JP3260538B2 - Control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for controlling the concentration of dissolved oxygen in an aeration tank of a sewage treatment plant to a constant value.

[0002]

2. Description of the Related Art Conventionally, the dissolved oxygen concentration in an aeration tank of a sewage treatment facility is controlled by a feedback control that determines the amount of aerated air from an error between the observed dissolved oxygen concentration and a target dissolved oxygen concentration. If the amount of inflow of sewage changes suddenly due to rainfall or the like, accurate control cannot be performed by feedback control, and the observer must manually adjust the aeration air volume. In addition, not only abrupt changes in the concentration of dissolved oxygen due to changes in the amount of inflow of sewage, but also the characteristics of the sewage treatment process greatly change depending on the season. At present, the gain constant of feedback control must be adjusted.

[0003] Although studies have been made to express the characteristics of sewage treatment processes in mathematical models, the sewage treatment process is a complex process involving various factors utilizing biochemically reactive ions of microorganisms. Therefore, it is difficult to adjust the parameters, and good results have not been obtained.

[0004]

As described above, the feedback control conventionally performed by controlling the dissolved oxygen concentration in the sewage treatment process cannot cope with a rapid change in the amount of sewage inflow due to rainfall or the like. It is necessary to adjust the gain constant for a gentle change in characteristics as in seasonal units.

The present invention eliminates the above-mentioned disadvantages of the prior art, and controls the dissolved oxygen concentration to a constant value by a calculation with a small number of repetitions in response to a rapid change in the amount of sewage inflow due to rainfall and the like. The purpose of the present invention is to provide a control device that can adapt to moderate changes in the characteristics of a sewage treatment process.

[0006]

According to a first aspect of the present invention, there is provided a control apparatus for controlling a control amount of a control target by changing an operation amount input to the control target. A virtual operation amount calculation unit that calculates a virtual operation amount based on the control amount measured in the control object and a target control amount, and a control amount measured in the control object and calculated by the virtual operation amount calculation unit. Prediction means for predicting the control amount after a predetermined time based on the virtual operation amount and the past operation amount, and the virtual amount based on the control amount and the target control amount after the predetermined time predicted by the prediction means. Correction means for correcting the virtual operation amount calculated by the operation amount calculation means, wherein the operation amount obtained by performing at least one of the prediction by the prediction means and the correction by the correction means is used. A control apparatus for and a manipulated variable correcting means for inputting the control target, the prediction means, Manabu
It consists of a function approximation device with learning function,
Corrective means for controlling the past control measured in the control object;
And the past operation amount input to the control object.
Learning means for learning the function approximation device.

[0007]

According to a second aspect of the present invention, there is provided a control apparatus for controlling a dissolved oxygen concentration in a sewage treatment process, wherein the dissolved oxygen concentration in the sewage treatment process is controlled by changing an aeration amount input to the sewage treatment process. Based on the measurement data measured in the process and the target dissolved oxygen concentration, the virtual aeration air volume calculation means for calculating a virtual aeration air volume, and a function approximation device having a learning function, and are measured in the sewage treatment process. Based on the measurement data, the virtual aeration air volume calculated by the virtual aeration air volume calculation unit and the past aeration air volume, a prediction unit that predicts the dissolved oxygen concentration after a predetermined time, and a prediction unit that predicts the dissolved oxygen concentration after the predetermined time. Based on the dissolved oxygen concentration and the target dissolved oxygen concentration, the virtual Correction means for correcting the aeration air volume, and learning means for learning the function approximation device based on past measurement data measured in the sewage treatment process and past aeration air volume input to the water treatment process, Aeration air volume correction means for inputting the aeration air volume obtained by performing at least one of the prediction by the prediction means and the correction by the correction means to the sewage treatment process.

According to a third aspect of the present invention, in the control apparatus for controlling the dissolved oxygen concentration in the sewage treatment process, the dissolved oxygen concentration in the sewage treatment process is controlled by changing the amount of aeration air input to the sewage treatment process. The sewage treatment comprises virtual aeration air volume calculation means for calculating a virtual aeration air volume based on the sewage inflow amount and dissolved oxygen concentration measured in the process and a target dissolved oxygen concentration, and a function approximation device having a learning function. Prediction means for predicting the dissolved oxygen concentration after a predetermined time based on the dissolved oxygen concentration measured in the process, the virtual aeration air flow calculated by the virtual aeration air flow calculation means and the past aeration air flow, The virtual aeration air volume calculating means based on the dissolved oxygen concentration after a predetermined time and the target dissolved oxygen concentration. Correction means for correcting the virtual aeration air volume calculated from the above, and the function based on the past aeration air volume input to the water treatment process and the past sewage inflow amount and dissolved oxygen concentration measured in the sewage treatment process. A learning unit for learning the approximating device, wherein at least one of the prediction by the prediction unit and the correction by the correction unit is performed.
Aeration air volume correction means for inputting the aeration air volume obtained by the rounding to the sewage treatment process.

