JP3240564B2 - Control knowledge generation method and apparatus - 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 knowledge generating method and apparatus suitable in the field of plant control.
In particular, if the automatic control system cannot achieve the expected performance unless manual correction is added due to disturbances, etc., the learning function automatically obtains the control knowledge necessary to reduce or eliminate the burden of manual correction. The present invention relates to a method and an apparatus for generating control knowledge which can be generated in a control knowledge.

[0002]

2. Description of the Related Art There are various cases in which a human being must be involved in the operation of a plant. One of the reasons is
Since the plant to be controlled is operated under various conditions, it is difficult to completely delegate control to a PLC (Programmable Logic Controller), a temperature controller, a computer, and the like.

[0003] Such a case is remarkably observed, for example, in the operation of a batch-type reactor (hereinafter, referred to as a jacket tank) frequently used in a chemical plant. The operation of the jacket tank greatly depends on the driving ability of the worker due to the difficulty of automatic control. That is,
When the polymerization reaction is carried out in a jacket tank, since the generation of heat of polymerization promotes the activation of the reaction, and thereby the generation of polymerization heat is accelerated, there is an internal positive heat feedback loop,
Controlling the reaction is not easy. In addition, various materials may be added during the manufacturing process. In such a case, there is a disturbance factor in the process in which the state of the reaction changes with the addition. However, control of the reaction is not easy.

For this reason, an intermediate parameter is used as a setting parameter of a PID controller used for automatic control of a jacket tank of this kind, and controllability is secured by manual correction of the parameter by an operator. . Under such circumstances, the psychological burden on the worker is enormous, and the quality of the product depends on the skill of the worker.

By the way, as an automatic operation method of such a jacket tank, it has been proposed to construct a model in advance for each product type and perform model reference control (for example, “System / Control / Information Vol. 35, No. 35”). No. 3 (1991) pp. 23-28 ”,“ Transactions of the Institute of Systems, Control and Information Engineers Vol.7 No.11p
p. 448-460 1994).

In this method, it is necessary to set the chemical reaction and the structure of the heat balance of the reactor, and to create a model having the desired accuracy. However, in a batch-type reactor, the number of types of products produced per reactor is on the order of several tens to several hundreds, and if all models are created, the number of steps becomes enormous. This is the model reference control,
Because the control performance depends greatly on the accuracy of the model, it is necessary to study the heat balance structure precisely for each product, and the number of model parameters also increases, and therefore, the task of estimating parameters also increases. Because it becomes. Therefore,
Although it is possible to apply model reference control to a line that manufactures a very small number of types, it was practically impossible to introduce it to a line with a large number of types.

On the other hand, Japanese Patent Application Laid-Open No. 5-127705 discloses that operation history data such as operation contents and plant values before operation are converted into operation knowledge data for outputting operation contents in response to input of plant values. Based on the operation knowledge data and the current plant values, the operation amount of the skilled operator is inferred.
A surveillance man-machine device that automates an unsteady operation by executing the method is disclosed. A learning method of converting such operation history data into operation knowledge data is applied to control of a jacket tank. For example, unlike the case of the model reference control, complicated model reference creation work may not be required. However, the water level control specifically described in the publication is that the valve is fully opened when the water amount of the tank becomes empty (0), and the valve is fully closed when the water amount of the tank becomes full (100). As described above, it is assumed that there is always a unique relationship between the operation amount and the control amount (water injection is started when the valve is opened, and water injection is stopped when the valve is closed). Therefore, it cannot be immediately applied to process control of a jacket tank or the like in which the relationship between the operation amount and the control amount may fluctuate.

[0008]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned technical background, and an object of the present invention is to provide an automatic control system having a desired automatic control system unless manual correction is added due to disturbance or the like. Provided is a control knowledge generation method and apparatus capable of automatically generating, using a learning function, control knowledge necessary for reducing or eliminating the burden of manual correction when performance cannot be exhibited. It is in.

