JPH06309005A - Controller - Google Patents

Abstract

PURPOSE:To speedily converge an input signal on a target value even when the target value is altered repeatedly in a short time. CONSTITUTION:A subtraction part 1 outputs a control deviation based upon the target value and an input signal. A proportional arithmetic part 3, a differential arithmetic part 5, an integral arithmetic part 7, and an integral value variation part 15 output P(n), D(n), and I(n) according to the control deviation. An addition part 9 adds them and sends a PID output to an output limiter 11, which outputs an output signal MVD. An integral control part 13 stops the arithmetic of the integral arithmetic part 7 when the control deviation exceeds a proportion band to fix the integral value. The integral value variation part 15 varies the integral value of the integral arithmetic part 7 to a specific integral value when the control deviation enters the proportion band from a value larger or smaller than the proportion band on the basis of variation in the target value.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control device for performing at least an integral operation, for example, a process control loop system for controlling process amounts such as temperature, pressure and flow rate, and a mechanical device for controlling mechanical devices such as industrial robots. Control loop system,
It is an improvement of a control device suitable for use in other general-purpose control loop systems, and more specifically, it relates to an improvement of an integral calculation part in the control device.

[0002]

2. Description of the Related Art As a control device of this type, a target value SV
The control deviation e between the control signal e from the control target and the control signal e is controlled by the output signal MV obtained by the proportional calculation (P), the integral calculation (I) and the differential calculation (D).
ID control is the mainstream. In this PID control, the integral calculation has a function of making the steady control deviation e between the input signal PV and the target value SV "zero", but always causes a large control deviation e due to the change of the target value SV. When the integration operation is executed, as shown in FIGS. 8 and 9, there is a so-called reset windup phenomenon in which the input signal PV exceeds the target value SV due to overintegration. Note that FIG. 8 shows the movement of the calculated values (P, I) when the control device is simulated, and FIG. 9 shows the movement of the input signal PV and the output signal MV corresponding to FIG.

Conventionally, in a control device including an integral calculation, in order to prevent the reset windup phenomenon, it is general to add a function of suppressing the integral calculation, that is, an anti-reset windup function. FIG. 10 is a block diagram showing an example of a conventional control device having the function.
In this control device, from the target value SV (n), the input signal PV
The control deviation e (n) obtained by subtracting (n) by the subtraction unit 1 is added to the proportional calculation unit 3, the differential calculation unit 5, and the integral calculation unit 7,
These proportional operation part 3, differential operation part 5 and integral operation part 7
The PID output PID (n) obtained by adding the calculation outputs P (n), D (n), and I (n) from the output limiter 11 to the output limiter 11 When the control deviation e (n) exceeds a certain range while the output signal MV (n) is output to the integral output unit MV (n), the integral calculation of the integral calculation unit 7 is stopped and the integral output I (n) is output. n) is fixed to a certain fixed value, for example, "0" or "1" to prevent the above reset windup phenomenon.

By the way, the control device shown in FIG. 10 has a discrete type configuration, and by adding a subscript (n) to each output, the n-th time point (reference numeral) when each output performs a discrete type operation. Although τ is a signal value of t = nτ when the calculation interval or the sampling period and the code t are times (the same applies hereinafter), the operating principle of the control device according to the present application is essentially Since it is irrelevant to the discrete (digital) and continuous (analog) calculation methods, the same effect can be obtained with either configuration.

Next, the operation of the integration controller 13 shown in FIG. 10 will be described in detail with reference to FIGS. 11 and 12. In addition, the code | symbol e in FIG. 10 is control deviation (e = SV-PV),
Reference numeral 1 / Kp is a proportional band (PB: reciprocal of proportional constant Kp),
Reference symbol P indicates a proportional output, reference symbol MV indicates an output signal, and reference symbol I indicates an integral output. Also, the proportional band 1 / Kp numerator “1”
Indicates the case where the input / output range of the control device is normalized to 0 to 1, and if the input / output range is set to 0 to 100%, the proportional band is 100 / Kp (%).

