JP2001242905A - Method and device for process controlling - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and a device for a process controlling, by which the model prediction control of a multivariable process can be executed with high accuracy of process control. SOLUTION: An estimated control value is obtained by using a process model 11 having plural control variables and operation variables. A revised value of operation value is obtained bared on the combination of an estimated deviation which is a deviation of estimated control value from desired value, and plural operation variables. At that time, combination of plural operation variable or plural control variable and operation variable are changed to obtain the changed revised value of operation value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a process control method and apparatus for optimally controlling a controlled object in a multivariable process while predicting a future behavior using a mathematical model.

[0002]

2. Description of the Related Art In recent years, in the field of process control such as petrochemicals, model predictive control is often employed in order to realize optimal control while satisfying restrictions on operation control of a controlled object. In this model predictive control method, the future behavior of a controlled object is predicted using a prediction formula based on a dynamic characteristic model of a plant, and a deviation between the predicted result and a target value, and an evaluation function for a future change of a manipulated variable are set. Then, an optimal operation amount for optimizing the evaluation function is sequentially calculated for each control cycle. This control method is described, for example, in “Transactions of the Institute of Electrical Engineers of Japan C”, Vol. 116, No. 10 (1996), No. 1089.
Page 10 to page 1096.

[0003] In the model predictive control method, an operation amount at the current time is determined so that a future control amount value approaches a target value under a constant operation amount as much as possible. The prediction formula is expressed by a mathematical expression related to the past control amount and operation amount. The prediction calculation is executed each time based on the past control amount and operation amount so that the future control amount value approaches the target value.

An evaluation function for model predictive control is set based on the prediction calculation result. Often this merit function is
It is expressed in the form of a sum of squares or a sum of absolute values of the “deviation between the control amount predicted value and the target value” and the “manipulated variable update value”. For the evaluation function, an optimal solution is obtained by applying a numerical optimization algorithm. This optimal solution corresponds to the manipulated variable update value, and the optimal manipulated variable is determined.

[0005] Restrictions related to a controlled object are used as a search range of an optimum solution in an optimization calculation.
Note that a linear programming method, a quadratic programming method, or the like is known as an optimization algorithm, and an optimal operation amount update value can be derived by applying these algorithms.

[0006]

It can be said that model predictive control is a suitable control method when a petrochemical process or the like is to be controlled. However, a petrochemical process to be controlled is a multivariable process and is a large-scale process. The model predictive control is a multivariable control and has many control variables and manipulated variables. The control variables and the operation variables used for obtaining one operation amount update value are determined in advance by simulation evaluation in the design stage.

However, in the field of petrochemical processes and the like, the technical causal relationship is often unknown, and it is often impossible to obtain the expected process control accuracy if control variables and manipulated variables are determined by simulation evaluation. are doing.

Further, during the actual operation of the process, the process equipment may be changed to manual control according to the operating conditions.
Also in this case, if the control variables and the operation variables are determined by the simulation evaluation, the expected control accuracy of the process cannot be obtained.

In the prior art, the control variables and the operation variables are determined by simulation evaluation as described above.
There is a problem that it often happens that the expected process control accuracy cannot be obtained.

An object of the present invention is to provide a process control method and apparatus capable of performing model predictive control of a multivariable process with high process control accuracy.

[0011]

A feature of the present invention is that a predicted value of a controlled variable is obtained by using a model of a process having a plurality of control variables and manipulated variables, and a deviation between the predicted value of the controlled variable and a target value is obtained. In order to determine the operation amount update value based on a plurality of combinations of the prediction deviation and the operation variable, the combination of the plurality of operation variables is changed.

Another feature of the present invention is that a predicted value of a controlled variable is obtained by using a dynamic characteristic model of a process having a plurality of control variables and manipulated variables, and a deviation between the predicted value of the controlled variable and a target value. Change the combination of multiple control variables and operation variables required when calculating the operation amount update value using the prediction deviation and the evaluation function obtained using the multiple control variables and operation variables, and I asked for it.