[0009]

According to a first aspect of the present invention, there is provided a control system for controlling a plant by adding an operation amount according to the state of the plant to the plant, which is used when trial control of the plant is performed. A time-series data acquisition step of acquiring time-series data relating to an operation amount to the plant and a control amount of the plant, a control state evaluation step of evaluating a control state based on the acquired time-series data, and the control state A time-series data extraction step of extracting time-series data (DATA-g) having a good evaluation and time-series data (DATA-b) having a bad evaluation from the time-series data; The time-series data with good evaluation (DATA-g) extracted by the time-series data extraction step and the time-series data with bad evaluation (DA-g) Characterized by comprising a knowledge data generation step of generating a knowledge data for determining the manipulated variable to correct a good state control amount based on A-b) from bad.

[0010] The invention according to claim 2 of the present application provides a control system for controlling a plant by adding an operation amount according to the state of the plant to the plant. A time-series data acquisition step of acquiring time-series data related to a control amount of a plant, a control state evaluation step of evaluating the quality of a control state based on the acquired time-series data, and a good evaluation in the control state evaluation step A time-series data extraction step of extracting the obtained time-series data (DATA-g) and time-series data (DATA-b) with a bad evaluation from the time-series data; The extracted time series data with good evaluation (DATA-g) and the time series data with poor evaluation (DA
TA-b) generating knowledge data for obtaining an operation amount for correcting the control amount from a bad state to a good state based on TA-b); and A time-series data reacquisition step of controlling and reacquiring time-series data relating to the control amount and the operation amount; and evaluating a control state indicated by the reacquired time-series data, when the evaluation value is good. If it is even better than the series data (DATA-g),
Regenerating knowledge data by using the reacquired time-series data instead of the time-series data (DATA-g) which has been evaluated as good until then.

According to a third aspect of the present invention, in the first or second aspect of the invention, the knowledge data is a control amount in time-series data (DATA-g) having good evaluation. Series data and time series data with poor evaluation (D
Data of the difference between the control amount time series data in ATA-b) and the operation amount in time series data (DATA-g) with good evaluation and the operation amount in time series data (DATA-b) with poor evaluation And a fuzzy rule created based on the difference data.

According to a fourth aspect of the present invention, in the invention according to any one of the first to third aspects, the plant is a system including a jacket tank and a temperature control device. The amount is an internal temperature and an external temperature of the jacket tank, and the operation amount is an external temperature set value.

According to a fifth aspect of the present invention, there is provided a control system for controlling a plant by adding an operation amount in accordance with a state of the plant to the plant. Time series data acquisition means for acquiring time series data relating to the control amount of the plant, control state evaluation means for evaluating the quality of the control state based on the acquired time series data, and good evaluation in the control state evaluation means Time-series data extracting means for extracting the obtained time-series data (DATA-g) and the time-series data (DATA-b) having a bad evaluation from the time-series data; The extracted time-series data with good evaluation (DATA-
g) and knowledge data generating means for generating knowledge data for obtaining an operation amount for correcting the control amount from a bad state to a good state based on the time series data (DATA-b) having a bad evaluation. It is characterized by doing.

According to a sixth aspect of the present invention, there is provided a control system for controlling a plant by adding an operation amount in accordance with the state of the plant to the plant. Time series data acquisition means for acquiring time series data relating to the control amount of the plant, control state evaluation means for evaluating the quality of the control state based on the acquired time series data, and good evaluation in the control state evaluation means Time-series data extracting means for extracting the obtained time-series data (DATA-g) and the time-series data (DATA-b) having a bad evaluation from the time-series data; The extracted time-series data with good evaluation (DATA-
g) and knowledge data generating means for generating knowledge data for obtaining an operation amount for correcting a control amount from a bad state to a good state based on the time series data (DATA-b) having a bad evaluation; Time series data reacquisition means for controlling the plant using the obtained knowledge data to reacquire time series data relating to the control amount and the operation amount, and evaluating a control state indicated by the reacquired time series data. , The evaluation value of which is the time series data with good evaluation (DATA-g)
In a case where the time series data is even better, a means for regenerating the knowledge data by using the reacquired time series data instead of the time series data (DATA-g) which has been evaluated so far is provided. A control knowledge generating apparatus characterized in that:

According to the invention described in claim 7 of the present application, in the invention described in claim 5 or claim 6, the knowledge data is obtained when the control amount in the time series data (DATA-g) with good evaluation is used. Series data and time series data with poor evaluation (D
Data of the difference between the control amount time series data in ATA-b) and the operation amount in time series data (DATA-g) with good evaluation and the operation amount in time series data (DATA-b) with poor evaluation And a fuzzy rule created based on the difference data.

