Urban sewage treatment process control method based on adaptive fuzzy terminal sliding mode

Abstract

The invention provides a municipal sewage treatment process control method based on a self-adaptive fuzzy terminal sliding mode, which realizes accurate control of dissolved oxygen concentration and nitrate nitrogen concentration. According to the method, an adaptive estimation fuzzy neural network is designed to predict dynamic changes of dissolved oxygen concentration and nitrate nitrogen concentration in a biochemical reaction process, and a weight adaptive updating mechanism based on an estimation error and a fuzzy terminal sliding mode controller based on a terminal sliding mode surface are established; the concentration of dissolved oxygen and the concentration of nitrate nitrogen are accurately controlled; the influence of the buffeting phenomenon on the system stability is reduced by designing a self-adaptive switching gain mechanism, and the problem that a control system is unstable in the urban sewage treatment process is solved. Experimental results show that the method can realize accurate control of the dissolved oxygen concentration and the nitrate nitrogen concentration, and ensures stable and efficient operation of the urban sewage treatment process.

Description

Technical Field

[0001] The present invention realizes precise control of dissolved oxygen concentration and nitrate nitrogen concentration in the process of urban sewage treatment based on the control method of adaptive fuzzy terminal sliding mode. The dissolved oxygen concentration and nitrate nitrogen concentration are key process variables in the biochemical reaction process of urban sewage treatment and have an important influence on the effluent water quality of the urban sewage treatment process. The invention belongs to both the field of water research and the field of intelligent control. Background Art

[0002] With the growth of population and economic development, the problem of water pollution in the country is becoming increasingly serious, and the water environment continues to deteriorate, posing a serious threat to the urban environment. The establishment of sewage treatment plants is of great significance to improving the urban environment. At present, the activated sludge method has been widely used in urban sewage treatment processes. Dissolved oxygen concentration and nitrate nitrogen concentration are key process variables in the urban sewage treatment process, and are closely related to the biochemical reaction process of all activated sludge microorganisms. Accurately controlling them within a reasonable range is a prerequisite for the normal operation of the urban sewage treatment process. The urban sewage treatment process is an industrial process with unknown disturbances and complex nonlinearity, which brings great challenges to the precise control of dissolved oxygen concentration and nitrate nitrogen concentration.

[0003] The present invention designs a control method for the urban sewage treatment process based on an adaptive fuzzy terminal sliding mode, which mainly predicts the dynamic changes of dissolved oxygen concentration and nitrate nitrogen concentration during the biochemical reaction by designing an adaptive prediction fuzzy neural network, establishes a weight adaptive update mechanism based on the prediction error and a fuzzy terminal sliding mode controller based on an adaptive switching gain, and realizes precise control of the dissolved oxygen concentration and nitrate nitrogen concentration. The present invention not only reduces the impact of a series of uncertain disturbances such as influent flow, water quality fluctuations, and changes in operating conditions and environment during the sewage treatment operation on the control performance, but also can approximate the dynamic changes of dissolved oxygen concentration and nitrate nitrogen concentration with higher accuracy, while reducing the system's chattering phenomenon, so that the dissolved oxygen concentration and nitrate nitrogen concentration meet the needs of the biochemical reaction, and ensure the smooth and efficient operation of the urban sewage treatment process. Summary of the invention

[0004] The present invention obtains a method for controlling the process of urban sewage treatment based on an adaptive fuzzy terminal sliding mode. The method adopts an adaptive prediction fuzzy neural network to predict the dynamic changes of dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process, establishes a weight adaptive update mechanism based on the prediction error and a fuzzy terminal sliding mode controller based on an adaptive switching gain, so as to realize the precise control of the dissolved oxygen concentration and the nitrate nitrogen concentration. The method solves the problem of unstable control system in the process of urban sewage treatment by designing an adaptive switching gain mechanism to reduce the influence of the chattering phenomenon on the system stability.