According to the present invention, when obtaining a manipulated variable update value based on a plurality of combinations of a predicted deviation, which is a deviation between a control variable predicted value and a target value, and a manipulated variable, a combination of a plurality of manipulated variables or a plurality of manipulated variables is determined. Since the setting of the combination of the control variable and the manipulated variable is changed, the model predictive control of the multivariable process can be performed with high process control accuracy.

[0014]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an embodiment of the present invention. In FIG. 1, the process control system includes a process control device 1, a control target process 2, and an input unit 3. The process control device 1 includes a model database 11, a prediction calculation unit 12, an evaluation function setting unit 13, an operation amount optimization unit 1,
4, variable setting means 15 and target value / constraint condition setting means 16.

The model database 11 stores model data including a mathematical model representing the dynamic characteristics of the process 2 to be controlled. The predicted value calculation means 12 reads the process data S11 of the control target process 2, calculates a future change of the control target process 2 using the model data of the model database 11, and outputs a free response f described later as a prediction result.

The evaluation function setting means 13 derives an evaluation function relating to the optimal operation amount in the model predictive control based on the prediction result of the prediction value calculation means 12. Further, the evaluation function setting means 13 is a function for setting the variable setting means 1 inputted by the input means 4.
The control variables and operation variables of No. 5 and the target values and constraint values in the plant operation of the target value / constraint condition setting means 16 are set when the evaluation function is constructed.

The operation amount optimizing means 14 receives the evaluation function determined by the evaluation function setting means 13 and determines an optimal operation amount by numerical optimization.

The process control device 1 outputs the determined operation amount to the control target process 2 as a control signal S12, and executes automatic control.

An example of the control target process 2 will be described with reference to FIG. FIG. 2 illustrates an example in which the control target process 2 is a petrochemical process.

The process 2 to be controlled includes a distillation column 21 as a chemical process device and a thermometer 21 as a process measuring device.
0, flow meters 22, 23, and control valves 221, 23
1 is comprised. Although the example of FIG. 2 has a simple configuration to facilitate understanding of the present invention and to simplify the description, the present invention can be used even in a complicated and large-scale process to be controlled. .

In the process shown in FIG. 2, a raw material such as ethylene is supplied to the distillation column 21 through a control valve 231, and steam is supplied through a control valve 221 for purification. The internal temperature of the distillation column 21 is measured with a thermometer 2
10, the temperature S210 is measured, and the temperature measurement value S
210 is a control variable.

The steam flow rate S22 measured by the flow meter 22
0 and the operation amount S221 of the opening degree of the control valve 221 are the operation variables, and the raw material flow rate S230 measured by the flowmeter 23.
And the opening operation amount S231 of the adjustment valve 231 is an operation variable.

FIG. 3 shows an example of the input means 3. The input unit 3 includes a display device having an operation screen 31 and an input device 32. In FIG. 3A, the operation screen 31 is
A plurality of pieces of process data related to the control target 2 are displayed. As the display contents, “variable name”, “process amount”, and “variable setting” indicating whether each variable is in the execution state (ON) or the release state (OFF) are displayed. The execution state (ON) of a variable means that the variable is used for model predictive control, and the cancellation state (OFF) means that the variable is not used.

The input device 32 is a keyboard, a mouse or the like, and includes a touch operation on the screen. FIG.
(B) shows an example of the display content of the operation screen 31 when a specific variable is released from the input unit 3. FIG. 3C shows an example of a display content of a screen for setting a target value and a constraint condition in the input means 3. As the display content, the target value, the upper limit value and the lower limit value of the constraint condition are displayed together with the "variable name". Next, the operation of the present invention will be described.

