A water resource multi-stage circulation control system and method

Abstract

The application discloses a water resource multi-stage circulation control system, which comprises a sewage treatment device, a temperature feedback controller, a flow feedback controller, a determinator and a feedback controller group. The output end of the sewage treatment device is connected with the input end of the temperature feedback controller, the flow feedback controller and the feedback controller group respectively. The output end of the temperature feedback controller and the flow feedback controller is connected with the input end of the determinator. The output end of the determinator is connected with the input end of the sewage treatment device and the feedback controller group respectively. The output end of the feedback controller group is connected with the input end of the sewage treatment device. The application aims at guaranteeing the water quality to reach the standard, realizing efficient sewage treatment, and solving the problems of insufficient intelligence, energy and material waste, lagged water quality data online acquisition, insufficient water quality online monitoring method and the like in the traditional sewage treatment technology.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of environmental engineering, and particularly relates to a water resource multi-stage circulation control system and method. BACKGROUND

[0002] In the prior art, sewage treatment technologies are mainly divided into physical method treatment, chemical method treatment and biological method treatment. The physical method treatment refers to a method for separating pollutants in sewage in a suspended state by using physical action, mainly including sedimentation, screening, air flotation, centrifugal and cyclone separation. The chemical method treatment refers to a method for adding chemicals to sewage to separate and recover pollutants in sewage by chemical reaction or to convert the pollutants into harmless substances, mainly including neutralization method, oxidation-reduction method, electrolysis method, adsorption method and chemical precipitation method. The biological method treatment refers to a method for creating an environment conducive to the growth and reproduction of microorganisms by taking certain artificial measures, so that microorganisms proliferate in large quantities to improve the oxidation and decomposition of organic pollutants by microorganisms, convert the organic pollutants into harmless substances, and purify sewage. The biological method treatment can be divided into aerobic treatment method and anaerobic treatment method.

[0003] Generally, flow rate and water temperature as reaction conditions of sewage treatment have a great influence on the removal rate of pollutants. Due to different uses and destinations of treated water resources, different water quality indexes need to be monitored during the sewage treatment process, and different effluent standards are required. Due to the relatively weak foundation of the sewage treatment industry in China, the traditional sewage treatment technology still has many defects, and there are phenomena such as imbalance and disorder in equipment installation and application.

[0004] For example, the sewage treatment intelligent control system construction idea based on the fusion of mechanism model and data model proposed by Yin Fengjun, Xu Zeyu and Liu Hong has a disconnection phenomenon between the design operation state of the sewage plant and the field water quality and working conditions in actual operation, resulting in waste of energy and material consumption and increase of operating cost. Intelligence is an inevitable trend of the development of sewage treatment technology, but the development of sewage treatment intelligence lacks innovation in promotion mode, needs to break through the difficulty of online acquisition of water quality data, and further establish and refine the intelligent control technology scheme of the fusion of mechanism and data driven models. Therefore, a more intelligent and automatic control system is needed at the present stage to continuously optimize and improve the treatment system to save the purposes of reagent cost, time cost and labor cost while meeting the water quality standards. SUMMARY

[0005] The purpose of the present application is to provide a water resource multi-stage circulation feedback control system and method, which adjusts sewage treatment conditions such as temperature, flow rate, chemical reagent dosage, aeration parameter and microbial dosage according to the real-time water inflow to ensure that the effluent water quality meets the standards, so as to realize efficient sewage treatment and solve the problems of insufficient intelligence of traditional sewage treatment technology, waste of energy and material consumption, lag of online acquisition of water quality data and insufficient online water quality monitoring method.

[0006] To solve the above technical problems, the technical scheme adopted by the present application is:

[0007] A water resource multi-stage circulation control system, comprising a sewage treatment device, a temperature feedback controller, a flow feedback controller, a determinator, a feedback controller group;

[0008] The output end of the sewage treatment device is connected with the input end of the temperature feedback controller, the flow feedback controller and the feedback controller group respectively, the output end of the temperature feedback controller and the flow feedback controller is connected with the input end of the determinator, the output end of the determinator is connected with the input end of the sewage treatment device and the feedback controller group respectively, and the output end of the feedback controller group is connected with the input end of the sewage treatment device.

