JPS63219002A - Adjusting device - Google Patents

Abstract

PURPOSE:To set an optimum PI arithmetic parameter in a short time by introducing an area ratio, related to a process quantity or a deviation signal, following to an interval between a prescribed points of time, and obtaining the PI arithmetic parameter based on an evaluation index, including the area ratio. CONSTITUTION:A waveform observing means 31 in a parameter tuning means 3 observes the waveform of the process quantity PV or of the deviation signal DV, and operates an area A1, related to the process quantity or the deviation signal from the point of time t1 of the generation of a first peak of a signal waveform till the point of time t2 of the generation of a sequential second peak, and the area A2, related to the process quantity or the deviation signal from the point of time t2 till the sequential point of time t2+(t2-t1), and obtains an area ratio AR=A2/A1, and outputs it as one of the evaluation indexes, representing a waveform. A parameter arithmetic means 33 calculates the PI arithmetic parameter based on the some number of the evaluation indexes including the ratio AR, given by the means 31, so as to be the responding target of a goal setting means 32. Thus, the PI arithmetic parameter is changed into the optimum value.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a regulating device used for process control;
More specifically, the present invention relates to an auto-tuning adjustment device that automatically adjusts at least proportional (P) and integral (I) calculation parameters to optimal values.

(Prior art) In the process P1 control device used for feedback control, the P+ calculation parameters are currently set manually based on the long-standing knowledge and experience of the process operator or instrumentation engineer. It is. However, manual settings may cause temporary or steady process operation disturbances during process startup, load fluctuations, unexpected disturbances, or systems with nonlinear gain characteristics. Depending on the situation, it may have caused economic losses.

Therefore, a controller has been proposed in which the P+ calculation parameter is automatically tuned. The main auto-tuning controllers that have been proposed so far are as follows.

(i) Connect the auxiliary controller in parallel to the main controller, increase the gain of the auxiliary controller to cause vibration, and from the two amplitudes and frequencies, 2iegl@r,
Determination of P1 calculation parameters based on the so-called z-N limit sensitivity method by Nieholg (1971 Proceedings of the Society of Instrument and Control Engineers Vol. 1. [k l'ss ~ P611
Research on adaptive control systems using the marginal sensitivity method.

Toshiyuki Kitamori).

(H) A limit cycle is generated using an on/off generator, and the optimal P1 calculation parameter is determined from its amplitude etc. (Proceedings of the 12th Academic Conference of the Society of Instrument and Control Engineers, 1971) HIT~PHJ I'
l1) Automatic setting type adaptive controller Sumi,
Fukuda).

(tii) A pattern recognition means (performance measurement) is provided to observe the behavior of the controlled variable, and this recognizes disturbances in the control system such as disturbances that occur randomly without causing any disturbance to the process, and uses this recognition result. The optimal P+ calculation parameters are determined so that the pattern recognition result approaches the response model by comparing it with the optimal response model (Hardcover ADAPTIVE C0NTR0
L SYSTEMS.

PERGAMOII PRESS, Published by 1i3 P1
~P mouth).

(iv) The technique described in (in) above is used, and the overshoot amount (excessive amount), the damping (attenuation rate) determined from the peak value, and the period (period) are used to recognize the response pattern due to disturbance in the control system. ) (US1'AT No, JJO2
, 326).

(Problems to be Solved by the Invention) In the above-mentioned prior art, the items (i) and (i) both require that the control system be It is necessary to forcibly apply a disturbance (identification signal), which poses a problem in that it has a considerable influence on the process.

Items (m) and (K) are superior in that they do not use identification signals and do not have any influence on the process. Here, the item (iv) provides one method for concretely realizing the item (ii), but the
As shown in the figure, the first deviation between the control set value and the process input value. Second. The third peak value E, , E, , L is detected, the overshoot amount is -L/L, and the damping is (L
-L)/(E, -E2), and the period is determined from the time between the first peak and the third peak.