FIGS. 4A and 4B correspond to the temperature in the tower and the steam amount in the process 2 to be controlled. FIG. 4A shows the trend of the control variable (temperature in the tower), and FIG. Quantity) is the trend. In the trend in FIG. 4, the vertical axis indicates the magnitude of the process amount, and the horizontal axis indicates time. On the time axis, the time at which the manipulated variable calculation is being performed is defined as the current time, and the trend to the right from this indicates the predicted trend of the process. Dotted lines in the graph represent the upper limit and lower limit of the operation constraint condition of the control variables and operation variables of the process.

The predicted value calculating means 12 predicts and derives a future change of the controlled variable (temperature in the tower) of the process 2 using a mathematical model representing the dynamic characteristic of the controlled process 2 stored in the model database 11. . The mathematical expression model represents the correlation of the temporal change between the manipulated variable and the disturbance variable of the process 2 and the control variable, and generally corresponds to a time series model or the like. In the prediction calculation of the control amount change, the past process data based on the measurement data S11 of the process 2 is substituted into the mathematical model, and the control amount change from the measured time is obtained.

The future change of the manipulated variable is determined by the manipulated variable optimizing means 14 described later.
A control variable change is predicted on the assumption that a future manipulated variable will maintain a constant value from the current time. This prediction result is called the free response f of the process 2, and corresponds to the control amount change 400 in FIG. The free response 400 is a prediction result of a control amount change when the operation amount is kept constant from the current time. The free response 400 is obtained from the prediction calculation means 12 to the evaluation function setting means 13
Is output to

The evaluation function setting means 13 derives an evaluation function for determining an operation amount update value in the model predictive control. Although there are various setting methods for the evaluation function, generally, a variable for optimizing the operation amount update value for the control target process 2 is represented by (control amount prediction value−target value) ² + (operation amount update value) ^2. .

The evaluation function setting means 13 calculates the manipulated variable update value 40 before the current time based on the free response 400 shown in FIG.
Then, a control amount future change 401 for 2 is derived. As an example of the derivation method, it is obtained by the calculation formula shown in Expression 1.

[0031]

Y = G × Δu + f (Expression 1) In Expression 1, f represents a free response 400, and Δu represents an operation amount update value (correction amount). G represents time response change rate data indicating the degree of influence of a change in the operation amount on a change in the control amount by a time change rate. As the time response change rate, generally, a step response change, an impulse response change, or the like is used. As shown in Equation 1, the control amount future change y is equal to the correction amount Δ
u can be expressed as an unknown variable.

The free response 400 is compared with the control amount target value over the prediction evaluation section shown in FIG. This deviation is called a prediction deviation, and in the model prediction control, the control purpose is to find an operation amount update value that minimizes the prediction deviation.
Here, the prediction evaluation section is a section in which the prediction deviation is evaluated in one operation amount calculation.

Next, as a feature of the process control, it is desirable that the change in the manipulated variable is small for the process 2 to be controlled. In FIG. 4, a section in which an update value is determined in one operation amount calculation is set as a prediction control section.

Since the evaluation function takes a quadratic form, it corresponds to the area sum corresponding to each section of the predicted deviation and the manipulated variable update value. In FIG. 4, the shaded portions correspond to the control variable trend and the operation variable trend. These areas use the manipulated variable update value Δu as a variable as described above. From the above, when the evaluation function is expressed using Expression 1, Expression 2 is obtained. here,
Let J be the evaluation function and r be the target value.

[0035]

J = (y−r) ² + (Δu) ² = (G × Δu + fr) ² + (Δu) ² (Expression 2) As shown in Expression 2, the evaluation function setting means 13 operates An evaluation function with the quantity update value Δu as an unknown variable is set.

The setting of the evaluation function will be specifically described with reference to the process 2 to be controlled shown in FIG. 2 as an example. However, the section length of each of the prediction evaluation section and the prediction control section is one step.

First, the time response change rate data G and the manipulated variable update value Δu can be set as follows.

[0038]

G = [GA GB] (Equation 3)

[0039]

Δu = [ΔuA ΔuB] (Equation 4) Here, the suffixes A and B correspond to the operation variables A and B in FIG. 2, respectively. From this, the prediction formula of the control variable is as follows. In the following equation, “′” means transposition of a vector.

