An inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment

Abstract

The present invention belongs to the field of urban sewage treatment and provides an inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment, comprising: constructing a dissolved oxygen concentration kinetic model, designing an inverse fuzzy dynamic multi-objective optimization algorithm, setting a judgment network and a long-term utility function, designing an execution network, designing an actual controller, determining an execution error function, optimizing the execution network and actual controller parameters, and controlling an aeration pump. The present invention achieves the acquisition of the optimal set value of the dissolved oxygen concentration through calculation using an inverse fuzzy dynamic multi-objective optimization algorithm; designing an execution network using the judgment network and the long-term utility function to identify unknown dynamic information of the dissolved oxygen concentration kinetic model; optimizing the execution network and actual controller using the judgment network and the execution error function; and achieving adaptive control of the aeration pump using an actual controller with optimal parameters, thereby improving the control accuracy of the aeration level and effectively reducing operating energy consumption.

Description

Technical Field

[0001] The present invention relates to the field of urban sewage treatment, in particular to an inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment. Background Art

[0002] The wastewater treatment process exhibits the following characteristics: High uncertainty: The wastewater treatment process is subject to uncertain interfering factors such as weather, water temperature, and pH, resulting in a high degree of uncertainty in its influent flow rate, composition, and pollutant concentrations. Strong nonlinearity: During the aeration reaction, the respiration rates and nitrification rates of anaerobic and aerobic bacteria exhibit a strong nonlinear relationship. Therefore, designing optimal control strategies to achieve high efficiency and low energy consumption under complex operating conditions is a key issue. Optimal control strategies are an effective means of precisely controlling equipment and reducing energy consumption during production. By obtaining optimal setpoints for control variables and implementing tracking control, operational energy consumption can be effectively reduced while ensuring effluent quality. However, due to the highly nonlinear nature of the wastewater treatment process, traditional control methods are unable to achieve precise tracking control. Fuzzy neural networks have demonstrated excellent performance in complex processes such as wastewater treatment, leading to their increasing application. Furthermore, reinforcement learning methods based on an actor-critic network learning structure possess strong online learning capabilities, making them an ideal choice for optimal control in certain high-dimensional and complex decision-making problems. The evaluation network and execution network utilize the policy gradient algorithm. By combining the policy gradient and value function update, they can effectively handle nonlinear problems in complex environments. Their application in the sewage treatment control process can enhance robustness and adaptability, improve the effluent water quality, and effectively reduce the energy consumption of equipment operation.

[0003] However, the existing technology does not have a technical solution for sewage aeration control that simultaneously utilizes optimization control strategy, fuzzy neural network, reinforcement learning method based on execution-criticism network learning structure and policy gradient algorithm. Summary of the Invention

[0004] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide an inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment, which can achieve the acquisition of the optimal set value of dissolved oxygen concentration, the identification of unknown dynamic information of the dissolved oxygen concentration kinetic model, the optimization of the execution network and the actual controller, and the adaptive control of the aeration pump.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] An inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment, comprising:

[0007] Construct a kinetic model of dissolved oxygen concentration;

[0008] Designing an inverse fuzzy dynamic multi-objective optimization algorithm, and using the inverse fuzzy dynamic multi-objective optimization algorithm to obtain the optimal setting value;

[0009] Setting a long-term utility function and a judgment network based on a first fuzzy neural network according to the dissolved oxygen concentration kinetic model and the optimal set value;

[0010] Designing an execution network based on a second fuzzy neural network according to the evaluation network and the long-term utility function;

[0011] The execution network is used to approximately obtain the unknown dynamic information of the dissolved oxygen concentration kinetic model; the calculation formula of the unknown dynamic information is:

[0012] ;

[0013] in, ; for Dissolved oxygen concentration set value at all times; is the calculated value of the execution network; For the moment; is the saturation coefficient of aerobic respiration of heterotrophic bacteria; is the heterotrophic biomass concentration; is the heterotrophic biomass production coefficient; is the growth respiration coefficient; is the actual concentration of dissolved oxygen in the fifth area of the biochemical reaction tank; is the flow rate of the 5th pool; is the volume of the biochemical reaction pool in the fifth zone; is the dissolved oxygen concentration tracking error; is the optimal output weight of the execution network; is the output of the normalization layer of the execution network; is the boundary approximation value of the execution network; is the flow rate of the fourth pool; is the actual concentration of dissolved oxygen in the fourth area of the biochemical reaction tank;