[0009] The feedback controller group comprises y1 feedback controller, y2 feedback controller, y3 feedback controller, …, ym feedback controller. m The feedback controller, wherein m is a positive integer;

[0010] The output end of the determinator is connected with the input end of the sewage treatment device and the y1 feedback controller respectively, the output end of the y1 feedback controller is connected with the input end of the sewage treatment device and the y2 feedback controller respectively, the output end of the y2 feedback controller is connected with the input end of the sewage treatment device and the y3 feedback controller respectively, and so on, the output end of the ym feedback controller is connected with the input end of the sewage treatment device. m The output end of the feedback controller is connected with the input end of the sewage treatment device.

[0011] The temperature feedback controller, the flow feedback controller and the feedback controller group are internally provided with a predictive regulator, the predictive regulator comprises a sensor, an optimizer, a simulator, a controller and a controlled element, the output end of the sensor is connected with the input end of the optimizer, the output end of the optimizer is connected with the input end of the simulator, and the output end of the controller is connected with the simulator and the controlled element respectively;

[0012] A first error regulator is arranged between the output end of the optimizer and the controller, the information output by the optimizer and the controller is transmitted to the first error regulator, the first error regulator adjusts according to the error between the optimizer and the controller and then acts on the controller to update the control variable;

[0013] A second error regulator is arranged between the output end of the simulator and the output end of the controlled element, the second error regulator adjusts according to the error between the output end of the simulator and the controlled element and then acts on the simulator to update the internal optimization model, so as to correct the expected state of the next round of optimization and continuously perform cyclic feedback adjustment; the internal control needs to be solved by repeated intelligent algorithm prediction and optimization in each time stage to obtain the optimal control step, and then the output of the solving controller is applied to the sewage treatment device.

[0014] The predictive regulator, in use, adopts the following steps:

[0015] Step S1: sensor t+t a detects water resource inflow index X(t+t a ), and transmits information to the optimizer to enter step S2;

[0016] Step S2: the optimizer performs optimization calculation according to the set water quality target value and the real-time simulation model input by the simulator, and gives the current control variable U(t+t a ), and enters steps S3 and S4;

[0017] Step S3: the control variable U(t+t a ) is applied to the simulator for simulation, and the output result Yd(t+t a +△t) is outputted, and step S7 is entered;

[0018] Step S4: after the controller receives the information transmitted by the optimizer, the actual output control variable U’(t+t a ) is outputted according to the existing data, and step S5 is entered;

[0019] Step S5: the optimizer calculates the control output U(t+t a ) and the information U’(t+t a ) of the actual output of the controller, and transmits them to the first error regulator, which adjusts the error between the two and then applies it to the controller for adaptive stability adjustment, and step S6 is entered;

[0020] Step S6: the actual output control variable U’(t+ta) is applied to the sewage treatment device to output the actual output Y(t+t a +△t); if the error between the actual output Y(t+t a +△t) and the target value is greater than the allowable error, step S7 is entered; if the error between the actual output Y(t+t a +△t) and the target value is less than the allowable error, step S8 is entered;

[0021] Step S7: the information outputted by the simulator and the controlled element is transmitted to the second error regulator, which adjusts the error between the two and then applies it to the simulator for updating the real-time simulation model, and step S2 is entered;

[0022] Step S8: the optimization target is continuously adjusted through loop feedback, and the optimization control process is realized when the error between the actual water treatment effect and the target value is less than the allowable error.

[0023] The specific real-time simulation model of the temperature feedback controller, the flow feedback controller and the internal simulator of the feedback controller group is as follows

[0024]

[0025] s.t.