The invention according to claim 4 is the invention according to claim 2 or 3.
The apparatus for controlling the concentration of dissolved oxygen in a sewage treatment process described above, wherein the predicting means is constituted by a neural circuit model, and the correcting means is realized by back propagation calculation of the neural circuit model.

According to a fifth aspect of the present invention, in the control apparatus of the dissolved oxygen concentration in the sewage treatment process according to any one of the second to fourth aspects, the aeration air volume correcting means includes a dissolved oxygen gas after a predetermined time predicted by the predicting means. Until the error between the concentration and the target dissolved oxygen concentration is within a predetermined error range, or a predetermined number of times or a predetermined range in which the amount of change in the dissolved oxygen concentration after a predetermined time predicted by the prediction means is determined. It is characterized in that the prediction by the predicting means and the correction by the correcting means are repeated until the value is within.

[0012]

In the control device according to the present invention, the prediction means predicts the control amount after a predetermined time based on at least the virtual operation amount calculated by the virtual operation amount calculation means and the past operation amount. Further, the correction means calculates a correction amount of the virtual operation amount based on the control amount after a predetermined time predicted by the prediction means and the target control amount. Then, the prediction by the prediction unit and the correction by the correction unit are repeated, and the resulting virtual operation amount is input as the operation amount of the control target.
Thereby, even if the parameter affecting the control amount changes rapidly, the control amount can be corrected accurately. In addition, by using the virtual operation amount calculated by the virtual operation amount calculation unit as the initial value of the prediction unit, the control amount can be corrected with a small number of repetitions even when the control amount is largely changed. . Further, the prediction means is constituted by a function approximation device having a learning function, and the learning means learns the function approximation device based on a past control amount measured in the control target and a past operation amount input to the control target. By continuing, it is possible to adapt to not only a sudden change of the parameter but also a gradual change of the parameter.

[0013]

FIG. 1 is a block diagram for explaining the principle of the present invention. In the figure, reference numeral 1 denotes a control target. The control amount of the control target 1 is the operation amount 3 input from the control device 2.
Is controlled by changing. The control device 2 includes a virtual operation amount calculation unit 4 and an operation amount correction unit 5. The virtual operation amount calculating means 4 calculates a virtual operation amount 8 based on the control amount 6 measured in the control target 1 and the target control amount 7.
Is calculated. The operation amount correction unit 5 predicts the control amount 9 after a predetermined time based on the control amount 6 measured in the control target 1, the virtual operation amount 8 calculated by the virtual operation amount calculation unit 4, and the past operation amount. Correction means for correcting the virtual operation amount 8 calculated by the virtual operation amount calculation means 4 based on the control amount 9 after a predetermined period of time predicted by the prediction means 10 and the target control amount 7 11 is provided. The operation amount correction unit 5 inputs the operation amount obtained by performing at least one of the prediction by the prediction unit 10 and the correction by the correction unit 11 as the operation amount 3 to the control target 1.

Thus, according to the present invention, the control amount can be accurately corrected even if the parameter affecting the control amount changes abruptly. In addition, by using the virtual operation amount 8 calculated by the virtual operation amount calculation unit 4 as an initial value of the prediction unit 10, even when the control amount is largely changed, the control amount can be corrected by calculation with a small number of repetitions. Can be.

Further, the prediction means 10 is constituted by a function approximation device having a learning function, and the function approximation device is configured based on a past control amount measured on the control target and a past operation amount input to the control target. By continuing to learn, it is possible to adapt not only to sudden changes in parameters but also to gradual changes in parameters. Next, an embodiment in which the present invention is applied to a dissolved oxygen concentration control device in a sewage treatment process for controlling the dissolved oxygen concentration in a sewage treatment process will be described.

FIG. 2 is a block diagram showing the configuration of the control device. In FIG. 2, reference numeral 12 denotes a sewage treatment process to be controlled, 13 denotes a feedback controller, 14 denotes an aeration air volume corrector, and 15 denotes a feedback loop.