[0014] The actual controller is designed according to the execution network and the calculation formula of the unknown dynamic information; the calculation formula of the actual controller is:

[0015] ;

[0016] in, is the input of the actual controller; is the control coefficient; For the execution network The latest weight value at all times; is the saturation value of dissolved oxygen concentration;

[0017] Determine an execution error function of the execution network; the execution error function is:

[0018] ;

[0019] in, is the execution error function; is the latest output value of the evaluation network;

[0020] The parameters of the execution network and the actual controller are iteratively optimized using the evaluation network and the execution error function to obtain the actual controller with optimal parameters; the calculation formula for iterative update of the parameters of the execution network is:

[0021] ;

[0022] in, ; is the loss function of the execution network; is the second learning rate;

[0023] The actual controller with the optimal parameters is used to control the preset aeration pump.

[0024] Preferably, the dissolved oxygen concentration kinetic model is:

[0025] ;

[0026] in, is the oxygen transfer coefficient of the last aerobic zone; is the growth respiration coefficient; is the heterotrophic biomass concentration; is the heterotrophic biomass production coefficient; is the saturation value of dissolved oxygen concentration.

[0027] Preferably, iteratively optimizing the parameters of the execution network and the actual controller using the evaluation network and the execution error function to obtain the actual controller with optimal parameters includes:

[0028] Using the historical Pareto optimal frontier as input of an inverse fuzzy neural network model based on a third fuzzy neural network, and using the historical Pareto optimal solution as output of the inverse fuzzy neural network model to train the inverse fuzzy neural network model, thereby obtaining a trained inverse model;

[0029] according to The Pareto optimal frontier of the environment and The Pareto optimal frontier of the environment is calculated using the central prediction method the Pareto optimal frontier of the environment;

[0030] Using the inverse model The Pareto optimal frontier of the environment is mapped to obtain The initial Pareto solution of the environment;

[0031] The decomposed multi-objective optimization algorithm is used to The Pareto initial solution of the environment is optimized to obtain multiple The Pareto optimal solution of the environment and the fuzzy membership function are used to The Pareto optimal solution of the environment is screened to obtain the optimal setting value;

[0032] Calculating the execution error function according to the optimal setting value to obtain an iterative identification error;

[0033] Iteratively optimizing the weights of the execution network according to the iterative identification error to obtain the iteratively completed execution network;

[0034] The parameters of the actual controller are updated using the weights of the iteratively executed execution network to obtain the actual controller with optimal parameters.

[0035] Preferably, the long-term utility function is:

[0036] ;

[0037] in, ; ; is the utility function; is the optimal setting value, is the actual concentration of dissolved oxygen in the fifth area of the biochemical reaction tank; is the discount factor; The range is 0 to 1.

[0038] Preferably, the historical Pareto optimal frontier is used as the input of an inverse fuzzy neural network model based on the third fuzzy neural network, and the historical Pareto optimal solution is used as the output of the inverse fuzzy neural network model to train the inverse fuzzy neural network model to obtain a trained inverse model, including:

[0039] Determine the calculation formula of the output of the inverse fuzzy neural network model; the calculation formula of the output of the inverse fuzzy neural network model is:

[0040] ;

[0041] in, is the output of the inverse fuzzy neural network model; is the input of the inverse fuzzy neural network model; is the third weight; is the number of neurons in the input layer; is the number of neurons in the membership layer and the regularity layer; and are the center value and width value of the membership layer respectively.

[0042] Preferably, the second weight update formula of the execution network is:

[0043] ;

[0044] in, is the second learning rate; is the latest weight value of the execution network.

[0045] Preferably, according to The Pareto optimal frontier of the environment and The Pareto optimal frontier of the environment is calculated using the central prediction method The Pareto optimal frontier of the environment, including:

[0046] Calculate the above The Pareto optimal frontier of the environment and the The center of mass of the Pareto optimal frontier of the environment is obtained to obtain the first center of mass and the second center of mass. The calculation formula of the first center of mass is: ; The calculation formula of the second center of mass is: ; is the first centroid; For the the Pareto optimal frontier of the environment; For the The Pareto optimal frontier of the environment corresponds to the objective function values; is the second centroid; For the the Pareto optimal frontier of the environment; For the The Pareto optimal frontier of the environment corresponds to the objective function values; is the number of objective function values corresponding to the Pareto optimal frontier;

[0047] The first centroid and the second centroid are calculated using the center prediction formula. The Pareto optimal frontier of the environment; the center prediction formula includes: 、 as well as ; is the dimension of all the objective function values;

[0048] get After finding the Pareto optimal frontier of the environment, the inverse model is used to map it and obtain The Pareto initial solution of the environment is obtained by using the decomposed multi-objective optimization algorithm. The Pareto initial solution of the environment is optimized to obtain multiple The Pareto optimal solution of the environment and the fuzzy membership function are used to The Pareto optimal solution of the environment is screened to obtain the optimal setting value.