[0026] y(t+1) = f(y(k), u(t)) (1)

[0027]

[0028]

[0029] External constraints: the objective function represents that the state y(t+k) of the system and the desired state y d (t+k) are as close as possible within the next N time steps, constraint (1) represents the dynamic characteristics of the controlled object, f is the prediction model, which is not limited to various machine learning algorithms including recurrent neural networks, constraints (2) and (3) represent the upper and lower limit constraints on the water treatment control parameter u(t) and the state parameter y(t) respectively.

[0030] Compared with the prior art, the present application has the following technical effects:

[0031] 1) The water resource multi-level circulation control method first adjusts the controllable variables such as temperature, flow, etc. of wastewater treatment, so that the use efficiency of reagents, the operation efficiency of equipment, etc. are maximized, thereby reducing the cost of reagents, electricity, equipment operation and maintenance, etc.

[0032] 2) Various feedback controllers are combined in series and in parallel to form a water resource multi-level circulation control system, which reduces the size of the minimum independent unit of the wastewater treatment process, reduces the complexity of the mechanism model, and enhances the controllability and predictability of the reaction unit;

[0033] 3) Control indicators can be expanded, and other influent indicators can be added to the control system, such as adding pH value to the first-level circulation, adding COD, BOD, NH3-N, NO - 3-N, NO - 2-N, TN, TP, DO concentration, color, turbidity, etc. multiple water quality indicators to the multi-level circulation;

[0034] 4) Easy to operate, with stronger applicability and flexibility, can be applied to physical, chemical, biological and other treatment methods and first, second, third and other treatment scenes, meeting various water resource index requirements;

[0035] 5) Online monitoring of water quality data, real-time monitoring of wastewater treatment process state and abnormal state of influent, achieving the purpose of rapid regulation and control to ensure that the effluent water quality meets the relevant standards and the system is always in the optimal state. BRIEF DESCRIPTION OF DRAWINGS

[0036] The application will be further described below in connection with the accompanying drawings and embodiments:

[0037] Fig. 1 is a water resource multi-stage circulation control system diagram;

[0038] Fig. 2 is a feedback controller block diagram;

[0039] Fig. 3 is a water resource multi-stage circulation control method flow chart. DETAILED DESCRIPTION

[0040] As Figs. 1 to 3 shown, a water resource multi-stage circulation control system, which comprises a sewage treatment device 1, a temperature feedback controller 2, a flow feedback controller 3, a determinator 4, a feedback controller group;

[0041] The output end of the sewage treatment device 1 is connected with the input end of the temperature feedback controller 2, the flow feedback controller 3 and the feedback controller group respectively, the output end of the temperature feedback controller 2 and the flow feedback controller 3 is connected with the input end of the determinator 4, the output end of the determinator 4 is connected with the input end of the sewage treatment device 1 and the feedback controller group respectively, and the output end of the feedback controller group is connected with the input end of the sewage treatment device 1.

[0042] The feedback controller group comprises a y1 feedback controller 5, a y2 feedback controller 6, a y3 feedback controller 7, …, a ym feedback controller 8, wherein m is a positive integer; m The feedback controller, wherein m is a positive integer;

[0043] The output end of the determinator 4 is connected with the input end of the sewage treatment device 1 and the y1 feedback controller 5 respectively, the output end of the y1 feedback controller 5 is connected with the input end of the sewage treatment device 1 and the y2 feedback controller 6 respectively, the output end of the y2 feedback controller 6 is connected with the input end of the sewage treatment device 1 and the y3 feedback controller 7 respectively, and so on, the output end of the ym feedback controller 8 is connected with the input end of the sewage treatment device 1. m The output end of the feedback controller is connected with the input end of the sewage treatment device 1.