In the sewage treatment process 12, an aeration air volume baq as an operation amount is input for each sampling time Δt, and a dissolved oxygen concentration do as a control amount and a sewage inflow amount iq as a parameter affecting the control amount are sampled. It is measured every time Δt. The subscript n indicates the time nΔt, for example, baq
_n means the aeration air volume input to the sewage treatment process at time nΔt, and do _n means the dissolved oxygen concentration observed at time nΔt. Subscripts will be used in the same meaning hereinafter. The feedback controller 13 controls the observed dissolved oxygen concentration do _n
Feedback aeration air volume from target dissolved oxygen concentration do ^d and b
Calculate aq ^f _n . The aeration air volume corrector 14 calculates a correction amount of the aeration air volume using a prediction model for predicting the dissolved oxygen concentration after a predetermined time and an error between the predicted dissolved oxygen concentration and the target dissolved oxygen concentration do ^d. An aeration air volume correction amount calculator is provided. Aeration amount corrector 14 is observed dissolved oxygen concentration do _n, sewage inflow iq _n, the target concentration of dissolved oxygen do ^d, feedback aeration amount BAQ ^f _n are input repeatedly feedback aeration amount BAQ ^f _n as an initial value The corrected virtual aeration air volume is output as the aeration air volume baq _n . The feedback loop 15 measures the dissolved oxygen concentration do _n
Is fed back to the feedback controller 13 and the aeration air volume corrector 14.

Next, the aeration air volume corrector 14 whose prediction model is constituted by a multilayer neural network model will be described. FIG. 3 is a configuration diagram of the aeration air volume corrector 14, in which 16 is a prediction model, 17 is an aeration air volume correction amount calculator, 18 is a first switch, 19 is a second switch, 20 is a time delay element, 21 is an adder.

The prediction model 16 calculates the time (n−m _do ) Δt
Dissolved oxygen concentration ( _{don -mdo} , do
_{n-mdo + 1, ...,} do n), the time (sewage inflow from n-m _iq) Δt to the time _{nΔt (iq n-miq, iq} n-miq + 1, ... iq
_n ), from time ( _nmbaq ) Δt to time nΔt,
When the aeration air volume (baq _n-mbaq , baq _{n-mbaq + 1} ,..., Baq _n ) is input, the neural network model trained to output the predicted dissolved oxygen concentration at time (n + 1) Δt I do. This learning method will be described later. m _do , m _iq , m
_baq is a positive integer. Now, the dissolved oxygen concentration do _n and the sewage inflow amount iq _n and the aeration air amount baq _n-1 input to the aeration air volume corrector 14 are input to the prediction model 16 by delaying the time through a multi _- stage time delay element 20. Is done. The prediction model 16 also includes a virtual aeration air volume baq (hat)
_n is entered.

The prediction model 16 calculates a predicted value do (存) _{n + 1} of the dissolved oxygen concentration at time (n + 1) Δt from these input signals.
Is calculated. This calculation is performed by forward calculation of the neural network model. The forward calculation of the neural circuit model will be described later. The predicted value do (〜) _{n + 1} of the dissolved oxygen concentration at the time (n + 1) Δt obtained by the prediction model 16 is input to the aeration air flow correction amount calculator 17.

When the dissolved oxygen concentration do (〜) _{n + 1} predicted by the prediction model 16 and the target dissolved oxygen concentration do _d are input to the aeration air flow correction amount calculator 17, the predicted dissolved oxygen concentration do (〜 The correction amount Δbaq _n of the virtual aeration air volume is calculated so that _{n + 1} matches the target dissolved oxygen concentration do _d . In this calculation, when the prediction model 16 is a multilayer neural network model, a correction amount can be calculated by backpropagation calculation.
The back propagation calculation of the multilayer neural network model will be described later together with the forward calculation.

The correction amount Δbaq _n of the virtual aeration air volume output from the aeration air volume correction amount calculator 17 is added to the virtual aeration air volume baq (hat) _n by the adder 21.

The first switch 18 is closed downward at the sampling time, and the feedback aeration air volume baq ^f _n output from the feedback controller 13 (see FIG. 2) is used as a virtual aeration air volume baq (hat) _n as a prediction model. 16 is input. Thereafter, the first switch 18 is closed upward until the next sampling time, and the corrected virtual aeration air volume baq _n (hat) is input to the prediction model 16.

The second switch 19 is closed at the sampling time, and the virtual aeration air volume baq (hat) _n is changed to the time nΔ
The aeration air volume corrector 14 outputs the aeration air volume baq _n of t to the sewage treatment process 12 (see FIG. 2).

Next, the calculation procedure of the entire control device will be described with reference to the flowchart shown in FIG. In the figure, from time 0 to N
Although it is described that control is performed up to Δt, this calculation procedure is repeated at each sampling time, and therefore, here, a procedure for obtaining the aeration air volume at time nΔt will be described.

(A) Target dissolved oxygen concentration set in advance
do ^d and the observed dissolved oxygen concentration do _n are input to the feedback controller 13 and the feedback aeration air volume baq ^f _n
Is output.

