JPH08161004A - Pid controller - Google Patents

Abstract

PURPOSE: To provide a scheduling function without initializing control when a PID constant is varied. CONSTITUTION: The PID controller which has a table 21b of PID constants corresponding to a predetermined finite number of controlled variables and performs PID control by using an inputted controlled variable and the PID constant in the table 21b corresponding to the inputted controlled variable performs interpolation by using the inputted controlled variable and the combination of two controlled variables which are quantities adjacent to the inputted controlled variable and defined in the table and the corresponding PID constants unless the PID constant corresponding to the inputted controlled variable is in the table. Further, the PID controller is equipped with an interpolation value calculating means 4 which calculates the PID constant corresponding to the inputted controlled variable.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a PID controller for industrial instrumentation used in, for example, a plant.

[0002]

2. Description of the Related Art Conventionally, controllers using PID control have been widely used from large chemical plants to small heat treatment furnaces. In a normal PID controller,
The PID constant is fixed and used. But for example ph
In control and the like, a PID controller having a function called gain scheduling that changes a PID constant based on a value such as a signal input value and a preset table is used.

FIG. 5 is a block diagram of a conventional PID controller, and FIG. 6 is an explanatory diagram for explaining a gain scheduling function in the PID controller. This PI
In the D controller, the control amount PV received from the control target
Signal inputs such as are input to the PID calculation unit 1 and the PID constant extraction unit 2.

The PID calculator 1 receives the input control amount PV
Based on the set value SV and the PID constant of the PID constant unit 11, PID calculation is performed and the manipulated variable MV is output. on the other hand,
The PID constant take-out unit 2 has P values corresponding to the reference value ranges.
A gain scheduling table 21 in which ID constants are stored is provided, and if the input manipulated variable MV, that is, the value of a signal is the reference value X and the value of the reference value is in a range different from that at the previous signal input, the gain scheduling table 21.
The corresponding PID constant is retrieved from

When the value of the reference value is in the range different from that at the time of the previous signal input, the PID constant extracting unit 2 extracts the P
The ID constant is input to the PID constant unit 11 in the PID calculation unit 1 and a command is issued to the control initialization unit 3.

When the control initialization unit 3 receives the command, the PI
The D operation unit 1 initializes the PID operation that has been performed so far. The reason is that the integral term in the PID control performs the next calculation based on the past calculation result, and therefore the control becomes discontinuous when the PID constant is changed.

Specifically, as shown in FIG. 6, the PID constant (parameter) in the PID constant section 11 takes a constant value within a range defined with respect to the reference value X and exceeds the threshold value. And, it changes discontinuously to the corresponding value in the next range. Then, when the PID constant changes discontinuously, the PID control is initialized (reset).

[0008]

As described above, the conventional PID controller having the gain scheduling function has a problem that the PID control must be initialized when changing the PID constant.

This problem becomes particularly serious when the controlled variable PV, that is, the reference value X, fluctuates in the vicinity of the threshold value for changing the PID constant. That is, in such a case, P
Since the ID control itself is always reset, there is a problem in that continuous PID control, particularly integration operation, cannot be performed.

The present invention has been made in consideration of such circumstances, and an object thereof is to provide a PID controller which realizes a gain scheduling function without initializing the control when changing the PID constant. And

[0011]

In order to solve the above-mentioned problems, the invention according to claim 1 has a table of PID constants corresponding to a predetermined finite number of control amounts, and the input control In a PID controller that performs PID control using an amount and a PID constant corresponding to the input control amount in the table, when the PID constant corresponding to the input control amount is not in the table, the input control amount and Interpolation value calculation for calculating a PID constant corresponding to the input control amount by performing interpolation using two sets of control amounts and corresponding PID constant sets defined in the table It is a PID controller equipped with means.

Further, the invention according to claim 2 corresponds to a plurality of ranges of a predetermined finite number of controlled variables, and has a table defining a constant PID constant for each of the ranges.
PID in the table corresponding to the range of input control amount
The PID constant used for control is calculated based on the constant, and PI
In the PID controller that performs D control, when the PID constant in the table corresponding to the range of the input control amount is different from the current PID constant used for control,
The PID controller is provided with a PID constant change rate limiting means for changing the current PID constant so as to approach the PID constant in the table within a range of a constant change rate.