[0044] The temperature feedback controller 2, the flow feedback controller 3 and the feedback controller group are internally provided with a predictive regulator, the predictive regulator comprises a sensor 8, an optimizer 9, an emulator 10, a controller 11 and a controlled element 12, the output end of the sensor 8 is connected with the input end of the optimizer 9, the output end of the optimizer 9 is connected with the input end of the emulator 10, and the output end of the controller 11 is connected with the emulator 10 and the controlled element 12 respectively;

[0045] A first error regulator 13 is provided between the output of the optimizer 9 and the controller 11, and the information output by the optimizer 9 and the controller 11 is transmitted to the first error regulator 13, which is adjusted according to the error between the optimizer 9 and the controller 11 and then acts on the controller 11 to update the control variable;

[0046] A second error regulator 14 is provided between the output of the simulator 10 and the output of the controlled element 12, and the second error regulator 14 is adjusted according to the error between the output of the simulator 10 and the controlled element 12 and then acts on the simulator 10 to update the internal optimization model, so as to correct the expected state of the next round of optimization and continuously perform cyclic feedback adjustment; the internal control needs to be solved by repeated intelligent algorithm prediction and optimization in each time stage to obtain the optimal control step, and then the output of the solving controller is applied to the sewage treatment device 1.

[0047] The following steps are adopted when the prediction regulator is used:

[0048] Step S1: The sensor 8t+t a detects the water resource inflow index X(t+t a ) at the moment, transmits the information to the optimizer 9, and enters step S2;

[0049] Step S2: The optimizer 9 optimizes and solves according to the set water quality target value and the real-time simulation model input by the simulator 10, and gives the current control variable U(t+t a ), and enters steps S3 and S4;

[0050] Step S3: The control variable U(t+t a ) is applied to the simulator 10 for simulation, and the output result Yd(t+t a +△t) is obtained, and enters step S7;

[0051] Step S4: After receiving the information transmitted by the optimizer 9, the controller 11 actually outputs the control variable U’(t+t a ) according to the existing data, and enters step S5;

[0052] Step S5: The optimizer 9 calculates the control output U(t+t a ) and the information U’(t+t a ) actually output by the controller 11, and transmits them to the first error regulator 13, which is adjusted according to the error between the two and then acts on the controller (11) for adaptive stability adjustment of the controller 11, and enters step S6;

[0053] Step S6: The actual output control variable U’(t+ta) acts on the sewage treatment device to output the actual output Y(t+ta +△t); if the actual output quantity Y(t+t) a If the error between the actual output Y(t+t) and the target value is greater than the allowable error, proceed to step S7; if the actual output Y(t+t) is greater than the target value, proceed to step S7. a If the error between +△t) and the target value is less than the allowable error, proceed to step S8;

[0054] Step S7: The information output by the simulator 10 and the controlled element 12 is transmitted to the second error regulator 14. The second error regulator 14 adjusts the information according to the error between the two and then applies it to the simulator 10 to update the real-time simulation model. Proceed to step S2.

[0055] Step S8: Continuously perform cyclical feedback to adjust and optimize the target. When the error between the actual water treatment effect and the target value is less than the allowable error, the optimized control process is achieved.

[0056] The specific real-time simulation models of temperature feedback controller 2, flow feedback controller 3, and internal simulator 10 of the feedback controller group are as follows:

[0057]

[0058] st

[0059] y(t+1)=f(y(k),u(t)) (1)

[0060]

[0061]

[0062] External constraints: The objective function represents the system's state y(t+k) and the desired state y. d (t+k) should be as close as possible to the target value in the next N time steps. Constraint (1) represents the dynamic characteristics of the controlled object. f is the prediction model, which is not limited to various machine learning algorithms including recurrent neural networks. Constraints (2) and (3) represent the upper and lower limits of the water treatment control parameter u(t) and the state parameter y(t), respectively.

[0063] In this invention, ym represents the target water inlet, which may include: physical indicators such as transparency, odor, turbidity, color, and temperature; and single-component indicators such as NH3-N and Cr. 6+ The indicators include: plasma or organic matter concentration; comprehensive component indicators such as total organic carbon, total phosphorus, total nitrogen, pH value, and total bacterial count; assessment indicators such as COD, hardness, alkalinity, and BOD; biotoxicity indicators such as the concentration of toxic substances such as cyanide, mercury, and lead; water quality transformation potential indicators such as chlorophyll, total phosphorus, total nitrogen, and permanganate index; and process indicators such as sludge volume index (SVI) and sludge density index (SDI). There are a total of m indicators.