[0005] The present invention adopts the following technical solutions and implementation steps:

[0006] 1. A control method for a municipal sewage treatment process based on an adaptive fuzzy terminal sliding mode, characterized in that: the change characteristics of dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of municipal sewage treatment are extracted, a prediction model of dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of municipal sewage treatment based on an estimated fuzzy neural network is established, and a terminal sliding mode controller based on an adaptive switching gain is designed to achieve stable control of dissolved oxygen concentration and nitrate nitrogen concentration; the method comprises the following steps:

[0007] (1) Extraction of the changing characteristics of dissolved oxygen concentration and nitrate nitrogen concentration during the biochemical reaction process of urban sewage treatment

[0008] The dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment fluctuate greatly over time and are easily affected by a series of disturbances such as influent flow, water quality fluctuations, influent composition and temperature changes. The control model of dissolved oxygen concentration in the fifth zone and nitrate nitrogen concentration in the second zone of the biochemical reaction tank is:

[0009]

[0010] Among them, S O,5 (t) represents the dissolved oxygen concentration in the fifth partition of the biochemical reaction pool at time t, f1(S O,5 (t)) represents the change in dissolved oxygen concentration in the fifth partition of the biochemical reaction tank at time t, K L a5(t) represents the oxygen transfer coefficient of the fifth partition of the biochemical reaction pool at time t, Q4(t) is the flow rate of the fourth partition, and its value range is [0.3m / s, 0.5m / s], m / s represents meter / second, S O,4 It represents the dissolved oxygen concentration in the fourth zone of the biochemical reaction pool, which is 2 mg / L. NO,2 (t) represents the nitrate nitrogen concentration in the second partition of the biochemical reaction pool at time t, f2(S NO,2 (t)) represents the change in the nitrate nitrogen concentration in the second partition of the biochemical reaction pool at time t, Q a (t) represents the internal reflux of the second partition of the biochemical reaction pool at time t, Q in (t) represents the flow from the first partition to the second partition at time t, and its value range is [100m 3 / d,1000000m 3 / d],m 3 / d means cubic meters per day;

[0011] (2) Establish a prediction model for dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment based on predictive fuzzy neural network

[0012] Design a prediction model based on the fuzzy neural network to predict the change of dissolved oxygen concentration f1(S O,5 (t)) and the change in nitrate nitrogen concentration in the second sub-area f2(S NO,2 (t)), the prediction model is structured into four layers: input layer, membership function layer, rule layer and output layer, and the network structure is 6-12-12-2; specifically:

[0013] Input layer: The input layer consists of 6 neurons, h(t) = [h1(t), h2(t), h3(t), h4(t), h5(t), h6(t)] T is the input vector of the prediction model at time t, h1(t) = S O,5 (t), h2(t)=K L a5(t), h3(t)=e1(t), h4(t)=S NO,2 (t), h5(t)=Q a (t), h6(t) = e2(t), e1(t) represents the set value of dissolved oxygen concentration S at time t O,5d (t) and the actual value S O,5 (t), e2(t) represents the set value S of nitrate nitrogen concentration at time t NO,2d (t) and the actual value S NO,2 (t) error, T represents the transpose of the vector;

[0014] e1(t)=S O,5d (t)-S O,5 (t) (3)

[0015] e2(t)=S NO,2d (t)-S NO,2 (t) (4)

[0016] Membership function layer: The membership function layer has l neurons, and the output of each neuron in the membership function layer is:

[0017]

[0018] Among them, μ ij (t) represents the output value of the jth neuron in the membership function layer at time t, i = 1, 2, ..., 6, j = 1, 2, ..., 12, h i (t) represents the i-th variable in the input layer input vector h(t), c ij (t) represents the center of the i-th neuron in the input layer and the j-th neuron in the membership function layer at time t, and the initial value range is [-0.5, 0.5], a ij(t) represents the width of the i-th neuron in the input layer and the j-th neuron in the membership function layer at time t, and the initial value range is [-0.5, 0.5];

[0019] Regular layer: The regular layer has l neurons, and the output of each neuron in the regular layer is:

[0020]

[0021] Among them, γ j (t) is the output value of the jth neuron in the regular layer at time t;

[0022] Output layer: The output of the first neuron in the output layer is the predicted value of the change in dissolved oxygen concentration in the fifth partition at time t, and the output of the second neuron in the output layer is the predicted value of the change in nitrate nitrogen concentration in the second partition at time t:

[0023]

[0024] in, is the predicted value of the change in dissolved oxygen concentration in the fifth subarea at time t, is the predicted value of the change in nitrate nitrogen concentration in the second sub-area at time t, w j (t) = [w j1 (t),w j2 (t)] is the weight vector between the jth neuron in the regular layer and the neuron in the output layer at time t, w j1 (t) is the weight vector between the jth neuron in the rule layer and the first neuron in the output layer at time t, with an initial value range of [-0.5, 0.5], and w j2 (t) is the weight vector between the jth neuron in the rule layer and the second neuron in the output layer at time t, with an initial value range of [-0.5, 0.5], and γ j (t) is the output vector of the jth neuron in the regular layer of the prediction model at time t;

[0025] Design the estimated dissolved oxygen concentration in the fifth zone and nitrate nitrogen concentration in the second zone during the biochemical reaction process of urban sewage treatment:

[0026]

[0027] in, Estimated dissolved oxygen concentration in the fifth subarea at time t The derivative of Estimated nitrate nitrogen concentration for the second zone The derivative of ρ1(t) and ρ2(t) are the estimated errors of dissolved oxygen concentration in the fifth partition and nitrate nitrogen concentration in the second partition at time t:

[0028]

[0029] in, and They are the estimated values ​​of dissolved oxygen concentration in the fifth sub-area and nitrate nitrogen concentration in the second sub-area at time t, respectively;

[0030] The prediction model parameter adaptive law is designed using the estimated errors ρ1(t) and ρ2(t):

[0031]

[0032] Among them, Γ j (t) = [w j (t),c j (t),a j (t)] represents the parameter vector of the jth neuron in the regular layer of the prediction model at time t, w j (t) represents the output weight of the jth neuron in the rule layer of the prediction model at time t, c j (t) represents the center of the jth neuron in the regular layer of the prediction model at time t, a j (t) represents the width of the jth neuron in the regular layer of the prediction model at time t, Γ j (t+1)=[w j (t+1),c j (t+1),a j (t+1)] represents the parameter vector of the jth neuron in the regular layer of the prediction model at time t+1, w j (t+1) represents the output weight of the jth neuron in the rule layer of the adaptive prediction model at time t+1, c j (t+1) represents the center of the jth neuron in the regular layer of the prediction model at time t+1, a j (t+1) represents the width of the jth neuron in the rule layer of the prediction model at time t+1, s(t)=[s1(t),s2(t)] represents the sliding surface vector, s1(t) represents the sliding surface constructed based on the tracking control error of dissolved oxygen concentration, and s2(t) represents the sliding surface constructed based on the tracking control error of nitrate nitrogen concentration:

[0033]

[0034] Among them, ι=1,2,α ι (t) = [α1(t), α2(t)], α1(t) is the sliding surface parameter for designing the dissolved oxygen concentration controller, and α2(t) is the sliding surface parameter for designing the nitrate nitrogen concentration controller:

[0035]

[0036] in, is the sliding surface parameter α ιThe derivative of (t);

[0037] (3) Design of terminal sliding mode controller based on adaptive switching gain:

[0038] ① At the initial time t=1 of the control action, the initial values ​​of the adaptive prediction model parameters are randomly generated;

[0039] ② Calculate the predicted value of the change in dissolved oxygen concentration in the fifth zone according to formula (7)

[0040] ③ Calculate the predicted value of the change in nitrate nitrogen concentration in the second sub-area according to formula (8)

[0041] ④ Solve the parameter adaptation law of the prediction model according to formula (13);

[0042] ⑤ Calculate the dissolved oxygen concentration control law u1(t) of the fifth partition and the nitrate nitrogen concentration control law u2(t) of the second partition at time t:

[0043] u1(t)=u e1 (t)+u s1 (t) (16)

[0044] Among them, u e1 (t) is the equivalent control law of dissolved oxygen concentration at time t, u s1 (t) is the switching control law of dissolved oxygen concentration at time t:

[0045]

[0046] Among them, sgn() is the symbol function, is the derivative of the set value of dissolved oxygen concentration at time t;

[0047] u2(t)=u e2 (t)+u s2 (t) (19)

[0048] Among them, u e2 (t) is the equivalent control law of nitrate nitrogen concentration at time t, u s2 (t) is the switching control law of nitrate nitrogen concentration at time t:

[0049]

[0050] in, is the derivative of the set value of nitrate nitrogen concentration at time t;

[0051] ⑥At time t, increase by 1. If t<200, return to step ②. If t=200, end the loop.