[0049] Preferably, it also includes:

[0050] Determine the evaluation network; the output expression of the evaluation network is: ;

[0051] Define the estimation error function of the evaluation network; the estimation error function is:

[0052] ;

[0053] in, ; is the estimation error function;

[0054] The loss function of the evaluation network is defined according to the estimation error function; the loss function of the evaluation network is: ;in, is the loss function of the evaluation network;

[0055] The first weight update formula of the evaluation network is determined according to the evaluation network, the estimation error function, and the loss function of the evaluation network; the first weight update formula is:

[0056] ;

[0057] in, is the first learning rate;

[0058] The evaluation network is iteratively optimized using the first weight update formula to obtain the iteratively completed evaluation network.

[0059] Preferably, it also includes:

[0060] The control effect of the actual controller with the optimal parameters is evaluated using the comprehensive square error and the comprehensive absolute error to obtain an evaluation result; the calculation formula of the comprehensive square error is:

[0061] ;

[0062] The calculation formula of the comprehensive absolute error is:

[0063] ;

[0064] in, is the comprehensive square error; is the comprehensive absolute error; is the total number of samples.

[0065] The present invention discloses the following technical effects:

[0066] The present invention provides an inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment. By utilizing a judgment network and a long-term utility function to design an execution network, the problem of limited reliability caused by updating the controller based on instantaneous tracking error in traditional aeration control in the prior art is solved, and accurate acquisition of unknown dynamic information of the dissolved oxygen concentration kinetic model is achieved; through the judgment network and the execution error function, the problem of low control accuracy of the controller is solved, and optimization of the execution network and the actual controller is achieved; by designing an inverse fuzzy dynamic optimization algorithm, the problem of optimal parameter acquisition is solved, and real-time calculation of the dynamic optimal set value of the dissolved oxygen concentration in the sewage treatment process is achieved; through the actual controller with optimal parameters, the problem of model-to-equipment correlation control is solved, and adaptive control of the aeration pump is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0068] Figure 1 A schematic diagram of the inverse fuzzy dynamic optimal adaptive aeration control process for sewage treatment provided by an embodiment of the present invention;

[0069] Figure 2 This is a statistical diagram of the inflow of sewage treatment process provided by the embodiment of the present invention. Figure 2 (a) is the statistical diagram of the inflow flow of the sewage treatment process under sunny conditions. Figure 2 (b) is a statistical diagram of the inflow of sewage treatment process under heavy rain conditions;

[0070] Figure 3 This is a statistical diagram of the tracking control effect under sunny conditions provided by an embodiment of the present invention. Figure 3 (a) is a statistical diagram of the control output and set value of dissolved oxygen concentration under sunny conditions. Figure 3 (b) is the tracking error statistics under sunny conditions;

[0071] Figure 4 This is a statistical diagram of the tracking control effect under rainstorm conditions provided by an embodiment of the present invention. Figure 3 (a) is a statistical diagram of the control output and set value of dissolved oxygen concentration under heavy rain conditions. Figure 3 (b) is the tracking error statistics under heavy rain conditions;

[0072] Figure 5 This is a statistical diagram of changes in the aeration pump controller provided by an embodiment of the present invention. Figure 5 (a) is the statistical diagram of the aeration pump controller changes under sunny conditions. Figure 5 (b) is a statistical chart of changes in the aeration pump controller under heavy rain conditions. DETAILED DESCRIPTION

[0073] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0074] The purpose of the present invention is to provide an inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment, which can achieve the acquisition of the optimal set value of dissolved oxygen concentration, the identification of unknown dynamic information of the dissolved oxygen concentration kinetic model, the optimization of the execution network and the actual controller, and the adaptive control of the aeration pump.