(B) In the aeration air volume corrector 14, the first switch 18 is closed downward, and the feedback aeration air volume baq ^f _n obtained in (a) becomes the virtual aeration air volume baq (hat) _n . (c) In addition to the virtual aeration air volume baq (hat) _n , the time (n−
m) Dissolved oxygen concentration from Δt to time nΔt (do _nm , do
_{n−m + 1} ,..., do _n ) and time (n−m) Δt to time n
Sewage inflow up to Δt (iq _n-miq , iq _{n-miq + 1} ,…, iq
_n ), aeration air volume from time (nm) Δt to time (n-1) Δt (baq _n-mbaq , baq _{n-mbaq + 1} , ..., baq _n-1 )
Is input to the prediction model 16, and the predicted dissolved oxygen concentration do (〜) _{n + 1} after the sampling time Δt is output.

(D) The aeration air flow correction amount calculator 17 calculates an error function from the predicted dissolved oxygen concentration do (〜) _{n + 1} and the target dissolved oxygen concentration do ^d.

(Equation 1) Find the value of

(E) A back propagation calculation is performed so that the error function value approaches 0, and a correction amount Δbaq _n of the virtual aeration air volume baq (hat) _n which is an input signal of the prediction model 16 is calculated.

(F) The adder 21 adds the correction amount Δbaq _n to the virtual aeration air volume baq (hat) _n .

(G) The first switch 18 is closed on the upper side until the error function is within a predetermined error range, or until the correction amount of the predicted dissolved oxygen concentration is within a predetermined number of times or within a predetermined range. , The corrected virtual aeration air volume baq (hat) _n is input to the prediction model 16 again, and from (c) to (f).
The procedure up to is repeated.

(H) At the sampling time, the second switch 19 is closed, and the virtual aeration air volume baq (hat) _n is input to the sewage treatment process 12 as the aeration air volume baq _n .

The above procedures (a) to (h) are repeated at each sampling time.

Next, the forward calculation, the backpropagation calculation and the learning calculation of the neural network model will be described in detail. When the prediction model 16 is composed of a multilayer neural network model, the prediction calculation is performed by forward calculation of the multilayer neural network model, and the calculation of the correction amount of the aeration air flow is performed by error back propagation calculation of the multilayer neural network model.

First, the forward calculation will be described.
The forward calculation is a calculation for obtaining an output signal when an input signal is given to a neural circuit model. Here, the description will be made using the most commonly used three-layer neural network model among the multilayer neural network models.

FIG. 5 is a configuration diagram of a prediction model 16 composed of a three-layer type neural circuit model. 23, 24, 2 in the figure
The circle shown in 5 is a nerve cell model called a unit, and the line segment connecting the units 23, 24, and 25 is a connection that transmits a signal between the units 23, 24, and 25.
The units 23, 24, and 25 are arranged from the left into three layers: an input layer, an intermediate layer, and an output layer. Coupling is provided between the unit 23 of the input layer and the unit 24 of the intermediate layer, and between the unit 24 of the intermediate layer and the unit 25 of the output layer. The coupling between the units 23, 24, 25 is directional in signal transmission, with forward calculations transmitting signals only in one direction from the input layer unit 23 to the intermediate layer unit 24 and from the intermediate layer unit 24 to the output layer unit 25. Is done.

The unit 23 of the input layer has a virtual aeration air volume.
baq (hat) _n, time (n-m _do) dissolved oxygen concentration of from Δt to _{nΔt (do n-mdo, do} n-mdo + 1, ..., do n)
, Sewage inflow from time (n-m _iq ) Δt to nΔt
(iq _n-miq , iq _{n-miq + 1} ,…, iq _n ) and time (nm
_baq ) Aeration air volume (baq) from Δt to time (n-1) Δt
_{n- mbaq} , baqn _{-mbaq + 1} ,..., baqn _-1 ) are input to each unit one signal at a time, and the predicted value of the dissolved oxygen concentration at time (n + 1) Δt is output. Here, m _do , m _iq , and m _baq each represent a positive integer. The forward calculation for calculating the output signal from the input signal is shown in an equation. Let xi _{i be} the input value of the ith unit 23 in the input layer,

(Equation 2) The input / output function of the input layer unit 23 is an identity function. The output yi _i (i = 1,..., Ni) of the input layer unit 23 is expressed as follows.

[0038]

(Equation 3) Ni is the number of input layer units (Ni = m _do + m _iq + m
_baq +3). Next, the input value xh _j to the j-th unit 24 of the hidden layer is obtained by inputting the output value of the input layer unit 23 −
It is a value obtained by subtracting the threshold value thh _j of the intermediate layer unit 24 from the total weighted by the coupling load between the intermediate layers. Nh represents the unit 24 of the intermediate layer.