[0052] (4) The solved u1(t) and u2(t) are used to track and control the dissolved oxygen concentration in the fifth zone and the nitrate nitrogen concentration in the second zone of the biochemical reaction process of urban sewage treatment, respectively. u1(t) is the oxygen transfer coefficient of the controller at time t, and u2(t) is the internal recirculation flow of the controller at time t. The output of the control system is the actual dissolved oxygen concentration value and the actual nitrate nitrogen concentration value. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is the tracking control effect diagram of the present invention when the dissolved oxygen concentration changes the set point in a sunny scene

[0054] Figure 2 This is a tracking control effect diagram of the nitrate nitrogen concentration change set point in a sunny scene of the present invention

[0055] Figure 3 This is the tracking control effect diagram of the present invention when the dissolved oxygen concentration changes the set point in a rainstorm scene

[0056] Figure 4 This is a tracking control effect diagram of the present invention for changing the set point of nitrate nitrogen concentration in a rainstorm scene DETAILED DESCRIPTION

[0057] 1. A control method for a municipal sewage treatment process based on an adaptive fuzzy terminal sliding mode, characterized in that: the change characteristics of dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of municipal sewage treatment are extracted, a prediction model of dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of municipal sewage treatment based on an estimated fuzzy neural network is established, and a terminal sliding mode controller based on an adaptive switching gain is designed to achieve stable control of dissolved oxygen concentration and nitrate nitrogen concentration; the method comprises the following steps:

[0058] (1) Extraction of the changing characteristics of dissolved oxygen concentration and nitrate nitrogen concentration during the biochemical reaction process of urban sewage treatment

[0059] The dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment fluctuate greatly over time and are easily affected by a series of disturbances such as influent flow, water quality fluctuations, influent composition and temperature changes. The control model of dissolved oxygen concentration in the fifth zone and nitrate nitrogen concentration in the second zone of the biochemical reaction tank is:

[0060]

[0061] Among them, S O,5 (t) represents the dissolved oxygen concentration in the fifth partition of the biochemical reaction pool at time t, f1(S O,5 (t)) represents the change in dissolved oxygen concentration in the fifth partition of the biochemical reaction tank at time t, K La5(t) represents the oxygen transfer coefficient of the fifth partition of the biochemical reaction pool at time t, Q4(t) is the flow rate of the fourth partition, and its value range is [0.3m / s, 0.5m / s], m / s represents meter / second, S O,4 It represents the dissolved oxygen concentration in the fourth zone of the biochemical reaction pool, which is 2 mg / L. NO,2 (t) represents the nitrate nitrogen concentration in the second partition of the biochemical reaction pool at time t, f2(S NO,2 (t)) represents the change in the nitrate nitrogen concentration in the second partition of the biochemical reaction pool at time t, Q a (t) represents the internal reflux of the second partition of the biochemical reaction pool at time t, Q in (t) represents the flow from the first partition to the second partition at time t, and its value range is [100m 3 / d,1000000m 3 / d],m 3 / d means cubic meters per day;

[0062] (2) Establish a prediction model for dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment based on predictive fuzzy neural network

[0063] Design a prediction model based on the fuzzy neural network to predict the change of dissolved oxygen concentration f1(S O,5 (t)) and the change in nitrate nitrogen concentration in the second sub-area f2(S NO,2 (t)), the prediction model is structured into four layers: input layer, membership function layer, rule layer and output layer, and the network structure is 6-12-12-2; specifically:

[0064] Input layer: The input layer consists of 6 neurons, h(t) = [h1(t), h2(t), h3(t), h4(t), h5(t), h6(t)] T is the input vector of the prediction model at time t, h1(t) = S O,5 (t), h2(t)=K L a5(t), h3(t)=e1(t), h4(t)=S NO,2 (t), h5(t)=Q a (t), h6(t) = e2(t), e1(t) represents the set value of dissolved oxygen concentration S at time t O,5d (t) and the actual value S O,5 (t), e2(t) represents the set value S of nitrate nitrogen concentration at time t NO,2d (t) and the actual value S NO,2 (t) error, T represents the transpose of the vector;

[0065] e1(t)=S O,5d(t)-S O,5 (t) (3)

[0066] e2(t)=S NO,2d (t)-S NO,2 (t) (4)

[0067] Membership function layer: The membership function layer has l neurons, and the output of each neuron in the membership function layer is:

[0068]

[0069] Among them, μ ij (t) represents the output value of the jth neuron in the membership function layer at time t, i = 1, 2, ..., 6, j = 1, 2, ..., 12, h i (t) represents the i-th variable in the input layer input vector h(t), c ij (t) represents the center of the i-th neuron in the input layer and the j-th neuron in the membership function layer at time t, and the initial value range is [-0.5, 0.5], a ij (t) represents the width of the i-th neuron in the input layer and the j-th neuron in the membership function layer at time t, and the initial value range is [-0.5, 0.5];

[0070] Regular layer: The regular layer has l neurons, and the output of each neuron in the regular layer is:

[0071]

[0072] Among them, γ j (t) is the output value of the jth neuron in the regular layer at time t;

[0073] Output layer: The output of the first neuron in the output layer is the predicted value of the change in dissolved oxygen concentration in the fifth partition at time t, and the output of the second neuron in the output layer is the predicted value of the change in nitrate nitrogen concentration in the second partition at time t:

[0074]

[0075] in, is the predicted value of the change in dissolved oxygen concentration in the fifth subarea at time t, is the predicted value of the change in nitrate nitrogen concentration in the second sub-area at time t, w j (t) = [w j1 (t),w j2 (t)] is the weight vector between the jth neuron in the regular layer and the neuron in the output layer at time t, w j1 (t) is the weight vector between the jth neuron in the rule layer and the first neuron in the output layer at time t, with an initial value range of [-0.5, 0.5], and w j2(t) is the weight vector between the jth neuron in the rule layer and the second neuron in the output layer at time t, with an initial value range of [-0.5, 0.5], and γ j (t) is the output vector of the jth neuron in the regular layer of the prediction model at time t;

[0076] Design the estimated dissolved oxygen concentration in the fifth zone and nitrate nitrogen concentration in the second zone during the biochemical reaction process of urban sewage treatment:

[0077]

[0078] in, Estimated dissolved oxygen concentration in the fifth subarea at time t The derivative of Estimated nitrate nitrogen concentration for the second zone The derivative of ρ1(t) and ρ2(t) are the estimated errors of dissolved oxygen concentration in the fifth partition and nitrate nitrogen concentration in the second partition at time t:

[0079]

[0080] in, and They are the estimated values ​​of dissolved oxygen concentration in the fifth sub-area and nitrate nitrogen concentration in the second sub-area at time t, respectively;

[0081] The prediction model parameter adaptive law is designed using the estimated errors ρ1(t) and ρ2(t):

[0082]

[0083] Among them, Γ j (t) = [w j (t),c j (t),a j (t)] represents the parameter vector of the jth neuron in the regular layer of the prediction model at time t, w j (t) represents the output weight of the jth neuron in the rule layer of the prediction model at time t, c j (t) represents the center of the jth neuron in the regular layer of the prediction model at time t, a j (t) represents the width of the jth neuron in the regular layer of the prediction model at time t, Γ j (t+1)=[w j (t+1),c j (t+1),a j (t+1)] represents the parameter vector of the jth neuron in the regular layer of the prediction model at time t+1, w j (t+1) represents the output weight of the jth neuron in the rule layer of the adaptive prediction model at time t+1, c j(t+1) represents the center of the jth neuron in the regular layer of the prediction model at time t+1, a j (t+1) represents the width of the jth neuron in the rule layer of the prediction model at time t+1, s(t)=[s1(t),s2(t)] represents the sliding surface vector, s1(t) represents the sliding surface constructed based on the tracking control error of dissolved oxygen concentration, and s2(t) represents the sliding surface constructed based on the tracking control error of nitrate nitrogen concentration:

[0084]

[0085] Among them, ι=1,2,α ι (t) = [α1(t), α2(t)], α1(t) is the sliding surface parameter for designing the dissolved oxygen concentration controller, and α2(t) is the sliding surface parameter for designing the nitrate nitrogen concentration controller:

[0086]

[0087] in, is the sliding surface parameter α ι The derivative of (t);

[0088] (3) Design of terminal sliding mode controller based on adaptive switching gain:

[0089] ① At the initial time t=1 of the control action, the initial values ​​of the adaptive prediction model parameters are randomly generated;

[0090] ② Calculate the predicted value of the change in dissolved oxygen concentration in the fifth zone according to formula (7)

[0091] ③ Calculate the predicted value of the change in nitrate nitrogen concentration in the second sub-area according to formula (8)

[0092] ④ Solve the parameter adaptation law of the prediction model according to formula (13);

[0093] ⑤ Calculate the dissolved oxygen concentration control law u1(t) of the fifth partition and the nitrate nitrogen concentration control law u2(t) of the second partition at time t:

[0094] u1(t)=u e1 (t)+u s1 (t) (16)

[0095] Among them, u e1 (t) is the equivalent control law of dissolved oxygen concentration at time t, u s1 (t) is the switching control law of dissolved oxygen concentration at time t:

[0096]

[0097] Among them, sgn() is the symbol function, is the derivative of the set value of dissolved oxygen concentration at time t;

[0098] u2(t)=u e2 (t)+u s2 (t) (19)

[0099] Among them, u e2 (t) is the equivalent control law of nitrate nitrogen concentration at time t, u s2 (t) is the switching control law of nitrate nitrogen concentration at time t:

[0100]

[0101] in, is the derivative of the set value of nitrate nitrogen concentration at time t;

[0102] ⑥At time t, increase by 1. If t<200, return to step ②. If t=200, end the loop.

[0103] (4) The solved u1(t) and u2(t) are used to track and control the dissolved oxygen concentration in the fifth zone and the nitrate nitrogen concentration in the second zone of the biochemical reaction process of urban sewage treatment, respectively. u1(t) is the oxygen transfer coefficient of the controller at time t, and u2(t) is the internal recirculation flow of the controller at time t. The output of the control system is the actual dissolved oxygen concentration value and the actual nitrate nitrogen concentration value. Figure 1 This is a tracking control effect diagram of the present invention when the dissolved oxygen concentration changes the set point in a sunny scene; Figure 2 This is a tracking control effect diagram of the nitrate nitrogen concentration changing set point in a sunny scene of the present invention; Figure 3 This is a tracking control effect diagram of the present invention when the dissolved oxygen concentration changes the set point in a rainstorm scene; Figure 4 This is a tracking control effect diagram of the present invention for changing the set point of nitrate nitrogen concentration in a rainstorm scene. X-axis: time, unit is day, Figure 1 and Figure 3 The Y axis in the chart shows the outlet dissolved oxygen concentration value and the dissolved oxygen concentration tracking error value, in mg / L. The blue dotted line is the actual dissolved oxygen concentration value, the black solid line is the dissolved oxygen concentration set value, and the black dotted line is the dissolved oxygen concentration tracking error value. Figure 2 and Figure 4 Y-axis: nitrate nitrogen concentration value and nitrate nitrogen concentration tracking error value, the unit is mg / L, the blue dotted line is the actual nitrate nitrogen concentration value, the black solid line is the nitrate nitrogen concentration set value, and the black dotted line is the nitrate nitrogen concentration tracking error value.