The invention according to claim 8 of the present application is the invention according to any one of claims 1 to 3, wherein the plant is a system comprising a jacket tank and a temperature control device. The amount is an internal temperature and an external temperature of the jacket tank, and the operation amount is an external temperature set value.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the method and apparatus of the present invention will be described below in detail with reference to the accompanying drawings. First,
First, the outline of the method of the present invention will be described together with the algorithm employed therein. In the method of the present invention, in a control system that controls the plant by adding an operation amount according to the state of the plant to the plant, first, as a first step, the operation amount to the plant when the plant is trial controlled and the plant (Step 101 in FIG. 1). For example, in the case of the above-described process control for the jacket tank, the trial control is performed until the acquired data reaches a specified number of data while manually correcting the operation amount in response to the influence of various disturbances. (FIG. 1, step 102)
reference).

Next, the quality of the control state is evaluated based on the time-series data acquired in this manner. After that, when the time-series data (DATA-g) obtained as a good evaluation and the bad evaluation are obtained, Sequence data (DATA-b) is extracted from the time-series data one by one (see step 103 in FIG. 1). FIG. 2 shows time-series data Yg of a control amount and time-series data Ug of an operation amount, which form time-series data (DATA-g) obtained as a good evaluation among a plurality of sets of time-series data obtained. And time-series data Yb of the control amount and time-series data Ub of the operation amount, which constitute the time-series data (DATA-b) with a bad evaluation. The specific evaluation method varies depending on the content of the plant to be controlled, and will be described later in detail in connection with the control of the jacket tank.

Next, the extracted time series data with good evaluation (DATA-g) and the time series data with poor evaluation (D
Based on ATA-b), knowledge data for calculating an operation amount for correcting the control amount from a bad state to a good state is created (see step 104 in FIG. 1). Here, as the above-mentioned knowledge data, it is preferable that the time-series data (Y) of the control amount in the time-series data (DATA-g) with good evaluation.
g) and the difference data (Yg-Y) between the time series data (Yb) of the control amount in the time series data (DATA-b) with poor evaluation.
b), the manipulated variable (Ug) in the time series data with good evaluation (DATA-g) and the time series data (DATA-g
b) Data (Ug-Ub) of difference from manipulated variable (Ub) in
Fuzzy rules created based on FIG. 3 shows the control amount difference data (Yg-Yb) and the operation amount difference data (Ug-Ub), respectively.

Here, obtaining knowledge data for obtaining an operation amount for correcting the above-mentioned control amount from a bad state to a good state is, in other words, a good control result (Yg, Ug).
g) and a difference component between the bad control result (Yb, Ug) and a function (hereinafter referred to as a correction function F ()) for correcting a bad control result to a good control result. () Is represented as equation (1). The operation amount (Ub) when the correction function F () and a bad control result are obtained.
(2) that a good control result (Yg) (Yg is a target trajectory) is obtained by using the equation (2). F (Yg−Yb) = Ug−Ub Equation (1) Yg = PLANT (Ug) = PLANT {Ub + F (Yg−Yb)} Equation (2) Each of the above equations (1) and (2) The meanings of the symbols are as follows. Yg: Control amount time series data with good control result Yb: Control amount time series data with poor control result Ug: Operation amount time series data with good control result Ub: Operation amount time series data with poor control result PLANT ( U): Time series data of a response when the time series (U) of the manipulated variable is given to the control target Yg = PLANT (Ug) which is the first equation of the equation (2)
Means that if the control object does not change at all, the highest level of control result can always be obtained if Ug is given. Further, it is considered that the correction function F () has the ability to correct a bad control result to a good control result. further,
When the control target changes and the response to Ug shifts, it is expected that the operation amount Ug is corrected so as to approach a better control result by the correction function F ().

