is solving the equation ;in Indicates the search direction. ; Represents the constraint matrix; Solve the quadratic programming subproblem using the effective set method or interior point method to obtain the search direction. ; The step size is found using the Armijio line search criterion. ; so that at the corresponding number of iterations, ; where constant ;(For example =10 -4 ); The optimal control input vector is obtained when the maximum number of iterations N is reached. ;in, This represents the control input obtained after the (N-1)th iteration. This indicates the search direction obtained after the (N-1)th iteration; because This refers to obtaining the control input for the dosage of various drugs.

[0027] For example: ;in, , , The changes in pH, total phosphorus, and nitrogen oxides were recorded separately. , and These represent the dosage of polyaluminum chloride, the amount of carbon source added, and Distribution volume.

[0028] if PAC hydrolysis reduces ; Slight increase in carbon source addition By increasing the amount of carbon source input, a balance can be achieved. Reduce the amount to ensure water quality.

[0029] The processes described above with reference to the flowcharts in the embodiments disclosed in this invention can be implemented as computer software programs. The embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), it performs the functions defined in the methods of this application. It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wire segments, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless segments, wire segments, optical fibers, RF, etc., or any suitable combination thereof.

[0030] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0031] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functions and structural principles of the present invention have been shown and explained in the embodiments. Without departing from the principles described, the implementation of the present invention may have any changes or modifications.

Claims

1. A sewage dosing automatic control method based on a wireless sensor network, characterized in that, The method comprises: The method comprises: Collecting water quality, flow and reagent state data through multiple types of sensors deployed at detection points; Processing measurement value drift caused by sensor probe pollution through a drift prevention model, and constructing a feedforward model to predict water quality in combination with the drift prevention model; Based on the water quality prediction result, the amount of reagent added is controlled, and the change information of water quality is fed back in real time.

2. The method according to claim 1, wherein, The method comprises: Deploying multiple types of sensors at the water inlet end, the reagent adding point, the mixing reaction zone, the sedimentation tank and the water outlet end; The multiple types of sensors comprise a flowmeter, a pH meter, a turbidimeter, a COD online analyzer, an ammonia nitrogen online analyzer, a total phosphorus online analyzer and a Zeta point instrument.

3. The method according to claim 2, wherein, The method comprises: The drift prevention model comprises: Gradual drift model ; Step drift model ; wherein , represents a sensor measurement value, represents a true measurement value of water quality, represents a drift rate (from historical data, lab calibration); random noise; represents a unit step function, represents a time of drift occurrence, represents a step amplitude; represents a measurement compensation term; Solving parameters by recursive least square method RLS Obtaining measurement compensation value ; Outputting a water quality true measurement value after drift compensation ; Setting a drift monitoring index, i.e. a sliding window average residual When , it is judged that a drift occurs a step, a step drift model is triggered, and the measurement value drift is processed through the step drift model; wherein, represents a prior error, a sliding window length; represents a switching threshold, = .

4. The method according to claim 3, wherein, The method comprises: Judging whether the water quality has mutated, comprising: Extracting water quality features and performing normalization processing; Performing mutation detection: ; wherein, represents a cumulative sum of water quality characteristics; wherein, represents a residual of the respective water quality characteristic, represents an allowed deviation (e.g. 0.5); ; represents a normal operating condition reference value, represents a normal fluctuation standard deviation; when ; wherein, represents a mutation threshold value; The water quality features comprise one or more of water inlet flow, pH value, turbidity, COD, ammonia nitrogen content and total phosphorus value; Fusing drift compensation to construct a feedforward model: ; wherein, denotes an autoregressive coefficient, denotes a difference order; denotes a moving average coefficient, denotes a reference offset; denotes a lag operator, ; denotes a water quality prediction value for a future time instant.

5. The method according to claim 4, wherein, The method comprises: Based on the water quality prediction value, the feedforward dosing amount is calculated ; wherein, represents the influent flow rate; represents the feedforward gain coefficient; Computational control increment ; wherein represents a control quantity conversion coefficient; the control increment includes an actuator action amplitude (valve opening, pump speed, etc.), a system energy change (change in output power); Collecting water quality data at the detection point and feeding back to the upper computer.

6. The method according to claim 5, wherein, The method further comprises: Considering the nonlinear characteristics of measurement value drift and the correlation influence between multiple sensor probes, optimizing the drift prevention model, comprising: Using a radial basis function kernel RBF to map nonlinear characteristics to obtain water quality true values: ; wherein, represents the real measurement value of water quality of the th probe; the multi-probe joint input vector ; represents the RBF center point vector; represents the kernel bandwidth; represents the weight coefficient; Considering the nonlinear characteristics of measurement value drift, optimizing the feedforward prediction model, comprising: Constructing the state space model: ; wherein, denotes the water quality state vector at the current time instant, denotes the control input at the current time instant, denotes the multi-probe data vector at the current time instant, ; denotes the number of probes; denotes the observation correction vector; , denotes the regression coefficient matrix, denotes the observation matrix; A feedforward prediction model is constructed by an LSTM-UKF algorithm, inputting a water quality state vector and outputting a predicted water quality prediction value .

7. The method according to claim 6, wherein, The method further comprises: Constructing a multi-reagent coordinated control model to establish the relationship between control input and water quality change; Fusing a nonlinear multi-objective optimization function of nonlinear drift compensation and multi-probe relationship to solve the multi-objective optimization function to obtain an optimal control input vector to coordinate the amount of multiple reagents added.

8. The method according to claim 7, wherein, The method comprises: Establishing a reagent interaction matrix: ;in, Indicates the first The water quality deviation of the first type of water is the first The amount of change in water quality characteristics Represents the drug interaction matrix. Indicates use for control The input for controlling the dosage of the seed drug; Represents the control input vector; The method further comprises: A multi-objective optimization function is established: ; wherein, represents a water quality characteristic weight, represents a Huber loss function, represents a measured value of a water quality characteristic, a set value of a water quality characteristic, represents a control increment vector, ; represents a cost vector corresponding to the control increment; ; , respectively represent a control increment weight and a control error weight; represents an ideal control input vector; Obtaining water quality prediction values , into a multi-objective optimization function , obtaining a nonlinear multi-objective optimization function ; solved iteratively by a sequential quadratic programming method SQP comprising: with a preset control parameter vector as a reference point, i.e. ; Computing current water quality bias and Huber loss function gradient ; ; wherein, represents the water quality deviation of the kth iteration of the kth water quality; and the kth water quality. constructing the hessian matrix of a quadratic programming subproblem and gradient vectors ; wherein ; ; denotes the identity matrix The quadratic programming subproblem is to solve the equation ; where denotes the search direction, ; denotes the constraint matrix; The search direction is obtained by solving a quadratic programming subproblem using an active set method or an interior point method ; Armijio line search criteria is used to find the step size ; such that at the corresponding iteration number, ; where the constant ; (e.g. = 10 -4 ); At the maximum number of iterations N, the optimal control input vector is obtained ; wherein denotes the control input obtained after the N-1 iteration, denotes the search direction obtained after the N-1 iteration; since the control input of the drug amount corresponding to the plurality of drugs is obtained.

9. The automatic control system for sewage dosing based on wireless sensor network, characterized in that, The system is used to execute the automatic control method of sewage reagent based on a wireless sensor network in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the automatic control method of sewage reagent based on a wireless sensor network in any one of claims 1-8.

Images