When the control deviation e exceeds a certain range, for example, the proportional band 1 / Kp, the integral control unit 13 stops the integral calculation of the integral calculation unit 7, and the integrated output I is changed depending on the direction of the excess. "0" (0%) or "1" (1
00%). That is, the control deviation e is e> 1 / Kp, in other words, PV <
In the case of [SV- (1 / Kp)], the integrated output I is "0".
Fixed to the input signal PV, target value SV, proportional output P,
The relationship between the integrated output I and the output signal MV is as shown in FIG. For convenience of explanation, the differential output D is omitted in FIG. 11 (the same applies to FIG. 12).

In FIG. 11, for example, at the point where the input signal PV is, the integral output I = 0, the proportional output P> 1, and the output signal MV = 1 is output. Therefore, even if the input signal PV increases, the integral The calculation is not performed until the input signal PV reaches the point of, which prevents overintegration. On the other hand, proportional output P
Becomes smaller than "1" when the input signal PV increases from the point of, and the output signal MV is "1" from the front of the target value SV by the proportional band 1 / Kp with respect to the increase of the input signal PV.
Since it starts to decrease from then, it moves so as to suppress the increase of the input signal PV, so that the overshoot of the input signal PV with respect to the target value SV can be suppressed. The control deviation e is e <-1 / K
In the case of p, in other words PV> [SV + (1 / Kp)], the integrated output I is fixed to “1”, so that the input signal P
The relationship among V, the target value SV, the proportional output P, the integral output I, and the output signal MV is as shown in FIG.

In FIG. 12, when the input signal PV is, for example, the integrated output I = 1 and the proportional output P <−1 and the output signal MV = 0 is output, so that the input signal PV is reduced. The integral operation is not performed until the input signal PV reaches the point of, thus preventing overintegration. On the other hand, proportional output P
Becomes larger than “−1” when the input signal PV decreases from the point of, and the target value SV decreases with the decrease of the input signal PV.
The output signal MV starts to increase from “0” by a proportional band 1 / Kp, and works to brake the decrease of the input signal PV to suppress the reset wind-up phenomenon, and to the target value SV of the input signal PV. You can prevent going too far. In this way, in the control device shown in FIG. 10, when the target value SV larger than the proportional band 1 / Kp is changed, the integral calculation is stopped and the integral calculation value is set to “0”.
Alternatively, by fixing the value to “1”, the overshoot of the input signal PV with respect to the target value SV is suppressed.

[0009]

However, as shown in FIG.
In the control device shown in FIG. 1, when the raising and lowering of the target value SV is repeated in a short time as compared with the changing speed of the input signal PV, FIG.
4 and the problem as shown in FIG. 14 and 15, in the control device of FIG. 10, the proportional band 1 / Kp = 1/3, and the target value SV is set between "0" and "0.5" for a short time as shown in FIG. FIG. 14 shows a simulation result of the PI control operation when changed in FIG. 14, FIG. 14 shows the response of the input signal PV and the output signal MV thereto, and FIG. 15 shows the response of the proportional output P and the integral output I. It is a thing.

First, when the target value SV is changed from "0" to "0.5" at time "0.5" in FIG. 13, while the control deviation e is larger than the proportional band 1 / Kp (e> 1 /). Kp)
Since the integral calculation is not performed and the integral output I is fixed to “0”, it can be seen from FIGS. 14 and 15 that the overshoot of the input signal PV with respect to the target value SV is suppressed. Next, when the input signal PV coincides with the target value SV (= 0.5) and a sufficient time has elapsed, the target value SV is changed from "0.5" to "0" at time "50" in FIG. , The control deviation e becomes e <-(1 / Kp), so that the integral calculation is not performed and the integral output I is fixed at "1".