[0013]

Therefore, first, in the PID controller of the invention according to claim 1, the PID constant is changed by the table, for example, when the control amount output from the plant process is a value defined on the table. It is like this.

Next, when a control amount not defined in the table, that is, a control amount between two defined control amounts is input by the interpolation value calculation means, when the control amount at the two points and the PID constant are set. Is used to calculate the PID constant corresponding to the input control amount by interpolation, for example, linear interpolation.

In general, when the rate of change of the PID constant is rapid, the factors affected by past PID constants such as the integral term become inaccurate, so it is necessary to initialize the PID control.
However, the PID calculated by the above interpolated value calculation means
Since the constant is a value that changes gently by interpolation,
The above inconvenience does not occur.

Therefore, when the PID constant is changed, the control can be continued without initializing the PID control. Next, the PID of the invention corresponding to claim 2
In the controller, the table defines the PID constants corresponding to each range of the control amount.
PI used for control while the controlled variable is within a certain range
The D constant is constant.

However, for example, when the controlled variable shifts from one range to another, at least the corresponding PID constant in the table changes abruptly. Therefore, for example, in such a case, the current PID constant is changed by the PID constant change rate limiting means so as to approach the PID constant in the table, and the PID constant is larger than the preset change rate in the change. The rate of change is limited so that it does not change.

Therefore, the change of the PID constant to be changed becomes gradual, and the control can be continued without initializing the PID control. By repeating the input of the controlled variable for a plurality of cycles, the current PID constant finally becomes a PI corresponding to the controlled variable range in the table.
It has the same value as the D constant.

[0019]

Embodiments of the present invention will be described below. FIG. 1 is a block diagram showing an embodiment of a PID controller according to the present invention. The same parts as those of the conventional example shown in FIG.

This PID controller is provided with a PID constant extracting section 2b and an interpolation value calculating section 4. PI
When a signal such as the controlled variable PV is input, the D constant extracting unit 2b checks which range on the gain scheduling table 21b the reference value X corresponding to the signal is, and becomes a threshold at both ends of the corresponding range. The PID constants corresponding to the point reference values X1 and X2 are input to the interpolation value calculation unit 4.

The interpolation value calculation unit 4 sets the PID constants of all P, I and D corresponding to the reference value X corresponding to the input signal to 2
It is calculated by linear interpolation from the reference values X1 and X2 of the points and the respective PID constants, and registered in the PID constant section 11 in the PID calculation section 1.

Therefore, for example, when the contents of the gain scheduling table 21b are as shown in FIG.
The proportional constant Kp corresponding to the reference value X continuously changes as shown in FIG. 2 according to the change of the control amount PV, that is, the reference value X.

Since the other PID constants, the integration constant Ti, and the differential constant Td are also the same, the PID constant gradually changes as the control amount PV changes, and the control is initialized in accordance with the change of the PID constant. No need.

Therefore, the PID calculation unit 1 uses the new PID constant registered in the interpolation value calculation unit 4 to calculate the manipulated variable M.
V is output to continuously control the controlled object. As described above, the PID controller according to the present embodiment is provided with the gain scheduling table 21b and the interpolation value calculation unit 4 so that the PID constant is continuously and gently changed according to the change of the control amount PV, that is, the reference value X. Because I chose
It is possible to perform continuous control as a whole without initializing the PID control.

Further, since the data is interpolated, the PID constant is continuously changed in the same manner as the actual change of the plant, and more detailed gain scheduling control can be realized.

In this embodiment, two reference values X1,
Although the linear interpolation is performed using X2, for example, if the reference values Xx of two or more points are used and the interpolation is performed using the Bessel function, the PID constant can be changed more smoothly.

FIG. 3 is a block diagram showing another embodiment of the PID controller according to the present invention. The same parts as those of the conventional example shown in FIG. This PID controller includes a PID constant take-out section 2c and a PI.
The D constant change rate limiting unit 5 is provided.

When the signal of the controlled variable PV or the like is input, the PID constant extracting section 2c checks which range on the gain scheduling table 21 the reference value X corresponding to the signal is.

When the two thresholds at both ends of the applicable range are reference values X1 and X2 (X1 <X2), the reference value X
The PID constant corresponding to 2 is taken out and input to the PID constant change rate limiting unit 5.

The PID constant change rate limiting unit 5 reads out the current PID constant from the PID constant unit 11, and if the two are different, changes so that the rate of change of the PID constant per unit time becomes a constant value. By limiting the rate, a new value is calculated so that the current PID constant approaches the PID constant input from the PID constant extracting unit 2c.

Then, the newly calculated PID constant is registered in the PID constant section 11 in the PID calculation section 1. 4A shows the proportional constant Kp extracted from the gain scheduling table 21 in correspondence with the control amount PV, that is, the reference value X, and FIG. 4B shows, for example, FIG.
In the case of (a), how the proportional constant Kp registered in the PID constant unit 11 changes with time when the reference value X changes from 5 to 15.

As a method of limiting the rate of change of the PID constant in the PID constant change rate limiting unit 5, for example, the magnitude of the value changed once may be fixed, or at the present time, for example. You may make it multiply the PID constant of 1 by a proportional constant.

As described above, when the change in the controlled variable PV crosses the threshold value on the gain scheduling table 21, the PID constant is as shown in FIG.
As shown in (b), it changes continuously and gently with respect to time.

The other PID constants, the integration constant Ti and the differential constant Td are the same as in the case of the proportional constant Kp. Therefore, the PID constant is a constant value when the control amount PV is within a certain range, and when the control amount PV exceeds the threshold value, the PID constant changes gently with the passage of time. There is no need to convert.

Therefore, the PID calculation unit 1 outputs the manipulated variable MV using the new PID constant registered in the PID constant change rate limiting unit 5 to continuously control the controlled object.
As described above, the PID controller according to the present embodiment is provided with the gain scheduling table 21 and the PID constant change rate limiting unit 5, and sets the PID constant used for control to a value different from the value currently used in the PID constant unit 11. When it is decided to use it, PID is limited by the rate of change.
Since the constant is made to change gradually and gradually, PI
It is possible to perform continuous control as a whole without initializing D control.

Further, the gain scheduling table 2
Even if the controlled variable fluctuates near the PID constant change point in 1, the change in the PID constant is limited by the time change rate, so gradually approach the PID constant in any range. become. Therefore, stable control can be performed.

Further, since the PID constants may be determined in units of areas (ranges) as in the conventional case, there is an advantage that the set value used in the gain scheduling table 21 can be easily obtained.

In the present embodiment, P at the present time
Although the ID constant is read from the PID constant unit 11, a storage unit may be provided in the PID constant change rate limiting unit 5 to store the current PID constant.
Further, the present invention can be variously modified without departing from the gist thereof.

[0039]

As described above in detail, according to the present invention, P
It is possible to provide a PID controller that realizes the gain scheduling function without initializing the control when changing the ID constant.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an embodiment of a PID controller according to the present invention.

FIG. 2 is a PID in the PID controller of the embodiment.
Explanatory drawing which shows the change of a constant.

FIG. 3 is a configuration diagram showing another embodiment of the PID controller according to the present invention.

FIG. 4 is a PID in the PID controller of the embodiment.
Explanatory drawing which shows the change of a constant.

FIG. 5 is a block diagram of a conventional PID controller.

FIG. 6 is a diagram illustrating a gain scheduling function in a PID controller.

[Explanation of symbols]

1 ... PID calculation unit, 2, 2b, 2c ... PID constant extraction unit, 4 ... Interpolation value calculation unit, 5 ... PID constant change rate limiting unit.

Claims (2)

[Claims] 1. A PID constant table corresponding to a predetermined finite number of control amounts, and a PID corresponding to the input control amount and the input control amount in the table.
In a PID controller that performs PID control using a constant, when the PID constant corresponding to the input control amount is not in the table, the input control amount and the amount adjacent thereto are in the table. Interpolation is performed using two sets of control amounts and corresponding sets of PID constants defined in paragraph 1, and interpolation value calculating means for calculating a PID constant corresponding to the input control amount is provided. PID controller. 2. A table corresponding to a plurality of ranges of a predetermined finite number of control variables, and having a table defining a PID constant of a constant value for each of these ranges, corresponding to the range of the input control variable. In a PID controller that calculates a PID constant used for control based on the PID constant in the table and performs PID control, the PID constant in the table corresponding to the range of the input control amount is currently used for control. PID constant change rate limiting means for changing the current PID constant so that the current PID constant approaches the PID constant in the table within a constant change rate range. PID controller.