[0039]

(Equation 4) ω1 _{j, i} is the ith unit of the input layer 23-jth layer of the intermediate layer
Thh _j is the middle layer j
The threshold of the 24th unit, Nh, is the unit 2 of the middle layer
4 is represented.

The input / output function of the intermediate layer unit 24 is represented by a sigmoid type function.

(Equation 5) Then, the output value y h _j of the intermediate unit 24 becomes the input value x
Using h _j ,

(Equation 6) Can be expressed as y h _j is the output value of the j-th unit 24 of the intermediate layer. The k-th input value to the unit 25 of x o _{k (k} = 1) is intermediate the output value of the unit 24 to the middle layer of the output layer - the threshold of the output layer unit 25 from the sum weighted by the connection weight output layers It becomes the subtracted value.

[0041]

(Equation 7) .omega.2 _{k, i} is the threshold value of the intermediate layer j-th unit 24 output layer k-th connection weights between units 25, th o _k is the k-th unit 25 in the output layer. The input / output function of the output layer unit 25 is an identity function.

[0042]

(Equation 8) The output value of the output layer unit 25 is the predicted dissolved oxygen concentration do
(~) _{N + 1} .

[0043]

(Equation 9) The above is the calculation formula of the forward calculation of the three-layer type neural circuit model.

In FIG. 5, a neural network model composed of three layers, an input layer, a middle layer, and an output layer, is shown as the prediction model 16, but a neural network model composed of two or more layers is used as the prediction model 16. May be.

Next, the error back propagation calculation will be described.
As described above, the aeration air flow correction amount calculator 17 calculates the dissolved oxygen concentration do (hat) predicted by the prediction model 16.
_{When n + 1} and the target dissolved oxygen concentration do ^d are input, the correction amount Δbaq _n of the virtual aeration air volume is calculated and output. This calculation can be performed by backpropagation calculation when the prediction model 16 is configured from a neural network model. The calculation formula of the error back propagation calculation by the aeration air flow correction amount calculator 17 is shown below.

First, the dissolved oxygen concentration do (hat) _{n + 1} predicted by the prediction model 16 and the target dissolved oxygen concentration do
^An error function E is defined as follows from an error from ^d .

[0047]

(Equation 10) Next, the correction amount of the unit output value for reducing the error function is sequentially obtained for each layer from the output layer to the input layer, and finally the correction amount of the input signal is obtained. First, the amount of correction of the output value of the output layer unit is calculated using a positive constant η.

[Equation 11] Is calculated. Next, the correction amount of the output value of the intermediate layer unit is expressed as follows using the correction amount of the output value unit.

[0048]

(Equation 12) Similarly, the correction amount of the output value of the input layer unit is calculated using the correction amount of the intermediate unit.

(Equation 13) It is expressed as f ′ (x) is a sigmoid function f ′ (x)
Represents the derivative of

[Equation 14] It is expressed as Since the input / output function of the input layer unit is an identity function, the correction amount of the output value becomes the correction amount of the input value as it is, and the correction amount Δbaq of the virtual aeration air volume baq (hat) _n
(Hat) _n

(Equation 15) Is expressed as This correction amount is added by the adder 21 to the virtual aeration air volume baq (hat) _n .

The above is the description of the error back propagation calculation and the correction of the virtual aeration air volume realized by the calculation.

Next, learning of the multilayer neural network model will be described with reference to FIG. The learning is performed so that the prediction model 16 adapts to a gradual change in characteristics of the sewage treatment process. In this embodiment, the learning data is generated from the past dissolved oxygen concentration, sewage inflow, and aeration air volume. In FIG. 6, 26 is a database holding past observation data and aeration air volume, 27 is a subtractor, 28 is an error back propagation learning device, and 29 is an error back propagation learning device that outputs the input layer unit of the neural circuit model. , An arrow indicating that the output of the intermediate layer unit of the neural network model is to be input to the backpropagation learning device, and 31 a coupling weight between the intermediate layer and the output layer of the neural circuit model. An arrow 32 indicating that the error is to be input to the backpropagation learning device, 32 outputs the correction amount of the connection weight between the input layer and the intermediate layer calculated by the backpropagation learning device to each connection weight, and corrects each connection weight. Arrows 33 indicate that the correction amount of the connection weight between the intermediate layer and the output layer calculated by the error back propagation learning device is output to each connection weight, and that each connection weight is corrected.