Thereafter, the next production is performed by controlling the plant using the obtained knowledge data (see step 105 in FIG. 1). Here, the control amount (Yg) obtained when a good control result is obtained is corrected to a better waveform, a correction function (F ()) using the corrected waveform (Yg ′) as a target trajectory, and a good function. Obtaining a better control result using the manipulated variable (Ug) at the time of obtaining a proper control result is represented by Expression (3). Yn = PLANT (Ug + F (Yg′−Yn)) ≦ Yg Equation (3) The meaning of each symbol in the above equation (3) is as follows. ≦: The left side has the same or better control result than the right side. Yn: Control amount obtained this time Yg ′: Yg modified to have better controllability In the above equation (3), a good control result ( Yg),
When the control result is changed to a better control result (Yg ′), if the degree of the change is small, Yg and Yg ′ are calculated by the correction function F ().
It is expected that an intermediate performance of will be obtained. Next, an example in which the control knowledge generation method and device according to the present invention are applied to a jacket tank temperature control system will be described.

In the hardware configuration diagram shown in FIG. 4, a jacket tank (A) is a double-structured container that generates a chemical reaction, and circulates a heat medium through an outer shell to adjust the temperature of the inner shell. It has been made like that. In the following description, the temperature in the inner shell is [inner temperature], and the temperature in the outer shell is [outer temperature].
Shall be called. The motor (B) serves as a drive source for stirring the reactant in the inner shell with a stirring fan.

The existing system (C) is a general term for existing temperature control systems attached to the jacket tank, and these are generally DCS (Distributed Control System).
Is constructed by The existing system (C) has a controller 1 (D) for controlling the internal temperature and a controller 2 (E) for controlling the external temperature.
And a process management unit (F). These functions are realized by software on the DCS.

Here, the controller 1 (D) for controlling the internal temperature includes:
It has a function of calculating a set value of the outside temperature from the measured value of the inside temperature and the set value of the inside temperature, and outputting this to the corrector [G] for the set value of the outside temperature. Also, a controller 2 (E) for adjusting the outside temperature
Calculates the opening of the steam and cooling water valves from the measured value of the outside temperature and the set value of the outside temperature. Further, the process management section (F) operates information related to the product type such as a product type, a lot number, and a production amount.

The external temperature set value corrector (G) is a part in which the control knowledge control method and apparatus according to the present invention are realized, and is constituted by a personal computer, a PLC or the like in terms of hardware. FIG. 5 shows the functional configuration of the external temperature set value corrector (G).
Shown in As shown in the figure, a data acquisition function (G1), a control knowledge creation function (G2), an automatic control function (G3), and control data (G) are provided in the external temperature set value corrector (G). A database (DB) for storing the operation amount and the control amount time-series data obtained by executing the control and the control knowledge for each product type is included. Data acquisition function (G1)
Is for acquiring control data from an existing temperature control system. The control knowledge creation function (G2) includes four functions including a target performance setting function, a performance evaluation function, a knowledge creation function, and a knowledge correction function.

The target performance setting function is for the user to set a target value and a gain for each control index, and may be registered in advance as fixed values, and is shown in FIG. You may make it input from a setting screen. The performance evaluation function is for calculating a control index of the acquired control data. The knowledge creation function is for creating control knowledge using the acquired control data. The knowledge correction function is for updating control knowledge using control data obtained by controlling a plant using the created control knowledge.

The database (DB) is for holding the acquired control data and the created control knowledge. The automatic control function (G3) calculates the set value of the outside temperature with reference to the created control knowledge. Specifically, FIG. 16 shows the configuration thereof.