Claims

1. A method for controlling a municipal sewage treatment process based on an adaptive fuzzy terminal sliding mode, characterized in that: Extract the variation characteristics of dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment, establish a prediction model of dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment based on the estimated fuzzy neural network, and design a terminal sliding mode controller based on adaptive switching gain to achieve stable control of dissolved oxygen concentration and nitrate nitrogen concentration; including the following steps: (1) Extraction of the changing characteristics of dissolved oxygen concentration and nitrate nitrogen concentration during the biochemical reaction process of urban sewage treatment The dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment fluctuate greatly over time and are easily affected by a series of disturbances such as influent flow, water quality fluctuations, influent composition and temperature changes. The control model of dissolved oxygen concentration in the fifth zone and nitrate nitrogen concentration in the second zone of the biochemical reaction tank is: Among them, S O,5 (t) represents the dissolved oxygen concentration in the fifth partition of the biochemical reaction pool at time t, f1(S O,5 (t)) represents the change in dissolved oxygen concentration in the fifth partition of the biochemical reaction tank at time t, K L a5(t) represents the oxygen transfer coefficient of the fifth partition of the biochemical reaction pool at time t, Q4(t) is the flow rate of the fourth partition, and its value range is [0.3m / s, 0.5m / s], m / s represents meter / second, S O,4 It represents the dissolved oxygen concentration in the fourth zone of the biochemical reaction pool, which is 2 mg / L. NO,2 (t) represents the nitrate nitrogen concentration in the second partition of the biochemical reaction pool at time t, f2(S NO,2 (t)) represents the change in the nitrate nitrogen concentration in the second partition of the biochemical reaction pool at time t, Q a (t) represents the internal reflux of the second partition of the biochemical reaction pool at time t, Q in (t) represents the flow from the first partition to the second partition at time t, and its value range is [100m 3 / d,1000000m 3 / d],m 3 / d means cubic meters per day; (2) Establish a prediction model for dissolved oxygen concentration and nitrate nitrogen concentration in the biochemical reaction process of urban sewage treatment based on predictive fuzzy neural network Design a prediction model based on the fuzzy neural network to predict the change of dissolved oxygen concentration f1(S O,5 (t)) and the change in nitrate nitrogen concentration in the second sub-area f2(S NO,2 (t)), the prediction model is structured into four layers: input layer, membership function layer, rule layer and output layer, and the network structure is 6-12-12-2; specifically: Input layer: The input layer consists of 6 neurons, h(t) = [h1(t), h2(t), h3(t), h4(t), h5(t), h6(t)] T is the input vector of the prediction model at time t, h1(t) = S O,5 (t), h2(t)=K L a5(t), h3(t)=e1(t), h4(t)=S NO,2 (t), h5(t)=Q a (t), h6(t) = e2(t), e1(t) represents the set value of dissolved oxygen concentration S at time t O,5d (t) and the actual value S O,5 (t), e2(t) represents the set value S of nitrate nitrogen concentration at time t NO,2d (t) and the actual value S NO,2 (t) error, T represents the transpose of the vector; e1(t)=S O,5d (t)-S O,5 (t) (3) e2(t)=S NO,2d (t)-S NO,2 (t) (4) Membership function layer: The membership function layer has l neurons, and the output of each neuron in the membership function layer is: Among them, μ ij (t) represents the output value of the jth neuron in the membership function layer at time t, i = 1, 2, ..., 6, j = 1, 2, ..., 12, h i (t) represents the i-th variable in the input layer input vector h(t), c ij (t) represents the center of the i-th neuron in the input layer and the j-th neuron in the membership function layer at time t, and the initial value range is [-0.5, 0.5], a ij (t) represents the width of the i-th neuron in the input layer and the j-th neuron in the membership function layer at time t, and the initial value range is [-0.5, 0.5]; Regular layer: The regular layer has l neurons, and the output of each neuron in the regular layer is: Among them, γ j (t) is the output value of the jth neuron in the regular layer at time t; Output layer: The output of the first neuron in the output layer is the predicted value of the change in dissolved oxygen concentration in the fifth partition at time t, and the output of the second neuron in the output layer is the predicted value of the change in nitrate nitrogen concentration in the second partition at time t: in, is the predicted value of the change in dissolved oxygen concentration in the fifth sub-area at time t, is the predicted value of the change in nitrate nitrogen