Then, when the target value SV is changed from "0" to "0.5" again at a short time "50.5" (see FIG. 13) where the input signal PV hardly moves, the control is performed this time. Since the deviation e is almost “0” and | e | ≦ 1 / Kp, the integration calculation is restarted, but since it starts from the initial value “1”, “1” remains for a while even though there is almost no control deviation. A problem occurs in which the input signal PV is held and greatly deviates from the target value SV. Such a problem is that the target value SV
Is likely to occur in a simple controller or the like having a structure in which the target value SV of a plurality of digits can be changed only for each digit by the up / down key, and a solution to it is particularly desired.

The present invention has been made in order to solve the above-mentioned conventional drawbacks, and it is possible to prevent the control amount from being overshooting with respect to a normal change of the target value, and also to make the target value due to an erroneous operation or the like. It is an object of the present invention to provide a control device capable of promptly converging to a target value by appropriately processing an integral value to minimize disturbance of control even if the value is repeatedly changed in a short time.

[0013]

In order to solve such a problem, the present invention performs at least a proportional calculation and an integral calculation so that an input signal from a controlled object coincides with a target value, and outputs the signal to the controlled object. A control device that outputs an output signal and, when the deviation between the input signal and the target value is outside the specified range, sets the integral value of the integration operation to a fixed value and outputs the output signal. If the deviation between the target value immediately before the change and the input signal at the time of the change is outside the specified range when the value changes from outside to within the range, the integral value of the integral calculation is changed to the specific integral value. It has a configuration including an integral value changing unit.

In the present invention, the specific range may be a proportional band in the proportional calculation. Further, in the present invention, when the deviation changes from within the specific range to outside the specific range, the integrated value immediately before the change can be set as the specific integrated value. Further, in the present invention, it is possible to use a value calculated from the steady relationship between the input signal and the output signal of the controlled object and the target value as the specific integrated value.

[0015]

According to the present invention having such means, when the deviation between the input signal and the target value is out of the specific range, the integrated value is fixed to a fixed value and the output signal is output to suppress the reset windup phenomenon. be able to. Moreover, when the deviation changes from outside the specified range to within the specified range, if the deviation between the target value immediately before the change and the input signal at the time of the change is outside the specified range, the integral changing unit calculates the integral. Since the integrated value of the calculation is changed to the specific integrated value, the output signal can be calculated with an appropriate integrated value even if the target value largely changes within a short time and the deviation returns to within the specific range.

In the structure in which the specific range is set to the proportional band, the specific range set to suppress the reset windup phenomenon can be used as it is. Also,
When the deviation changes from within the specific range to outside the specific range, the integrated value immediately before the change is used as the specific integrated value, or is calculated from the steady input / output signal relationship of the controlled object and the target value. With the configuration in which the value is the specific integral value, the specific integral value can be easily calculated.

[0017]

Embodiments of the present invention will be described below with reference to the drawings. The same parts as those in the conventional example are designated by the same reference numerals. FIG. 1 is a block diagram showing an embodiment of a control device according to the present invention. In FIG. 1, the subtraction unit 1 uses the target value SV.
The control deviation e is obtained by subtracting the input signal PV (n) from (n).
(N) is obtained, and is connected to the proportional calculation unit 3, the differential calculation unit 5, the integral calculation unit 7, and the integral value changing unit 15.
The target value SV (n) is also added to the integral value changing unit 15, and the integral value changing unit 15 has the main function of the present invention. Details will be described later.

The proportional calculation unit 3, the differential calculation unit 5 and the integral calculation unit 7 calculate the proportional output P (n) from the control deviation e (n), respectively.
The differential output D (n) and the integrated output I (n) are calculated and connected to the adder 9. The adder 9 adds the respective operation outputs P (n), D (n) and I (n) to PI
The D output PID (n) is obtained and is connected to the output limiter 11. This output limiter 11 has a PID output P
Output signal MV from ID (n) to a control target (not shown)
(N) is output. When the control deviation e (n) exceeds a specific range (x or -x), for example, a proportional band (1 / Kp or -1 / Kp), the integral control unit 13 stops the integral calculation of the integral calculation unit 7. And the integral output I
(N) is fixed to a certain fixed value, for example, "0" or "1", and has a function described later.

Next, the basic concept of such a control device of the present invention will be described. First, the control deviation e (n) is compared with, for example, the proportional band 1 / Kp, and the execution or stop of the integration operation or the initialization of the integration value is performed in the same manner as in the conventional example according to the magnitude relationship. is there. (1) In case of control deviation ≥ proportional band, integration value is 0% when integration is stopped
(Output conversion) Fixed. (2) When control deviation ≤ (-) proportional band, integration value is fixed to 100% (output conversion) when integration is stopped. (3) (-) If proportional band <control deviation <proportional band, execute normal integration calculation. However, in the configuration of the present invention, the relationship between the change in the control deviation e (n) and the target value SV is further calculated from the current control deviation e (n), the deviation between the immediately preceding target value SV and the current input signal PV. As a result, when it is determined that the control deviation e has changed from (1) or (2) to (3) due to the change in the target value SV, the integration calculation is not performed immediately as in the conventional example, but the integration is performed. The feature is that once the value is changed to an appropriate value, the integral calculation is started.

With this configuration, the control deviation e (n) instantaneously becomes the above (1) or (2) due to an erroneous operation change of the target value SV and the integral value becomes 0% or 100%. Even if the state becomes (3) immediately after being initialized to 0, the integration operation is started after the integration value is changed to an appropriate value, so that the control disturbance is minimized and the input signal PV is promptly changed. It becomes possible to converge to the target value SV. In the present invention, the integral value changing unit 15 determines that the control deviation e (n) enters the proportional band 1 / Kp from a value larger or smaller than the proportional band 1 / Kp due to the change of the target value SV. At this time, the integration calculation unit 7 has a function of temporarily changing the integration value to a specific integration value.

As the specific integral value, for example, firstly, the control deviation e (n) changes from the state within the proportional band 1 / Kp to the proportional band 1
When the value changes to a value larger or smaller than / Kp, there is an integrated value immediately before the change, and secondly, there is an integrated value calculated from the steady input / output relationship of the controlled object and the target value. This point will be described later. Such a control device of the present invention is an R which stores a CPU and an operation program of the CPU.
It is generally composed of a microcomputer mainly composed of OM and I / O which is an input / output interface.

FIG. 2 is an overall flow chart when the control device of the present invention is realized by using the microcomputer, which is executed at every calculation cycle. That is, when the processing is started, the control deviation e (n) is calculated in step 201, the proportional output P (n) value is calculated in step 202, and the differential output D (n) value is calculated in step 203. Calculation is performed. In the following step 204, integration processing (see FIG. 3 described later) is performed, and in step 205, PID output PI
The calculation of D (n) is performed, the output limiter process is performed in step 206, and the process ends.

Step 204 is a processing step which characterizes the present invention, and a detailed flowchart is shown in FIG. First, steps 301 to 303 are performed as a part corresponding to the process of the integration controller 13.
In step 301, the control deviation e (n) is set to a specific range x
For example, comparing with the proportional band 1 / Kp, and when e (n)> x (outside the proportional band), the integrated output I (n) is 0% in step 302.
When e (n) <− x (outside the proportional band), the integral output I (n) is fixed to 100% in step 303. When | e (n) | ≦ x, The process proceeds to step 305 without performing integration.

When the integrated output I (n) is fixed in steps 302 and 303, the reset windup phenomenon due to overintegration is suppressed, and in step 304, the control deviation e (n-1) before the temporary point and the target value SV ( n-1) is the current control deviation e (n) and the target value SV, respectively.
Update with (n) and end. The control deviation e (n-1) and the target value SV (n-1) before the temporary point are used by the integral changing unit 15 as described below. Step 301
As a result of comparing the control deviation e (n) with the specific range x,
When ≦ e (n) ≦ x (in the proportional band), the integration changing unit 15
Step 304 to Step 308 and Step 310 corresponding to the above or Step 309 corresponding to the integral calculation unit 7
Is processed.

First, in step 305, the control deviation e (n-1) before the temporary point is compared with the specific range x, and | e (n-
1) |> x (outside the proportional band) and step 305 returns YE
If S, step 306 is executed. If it is within the proportional band and step 305 is NO, normal integration processing is performed in step 309. In step 306, a deviation e ′ (n) is calculated from the target value SV (n−1) before the temporary point and the current input signal PV (n), and this deviation e ′ is calculated in step 307.
(N) is compared with the specific range x, and if | e ′ (n) |> x and step 307 is YES, the integrated output I (n) is changed to the specific integrated value IMEM in step 308, and | e '
If (n) | ≦ x and step 307 is NO, normal integration processing is performed in step 309. Here, the specific integral value IMEM is an integral value when the absolute value of the last control deviation before the change of the target value SV is smaller than the proportional band.

When the normal integration process of step 309 is performed, the specific integration value IMEM is updated by the integration output I (n) in step 310. After step 308 or step 310, the deviation e (n-1) is calculated in step 305.
And SV (n-1) are e (n) and SV (n), respectively.
Is updated, and the integration process ends. This step 305
~ The processing from step 308 to step 310 has the following meaning. That is, in step 305, the absolute value of the control deviation e (n-1) before the temporary point is set to the specific range x
If the target value SV is not the current value but the value before the temporary point, the absolute value of the control deviation e ′ (n) is larger than the specific range x. It is to judge whether.

Therefore, the process of changing the integral value in step 308 is performed by (a) the control deviation e (n-
1) The absolute value of the absolute value is larger than the specific range x, in other words, the integrated output I (n-1) is initialized to 0% or 100%, and (b) the target value SV instead of the input signal PV.
This is the case where the absolute value of the control deviation e becomes smaller than the specific range x due to the change of. Since the conventional control device does not perform this part of judgment and processing and immediately starts the integration calculation, as shown in FIGS. 14 and 15 described above, the control caused by the improper integration initial value at the start of integration is performed. Disturbance had occurred. However, according to the control device of the present invention, since the integrated value is changed by judging such a change state, it is possible to suppress the control disturbance to the minimum.

By the way, in the above-mentioned flowchart, the specific range x to be compared with the control deviation e (n) can be arbitrarily determined. In general, in the integral control for suppressing the reset windup phenomenon (steps 301 to 303 in FIG. 3) in the control device, when the proportional gain which is the constant of the proportional operation part 7 is Kp, it is defined as the reciprocal thereof. The proportional band PB is used. PB = 1 / Kp (or PB [%] = 100 / Kp) Then, in steps 305 to 308 and step 310 which are the main part of the present invention, the proportional band PB can be used as the specific range x in the same manner. . In addition,
The proportional output P (n) is calculated as P (n) = Kp · e (n) = e (n) / PB = e (n) [%] / PB [%], and the proportional band PB is the proportional output P. (N) is equal to the control deviation e (n) when exactly 100% is output.

Next, with respect to the specific integral value IMEM shown in the flow chart of FIG. 3, the reason why the integral value immediately before the control deviation changes outside the specific range is used will be described with reference to FIG. In FIG. 4, the input signal PV matches the target value SV1 until time t1, normal integration is performed, and the integrated value is stable at I1. Then, at time t1, a large target value change (SV1 → SV2) in which the control deviation e (= SV-PV) exceeds the specific range x is performed, so the integrated value is 10
It is fixed at 0%, and the input signal PV starts to drop gradually.

However, the input signal PV is a new target value SV.
The target value is changed again at time t2 before reaching 2 (S
As a result of V2 → SV3), the control deviation e is again in the specific range x
It is getting smaller. In this pattern of changing the target value, the second target value SV2 is switched to the next target value SV3 in a short time before the input signal PV arrives as described above. This is likely to occur when the operator makes an erroneous setting or the target value can be set only for each digit.

However, since the integral value is fixed to 100% by this second change of the target value, in the conventional controller, the integral is started with 100% as the initial value for the change to the next target value SV3. Therefore, an undesired response as shown in FIG. Further, FIG. 4c shows the case where the integrated value is once cleared and started from 0%, but the response in this case is also undesirable. Therefore, the integration initial value when the target value SV3 is changed is a certain value between 0% and 100%, but the change to the target value SV2 is caused by an erroneous setting, etc. Is SV2 for a short time, so that when the input signal PV does not change much from the original target value SV1 and the target value SV3 is changed for such an input signal PV, the control deviation e is specified. Since it is smaller than the range x, it can be considered that the target values SV1 and SV3 are close to each other, and therefore the integration initial value is the input signal P.
It is appropriate to adopt the integral value I1 when V is stable at the target value SV1.

FIG. 4b shows the response when I1 is used as the initial integration value when the target value is changed to SV3, which is clearly better than the responses a and c described above. In the flowchart of FIG. 3, the control deviation e is smaller than the specific range x,
Further, since the specific integral value IMEM is updated with the integral result I (n) during normal integration, the integral value (≈I1) immediately before the time t1 in FIG. Used as the initial value of integration. further,
As shown in FIG. 5, the target value S when the control is stable
Two or more sets of V and output signal MV are stored (SV1 and MV1, SV2 and MV2 in FIG. 5), the output signal MVx for the current target value SVx is calculated from the relationship, and the value is used as the specific integral value IMEM. May be. The calculation formula is as follows. IMEM = MVx = [(SVx-SV1) (MV2-MV1) / (SV2
-SV1)] + MV1

FIGS. 6 and 7 are control simulation results by the control device of the present invention according to the flowchart of FIG. 3, and the same conditions as those of the conventional control device shown in FIG. 10 described above, that is, proportional gain Kp = 3. , Proportional band PB =
1/3 = 0.33, specific range x = PB = 0.33,
Further, as shown in FIG. 13, the target value SV is set to "0" and "0.
The input signal PV is changed to the target value SV in a short time between 5 ".
(= 0.5) and when a sufficient time has passed "5
Change the target value SV from "0.5" to "0" at 0 ",
FIG. 9 is a characteristic diagram in the sequence of changing the target value SV.

FIG. 6 shows the response of the input signal PV and the output signal MV, and FIG. 7 shows the response of the proportional output P and the integral output I. According to FIGS. 6 and 7,
At the time 0.5 hour when the first target value SV is changed, the normal target value is changed. Since the change amount 0.5 is larger than the proportional band 0.33, the integral value I is initialized to (0.0). Then, the integration calculation is stopped until the input signal PV becomes larger than [0.5−0.33 = 0.17], so the reset windup phenomenon is suppressed, and the target value SV of the input signal PV is suppressed.
The overshoot for is suppressed. This is similar to the conventional example of FIG.

Next, the target value SV is raised and lowered in a short time at the time points 50 and 50.5 when the second and third target values SV are changed, which is an abnormal target value change such as erroneous setting. In this case, in the conventional configuration of FIG. 10, the sign of the control deviation e is negative and the absolute value (| SV-PV | = 0.5) becomes larger than the proportional band (0.33) by the second change of the target value. Therefore, when the integration is stopped, the integrated value I is initialized to (1.0). And the third target value S
By changing V, the absolute value of the control deviation e (| SV-P
V |) becomes smaller than (0.5) and the integral calculation is restarted, but since the integral value starts from (1.0) by the previous initialization, the input signal PV subsequently rises and the target value is increased. It takes time to converge to SV.

On the other hand, in the control device of the present invention, 2
The integration value I is initialized to (1.0) by the change of the target value SV of the third time as in the conventional example, but the target value S of the third time is changed.
When V is changed, the absolute value of the current control deviation is smaller than the proportional band (0.33), the absolute value of the previous control deviation is larger than the proportional band (0.33), the immediately preceding target value (0.0) Since the difference between the current input PV and the current input PV is larger than the proportional band is satisfied, the integrated value is the target value SV.
Specific integrated value I when the absolute value of the last control deviation before the change of is smaller than the proportional band (immediately before the second change of the target value)
Changed to MEM. Therefore, since the integrated value is started from a substantially appropriate value, the disturbance of the input signal PV is small, and the convergence time to the target value SV is fast compared to the conventional example, and a fast control result can be obtained.

[0037]

As described above, in the constitution of the present invention,
When the deviation between the input signal and the target value is outside the specified range, the integrated value is fixed and the output signal is output to suppress the reset windup phenomenon. When the deviation between the target value immediately before the change and the input signal at the time of the change is out of the specific range, the integrated value changing unit changes the integrated value by the integration operation to the specific integrated value. Since the,
For normal target value changes, it is possible to suppress the overshoot amount of the input signal with respect to the target value, and even if the target value is repeatedly changed up and down in a short time due to an erroneous operation, the integrated value is appropriate. Since it can be changed to such a specific integral value, it is possible to promptly converge the input signal to the target value while minimizing the control disturbance. In the configuration in which the specific range is set to the proportional band, the specific range set to suppress the reset windup phenomenon can be used as it is, so that the simplification of the configuration can be maintained. Further, when the deviation changes from within the specific range to outside the specific range, the integrated value immediately before the change is used as the specific integrated value, or is calculated from the steady input / output signal relationship or the target value of the control target. With the configuration in which the determined value is used as the specific integral value, the specific integral value at the time of restarting the integration calculation can be easily calculated, further optimization can be performed, and the configuration does not become complicated.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a control device according to the present invention.

FIG. 2 is a schematic flowchart illustrating the operation of the control device of FIG.

FIG. 3 is a flowchart illustrating a main operation of the control device of FIG.

FIG. 4 is a diagram illustrating a procedure for setting a specific integral value of the control device of the present invention.

FIG. 5 is a diagram illustrating a procedure for setting a specific integral value of the control device of the present invention.

FIG. 6 is a characteristic diagram when the control device of the present invention is operated in simulation.

FIG. 7 is a characteristic diagram when the control device of the present invention is operated in simulation.

FIG. 8 is an operating characteristic diagram of a general control device.

FIG. 9 is an operating characteristic diagram of a general control device.

FIG. 10 is a block diagram showing a conventional control device in which a reset windup phenomenon is suppressed.

11 is a diagram illustrating an operation of the control device in FIG.

12 is a diagram illustrating the operation of the control device in FIG.

13 is a diagram for explaining the operation of the control device in FIG.

14 is a characteristic diagram when the control device of FIG. 10 is operated in simulation.

15 is a characteristic diagram when the control device of FIG. 10 is operated in a simulation manner.

[Explanation of symbols]

 1 Subtractor 3 Proportional Calculator 5 Derivative Calculator 7 Integral Calculator 9 Adder 11 Output Limiter 13 Integral Control 15 Integrated Value Changer

Claims (4)

[Claims] 1. An output signal to the controlled object is output by performing at least a proportional operation and an integral operation so that an input signal from the controlled object matches a target value, and a deviation between the input signal and the target value is calculated. In a control device that outputs the output signal by setting the integral value of the integral operation to a fixed value when the deviation is outside the specific range, when the deviation changes from outside the specific range to within the specific range, immediately before the change. A control device comprising an integral value changing unit that changes an integral value of the integral operation to a specific integral value when a deviation between the target value and the input signal at the time of change is outside the specific range. 2. The control device according to claim 1, wherein the specific range is a proportional band in the proportional calculation. 3. The control device according to claim 1, wherein when the deviation changes from within the specific range to outside the specific range, the specific integrated value is an integrated value immediately before the change. 4. The control device according to claim 1, wherein the specific integration value is a value calculated from the steady relationship between the input signal and the output signal of the controlled object and the target value.