[0040]

Y = G × Δu + f = [GA GB] [ΔuA ΔuB] ′ + f = GA × ΔuA + GB × ΔuB + f (Equation 5) From Equation 5, an example of the evaluation function set by the evaluation function setting means 13 is as follows. Shown in

[0041]

J = (y−r) ² + (Δu) ² = (GA × ΔuA + GB × ΔuB + fr) ² + (ΔuA) ² + (ΔuB) ² (Equation 6) The evaluation function set by the evaluation function setting means 13 can be expressed in a quadratic form (quadratic function) relating to the operation amount update values of the operation variables A and B. The evaluation function setting means 13 outputs coefficient data (matrix data) of the evaluation function to the manipulated variable optimization means 14.

The evaluation function setting means 13 may set constraints on the process operation in some cases. In process control,
Generally, an upper limit value and a lower limit value are often set for a control variable, an operation variable, and an operation amount update value, respectively, as shown in FIG. In this case, a mathematical expression using the manipulated variable update value as an unknown variable is compared with the constraint condition data, and the inequality condition can be described as shown below.

(Constraint lower limit) <(control amount predicted value) <(constraint upper limit) (constraint lower limit) <(operation amount future value) <(constraint upper limit) (update value lower limit) <(operation amount update) Value) <(upper limit of update value) These inequalities are used as a solution search range. These target values and constraint conditions are input to the target value and constraint condition setting means 16 by the input means 3 using information stored in a process database (not shown).

The manipulated variable optimizing means 14 receives the evaluation function J and the inequalities relating to the constraint conditions, and determines an optimal manipulated variable update value Δu for each control cycle using a numerical optimization algorithm. As the numerical optimization algorithm, a linear programming, a quadratic programming, or the like can be applied according to the mathematical expression of the evaluation function. In the search for the solution, for example, an operation amount update value Δu that minimizes the above-described area sum shown in FIG. 4 is derived.

The manipulated variable optimizing means 14 monitors the calculation load related to the solution search in the numerical optimization algorithm. As the calculation load, an index that can be easily monitored, such as the number of repetitive operations related to the above-described solution search and the occupancy of a computer, is used. Further, the manipulated variable optimization unit 14 monitors the above-mentioned calculation load and determines whether or not an optimal solution exists. For the determination of the existence of the optimal solution, the above-described algorithms of the linear programming and the quadratic programming generally have determination parameters, which can be used.

The variable setting means 15 includes the evaluation function setting means 1
3 manages the time-response change rate data G shown in the equation (1). The input means 3 outputs an execution cancellation command for each variable determined by the operator or another control program to the variable setting means 15, and manages the state of control execution cancellation for each variable.

An example of a method of managing the time response change rate data in the variable setting means 15 will be described with reference to FIG.

Since the control target process 2 is a multivariable process, the time response change rate data G is set for each control variable and each operation variable. FIG. 5 shows an example of the data configuration. This example shows a case where the number of control variables is cvs and the number of operation variables is mvs.

When all the control variables and operation variables are in the execution state, the time response change rate data G is as shown in FIG. On the other hand, when either the control variable or the operation variable is released by the input means 3, the time response change rate data G
Changes the data configuration by the variable setting means 15. That is, in response to a variable release command from the input means 3, the corresponding data is changed to "0" as shown in FIG. 5B. FIG. 5B shows, as an example, a configuration change result of the time-response change rate data G when a release command of the second registered operation variable is input. The configuration change result of the time response change rate data G is reflected on the evaluation function setting means 13.

How the evaluation function setting means 13 operates according to the configuration change result of the time response change rate data G will be described using the evaluation function of Expression 6 as an example.

First, for ease of understanding, the above equation (6) is again shown below.

[0052]

J = (y−r) ² + (Δu) ² = (GA × ΔuA + GB × ΔuB + fr) ² + (ΔuA) ² + (ΔuB) ² (Equation 6) Here, the operation variable A is released. Here is an example. In the above equation 3, GA = 0 by the variable setting means 15.