Next, the functions (G1) to (G3) constituting the external temperature set value corrector (G) described above will be described in more detail. FIG. 7 shows an example of data acquired by the data acquisition function (G1), and FIG.
Are shown in tables. As is clear from these figures, the data acquisition function (G1) uses the set value of the internal temperature, the set value of the external temperature obtained from the existing control system, the current value of the internal temperature, the current value of the external temperature, and the type of the variety. 10 kinds of data including lot identification, production quantity, production time (unit: time), contact signal indicating production start / end state, and external temperature set value calculated by external temperature set value corrector We are handling,
These data are input / output at a specified timing, and each has a specified domain.

In the performance evaluation function (G2), the following 7
The performance is evaluated using the three indices. That is, in the performance evaluation function (G2), a value obtained by linearly combining these seven indices is used as an evaluation value (performance evaluation value). The smaller the evaluation value, the better the control result is determined. Overshoot (OVER) This is the difference between the maximum value of the control amount and the control target value in a section from when the control amount first reaches the control target value to when the slope becomes negative for the first time. The initial value is -1.0. Undershoot amount (UNDER) This is the difference between the minimum value of the control amount and the control target value in a section from the end of the section for calculating the overshoot amount until the slope becomes first positive. The initial value is the control target value x -1 The maximum value at the time of settling (MAX) The difference between the maximum value of the control amount and the control target value in the section from the end of the calculation section of the undershoot amount until the control amount target value changes is there. Setting minimum value (MIN) This is the difference between the control value minimum value and the control target value in a section from the time when the setting maximum value is first calculated to the time when the control amount target value changes. Hunting frequency (HUMP) This is the number of times that the upper / lower limit value at the time of stabilization is exceeded in the section from the end of the undershoot amount calculation section to the change of the control amount target value. Error area ratio (ERROR) After the control amount reaches the defined ratio (START) of the control target value,
In the section up to the end of the control, it is calculated from the following equation. Error area ratio = Σ (control amount-control target value) ² / section time Rise time (STAND) After the control amount reaches a predetermined value (predefined: constant), until the control amount reaches the control target value for the first time It's time. The performance evaluation value is calculated by the following equation (4). Performance evaluation value = Σ (target value−index value) ² × gain (Equation (4)) The target value and gain for each index are set in advance. The index that could not be calculated is assumed to be [0] and the equation (4) is calculated.

The control knowledge creation function (G2) creates control knowledge in the following procedure. The one with the smallest performance evaluation value (DATA-g) is selected. The performance evaluation value at that time is defined as α. The performance evaluation value is α + δ or more, and the minimum value (DA
Select TA-b). Here, δ is a constant and is set in advance. The reason why the performance evaluation value with the largest performance evaluation value is not selected as DATA-b is that data in a state where an error has occurred in the system including the plant is not registered.
If the system is in the normal range and has poorer control than the best control, but not too bad
Register as DATA-b. By doing so, the correction of the operation amount by the control knowledge becomes appropriate. FF knowledge = U _DATA-g . When calculating the FB knowledge, the correction function F () is calculated according to the following procedure. i) Calculate the difference component β. β = Yg−Yb OUT = Ug−Ub ii) Differentiation of β (dβ), integral (iβ), second derivative (dd)
β) Calculate the component. Note that β, dβ, d
dβ, iβ, and teacher data consisting of OUT as an output signal. That is, the input to the correction function F () is β, d
β, ddβ, iβ, and the output is Ug-Ub. iii) The correction function F () is described by fuzzy rule inference according to the following procedure. It is assumed that the number of MFs (odd number) for input / output signals is set in advance. 1) Acquire maximum / minimum values from input / output data. 2) The MF is defined by equally dividing the range between the maximum value and the minimum value by the defined number of MFs. 3) Calculate the degree of coincidence between the input / output data and the corresponding MF. 4) A set of MFs with the highest degree of coincidence is created as a rule. The product of the degree of coincidence is registered as the weight of the rule. 5) MFs with the same degree of coincidence are smaller MF when data is close to the maximum value, and larger MF when data is close to the minimum value.
Is adopted. 6) Execute 3, 4, and 5 for all input / output data. 7) If a rule with the same antecedent / consequent part occurs, the weight is updated to the one with the greater degree of coincidence. 8) If a rule with the same antecedent occurs, delete the rule with the smaller weight. 9) After the creation is completed, change the MF width of the consequent part to 0
(Singleton).