[0064] In the present application, y refers to the output quantity, which is the water output index, i.e. the optimal water inflow, optimal temperature, removal rate of each pollutant, etc., corresponding to the water inflow index, and there are m of them; u refers to the controllable variable, which is the various operating conditions used in the actual wastewater treatment process, which can include: the amount of certain chemical agents such as flocculants, demulsifiers, and oxidation-reduction agents; aeration parameters such as aeration intensity, air intake, and pump flow; the operating power of machines such as air blowers, filter presses, and water pumps; the amount of certain microorganisms; the amount of activated sludge, etc., and there are n of them; X refers to the m-order vector of input signals; U refers to the n-order vector of controllable variables; and Y refers to the m-order vector of output quantities.

[0065] Embodiment: A multi-stage circulation feedback control system for water resources, which comprises a wastewater treatment device, a temperature feedback controller, a flow feedback controller, a determinator, and m-2 water quality index feedback controllers. The output end of the wastewater treatment device is connected to the input ends of the temperature feedback controller, the flow feedback controller, and the m-2 water quality index feedback controllers; the output ends of the temperature feedback controller and the flow feedback controller are connected to the input end of the determinator; the output end of the determinator is connected to the input ends of the wastewater treatment device and the y3 feedback controller; the output end of the y3 feedback controller is connected to the input ends of the wastewater treatment device and the y4 feedback controller; the output end of the y4 feedback controller is connected to the input ends of the wastewater treatment device and the y5 feedback controller; and so on, until the output end of the ym-2 feedback controller is connected to the input ends of the wastewater treatment device and the ym-1 feedback controller. m The output end of the feedback controller is connected to the input end of the wastewater treatment device.

[0066] In use, the system comprises the following steps:

[0067] Step one: Obtain the target value of the water inflow index, and establish an optimization control model for the water inflow index.

[0068] Step two: Adjust the temperature and flow.

[0069] Step three: Determine whether the temperature and flow conditions meet the requirements.

[0070] Step four: Adjust other operating conditions.

[0071] Step five: Achieve the target value of the water quality index.

[0072] In step two, the temperature feedback controller and the flow feedback controller detect the water inflow index X1(t) at time t, and first perform feedback adjustment on the water temperature and flow.

[0073] In step three, the determinator determines whether the difference between the output result of the temperature feedback controller or the flow feedback controller this time and the output result of the temperature feedback controller or the flow feedback controller last time meets the minimum requirement Δy 1min or Δy 2minIf the requirements are met, the output temperature and flow rate results will be applied to the wastewater treatment device for further water quality adjustment; if the requirements are not met, feedback adjustment of temperature and flow rate will need to be performed cyclically.

[0074] In step four, after optimizing the water temperature and flow rate in the influent parameters, the other operating conditions in the influent parameters are then optimized through multi-stage cyclic feedback adjustment. When adjusting other operating conditions, the external control system employs the following steps:

[0075] S1: A multi-level loop y a The feedback controller only responds to one water quality index y. a Feedback adjustments are made, and through optimization and regulation, an output is generated for this water quality indicator y. a Optimal operating conditions U a ;

[0076] S2: Set the optimal operating condition U a When applied to a wastewater treatment device, the overall water quality index output Y under these operating conditions is obtained. a And output this overall water quality index Y a As the next multi-level loop y a+1 The input X of the feedback controller a+1 .

[0077] The internal controller employs the following steps:

[0078] S1: The water inflow index X(t+t) is detected by the sensor at time t+ta. a The system sends information to the optimizer and enters S2;

[0079] S2: The optimizer performs optimization based on the set water quality target value and the real-time simulation model input by the simulator, and gives the current control variable U(t+t). a ), then enter S3 and S4;

[0080] S3:U(t+t a The simulation is performed on the simulator, and the output result Y is obtained. d (t+t a +△t);

[0081] S4: After the optimizer sends information to the controller, the controller actually outputs the control variable U'(t+t) based on the existing data. a Enter S5;

[0082] S5: The optimizer calculates the information between the control output and the actual output of the controller and transmits it to the first error regulator. The first error regulator adjusts the controller based on the error between the two and then applies it to the controller for adaptive stability adjustment.