[0053]

G = [0 GB] (Expression 7) As a result, Expression 6 is set in the evaluation function setting means 13 as follows.

[0054]

J = (0 × ΔuA + GB × ΔuB + f−r) ² + (ΔuA) ² + (ΔuB) ² = (GB × ΔuB + fr) ² + (ΔuA) ² + (ΔuB) ² (Equation 8) From equation 8, the predictive calculation y of the control variable in the evaluation function J
In contrast, there is no influence of ΔuA, and only the square value of ΔuA itself remains in the evaluation function. On the other hand, in the operation amount optimizing means 14, ΔuA that minimizes the evaluation function J,
Since the operation is performed to obtain ΔuB, the optimal solution of the update value ΔuA of the operation variable A is 0.

Since ΔuA = 0, as a result, the operation variable A is not updated, and the release state can be realized. In addition, to set the operation variable A to the execution state again, it can be realized by returning GA to the initial setting value in Expression 7.

FIG. 6 is a flowchart of a calculation process for each control cycle for a series of calculation operations related to the execution / cancellation of control variables and manipulated variables as described above. FIG. 6 shows the calculation processing at the k-th time.

First, following the start 600 of the k-th time of the control cycle, the variable setting means 15 checks the command of the operation variable set by the input means 3 of the process control device 1 (step 6).
01) and its determination (step 602). If the result of the check in step 602 is execution, the evaluation function setting means 13 executes an evaluation function construction process (step 604). If the confirmation result in step 602 is cancelled, the variable setting means 15 shifts to step 603 to execute a zero value setting process on the time response change rate data, and then executes step 604.
Execute the evaluation function construction processing of. Operation amount optimization means 14
In step 605, the optimum operation amount update value is obtained using the evaluation function obtained as described above, and the optimum operation amount is output. In this way, the calculation operation at the k-th time of the control cycle is performed.

Next, the control target process 2 is a multivariable process. In a multivariable process, the combination of control variables in addition to manipulated variables may be changed. In this case, the configuration of the data relating to the corresponding control variable in the time response change rate data G is changed in the same manner as the method shown in FIG. That is, the corresponding data is set to “0” in response to the variable release command. As an example,
FIG. 5C shows a case where the control variable registered first as the control variable is released.

By using the time response change rate data G to construct the evaluation function J in the evaluation function setting means 13, the control variable 1 performs an operation released from the target value follow-up control. become.

FIG. 7 shows a control operation characteristic according to the present invention.
FIG. 7 shows an execution result of temperature control in the distillation column 21 which is one of the chemical processes shown in FIG. The process control device 1 is subjected to trial operation and local adjustment, and a combination of control variables and operation variables, a process model, and the like are determined in advance by simulation evaluation.

The distillation column 21 has such a characteristic that the temperature in the column easily rises due to the increase in the amount of steam, but conversely, the temperature in the column hardly decreases even if the amount of steam is reduced. The control valves 221 and 231 installed in each pipe adjust the steam amount and the raw material flow rate.

It is assumed that the control amount S210 (tower temperature) shown in FIG. 7 (a) is reduced at time 0:00 due to some disturbance (for example, the tower temperature is lowered).

The model predictive control system of the process control device 1 of the present invention is adjusted in advance so that the operation variable A (steam amount S220) shown in FIG. (For example, an increase in the amount of steam), the decrease in the control amount S210 (column temperature) is recovered. After the time 2:00 has elapsed, the model predictive control using the operation variable A is released (OFF), and the operation variable B shown in FIG.
The control is switched to only (example: feed amount S230).

As a result, the variable setting means 15 for setting the time response change rate changes the data corresponding to the operation variable A of the time response change rate data G used for constructing the evaluation function of the model predictive control. The manipulated variable A is separated from the evaluation function of the model predictive control, and the control is switched to the control by the manipulated variable B. In FIG. 7, the deviation between the control variable and the target value, which has occurred at the time of 2:00, has been eliminated.