As a specific example, a case of two inputs and one output in which the deviation (β) and the deviation derivative (dβ) are input and the operation amount difference (Ug-Ub) is output will be described below. Input: x1 Defined MF number: 3 Maximum value: 10 Minimum value: -10 x2 Defined MF number: 3 Maximum value: 10 Minimum value: 0 Output: y Definition MF number: 3 Maximum value: 5 Minimum value: -5

The MF defined in this case is shown in FIG. 9 and FIG.
0 is shown. Input / output data A: x1 = 7.50, x2 = 5.00, y = 5.00 → A11 (x1) = 0.00, A12 (x1) = 0.25, A13 (x1) = 0.75 A21 (x2) = 0.00, A22 (x2) = 1.00 , A23 (x2) = 0.00 Ay1 (y) = 0.00, Ay2 (y) = 0.00, Ay3 (y) = 1.00 → IF x1 = A13 & x2 = A22 THEN y = Ay3 Weight = A13 (x1) ・ A22 (x2 Ay3 (y) = 0.75 B: x1 = 7.50, x2 = 2.50, y = 1.25 → A11 (x1) = 0.00, A12 (x1) = 0.25, A13 (x1) = 0.75 A21 (x2) = 0.50, A22 (x2) = 0.50, A23 (x2) = 0.00 Ay1 (y) = 0.00, Ay2 (y) = 0.75, Ay3 (y) = 0.25 → IF x1 = A13 & x2 = A22 THEN y = Ay2 Weight = A13 (x1 ) .A22 (x2) .Ay2 (y) = 0.28125 Rule A created with input / output data A and rule B created with B have the same antecedent part. In this case, since the weight of the rule A is larger, the rule B is deleted.

As another specific example, a deviation (β), a deviation derivative (dβ), a deviation second derivative (ddβ), and a deviation integral (iβ) are input, and an operation amount difference (Ug−Ub) is output. The data structure of the fuzzy rule in the case of four inputs and one output is shown in FIG. 11, and examples of the rules are shown in FIGS.

In the data structure shown in FIG. 11, the numeral "10" following "; rule definition" is the number of rules, and the symbol string "N: N: *: P; OUT1" following it.
Means IF IN0 = N and IN1 = N and 1IN2 = * and IN3 = P then OUT = OUT1. However, "*" indicates Don't Care
Means

Next, the knowledge correction function will be described. The correction of the knowledge is performed in accordance with the following two types of correction methods (step 150) according to the control performance of the control result obtained using the knowledge.
5 or step 1506) (FIG. 15)
reference).

That is, when the process is started in FIG. 15, the plant is controlled using the generated knowledge data, and accordingly, the time series data relating to the control amount and the operation amount is reacquired (step 1501). ). Thereafter, the control state indicated by the reacquired time-series data is evaluated (step 1502). Here, if the current evaluation result is better than the specifications required for the plant (step 1503 YES), the shift to the next production is performed as it is (step 150).
7). On the other hand, if the current evaluation result is worse than the required specification (NO in step 1503), and if the obtained control performance is worse than the control performance of the past best data (NO in step 1504), the obtained control data FB knowledge is created from among the control knowledge using the best control data in the past (step 1506). On the other hand, if the obtained control performance is equivalent to the control performance of the best data used at the time of creating the knowledge (YES in step 1504), the control amount data of the obtained control data is changed so as to further improve the accuracy. (Step 1505). Thereafter, the changed data is used as a reference trajectory for the next control. Note that the above-described changing method is performed by detecting a portion where the control amount reaches the target value for the first time, and correcting the hunting component in the settling range according to Expression (5). Y1 = Control amount + Deviation × θ Expression (5)

The automatic control function (G3) has the structure shown in FIG.
F) and a feedback section (FB). FIG. 17 shows a processing flow of this automatic control function.