[0075] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0076] Figure 1 The schematic diagram of the inverse fuzzy dynamic optimal adaptive aeration control process for sewage treatment provided by the embodiment of the present invention is as follows: Figure 1 As shown, the present invention provides an inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment, comprising:

[0077] Step 100: constructing a dissolved oxygen concentration kinetic model;

[0078] Step 200: Design an inverse fuzzy dynamic multi-objective optimization algorithm, and use the inverse fuzzy dynamic multi-objective optimization algorithm to obtain the optimal setting value;

[0079] Step 300: setting a long-term utility function and an evaluation network based on a first fuzzy neural network according to a dissolved oxygen concentration kinetic model and an optimal setting value;

[0080] Step 400: Designing an execution network based on a second fuzzy neural network according to the evaluation network and the long-term utility function;

[0081] Step 500: Utilize the execution network to approximate the unknown dynamic information of the dissolved oxygen concentration kinetic model; the calculation formula of the unknown dynamic information is:

[0082] ;

[0083] in, ; for Dissolved oxygen concentration set value at all times; The calculated value for executing the network; For the moment; is the saturation coefficient of aerobic respiration of heterotrophic bacteria; is the heterotrophic biomass concentration; is the heterotrophic biomass production coefficient; is the growth respiration coefficient; is the actual concentration of dissolved oxygen in the fifth area of the biochemical reaction tank; is the flow rate of the 5th pool; is the volume of the biochemical reaction pool in the fifth zone; is the dissolved oxygen concentration tracking error; is the optimal output weight of the execution network; The output of the normalization layer of the execution network; is the boundary approximation value of the execution network; is the flow rate of the fourth pool; is the actual concentration of dissolved oxygen in the fourth area of the biochemical reaction tank;

[0084] Step 600: Design an actual controller based on the calculation formula of the execution network and unknown dynamic information; the calculation formula of the actual controller is:

[0085] ;

[0086] in, is the input of the actual controller; is the control coefficient; To execute the network The latest weight value at all times; is the saturation value of dissolved oxygen concentration;

[0087] Step 700: Determine the execution error function of the execution network; the execution error function is:

[0088] ;

[0089] in, is the execution error function; To judge the latest output value of the network;

[0090] Step 800: Optimize the parameters of the execution network and the actual controller iteratively using the evaluation network and the execution error function to obtain the actual controller with the optimal parameters. The calculation formula for iterative update of the parameters of the execution network is:

[0091] ;

[0092] in, ; is the loss function of the execution network; is the second learning rate;

[0093] Step 900: Using the actual controller with optimal parameters to control the preset aeration pump.

[0094] Preferably, the dissolved oxygen concentration kinetic model is:

[0095] ;

[0096] in, is the oxygen transfer coefficient of the last aerobic zone; is the growth respiration coefficient; is the heterotrophic biomass concentration; is the heterotrophic biomass production coefficient; is the saturation value of dissolved oxygen concentration.

[0097] Furthermore, the judgment network and the execution error function are used to iteratively optimize the parameters of the execution network and the actual controller to obtain the actual controller with the optimal parameters, including:

[0098] The historical Pareto optimal frontier is used as the input of the inverse fuzzy neural network model based on the third fuzzy neural network, and the historical Pareto optimal solution is used as the output of the inverse fuzzy neural network model to train the inverse fuzzy neural network model and obtain a trained inverse model;

[0099] according to The Pareto optimal frontier of the environment and The Pareto optimal frontier of the environment is calculated using the central prediction method the Pareto optimal frontier of the environment;

[0100] Using the inverse model The Pareto optimal frontier of the environment is mapped to obtain The initial Pareto solution of the environment;

[0101] Using decomposed multi-objective optimization algorithm to The Pareto initial solution of the environment is optimized to obtain multiple The Pareto optimal solution of the environment and the fuzzy membership function are used to The Pareto optimal solution of the environment is screened to obtain the optimal setting value;

[0102] The execution error function is calculated according to the optimal setting value to obtain the iterative identification error;

[0103] Iteratively optimize the weights of the execution network according to the iterative identification error to obtain the iteratively completed execution network;

[0104] The weights of the iterative execution network are used to update the parameters of the actual controller to obtain the actual controller with optimal parameters.

[0105] Preferably, the long-term utility function is:

[0106] ;

[0107] in, ; ; is the utility function; is the optimal setting value, is the actual concentration of dissolved oxygen in the fifth area of the biochemical reaction tank; is the discount factor; The range is 0 to 1.