FIG. 8 shows the control operation characteristics when the control is executed by the conventional method. In the conventional method, the combination of the control variable and the manipulated variable is fixed, and the change tendency of the variable is adjusted by the adjustment coefficient (weighting coefficient) for each variable. FIG. 8 illustrates a case where the model predictive control is used as multivariable control under such a conventional method.

It is assumed that the control amount S210 (tower temperature) shown in FIG. 8A is reduced at time 0:00 due to some disturbance (for example, the tower temperature is lowered). On the other hand, the model prediction control system is adjusted in advance so that the operation variable A (steam amount S220) shown in FIG.
The decrease in the control amount S210 (the temperature in the tower) is recovered by increasing the operation amount A (eg, increasing the steam amount).

When the time 2:00 has elapsed, the operation amount optimizing means 14 sets the operation variable A as shown in FIG.
(Steam amount) is reduced.
However, in the distillation column 21, it takes time to lower the temperature in the column when the amount of steam is reduced, and the change in the flow rate of the raw material shown in FIG. As a result, it takes time to lower the temperature. As a result, it takes time to reduce the deviation between the control variable (the temperature in the tower) and the target value, which eventually leads to deterioration of the plant operation characteristics.

Here, the above-described evaluation function can be realized by another expression as shown below.

Evaluation function J = | Control amount predicted value−Target value | +
| Manipulated variable update value] The model predictive control of the process is performed as described above, and the manipulated variable update value is calculated based on a combination of a predicted deviation which is a deviation between the controlled variable predicted value and the target value and a plurality of manipulated variables. At the time of calculation, a combination of a plurality of operation variables or a combination of a plurality of control variables and operation variables is set and changed, so that model predictive control of a multivariable process can be performed with high process control accuracy.

In the above-described embodiment, since the execution state and the release state of the control variable and the operation variable of the control object are displayed on the display device, which control variable of the control object is controlled and Whether or not the process control is being executed by the operation variable can be visually confirmed.

The process control device 1 is stored as a program processing procedure in a storage medium such as a CD-ROM.
It is obvious that the present invention can be implemented using an arithmetic processing device such as a computer.

[0072]

According to the present invention, when calculating the manipulated variable update value based on a plurality of combinations of the predicted deviation, which is a deviation between the control value predicted value and the target value, and the manipulated variable, a combination of the plurality of manipulated variables is used. Alternatively, since a combination of a plurality of control variables and operation variables is set and changed, model predictive control of a multivariable process can be performed with high process control accuracy.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing one embodiment of the present invention.

FIG. 2 is a configuration diagram illustrating an example of a process to be controlled according to the present invention.

FIG. 3 is a diagram showing an example of a setting display screen according to the present invention.

FIG. 4 is a characteristic diagram for explaining the operation of the present invention.

FIG. 5 is an explanatory diagram of setting of time response change rate data according to the present invention.

FIG. 6 is a flowchart for explaining the operation of the present invention.

FIG. 7 is a control operation characteristic diagram according to the present invention.

FIG. 8 is a graph showing a control operation characteristic according to the related art.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Process control apparatus, 2 ... Control target process, 3 ... Input means, 11 ... Model database, 12 ... Predicted value calculation means, 13 ... Evaluation function setting means, 14 ... Operation amount optimization means, 15 ... Variable setting means, 16. Target value / constraint condition setting means.