In the figure, when the process is started, after the process of setting the type (step 1701) and the process of reading control knowledge (step 1702), the sensor signal is read (step 1703). It waits for the activation signal to be turned on (step 1704). In this state, when the start signal is turned on (step 17).
04YES), FF operation amount read processing (step 1)
705), target (Y REF) read processing (step 1)
706) and FB operation amount calculation processing (step 170)
After performing 7) in order, an operation amount (= FF operation amount + FB operation amount) is obtained based on them (Step 1707), and is applied to the plant (Step 1709). Thereafter, the processing of the abnormality (steps 1705 to 1709) is repeated.

[0039]

According to the present invention, when the automatic control system cannot exhibit the expected performance unless manual correction is added due to disturbance or the like, the control required to reduce or eliminate the burden of manual correction is eliminated. Knowledge can be automatically generated using a learning function.

[Brief description of the drawings]

FIG. 1 is a flowchart showing an outline of a control knowledge creation process.

FIG. 2 is a graph showing a relationship between a control result and an operation amount.

FIG. 3 is a graph showing a difference sentence between a case where a control state is good and a case where a control state is bad for each of a control amount and an operation amount.

FIG. 4 is a block diagram illustrating a hardware configuration of a jacket tank temperature control system.

FIG. 5 is a block diagram showing a functional configuration of an external temperature set value corrector.

FIG. 6 is a diagram showing an example of a setting screen for setting a target performance.

FIG. 7 is a table showing a list of input / output signals in a data acquisition function.

FIG. 8 is a table showing a list of signal variations in a data acquisition function.

FIG. 9 is a graph showing an initial shape of a defined membership function.

FIG. 10 is a graph showing the final form of a defined membership function.

FIG. 11 is a diagram illustrating a data structure of a correction function of FB knowledge.

FIG. 12 is a diagram showing an example of a fuzzy rule constituting control knowledge.

FIG. 13 is a graph illustrating the meaning of the rule shown in FIG.

FIG. 14 is a graph illustrating the meaning of the rule shown in FIG.

FIG. 15 is a flowchart showing a process at the time of operation using control knowledge.

FIG. 16 is a block diagram illustrating a structure of an automatic control unit.

FIG. 17 is a flowchart showing the contents of an automatic control function.

[Explanation of symbols]

 A Jacket tank B Motor C Existing system D Controller for internal temperature control 1 E Controller for external temperature control 2 F Process control section G Corrector for external temperature setting device G1 Data acquisition function G2 Control knowledge creation function G3 Automatic control Function DB database

Continuation of the front page (56) References JP-A-5-127705 (JP, A) JP-A-6-214611 (JP, A) JP-A-1-217601 (JP, A) JP-A-7-210585 (JP) , A) JP-A-62-42202 (JP, A) JP-A-4-302305 (JP, A) JP-A-4-303202 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB G05B 13/00-13/04 G05B 9/02 G05B 23/02 G06F 9/44 554

Claims (8)