[0108] Specifically, the historical Pareto optimal frontier is used as the input of the inverse fuzzy neural network model based on the third fuzzy neural network, and the historical Pareto optimal solution is used as the output of the inverse fuzzy neural network model to train the inverse fuzzy neural network model, thereby obtaining a trained inverse model, including:

[0109] Determine the calculation formula of the output of the inverse fuzzy neural network model; the calculation formula of the output of the inverse fuzzy neural network model is:

[0110] ;

[0111] in, is the output of the inverse fuzzy neural network model; is the input of the inverse fuzzy neural network model; is the third weight; is the number of neurons in the input layer; is the number of neurons in the membership layer and the regularity layer; and are the center value and width value of the membership layer respectively.

[0112] Specifically, the second weight update formula of the execution network is:

[0113] ;

[0114] in, is the second learning rate; The latest weight values for the executing network.

[0115] Preferably, in order to obtain The Pareto initial solution of the environment is used to The Pareto optimal frontier of the environment is mapped. Then the decomposed multi-objective optimization algorithm is used to The Pareto initial solution of the environment is optimized to obtain The Pareto optimal solution of the environment and the fuzzy membership function are used to The optimal setting value is selected from the Pareto optimal solution of the environment.

[0116] Further, according to The Pareto optimal frontier of the environment and The Pareto optimal frontier of the environment is calculated using the central prediction method The Pareto optimal frontier of the environment, including:

[0117] Calculate separately The Pareto optimal frontier of the environment and The center of mass of the Pareto optimal frontier of the environment is obtained, and the first center of mass and the second center of mass are obtained; the calculation formula of the first center of mass is: ; The calculation formula for the second centroid is: ; is the first centroid; for the Pareto optimal frontier of the environment; for The Pareto optimal frontier of the environment corresponds to the objective function values; is the second center of mass; for the Pareto optimal frontier of the environment; for The Pareto optimal frontier of the environment corresponds to the objective function values; is the number of objective function values corresponding to the Pareto optimal frontier;

[0118] Calculated based on the first and second centroids using the center prediction formula The Pareto optimal frontier of the environment; the central prediction formula includes: 、 as well as ; is the dimension of all objective function values;

[0119] get After finding the Pareto optimal frontier of the environment, we use the inverse model to map it and get The Pareto initial solution of the environment is used to optimize the The Pareto initial solution of the environment is optimized to obtain multiple The Pareto optimal solution of the environment and the fuzzy membership function are used to The Pareto optimal solution of the environment is screened to obtain the optimal setting value.

[0120] Specifically, it also includes:

[0121] Determine the evaluation network; the long-term utility function is approximated by the evaluation network, and the output expression of the evaluation network is: ;

[0122] Define the estimation error function of the judgment network; the estimation error function is:

[0123] ;

[0124] in, ; is the estimation error function;

[0125] The loss function of the evaluation network is defined according to the estimated error function; the loss function of the evaluation network is: ;in, is the loss function for judging the network;

[0126] The first weight update formula of the evaluation network is determined according to the evaluation network, the estimated error function and the loss function of the evaluation network; the first weight update formula is:

[0127] ;

[0128] in, is the first learning rate;

[0129] The judgment network is iteratively optimized using the first weight update formula to obtain an iteratively completed judgment network.

[0130] Furthermore, it also includes:

[0131] The comprehensive square error and comprehensive absolute error are used to evaluate the control effect of the actual controller with optimal parameters to obtain the evaluation results; the calculation formula of the comprehensive square error is:

[0132] ;

[0133] The calculation formula for the comprehensive absolute error is:

[0134] ;

[0135] in, is the comprehensive square error; is the comprehensive absolute error; is the total number of samples.

[0136] Specifically, the kinetic model for determining dissolved oxygen concentration is:

[0137]

[0138] in, For the moment, The biochemical reaction pool The concentration of the region, is the heterotrophic biomass concentration, represents the heterotrophic biomass production coefficient, is the growth respiration coefficient, is the oxygen transfer coefficient of the last aerobic zone, For the The flow rate of the pool, is the volume of the biochemical reaction pool in the fifth zone, is the saturation coefficient of aerobic respiration of heterotrophic bacteria, is the saturation value of dissolved oxygen concentration.

[0139] Furthermore, the optimal set value of dissolved oxygen concentration is dynamically obtained: In the dynamic multi-objective optimization algorithm, when the operating environment changes, the generation of the initial solution is crucial. In order to generate an initial population with good convergence and diversity, an inverse fuzzy dynamic multi-objective optimization algorithm is designed. The Pareto optimal solution under and Pareto optimal frontier and inverse fuzzy neural network modeling methods to predict new environments The initial population of .