Claims (10)

[Claims] 1. A control variable predicted value is obtained by using a process model having a plurality of control variables and manipulated variables, and a predicted deviation which is a deviation between the controlled variable predicted value and a target value; A process control method for obtaining an operation amount update value based on a combination of the plurality of operation variables when obtaining the operation amount update value. 2. A control amount predicted value is obtained by using a dynamic characteristic model of a process having a plurality of control variables and manipulated variables, and a predicted deviation between the control amount predicted value and a target value and the plurality of manipulated variables are used. An operation amount update value is obtained by the obtained evaluation function, and the plurality of operation variables required for obtaining the operation amount update value are set and changed to obtain the operation amount update value. A process control method characterized by the above-mentioned. 3. A predicted value of a controlled variable based on a manipulated variable at a prediction point in time using a mathematical model of dynamic characteristics of a process having a plurality of controlled variables and controlled variables, and a predicted deviation between the predicted value of the controlled variable and a target value. And calculating an operation amount update value by an evaluation function obtained by using the plurality of operation variables, and displaying the plurality of operation variables required for obtaining the operation amount update value on a display screen. A process control method for obtaining the manipulated variable update value. 4. A control variable predicted value is obtained by using a dynamic characteristic model of a process having a plurality of control variables and controlled variables, and a predicted deviation between the controlled variable predicted value and a target value is calculated. Is used to calculate an operation amount update value by an evaluation function obtained by using the above-described method, and the operation amount is changed by setting and changing a combination of the plurality of control variables and operation variables required when obtaining the operation amount update value. A process control method characterized in that an update value is obtained. 5. A control variable predicted value is obtained using a dynamic characteristic model of a process having a plurality of control variables and manipulated variables, and a predicted deviation between the controlled variable predicted value and a target value; And calculating an operation amount update value by an evaluation function obtained using the constraint condition, wherein the plurality of control variables and the operation are expressed by a time response change rate required when obtaining the operation amount update value. A process control method, wherein a variable is displayed on a display screen and set to obtain the manipulated variable update value. 6. A predicted value calculating means for calculating a predicted value of a controlled variable using a process model and process data having a plurality of control variables and controlled variables, a predicted deviation between the predicted controlled variable and a target value, and the controlled variable Evaluation function setting means for obtaining an evaluation function based on a plurality of combinations of the above, operation amount optimization means for inputting the evaluation function to obtain an operation amount update value, and the plurality of operation variables used in the evaluation function setting means And a variable setting means for setting the following. 7. Predicted value calculating means for obtaining a predicted value of a controlled variable based on a process manipulated variable at a prediction time point using a dynamic characteristic model of a process having a plurality of controlled variables and controlled variables; Evaluation function setting means for obtaining an evaluation function using a prediction deviation which is a value deviation and the plurality of operation variables; operation amount optimization means for obtaining an operation amount update value by the evaluation function; and the evaluation function setting means. A variable setting means for displaying and setting the plurality of operation variables to be used on a display screen. 8. A predicted value calculating means for obtaining a predicted value of a controlled variable using a dynamic characteristic model of a process having a plurality of control variables and manipulated variables, and a predicted deviation which is a deviation between the predicted value of the controlled variable and a target value. Evaluation function setting means for obtaining an evaluation function using a plurality of control variables and operation variables; operation amount optimization means for obtaining an operation amount update value by the evaluation function; and the plurality of controls used in the evaluation function setting means A process control device comprising: a variable setting unit that sets a combination of a variable and an operation variable. 9. A predicted value calculating means for obtaining a controlled variable predicted value using a dynamic characteristic model of a process having a plurality of control variables and manipulated variables; a predicted deviation between the controlled variable predicted value and a target value; Evaluation function setting means for obtaining an evaluation function using control variables, operation variables, and constraints; operation amount optimization means for obtaining an operation amount update value using the evaluation function; and the plurality of controls used in the evaluation function setting means. A process control device comprising: a variable setting unit that displays a variable and an operation variable on a display screen and sets the variable and the operation variable. 10. A control variable predicted value is obtained using a dynamic characteristic model of a process having a plurality of control variables and manipulated variables, and a predicted deviation between the control variable predicted value and a target value and the plurality of manipulated variables are used. A program for obtaining the manipulated variable update value by the obtained evaluation function, wherein a program for setting and changing the plurality of operation variables required for obtaining the manipulated variable update value and obtaining the manipulated variable update value is stored. A storage medium characterized by the following.