(57) [Claims] 1. A control system for controlling a plant by adding an operation amount according to the state of the plant to the plant, wherein time series data on the operation amount to the plant and the control amount of the plant when trial-controlling the plant is acquired. Time series data acquisition step, a control state evaluation step of evaluating the quality of the control state based on the acquired time series data, and a time series data (DATA-g) obtained in the control state evaluation step. ) And time-series data (DATA-b) with bad evaluation obtained from the time-series data, and time-series data with good evaluation extracted by the time-series data extraction step ( Based on DATA-g) and the time series data (DATA-b) with poor evaluation, the control amount is changed from bad to good. Control knowledge generation method characterized by comprising: a knowledge data generation step of generating a knowledge data for determining the manipulated variable to correct the. 2. A control system for controlling a plant by adding an operation amount according to the state of the plant to the plant, wherein time series data on the operation amount to the plant and the control amount of the plant when the plant is trial-controlled are acquired. Time series data acquisition step, a control state evaluation step of evaluating the quality of the control state based on the acquired time series data, and a time series data (DATA-g) obtained in the control state evaluation step. ) And time-series data (DATA-b) with bad evaluation obtained from the time-series data, and time-series data with good evaluation extracted by the time-series data extraction step ( Based on DATA-g) and the time series data (DATA-b) with poor evaluation, the control amount is changed from bad to good. A knowledge data generating step of generating knowledge data for obtaining an operation amount for correcting the operation amount, and re-acquiring time-series data relating to the control amount and the operation amount by controlling the plant using the generated knowledge data. A time-series data reacquisition step, and evaluating a control state indicated by the reacquired time-series data, wherein the evaluation value is a time-series data (DA
TA-g), regenerating the knowledge data by using the re-acquired time-series data instead of the time-series data (DATA-g) which has been evaluated as good until then. A control knowledge generating method, comprising: 3. The knowledge data includes time-series data of a control amount in time-series data (DATA-g) with good evaluation and time-series data of control amount in time-series data (DATA-b) with poor evaluation. Fuzzy data created based on the difference data between the operation amount in the time series data with good evaluation (DATA-g) and the operation amount in the time series data (DATA-b) with poor evaluation. 3. The control knowledge generation method according to claim 1, wherein the control knowledge is a rule. 4. The system according to claim 1, wherein the plant is a system including a jacket tank and a temperature control device, wherein the controlled variable is an internal temperature and an external temperature of the jacket tank, and the manipulated variable is an external temperature set value. The control knowledge generation method according to claim 1, wherein: 5. A control system for controlling a plant by adding an operation amount according to the state of the plant to the plant, wherein time series data on the operation amount to the plant and the control amount of the plant when the plant is trial-controlled are acquired. Time-series data acquisition means, control-state evaluation means for evaluating the quality of the control state based on the acquired time-series data, and time-series data (DATA-g) obtained by the control state evaluation means. ) And time-series data (DATA-b) obtained with a bad evaluation from the time-series data, and time-series data with good evaluation extracted by the time-series data extraction means (DATA-b). An operation for correcting the control amount from a bad state to a good state based on DATA-g) and time-series data (DATA-b) with poor evaluation. Control knowledge generating apparatus characterized by comprising a knowledge data generating means for generating a knowledge data for determining the amount of a. 6. A control system for controlling a plant by adding an operation amount according to a state of the plant to the plant, wherein time series data relating to the operation amount to the plant and the control amount of the plant when the plant is trial-controlled are acquired. Time-series data acquisition means, control-state evaluation means for evaluating the quality of the control state based on the acquired time-series data, and time-series data (DATA-g) obtained by the control state evaluation means. ) And time-series data (DATA-b) obtained with a bad evaluation from the time-series data, and time-series data with good evaluation extracted by the time-series data extraction means (DATA-b). An operation for correcting the control amount from a bad state to a good state based on DATA-g) and time-series data (DATA-b) with poor evaluation. Knowledge data generating means for generating knowledge data for obtaining an amount; time-series data re-acquiring means for controlling the plant using the generated knowledge data to re-acquire time-series data relating to a control amount and an operation amount And the control state indicated by the reacquired time-series data is evaluated, and the evaluation value is evaluated as the time-series data (DA
TA-g), means for regenerating knowledge data by using the re-acquired time-series data instead of the time-series data (DATA-g) which had been evaluated as good until then. A control knowledge generation device, comprising: 7. The knowledge data includes time series data of a control amount in time series data with good evaluation (DATA-g) and time series data of control amount in time series data (DATA-b) with poor evaluation. Fuzzy data created based on the difference data between the operation amount in the time series data with good evaluation (DATA-g) and the operation amount in the time series data (DATA-b) with poor evaluation. The control knowledge generating device according to claim 5, wherein the control knowledge generating device is a rule. 8. The system according to claim 1, wherein the plant is a system including a jacket tank and a temperature control device, wherein the controlled variable is an internal temperature and an external temperature of the jacket tank, and the manipulated variable is an external temperature set value. The control knowledge generation device according to any one of claims 5 to 7, characterized in that: