JPH02186402A - Pid controller - Google Patents

Abstract

PURPOSE:To save a regulating work, to cancel individual difference, and to optimally tune a control parameter by obtaining an evaluation index based on the target value of a control object and a control quantity, and assuming the corrected value of a PID control parameter. CONSTITUTION:A target value SV of a control object 2 and a control quantity PV are inputted, the three evaluation indexes are obtained by a recognizing means 4 from a control quantity responding shape when the target value is changed or a disturbance is impressed, the three evaluation indexes are respectively quantitatively evaluated, and the corrected value of the PID control parameter is assumed by a correcting means 6 based on the evaluation. Further according to the unsatisfaction degree of the two evaluating indexes for two target control specifications out of the three evaluation indexes, the assumed corrected value of the PID control parameter is evaluated by an evaluating means 7. On the other hand, when the two evaluating indexes do not satisfy the target control specifications, the means 6 and 7, and a control parameter regulated value arithmetic function 8 are respectively made to function by a controllability deciding means 5.

Description

Detailed Description of the Invention [Industrial Application Field] The present invention is directed to PID (proportional,
The present invention relates to an integral (integral, differential) controller, and particularly to a PID controller that enables automatic tuning of I'ID control parameters.

[Conventional technology]

Conventionally, tuning of PID control parameters in a T'ID controller has been performed manually by a controller while observing fluctuations in the controlled variable. on the other hand.

From the viewpoint of control theory, many methods have been proposed in which the dynamic characteristics of the controlled object are identified by applying an identification test signal to the controlled object, and the control parameters are tuned to optimal values based on the results.

Furthermore, as a hyuristic method (expert method) for tuning control parameters by looking at the control variable response shape, r measurement technology: (110, September 1961), Expert Self-Tuning! gIffl
I total (P, 66-P, 72) J. This method compares the actual response shape with a plurality of basic response shapes prepared in advance, selects the optimal rule from among the plurality of adjustment rules for the matched basic response shapes according to the actual response shape or its transition tendency, and I l) A method of modifying control parameters.

Incidentally, as this type of method, for example, ``Measure'' Technique (September 1988) and +310 Self-Tuning by Expert Method (P, 52 to P, 59) J are known.

[Problem to be solved by the invention]

However, among the above conventional techniques, when the PID control parameters are manually tuned.

There were problems such as the adjustment work took a long time and the tuning results were subject to individual differences among the adjusters. In addition, when tuning is performed using an identification test signal, the control amount fluctuates due to the application of the identification test signal, resulting in problems such as quality deterioration and abnormal conditions, especially in plants with high nonlinearity. occurs. Furthermore, in response to changes in the dynamic characteristics of the controlled object, the optimum values of control parameters cannot be obtained unless an identification test is performed each time, resulting in poor operability. Further, the expert method has the disadvantage that the number of adjustment rules increases and the memory capacity increases.

The present invention has been made in view of the above circumstances, and provides an auto-tuning method for automatically optimizing control parameters without requiring human labor and without applying an identification test signal to a controlled object. The purpose of this invention is to provide a PID controller.

[Means to solve the problem]

In order to achieve the above object, the present invention provides a PTI controller that performs proportional control, integral control, and differential control of a controlled object, by inputting a target value and a controlled amount of the controlled object, and controlling the controller when the target value changes or when a disturbance occurs. A control response shape recognition means that obtains three evaluation indices from the control amount response shape at the time of application, qualitatively evaluates each of the three evaluation indices, and based on this evaluation, a correction value of the I''ID control parameter is determined. A control parameter correction means to be estimated and two of the three evaluation indicators.
The control performance dissatisfaction degree evaluation means evaluates the estimated correction value of the PID control parameter according to the degree of dissatisfaction of the two evaluation indicators with respect to the one-standard control specification.

Further, when the two evaluation indicators do not satisfy the target control specification, the control parameter modification means.

The present invention is characterized in that a controllability determining means is added to each of the above-mentioned means to function as a control performance dissatisfaction evaluation means and a control parameter adjustment value calculation function, respectively.

[Effect]

” The controlled variable response shape recognition means detects l31i! of the controlled variable response shape when the control deviation exceeds a predetermined value. l'l is started, thereby making it possible to continuously detect changes in the IJh characteristics of the controlled object. The control parameter modification means applies fuzzy inference to tune the control parameters in a manner similar to the thinking of a skilled operator, and adjusts the previous and current values of the damping ratio and period to overshoe 1 of the control variable response shape. By qualitatively evaluating the period ratio, which is the ratio to the
A method is used to estimate the optimal values of control parameters.

As a result, optimal tuning of control parameters can be achieved in the same way as if it were done by a skilled operator. The control performance dissatisfaction level evaluation means evaluates the extent to which the current control performance is unsatisfactory with respect to the 11 standard control specifications, and weights the fuzzy inference results according to the degree. Thereby, the control parameters can be stably converged. Further, the control performance determining means repeatedly executes the control parameter modifying means until the overshoot amount and the damping ratio satisfy the target control specifications. Thereby, control performance can be maintained in an optimal state.

〔Example〕

An embodiment of a PID controller according to the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the functional configuration of the automatic tuning method for control parameters of this embodiment. In FIG.
The control deviation obtained by comparing with V is calculated by I) ID, and the result is input to the controlled object 2 as the manipulated variable MV. Pl: D controller 1 auto tuning machine f+
13 is a control amount response shape recognition function 4. Control performance judgment function 5. Control parameter modification function 6. It is comprised of a control performance dissatisfaction evaluation function 7 and a control parameter adjustment value calculation function 8.

Further, the control parameter modification function 6 includes a control parameter modification coefficient estimation function 6a and an adjustment rule 6b.

Next, each of the above functions will be explained. The control response shape recognition means 4 keeps the target value SV and the control 11) V open at all times, and after the control PV has settled to the target value S■, the control response shape is changed to 411111 when the control deviation exceeds a predetermined value. Begin to. At the same time as the start of observation, a search for the extreme value of the control variable pv is performed, and the observation is ended when the control value pv has settled to ]-1 target value Sv.8. The overshoot 111° damping ratio and period ratio are determined from the obtained plurality of extreme values and occurrence times. How to obtain these values will be explained with reference to FIGS. 2 and 3. FIG. 2 is an example of the time response of the control ipv when the 11 target value S■ changes stepwise from Yo to Yl at time to.

At times tt* tz and ts, the extreme values X, 1. X
2. A case is shown in which X3 is cursed and stabilizes in one hour t4.

The overshoot ff1E, damping ratio [], and period T at this time are determined by the following equations.

E: (XL-Yl)/(Yt-Yo) D=(Xt-Xl)/(X3 Xl)T=t s-
t 1 Therefore, the period ratio R is determined by the following equation, assuming that the previous value of the period is T1.

R=　T　h　/　T Furthermore, in Fig. 5, the 11 target value Sv is below YO.

υj is an example of the time response of the WIIt control amount Pv when a step disturbance DB is applied to the input end of the control target 2, and the time jl
+j2+j3+j4 are extreme values Xt and X21, respectively
The case where X5yx1 appears and settles at 1 hour t6 is shown. At this time, the overshoot amount E + damping ratio and period'
r is determined by the following formula.

E=CYo-Xl)/(Xj-Yo) D=(Xa-X4)/(Xs-Xz)T==
t4-tz If an extreme value does not appear in the controlled variable response shape when the target value changes, or if a second extreme value does not appear in the controlled variable response shape when a disturbance is applied, set the overshoot amount to a negative value and the period. are each set to zero. In addition, if only one extreme value appears in the controlled variable response shape when the target value changes, or if only two extreme values appear in the controlled variable response shape when a disturbance is applied, 1j() and The time difference between is used as the period.

The control performance judgment function 5 determines that the obtained overshoot amount and damping ratio are respectively [if one standard control specification is added to i+'11,
The control parameters are assumed to be at their optimum values, and tuning is completed. If the overshoot Lt and damping ratio do not satisfy any one of the target control specifications, modify the control parameters (operate function 6 and control performance dissatisfaction evaluation function 7).

Next, the control parameter modification coefficient estimation function 6a using fuzzy inference will be explained. overshoot J11
．． In order to qualitatively evaluate the magnitudes of the damping ratio and the period ratio, membership functions as shown in FIGS. 4 to 6 are defined. E(i) (i=1-5), D(j,) in the figure
(j = 1 to 3) and R(i) (i = 1 to 3) are constants that define membership functions, respectively, and PB
, IBM, ZE, and NB are names given to the membership functions to qualitatively evaluate the size, and each has the meaning described below.

PB: Po5itiVe BigP M: P
o5iLjve MediumZ I", : Zsr
.

N T3 : Negat-ive [3jgThe vertical axis in the figure is the membership value, which represents a qualitative degree. P for various control response shapes created using these membership functions.

An example of the adjustment rule 6b for each control parameter I and D is shown in FIG.

[In 21, for example, in the case of rule 2, E, D.

CKP if R is PB, PM, and PB, respectively.

This means that CTT and CTD become NB, Nrf, and ZE, respectively, and the former is called the condition part and the latter part is called the conclusion part. Here, E is an overshoot amount, D is a damping ratio, R is a period ratio, CKP is a proportional gain, CTI is an integral time, and CTD is a correction coefficient related to a differential time.

FIG. 8 shows membership functions for converting qualitatively determined correction coefficients of control parameters into quantitative values. C(i) (i = 1 to 4) in the figure is a constant that defines the form of the membership function.

PB, ZE, and NB are names given to membership functions to qualitatively represent the magnitude of the control parameter correction coefficient, and correspond to the names used in FIG. 3.

Moreover, the vertical axis of FIG. 5 is the membership value.

Hereinafter, the method for determining the correction coefficient of the control parameter will be explained using the case where Rule 2 and Rule 3 shown in FIG. 7 are applied as an example. FIG. 9 shows a method for determining the proportional gain correction coefficient CKP by fuzzy inference. Controlled variable response shape recognition function 4 1j) Rare overshoot t" ri
Qualitative degrees of E n, damping ratio Do, and period ratio Ro are determined using the membership functions shown in FIGS. 4 to 6. In rule 2, Gcp, Gdm, Grp respectively.
, in rule 3, Gep, Gdra, Grz, respectively.
becomes. An intersection set (minimum value) operation is performed for each rule to determine the degree of suitability of each rule.

In rule 2, Grp and in rule 3, Grz are obtained as respective degrees of fitness. Next, the membership function of the conclusion part of each rule is weighted by the suitability of each rule, their union (maximum value) calculation is performed, and the value of the center of gravity is set as the proportional gain correction coefficient CKPo.

The respective correction coefficients CTI and CTD for the integral time and differential time are determined in the same manner.

Next, the control performance dissatisfaction evaluation function 7 will be explained.

FIG. 10 shows a processing block diagram.

The degree of dissatisfaction with the target control specifications for the amount of overshoot and the damping ratio is defined by membership functions as shown in blocks 7a and 7b. value, DR is the target control specification upper limit value of the damping ratio (lower limit value is zero)
It is. The vertical axis of the figure is the membership value, and the overshoot Iit E and damping ratio ■ obtained by the control variable response shape recognition function 4, which represent the degree of dissatisfaction, are input into blocks 7a and 7b, respectively. In step 6rt4, the degree of dissatisfaction with respect to the specifications is determined, and in block 7c, the maximum value of these is selected to obtain the degree of dissatisfaction with control performance. This control performance dissatisfaction degree and the weighting coefficient lower limit value WL are compared in block 7d, and the maximum value thereof is output as the weighting coefficient W.

The lower limit value WL is provided to prevent slow convergence of the control parameters.

The control parameter adjustment value calculation function 8 calculates the value obtained by multiplying the weighting coefficient obtained by the control performance dissatisfaction degree evaluation function 7, the control parameter correction coefficient obtained by the control parameter correction function 6, and the current value of the control parameter. The current value of the PID control parameter is added to the corrected value of the control parameter. A! 6 Determine the value.

FIG. 11 shows a schematic processing flow diagram of the auto-tuning function 3. In block 10, Sv and PV are manually input at regular intervals, and each time, in block 11, a status flag indicating the processing status of the auto-tuning function 3 is determined. When the status flag is O, the monitoring status of the control response, when the status flag is 1, the station measurement evaluation status of the control response, and when the status flag is 2, it is the control parameter calculation status. If the status flag is 0, block 12 determines whether the control deviation exceeds a predetermined value. When the control deviation exceeds a predetermined value, the state flag is set to 1 in block 13, and the control response i-measurement state is entered. When the control deviation does not exceed a predetermined value, the control response monitoring state is maintained.

If the status flag is 1 in block 11, an extreme value search for Pv is performed in block 14, and this process is continued until PV is settled to Sv in block 15 (if! measurement complete).

When the evaluation is completed, the results of the extreme value search in block 14 are used to determine the evaluation index (overshoot amount, damping ratio).

(period ratio, etc.) is determined in block 16, and state flag 2 is set in block 17, and the process moves to a control parameter calculation state. The processing flow up to this point corresponds to the control response shape recognition function 4 described above.

Next, if the status flag is 2 as a result of the determination in block 11, then block 18 determines whether the observed control response is in the optimal state or not, and whether the evaluation index determined in block 16 satisfies the target control specification. Determine whether or not. This process corresponds to the control performance determination function 5 described above. Only when the control response has not reached the optimal state, block 19 sequentially obtains a weighting coefficient, block 20 a control parameter correction coefficient, and block 21 a control parameter adjustment value. The obtained control parameter adjustment values are used for control calculations of the PID controller 1.

Block 19 is the control performance dissatisfaction evaluation function 7. Block 20 is the control parameter correction coefficient inference function 6a.
and adjustment rule 6b, and block 21 corresponds to the control parameter adjustment value calculation function 8. block 18
If it is determined that the state is the optimum state and the processing in block 21 is completed, the state flag is reset to O in block 22.

Return to the monitoring state of the III control response.

Next, the PID controller according to the present invention is
The results of applying the dead time characteristics to the controlled object are shown in the 12th section.
As shown in the figure. This figure shows the time response of the controlled variable Pv when the target value S V is changed, and (a) shows the PID according to the present invention.
When using a controller.

(b) shows each O1-tuning result when the control performance dissatisfaction evaluation function is not used.

In (a), it can be seen that the target control specifications (overshoot: 3 to 7%, damping ratio: O to 0.5) are stably reached after three trials. On the other hand, in (b), the control parameters were excessively corrected in the second trial, so that subsequent trials are cursed with an oscillatory tuning process.

As described above, according to this embodiment, it is possible to auto-tune the control parameters with a small number of adjustment rules, and there is an effect that the convergence of the control parameters can be stably performed during repeated trials.

FIG. 13 shows a functional configuration regarding a control parameter straight tuning method in a PID controller according to another embodiment of the present invention.

FIG. 14 shows a schematic processing flow diagram of this auto-tuning function. In FIGS. 13 and 14, parts that are the same as or equivalent to those in the embodiment shown in FIGS. 1 and 11 are designated by the same reference numerals. The difference from the embodiment shown in FIG. 1 is that, except for the control performance determination function 5, the control parameter correction function 6. A control performance dissatisfaction level evaluation function 7 and a control parameter adjustment value calculation function 8 are made to function.

According to this embodiment, it is possible to obtain an effect equivalent to that of the above-mentioned embodiment, but in order to execute all functions even when the R,l control performance satisfies the [1 standard control specification], the above-mentioned embodiment Microcontroller load compared to ! The disadvantage is that the C becomes a'G.

In addition, in each of the above embodiments, as the third evaluation index of the control response waveform recognition function, the Xγ'' time of the controlled variable (for example, the first extreme value generation time when the target value changes) or the settling time of the controlled number A ratio with the previous +01 value may be used. Also,
In the control parameter correction function 6, the membership function is triangular, but it is not necessarily limited to this shape.
A quadratic curve or an exponential curve may also be used. Moreover, not only the form of the membership function but also the number thereof may be arbitrarily set.

〔Effect of the invention〕

As described in detail below, the present invention provides the following effects.

(1.) Since the PID control parameters in the PID controller can be automatically tuned, the adjustment work by the operator can be significantly reduced, and individual differences in adjustment results can be eliminated.

(2) Since no identification signal is used, control parameters can be optimally tuned without disturbing the controlled object.

(:3) Changes in the dynamic characteristics of the controlled object can be detected quickly without any manual effort, so the optimal control characteristics can be maintained at all times.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the functional configuration of an auto-tuning method for control parameters in an embodiment of the PIr) controller according to the present invention, and FIG. An example of the controlled variable response shape when a step change in value, FIG. 3 is an example of the controlled variable response shape when a step disturbance is applied, and FIGS. 4 to 6 are membership functions for overshoot amount evaluation and members for period ratio evaluation. Figure 7 is a graph showing an example of an adjustment rule, Figure 8 is a graph showing a membership function for control parameter correction coefficients, and Figure 9 shows how to obtain control parameter correction coefficients by fuzzy inference. Graph, FIG. 10 is a block diagram showing the processing of the control performance dissatisfaction evaluation function, FIG. 11 is a schematic processing flow diagram of the auto-tuning function according to this embodiment, and FIG. 12 is a graph showing the auto-tuning results according to this embodiment. , 13th [4
14 is a block diagram showing a functional configuration of an auto-tuning method for control/control parameters in a controller according to another embodiment of the present invention, and FIG. 14 is a schematic processing flow diagram of the auto-tuning function. 1... PID controller, 2... Controlled object, 3.
...Control parameter auto-tuning mechanism, 4...
Controlled amount response shape recognition function, 5... Control performance judgment function,
6... Control parameter correction function, 6a... Control parameter correction coefficient estimation function, 6b... Adjustment rule, 7.
...Control performance dissatisfaction evaluation function, 8...Control parameter adjustment agent PtJ! l! r Small I eye j force f Face inspection 2 Diagram -・- Anatomy one pin imperial sentry 1 Ki i 已! 1st stitch fV ~-
-1'J4fF (jL Sheng's #tOr:jJ $40 #50 formula% formula% me1Z's dreamω tea I4aki

Claims (12)

[Claims] 1. In a PID controller that performs proportional control, integral control, and differential control on a controlled object, the target value and control amount of the controlled object are input, and three evaluation indicators are calculated from the control amount response shape when the target value changes or when a disturbance is applied. control response shape recognition means to seek; control parameter modification means for qualitatively evaluating each of the three evaluation indicators and estimating a modified value of the PID control parameter based on the evaluation; and two of the three evaluation indicators. and a control performance dissatisfaction level evaluation means for evaluating the estimated correction value of the PID control parameter according to the dissatisfaction level of the two evaluation indicators with respect to the target control specifications. 2. Estimating the corrected value of the PID control parameter based on the relationship between three evaluation index evaluation means for qualitatively evaluating the magnitude of each of the three evaluation indices and the three qualitatively evaluated evaluation means. 2. The PID controller according to claim 1, further comprising an adjustment rule for controlling the PID controller. 3. The control performance dissatisfaction evaluation means compares the dissatisfaction levels of the first and second evaluation indicators with respect to the respective target indicators among the three evaluation indicators, and determines the weighting coefficient as the maximum value thereof. The PID controller according to claim 1. 4. Among the three evaluation indexes, the first evaluation index is the amount of overshoot, the second evaluation index is the damping ratio, and the third evaluation index is the cycle ratio, which is the ratio of the previous value of the cycle to the current value. The PID controller according to claim 1, characterized in that: 5. 5. The PID controller according to claim 4, wherein the third evaluation index is a rise time ratio defined as a ratio between a previous value and a current value of the rise time. 6. 5. The PID controller according to claim 4, wherein the third evaluation index is a provisional time ratio defined as a ratio between a previous value of provisional time and a current value of provisional time. 7. In a PID controller that performs proportional control, integral control, and differential control on a controlled object, the target value and control amount of the controlled object are input, and three evaluation indicators are calculated from the control amount response shape when the target value changes or when a disturbance is applied. control response shape recognition means for determining the value of the PID control parameter; control parameter modification means for qualitatively evaluating each of the three evaluation indicators and estimating a modified value of the PID control parameter based on the evaluation; control performance dissatisfaction evaluation means for evaluating the estimated correction value of the PID control parameter according to the degree of dissatisfaction of the two evaluation indicators with respect to the two target control specifications; A PID characterized in that it is provided with controllability determining means for functioning the control parameter correction means, control performance dissatisfaction level evaluation means, and control parameter adjustment value calculation function when the control parameters are not satisfied.
controller. 8. Estimating the corrected value of the PID control parameter based on the relationship between three evaluation index evaluation means for qualitatively evaluating the magnitude of each of the three evaluation indices and the three qualitatively evaluated evaluation means. 8. The PID controller according to claim 7, further comprising an adjustment rule for controlling the PID controller. 9. The control performance dissatisfaction evaluation means compares the dissatisfaction levels of the first and second evaluation indicators with respect to the respective target indicators among the three evaluation indicators, and determines the weighting coefficient as the maximum value thereof. The PID controller according to claim 7. 10. Among the three evaluation indexes, the first evaluation index is the amount of overshoot, the second evaluation index is the damping ratio, and the third evaluation index is the cycle ratio, which is the ratio of the previous value of the cycle to the current value. The PID controller according to claim 7, characterized in that: 11. 11. The PID controller according to claim 10, wherein the third evaluation index is a rise time ratio defined as a ratio between a previous value of rise time and a current value. 12. Claim 1, wherein the third evaluation index is a provisional time ratio defined as a ratio between a previous value of the provisional time and a current value.
PID controller described in 0.