For a past time lΔt, the time (l
−m _do ) Dissolved oxygen concentration (do from time ΔT to time lΔt
_{l-mdo, do l-mdo} +1, ... do l), time (l -m _iq)
Sewage inflow from Δt to time lΔt (iq _l-miq , iq
_{l-miq + 1} , ... iql), time (from l-m _bag Δt to time lΔt
Aeration volume up to (baq _l-mbaq , baq _{l-mbaq + 1} , ..., baql)
Is output as an input signal, and the dissolved oxygen concentration do _{l + 1 at} the time (l + 1) Δt is output from the database 26 to the neural circuit model as a teacher signal. In the neural network model, a forward calculation is performed on an input signal, and an output signal do _{l + 1} is output. In the subtracter 27, an error signal between the teacher signal do _{l + 1} and the output signal do _{l + 1} of the neural circuit model is obtained. The error signal is input to the error back propagation learning device 28. In addition to the error signal, the error back propagation learning unit 28 outputs the output 29 of each input layer unit of the neural network model and the output 3 of each intermediate layer unit.
0, the coupling load 31 between each intermediate layer and the output layer is input, and correction amounts 32 and 33 of the coupling load are calculated. The calculation of the correction amount in the error back propagation learning unit 28 will be described on the assumption that learning data is created from past data from the time L1Δt to the time L2Δt.

Assuming that the output of the neural network model is do (〜) _{l + 1} with respect to the input signal at time l Δt, the error function E becomes

(Equation 16) According to the backpropagation learning method, the connection weight of the neural network model and the correction amount of the threshold are calculated as follows.

[0053]

[Equation 17] It is.

In the neural network model, the connection weight and the threshold are corrected based on these correction amounts. As described above, in the control device according to the present embodiment, the virtual aeration air volume obtained by the feedback controller 13 and the observed sewage inflow volume and dissolved oxygen concentration, and the past aeration air volume, a time after a predetermined time is determined by the prediction model. The dissolved oxygen concentration can be predicted. Further, the correction amount of the virtual aeration air volume can be calculated by the aeration air volume correction amount calculator 17 from the predicted value of the dissolved oxygen concentration and the target dissolved oxygen concentration. Then, the prediction of the dissolved oxygen concentration after a predetermined time by the prediction model 16, the calculation of the virtual aeration air flow correction amount by the aeration air flow correction amount calculator 17, the correction of the virtual aeration air flow by this correction amount, and the like. Is repeated and the resulting virtual aeration air volume is input to the sewage treatment process 12 as the aeration air volume, whereby the dissolved oxygen concentration can be controlled to a target value with respect to a rapid change in the sewage inflow volume.

By using the aeration air volume calculated by the feedback controller 13 as the initial value of the virtual aeration air volume of the prediction model 16, even when the aeration air volume changes greatly, the dissolved oxygen concentration can be estimated with a small number of repetitions. The value corresponds to the target dissolved oxygen concentration.

Further, the prediction model is composed of a function approximation device having a learning function, and the learning means shown in FIG. 6 continuously learns the function approximation device using observation data, thereby adapting to a gradual characteristic change. it can.

That is, in the conventional dissolved oxygen concentration control in the aeration tank of the sewage treatment facility, feedback control is performed to determine the amount of aerated air from the error between the observed dissolved oxygen concentration in the sewage and the target dissolved oxygen concentration. However, if the amount of sewage inflow suddenly changes due to rainfall or the like, accurate control cannot be performed by feedback control, and the observer must manually adjust the aeration air volume. In addition, because the characteristics of the sewage treatment process change greatly depending on the season as well as the sudden change in the dissolved oxygen concentration due to changes in the amount of sewage inflow, the gain constant of the feedback control has to be adjusted at present. . In addition, research has been conducted to express the characteristics of sewage treatment processes in mathematical models.However, since sewage treatment processes are complex processes involving various factors that utilize the biochemical reactions of microorganisms, parameters must be adjusted. Was difficult, and good results were not obtained.

Further, for controlling the concentration of dissolved oxygen in the aeration tank of such a sewage treatment facility, Japanese Patent Application No. 3-2044 proposed by the present inventors.
Although it is conceivable to apply the invention described in Japanese Patent Application No. 94 (refer to the prior art), in the invention described in Japanese Patent Application No. 3-204494, the amount of operation when the first to third steps are repeated is considered. Because the manipulated variable before one sampling time is used for the initial value of, when a large change in the aeration air flow value is required for a sudden change in the sewage inflow, it is corrected to an accurate manipulated variable. Before being corrected, it may stop at a local solution and not be corrected. Further, even when the correction is performed up to an accurate solution, it is necessary to repeat the first to third steps many times.

Therefore, if the present invention is applied to the control of the dissolved oxygen concentration in the aeration tank of such a sewage treatment facility, it is possible to control the dissolved oxygen concentration to a constant value even when the sewage inflow rate changes rapidly. Therefore, even when the aeration airflow value changes greatly, the aeration airflow that makes the dissolved oxygen concentration a constant value can be obtained with a small number of repetitions, and the invention can be applied to a more gradual change in characteristics.