concentration in the second sub-area at time t, w j (t) = [w j1 (t),w j2 (t)] is the weight vector between the jth neuron in the regular layer and the neuron in the output layer at time t, w j1 (t) is the weight vector between the jth neuron in the rule layer and the first neuron in the output layer at time t, with an initial value range of [-0.5, 0.5], and w j2 (t) is the weight vector between the jth neuron in the rule layer and the second neuron in the output layer at time t, with an initial value range of [-0.5, 0.5], and γ j (t) is the output vector of the jth neuron in the regular layer of the prediction model at time t; Design the estimated dissolved oxygen concentration in the fifth zone and nitrate nitrogen concentration in the second zone during the biochemical reaction process of urban sewage treatment: in, Estimated dissolved oxygen concentration in the fifth subarea at time t The derivative of Estimated nitrate nitrogen concentration for the second zone The derivative of ρ1(t) and ρ2(t) are the estimated errors of dissolved oxygen concentration in the fifth partition and nitrate nitrogen concentration in the second partition at time t: in, and They are the estimated values ​​of dissolved oxygen concentration in the fifth sub-area and nitrate nitrogen concentration in the second sub-area at time t, respectively; The prediction model parameter adaptive law is designed using the estimated errors ρ1(t) and ρ2(t): Among them, Γ j (t) = [w j (t),c j (t),a j (t)] represents the parameter vector of the jth neuron in the regular layer of the prediction model at time t, w j (t) represents the output weight of the jth neuron in the rule layer of the prediction model at time t, c j (t) represents the center of the jth neuron in the regular layer of the prediction model at time t, a j (t) represents the width of the jth neuron in the regular layer of the prediction model at time t, Γ j (t+1)=[w j (t+1),c j (t+1),a j (t+1)] represents the parameter vector of the jth neuron in the regular layer of the prediction model at time t+1, w j (t+1) represents the output weight of the jth neuron in the rule layer of the adaptive prediction model at time t+1, c j (t+1) represents the center of the jth neuron in the regular layer of the prediction model at time t+1, a j (t+1) represents the width of the jth neuron in the rule layer of the prediction model at time t+1, s(t)=[s1(t),s2(t)] represents the sliding surface vector, s1(t) represents the sliding surface constructed based on the tracking control error of dissolved oxygen concentration, and s2(t) represents the sliding surface constructed based on the tracking control error of nitrate nitrogen concentration: Among them, ι=1,2,α ι (t) = [α1(t), α2(t)], α1(t) is the sliding surface parameter for designing the dissolved oxygen concentration controller, and α2(t) is the sliding surface parameter for designing the nitrate nitrogen concentration controller: in, is the sliding surface parameter α ι The derivative of (t); (3) Design of terminal sliding mode controller based on adaptive switching gain: ① At the initial time t=1 of the control action, the initial values ​​of the adaptive prediction model parameters are randomly generated; ② Calculate the predicted value of the change in dissolved oxygen concentration in the fifth zone according to formula (7) ③ Calculate the predicted value of the change in nitrate nitrogen concentration in the second sub-area according to formula (8) ④ Solve the parameter adaptation law of the prediction model according to formula (13); ⑤ Calculate the dissolved oxygen concentration control law u1(t) of the fifth partition and the nitrate nitrogen concentration control law u2(t) of the second partition at time t: u1(t)=u e1 (t)+u s1 (t) (16) Among them, u e1 (t) is the equivalent control law of dissolved oxygen concentration at time t, u s1 (t) is the switching control law of dissolved oxygen concentration at time t: Among them, sgn() is the symbol function, is the derivative of the set value of dissolved oxygen concentration at time t; u2(t)=u e2 (t)+u s2 (t) (19) Among them, u e2 (t) is the equivalent control law of nitrate nitrogen concentration at time t, u s2 (t) is the switching control law of nitrate nitrogen concentration at time t: in, is the derivative of the set value of nitrate nitrogen concentration at time t; ⑥At time t, increase by 1. If t<200, return to step ②. If t=200, end the loop. (4) The solved u1(t) and u2(t) are used to track and control the dissolved oxygen concentration in the fifth zone and the nitrate nitrogen concentration in the second zone of the biochemical reaction process of urban sewage treatment, respectively. u1(t) is the oxygen transfer coefficient of the controller at time t, and u2(t) is the internal recirculation flow of the controller at time t. The output of the control system is the actual dissolved oxygen concentration value and the actual nitrate nitrogen concentration value.

Images