[0140] Specifically, in the environment Down, and is used as the input and output value training of the inverse fuzzy neural network model, where For the environment The dissolved oxygen concentration under is the objective function value of operating energy consumption and effluent water quality, yes and Then, in the environment By directly predicting the individual generate In order to improve the prediction accuracy of the population and reduce the computational complexity, a method based on center prediction is designed to generate , as follows:

[0141] 1) Calculate separately The center of mass and The center of mass :

[0142]

[0143] 2) It can be predicted by the following formula:

[0144]

[0145] in is a Gaussian perturbation with a mean of 0 and a standard deviation of λ, which can be defined as:

[0146]

[0147] in, .

[0148] 3) Use the trained fuzzy neural network model to Mapping to the decision space to obtain the environment The initial solution At the same time, a multi-objective optimization algorithm based on decomposition is selected to obtain Optimize to find the environment Pareto optimal solution Finally, the fuzzy membership function method is used to Select the optimal setting value, that is, the aeration controller of the sewage treatment process value.

[0149] Preferably, the evaluation network is designed: first, the tracking performance is evaluated by the reinforcement learning strategy, and then the control decision is determined by analyzing the evaluation results. Define the dissolved oxygen concentration tracking error :

[0150]

[0151] in is the actual concentration of dissolved oxygen, is the optimal setting value.

[0152] Then, the derivative of the dissolved oxygen concentration tracking error formula is for:

[0153]

[0154] During the iterative learning phase, the current controller policy can be adjusted by providing rewards or penalties to the critic network. Defined as:

[0155]

[0156] Long-term utility function as follows:

[0157]

[0158] in is the discount factor.

[0159] Furthermore, a fuzzy neural network is chosen to approximate the long-term cost function:

[0160]

[0161] in is the optimal weight of the first fuzzy neural network, is the output value of the first fuzzy neural network normalization layer, is the boundary approximation value of the first fuzzy neural network. Therefore, the current output value of the judgment network is , is the weight value of the current network.

[0162] Function estimation error Defined as:

[0163]

[0164] Define the cost function of the critic network , therefore, the update rule of the weights in the network is designed as:

[0165]

[0166] Where 0< <1 is the learning rate.

[0167] Specifically, controller design based on execution network: the control strategy is formulated by gradually accumulating control system experience, and the execution network is composed of fuzzy neural networks. Consider the following Lyapunov function of the execution network :

[0168]

[0169] Then the derivative of the Lyapunov function of the execution network is for:

[0170]

[0171] Since there is unknown dynamic information in the sewage treatment process, the execution network is introduced to approximate the system uncertainty. , expressed as:

[0172]

[0173] in is the optimal output weight of the second fuzzy neural network, is the output of the second fuzzy neural network normalization layer, is the second fuzzy neural network boundary approximation value.

[0174] Then enter the desired controller into Designed to:

[0175]

[0176] is the optimal weight expected by the second fuzzy neural network;

[0177] in is the control coefficient, and the weight change is Therefore, the actual controller input Designed to:

[0178]

[0179] The error of the execution network Defined as:

[0180]

[0181] Define the cost function of the execution network as , then the update rule of the weights in the execution network is designed as:

[0182]

[0183] in is the learning rate.

[0184] Further, system stability analysis: Lyapunov function of the evaluation network Designed to:

[0185] ;

[0186] Then the derivative of the Lyapunov function of the evaluation network is for:

[0187]

[0188] in, .definition ,but satisfy:

[0189]

[0190] Execute network Lyapunov function Designed to:

[0191]

[0192] Then the derivative of the network Lyapunov function is performed for:

[0193]

[0194] in .definition ,Then satisfy:

[0195]

[0196] Introducing Lyapunov candidate function:

[0197]

[0198] In summary, we can get the derivative of the Lyapunov candidate function for:

[0199]

[0200] According to Young's inequality, we have:

[0201]

[0202]

[0203] Can be further rewritten as:

[0204]

[0205] in

[0206]

[0207]

[0208]

[0209] Furthermore, considering the dynamic time-varying conditions of the sewage treatment process, the inverse fuzzy dynamic optimal aeration adaptive tracking controller and two network weight updating rules can ensure ideal control performance.

[0210] Optionally, control performance evaluation: To fairly evaluate the control performance, the integrated square error ( ), comprehensive absolute error ( ) as the evaluation index.