Next, the results of computer simulation of this embodiment are shown. In this simulation, neural networks with different numbers of inputs are used for the sewage treatment process and the prediction model. The simulation was performed with a sampling time of 0.25 hours, a target dissolved oxygen concentration of 1.6 (mg / l), and an error range of 0.1 (mg / l). Figure 7 shows sewage inflow (m ³ / min)
8 shows the aeration air volume (m ³ / m) output from the aeration air volume corrector.
min) and FIG. 9 shows the dissolved oxygen concentration (mg /
l), FIG. 10 shows the time change of the dissolved oxygen concentration (mg / l) output from the sewage treatment process model. The horizontal axis is time and the unit is time. From this result, it can be seen that even if the inflow of sewage changes rapidly, the dissolved oxygen concentration can be controlled without largely changing. For comparison, FIGS. 11 to 13 show control simulation results using only the feedback controller without using the aeration air volume corrector. FIG. 11 is the same as FIG. 7 of sewage inflow (m ³ / min).
FIG. 12 shows the aeration air volume (m ³ / min) output from the feedback controller, and FIG. 13 shows the dissolved oxygen concentration (mag / l) output from the sewage treatment process model. It can be seen that the vibration of the dissolved oxygen concentration is smaller in the control result using the aeration air volume corrector than in the control result using only the feedback controller. Furthermore, if the neural network model, which is a prediction model, sufficiently learns the characteristics of the sewage treatment process, it can be expected that the results will be better.

It should be noted that the present invention is not limited to the above-described embodiment, but can be variously modified without departing from the gist of the present invention. For example, in the above-described embodiment, a three-layer neural network model has been described as an example of a prediction model. However, the present invention can be similarly applied to a multilayer neural network model.
Furthermore, any function approximation model having a learning function can be similarly applied. In addition, regarding the input / output of the prediction model, in the embodiment, the dissolved oxygen concentration after the sampling time is predicted from the past and current dissolved oxygen concentration, the amount of sewage inflow, and the amount of aeration air. For example, it is conceivable to reduce the input signal or to input another observation amount. For the output signal, not only the dissolved oxygen concentration after the sampling time is predicted, but also the dissolved oxygen concentration at a further earlier time, such as the dissolved oxygen concentration after two sampling times, is predicted. The correction amount of the virtual aeration air volume can also be obtained from the error between the target and the dissolved oxygen concentration.

[0062]

According to the first aspect of the present invention, the control amount is affected by repeating the prediction by the prediction unit and the correction by the correction unit, and inputting the resulting virtual operation amount as the operation amount of the control target. even parameters change rapidly give control amount can be corrected accurately, also, the operation of the virtual calculated by the virtual operation amount-calculating means, by the initial value of the prediction means, control Even when the amount is largely changed, the control amount can be corrected by calculation with a small number of repetitions. Furthermore, by learning means
This function approximation device is
Based on the control amount and the past operation amount input to the control target.
Rapidly changing parameters by continuing learning
Not only for gradual changes in parameters
I can respond.

[0063]

According to the second, third, and fifth aspects of the present invention, the dissolved oxygen concentration can be controlled to a constant value even when the sewage inflow rate changes abruptly. It is possible to obtain the amount of aeration air to keep the dissolved oxygen concentration at a constant value with a small number of repetitions.

According to the fourth aspect of the present invention, it is possible to provide a dissolved oxygen concentration control apparatus in a sewage treatment process which can be applied to a gradual change in characteristics.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining the principle of the present invention.

FIG. 2 is a block diagram of one embodiment of the present invention.

FIG. 3 is a diagram showing an aeration air volume corrector in FIG. 1 in detail.

FIG. 4 is a flowchart for explaining a calculation procedure of the entire control device.

FIG. 5 is a diagram illustrating a three-layer neural circuit model used for a prediction model as an example.

FIG. 6 is a diagram illustrating learning of a three-layer neural network model that is a prediction model.

FIG. 7 is a diagram showing the results of computer simulation of an example for explaining the operation and effect of the present invention.

FIG. 8 is a diagram showing the results of computer simulation of the example for explaining the operation and effect of the present invention.

FIG. 9 is a diagram showing the results of computer simulation of the example for explaining the operation and effect of the present invention.

FIG. 10 is a diagram showing the results of computer simulation of the example for explaining the operation and effect of the present invention.

FIG. 11 is a diagram illustrating a result of a control simulation using only a feedback controller.

FIG. 12 is a diagram illustrating a result of a control simulation using only a feedback controller.