[0083] S6: the actual output control variable U'(t+t a ) acts on the water treatment device to output the actual output Y(t+t a +△t) into S7, S8;

[0084] S7: the information output by the simulator and the valve is transmitted to the second error regulator, which adjusts according to the error between the two and then acts on the simulator to update the prediction model, and enters S2;

[0085] S8: continuously perform cyclic feedback adjustment and optimization of the target, and when the error between the actual output and the target value is less than the allowable error, the optimization control process is realized.

[0086] In step five, the last multi-stage cycle y m After the feedback controller completes the optimization adjustment and the sewage treatment device executes the corresponding operation condition, the water quality indicators of the entire water resource multi-stage cycle control system have reached the target value. The control system runs throughout the sewage treatment process, and adjusts the temperature, flow rate, chemical dosage, aeration parameters, and microbial dosage in real time according to changes in the influent water quality indicators. After the sewage treatment is completed, the entire multi-stage cycle control system is ended.

[0087] It should be noted that one operation condition does not correspond to one or several water quality indicators, but has cross effects and mutual effects, so for any multi-stage cycle feedback controller, the influent water quality indicators, effluent water quality indicators, and operation conditions need to be input or output together.

Claims

1. A water resource multi-level circulation control system, characterized in that, It includes sewage treatment device (1), temperature feedback controller (2), flow feedback controller (3), determinator (4), feedback controller group; The output end of the sewage treatment device (1) is connected with the input end of the temperature feedback controller (2), the flow feedback controller (3) and the feedback controller group respectively, the output end of the temperature feedback controller (2) and the flow feedback controller (3) is connected with the input end of the determinator (4), the output end of the determinator (4) is connected with the input end of the sewage treatment device (1) and the feedback controller group respectively, and the output end of the feedback controller group is connected with the input end of the sewage treatment device (1); The system comprises the following steps when in use: Step one: obtaining the target value of the water inlet index, and establishing an inlet index optimization control model; Step two: adjusting temperature and flow; Step three: determining whether the temperature and flow conditions meet the requirements; Step four: adjusting other operating conditions; Step five: reaching the water quality index target value; In step two, the temperature feedback controller and the flow feedback controller detect the water resource inlet index at all times X 1 (t) , and first feedback adjust the water temperature and flow In step three, the determinator determines whether the difference between the current output result of the temperature feedback controller or the flow feedback controller and the last output result meets the minimum requirement or ; if the requirement is met, the output temperature and flow results are applied to the sewage treatment device to proceed to the next water quality adjustment; if the requirement is not met, the feedback adjustment of the temperature and flow needs to be recycled. In step four, after the water temperature and flow index in the water inlet index are feedback adjusted to be optimal, other operating conditions in the water inlet index are feedback adjusted to be optimal through multi-stage circulation; When adjusting other operating conditions, the following steps are adopted: S1: one feedback controller for one water quality index y a carries out feedback regulation, through optimization and adjustment, outputs the optimal operation condition for this water quality index U a ; S2: the optimal operating conditions U a acting on the wastewater treatment plant, resulting in an overall water quality index output at these operating conditions Y a and using this overall water quality index output Y a as input to the next feedback controller X a+1 ; In step five, after the last feedback controller optimizes and adjusts, executes the corresponding operating conditions to the sewage treatment device, and at this time, the water quality index of the whole water resource multi-stage circulation control system has reached the target value; the control system runs all the time in the sewage treatment process, adjusts the temperature, flow, chemical agent dosage, aeration parameter and microbial dosage in real time with the change of the water inlet water quality index; after the sewage treatment is completed, the whole multi-stage circulation control system is completed.