[0211]

[0212]

[0213] in is the actual concentration of dissolved oxygen, which is measured by the sensor during the sewage treatment process. is the desired dissolved oxygen concentration, is the total number of samples.

[0214] Specifically, Figure 2 The influent flow rate of the wastewater treatment process under two conditions, sunny day and rainy day, is shown. Figure 3 The tracking control results under sunny conditions are shown. Figure 3 (a) It can be seen that the control output of dissolved oxygen concentration is closely matched to the set value. Figure 3 As can be seen in (b), the tracking control error remains within the range of [-0.035, 0.055] mg / L. The results show that the designed controller can accurately track the dynamic optimal setpoint in real time. Figure 4 The tracking control results under heavy rain conditions are shown. Figure 4 (a) It can be seen that the control output of dissolved oxygen concentration is closely matched to the set value. Figure 4 As can be seen in (b), the tracking control error remains within the range of [-0.1, 0.16] mg / L. The results show that the designed controller can accurately identify the current state and accurately track the dynamic optimal set value even under heavy rainfall interference. Figure 5 Shows the aeration pump controller changes under sunny and rainy conditions. All of them float between [0, 300].

[0215] The beneficial effects of the present invention are as follows:

[0216] The present invention achieves the acquisition of the optimal set value of dissolved oxygen concentration through calculation using an inverse fuzzy dynamic multi-objective optimization algorithm; designs an execution network by utilizing a judgment network and a long-term utility function, thereby achieving the identification of unknown dynamic information of a dissolved oxygen concentration kinetic model; optimizes the execution network and an actual controller through the judgment network and the execution error function; and achieves adaptive control of an aeration pump through an actual controller with optimal parameters, thereby improving the control accuracy of the aeration degree and effectively reducing operating energy consumption.

[0217] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0218] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. An inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment, characterized in that: include: Construct a kinetic model of dissolved oxygen concentration; Designing an inverse fuzzy dynamic multi-objective optimization algorithm, and using the inverse fuzzy dynamic multi-objective optimization algorithm to obtain the optimal setting value; Setting a long-term utility function and a judgment network based on a first fuzzy neural network according to the dissolved oxygen concentration kinetic model and the optimal set value; Designing an execution network based on a second fuzzy neural network according to the evaluation network and the long-term utility function; The execution network is used to approximately obtain the unknown dynamic information of the dissolved oxygen concentration kinetic model; the calculation formula of the unknown dynamic information is: ； in, ; for Dissolved oxygen concentration set value at all times; is the calculated value of the execution network; For the moment; is the saturation coefficient of aerobic respiration of heterotrophic bacteria; is the heterotrophic biomass concentration; is the heterotrophic biomass production coefficient; is the growth respiration coefficient; is the actual concentration of dissolved oxygen in the fifth area of the biochemical reaction tank; is the flow rate of the 5th pool; is the volume of the biochemical reaction pool in the fifth zone; is the dissolved oxygen concentration tracking error; is the optimal output weight of the execution network; is the output of the normalization layer of the execution network; is the boundary approximation value of the execution network; is the flow rate of the fourth pool; is the actual concentration of dissolved oxygen in the fourth area of the biochemical reaction tank; The actual controller is designed according to the execution network and the calculation formula of the unknown dynamic information; the calculation formula of the actual controller is: ； in, is the input of the actual controller; is the control coefficient; For the execution network The latest weight value at all times; is the saturation value of dissolved oxygen concentration; Determine an execution error function of the execution network; the execution error function is: ； in, is the execution error function; is the latest output value of the evaluation network; The parameters of the execution network and the actual controller are iteratively optimized using the evaluation network and the execution error function to obtain the actual controller with optimal parameters; the calculation formula for iterative update of the parameters of the execution network is: ； in, ; is the loss function of the execution network; is the second learning rate; The actual controller with the optimal parameters is used to control the preset aeration pump.

2. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to claim 1 is characterized in that: The dissolved oxygen concentration kinetic model is: ； in, is the oxygen transfer coefficient of the last aerobic zone; is the growth respiration coefficient; is the heterotrophic biomass concentration; is the heterotrophic biomass production coefficient; is the saturation value of dissolved oxygen concentration.

3. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to claim 1 is characterized in that: Iteratively optimizing the parameters of the execution network and the actual controller using the evaluation network and the execution error function to obtain the actual controller with optimal parameters includes: Using the historical Pareto optimal frontier as input of an inverse fuzzy neural network model based on a third fuzzy neural network, and using the historical Pareto optimal solution as output of the inverse fuzzy neural network model to train the inverse fuzzy neural network model, thereby obtaining a trained inverse model; according to The Pareto optimal frontier of the environment and The Pareto optimal frontier of the environment is calculated using the central prediction method the Pareto optimal frontier of the environment; Using the inverse model Mapping the Pareto optimal frontier of the environment, we get The initial Pareto solution of the environment; The decomposed multi-objective optimization algorithm is used to The Pareto initial solution of the environment is optimized to obtain multiple The Pareto optimal solution of the environment and the fuzzy membership function are used to The Pareto optimal solution of the environment is screened to obtain the optimal setting value; Calculating the execution error function according to the optimal setting value to obtain an iterative identification error; Iteratively optimizing the weights of the execution network according to the iterative identification error to obtain the iteratively completed execution network; The parameters of the actual controller are updated using the weights of the iteratively executed execution network to obtain the actual controller with optimal parameters.

4. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to claim 1 is characterized in that: The long-term utility function is: ； in, ; ; is the utility function; is the optimal setting value, is the actual concentration of dissolved oxygen in the fifth area of the biochemical reaction tank; is the discount factor; The range is 0 to 1.

5. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to claim 3 is characterized in that: The method uses the historical Pareto optimal frontier as the input of an inverse fuzzy neural network model based on the third fuzzy neural network, uses the historical Pareto optimal solution as the output of the inverse fuzzy neural network model, and trains the inverse fuzzy neural network model to obtain a trained inverse model, including: Determine the calculation formula of the output of the inverse fuzzy neural network model; the calculation formula of the output of the inverse fuzzy neural network model is: ； in, is the output of the inverse fuzzy neural network model; is the input of the inverse fuzzy neural network model; is the third weight; is the number of neurons in the input layer; is the number of neurons in the membership layer and the regularity layer; and are the center value and width value of the membership layer respectively.

6. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to claim 3 is characterized in that: The second weight update formula of the execution network is: ； in, is the second learning rate; is the latest weight value of the execution network.

7. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to any one of claims 4 and 5, characterized in that: according to The Pareto optimal frontier of the environment and The Pareto optimal frontier of the environment is calculated using the central prediction method The Pareto optimal frontier of the environment, including: Calculate the above The Pareto optimal frontier of the environment and the The center of mass of the Pareto optimal frontier of the environment is obtained to obtain the first center of mass and the second center of mass. The calculation formula of the first center of mass is: ; The calculation formula of the second center of mass is: ; is the first centroid; For the the Pareto optimal frontier of the environment; For the The Pareto optimal frontier of the environment corresponds to the objective function values; is the second centroid; For the the Pareto optimal frontier of the environment; For the The Pareto optimal frontier of the environment corresponds to the objective function values; is the number of objective function values corresponding to the Pareto optimal frontier; The first centroid and the second centroid are calculated using the center prediction formula. The Pareto optimal frontier of the environment; the center prediction formula includes: 、 as well as ; is the dimension of all the objective function values; get After finding the Pareto optimal frontier of the environment, the inverse model is used to map it and obtain The Pareto initial solution of the environment is obtained by using the decomposed multi-objective optimization algorithm. The Pareto initial solution of the environment is optimized to obtain multiple The Pareto optimal solution of the environment and the fuzzy membership function are used to The Pareto optimal solution of the environment is screened to obtain the optimal setting value.

8. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to claim 4 is characterized in that: Also includes: Determine the evaluation network; the output expression of the evaluation network is: ; Define the estimation error function of the evaluation network; the estimation error function is: ； in, ; is the estimation error function; The loss function of the evaluation network is defined according to the estimation error function; the loss function of the evaluation network is: ;in, is the loss function of the evaluation network; The first weight update formula of the evaluation network is determined according to the evaluation network, the estimation error function, and the loss function of the evaluation network; the first weight update formula is: ； in, is the first learning rate; The evaluation network is iteratively optimized using the first weight update formula to obtain the iteratively completed evaluation network.

9. The inverse fuzzy dynamic optimal adaptive aeration control method based on sewage treatment according to claim 4, characterized in that: Also includes: The control effect of the actual controller with the optimal parameters is evaluated using the comprehensive square error and the comprehensive absolute error to obtain an evaluation result; the calculation formula of the comprehensive square error is: ； The calculation formula of the comprehensive absolute error is: ； in, is the comprehensive square error; is the comprehensive absolute error; is the total number of samples.

Images