FIG. 13 is a diagram illustrating a result of a control simulation using only a feedback controller.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Control object, 2 ... Control device, 4 ... Virtual operation amount calculation means, 5 ... Operation amount correction means 5, 10, ... Prediction means, 11 ... Correction means, 12 ... Sewage treatment process, 13 ... Feedback controller, 14 ... Aeration air volume corrector, 15: feedback loop, 16: prediction model, 17: aeration air volume correction amount calculator.

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Ryosuke Miura 1 Toshiba-cho, Fuchu-shi, Tokyo Toshiba Corporation Fuchu Plant (56) References JP-A-7-210207 (JP, A) (58) Investigated Field (Int.Cl. ⁷ , DB name) G05B 11/00-13/04

Claims (5)

(57) [Claims] 1. A control device for controlling a control amount of a control target by changing an operation amount input to the control target, wherein a virtual amount based on a control amount measured in the control target and a target control amount. A virtual operation amount calculating means for calculating an operation amount; a control amount after a predetermined time based on a control amount measured in the control target, a virtual operation amount calculated by the virtual operation amount calculation means, and a past operation amount. Prediction means for predicting the control amount and a correction means for correcting the virtual operation amount calculated by the virtual operation amount calculation means based on the control amount after a predetermined time and the target control amount predicted by the prediction means. comprising, a manipulated variable obtained by performing at least once modified by the prediction and correction means by these prediction means in the control device including an operation amount compensation means for inputting to said control object Te, the prediction means, composed of function approximation device having a learning function
The operation amount correction unit may further include a learning unit that learns the function approximation device based on a past control amount measured in the control target and a past operation amount input to the control target. Characteristic control device. 2. An apparatus for controlling the concentration of dissolved oxygen in a sewage treatment process, which controls the concentration of dissolved oxygen in the sewage treatment process by changing the amount of aeration air input to the sewage treatment process. Virtual aeration air volume calculation means for calculating a virtual aeration air volume based on data and a target dissolved oxygen concentration; and a function approximating device having a learning function, the measurement data measured in the sewage treatment process, and the virtual aeration. Prediction means for predicting the dissolved oxygen concentration after a predetermined time based on the virtual aeration air flow calculated by the air flow calculation means and the past aeration air flow; and a dissolved oxygen concentration and a target after the predetermined time predicted by the prediction means. The virtual aeration air volume calculated by the virtual aeration air volume calculation means is corrected based on the dissolved oxygen concentration to be calculated. Correct means, and learning means for learning the function approximation device based on past measurement data measured in the sewage treatment process and the past aeration air volume input to the water treatment process. A control device for a dissolved oxygen concentration in a sewage treatment process, comprising: an aeration air volume correction device for inputting an aeration air volume obtained by performing correction by a correction device at least once to the sewage treatment process. 3. A control device for a dissolved oxygen concentration in a sewage treatment process for controlling a dissolved oxygen concentration in a sewage treatment process by changing an aeration air volume input to the sewage treatment process, wherein the sewage measured in the sewage treatment process is provided. Virtual aeration air volume calculation means for calculating a virtual aeration air volume based on the inflow amount, the dissolved oxygen concentration, and the target dissolved oxygen concentration, and a function approximation device having a learning function, and the dissolved gas measured in the sewage treatment process is provided. Prediction means for predicting the dissolved oxygen concentration after a predetermined time based on the oxygen concentration, the virtual aeration air flow calculated by the virtual aeration air flow calculation means and the past aeration air flow; Based on the dissolved oxygen concentration and the target dissolved oxygen concentration, the virtual Correction means for correcting the aeration air volume; learning to learn the function approximation device based on the past sewage inflow amount and dissolved oxygen concentration measured in the sewage treatment process and the past aeration air volume input to the water treatment process. Aeration air volume correction means for inputting the aeration air volume obtained by performing the prediction by the prediction device and the correction by the correction device at least once to the sewage treatment process. Control device for dissolved oxygen concentration. 4. The control device for a dissolved oxygen concentration in a sewage treatment process according to claim 2 or 3 , wherein the predicting means is constituted by a neural network model, and the correcting means is configured to calculate an error back propagation of the neural circuit model. A control device for a dissolved oxygen concentration in a sewage treatment process, which is realized. 5. The control apparatus of dissolved oxygen concentration in the claims 2 to 4 wherein the sewage treatment process, dissolved said aeration amount correcting means, and dissolved oxygen concentration and the target after the predetermined predicted time by the prediction means Until the error with the oxygen concentration is within a predetermined error range, or until the amount of change in the dissolved oxygen concentration after a predetermined time predicted by the prediction means is within a predetermined number of times or within a predetermined range, A control device for a dissolved oxygen concentration in a sewage treatment process, wherein a prediction by a prediction unit and a correction by the correction unit are repeated.