2. The system of claim 1, wherein, The feedback controller group comprises The feedback controller (5), The feedback controller (6), The feedback controller (7), …, The feedback controller, wherein m is a positive integer. The output of the determinator (4) is connected to the input of the sewage treatment device (1) and The input of the feedback controller (5) is connected to the output of the determinator (4), The output of the feedback controller (5) is connected to the input of the sewage treatment device (1) and The input of the feedback controller (6) is connected to the output of the feedback controller (5), The output of the feedback controller (6) is connected to the input of the sewage treatment device (1) and The input of the feedback controller (7) is connected to the output of the feedback controller (6), and so on, The output of the feedback controller is connected to the input of the sewage treatment device (1).

3. The system of claim 1, wherein, The temperature feedback controller (2), the flow feedback controller (3) and the feedback controller group are internally provided with a prediction regulator, the prediction regulator comprises a sensor (8), an optimizer (9), a simulator (10), a controller (11) and a controlled element (12), the output end of the sensor (8) is connected with the input end of the optimizer (9), the output end of the optimizer (9) is connected with the input end of the simulator (10), and the output end of the controller (11) is connected with the simulator (10) and the controlled element (12) respectively; A first error regulator (13) is arranged between the output ends of the optimizer (9) and the controller (11), the information output by the optimizer (9) and the controller (11) is transmitted to the first error regulator (13), the first error regulator (13) adjusts according to the error between the optimizer (9) and the controller (11) and then acts on the controller (11) to update the control variable; A second error regulator (14) is arranged between the output end of the simulator (10) and the output end of the controlled element (12), and the second error regulator (14) is used to adjust the error between the output end of the simulator (10) and the controlled element (12) and then to act on the simulator (10) to update the internal optimization model, so as to correct the expected state of the next round of optimization and to continuously perform cyclic feedback adjustment; the internal control needs to be solved by repeated intelligent algorithm prediction and optimization in each time stage to obtain the optimal control step, and then the output of the solving controller is applied to the sewage treatment device (1).

4. The system of claim 1, wherein, In use, the prediction regulator adopts the following steps: Step S1: sensor (8) t+t a detects water resource inflow index X(t+t a ), and passes information to optimizer (9), and enters step S2; Step S2: The optimizer (9) performs optimization according to the set water quality target value and the real-time simulation model input by the simulator (10), and gives the current control variable U(t+t a ), and simultaneously enters step S3 and step S4; Step S3: Apply the control variable U(t+t a ) to the simulator (10) for simulation, output the result Yd(t+t a +△t), and go to Step S7. Step S4: After the controller (11) receives the information delivered by the optimizer (9), the actual output control variable U'(t+t a ) is output according to the existing data, and step S5 is entered. Step S5: The optimizer (9) calculates the control output U(t+t a ) and the information U'(t+t a ) actually output by the controller (11) is transmitted to the first error regulator (13), which acts on the controller (11) for adaptive stability adjustment of the controller (11) according to the error between the two, and enters step S6; Step S6: the actual output control variable U'(t+ta) acts on the sewage treatment device to output an actual output Y(t+t a +△t); if the error between the actual output Y(t+t a +△t) and the target value is greater than the allowable error, then go to step S7; If the error between the actual output Y(t+t a +△t) and the target value is less than the allowable error, then step S8 is entered. Step S7: The information output by the simulator (10) and the controlled element (12) is transmitted to the second error regulator (14), the second error regulator (14) is adjusted according to the error between the two and then acts on the simulator (10) to update the real-time simulation model of the simulator (10), and step S2 is entered; Step S8: The optimization target is continuously adjusted by cyclic feedback, and the optimization control process is realized when the error between the actual water treatment effect and the target value is less than the allowable error.

5. The system of claim 1, wherein, The specific real-time simulation model of the temperature feedback controller (2), the flow feedback controller (3) and the internal simulator (10) of the feedback controller group is as follows s. t. ； ； ； External constraints: the objective function represents the state of the system and the desired state The constraint (1) represents the dynamic characteristics of the controlled object, which should be approached as closely as possible in the next N time steps, For the prediction model, various machine learning algorithms, not limited to recurrent neural networks, the constraints (2) and (3) represent the upper and lower limits of the water treatment control parameters and state parameters respectively.

Images