For this reason, it is necessary to search for the third peak, but searching for the third peak takes a lot of time, and
If the third peak does not appear, a pseudo peak is determined by estimation, which poses a problem in that the response pattern cannot necessarily be recognized faithfully. Also, the third
Since the amplitude value of the peak is generally smaller than that of the 1j-th second peak, there is also the problem that it is easily affected by noise unrelated to disturbances from the control system.

The present invention basically utilizes the technique of the above-mentioned (in) term, and eliminates the need to search for the third peak to recognize response patterns due to disturbances in the control system. The response pattern can be recognized accurately in a short time and without being affected by noise, etc., and the optimal P1
The present invention provides an auto-tuning adjustment device that can set calculation constants.

(Means for Solving the Problems) FIG. 1 is a basic functional block diagram of the apparatus of the present invention. In the figure, l is a block indicating a controlled object (process), and its dynamic characteristics change due to changes in production volume, changes in control target values, various disturbances, etc. 2 performs at least proportional (P) and lf1 (1) calculations on the deviation signal DV between the process amount Pv from the controlled object 1 and the control target value Sv, and outputs the obtained manipulated variable (MY) to the controlled object 1. PIft+control means, 3, P1 control means 2 P, 3 according to the waveform pattern of the process amount PV or deviation signal DV;
This is a parameter tuning means for tuning I calculation parameters.

In this parameter tuning means 3, 31 observes the waveform of the process signal or deviation signal Dv, detects the occurrence time t1 of the first peak Dvl of the signal waveform, the occurrence time t of the second peak DV, and 32 is a target setting means for setting a target value (model) serving as a controllability target; 33 is an evaluation from the waveform observing means 31; This is a parameter calculation means that calculates PI calculation parameters such that the index becomes a response target set in advance in the goal setting means 32.

(Operation) The waveform observation means 31 in the parameter tuning means 3 observes the waveform of the process quantity Pv or the deviation signal Dv, and determines the time point t1 (t, -t1) when the first peak of the signal waveform occurs.
) and output it as one of the evaluation indicators representing the waveform pattern.

The parameter calculation means 33 calculates P and I calculation parameters based on several evaluation indicators including the area ratio AR given from the waveform observation means 31.

(Embodiments) Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 2 is a configuration block diagram of the adjustment device according to the present invention. In the figure, reference numeral 4 denotes an adjustment device, which inputs the process amount PV and control target value Sv from the controlled object 1, and receives the operation signal M.
Output v to the controlled object 1.

In this adjustment device 4, 41 is a multiplexer that sequentially selects and inputs the process amount Pv, various analog signals el + control target value Sv, 42 is a comparator that receives the signal selected by the multiplexer 41 as one input, and 43 is a microprocessor. , signals from the comparator 4z, etc. are input. 44 is a D-converter that converts the digital signal from the processor 43 into an analog signal; 45 is a sample-hold circuit that holds the output of the D-converter at a predetermined timing; the output is the operation signal Mv to the controlled object 1;
becomes. 46 is a 110 port, 47 is a RAM that stores various data, etc., 48 is a system ROM that stores programs for main operations performed by the microprocessor 43,
Reference numeral 49 is a ROM storing, for example, a program created by a user according to a process, and 50 is a display keyboard, which is coupled to the microprocessor 43 via a data bus BS.

Here, the system ROM 4g includes a microprocessor 43, a waveform observation means 31, a parameter calculation means 33, a P1 control means 2. Data functioning as the goal setting means 32 is stored, which includes a program for performing a closing function, data of an optimal response model of the controlled object 1, and the like. In addition, system AROM 48. R
OM 49 is one person'') (7) It is also good to share the ROM.

FIG. 3 is a flowchart showing an example of the operation of the apparatus configured as described above.

The microprocessor 43 starts with the system ROM
48, the process variable PV, control target value Sv, and other analog signals e applied to the multiplexer 41 are sequentially selected and extracted, and these signals are sent to the comparator 42, microprocessor 43, and D. /A/D formed by A converter 44
Each is converted into a digital signal by a conversion loop (step 1).

Next, it is determined whether the deviation DV between the process amount PV and the control target value Sv is within a predetermined setting (1i111DB) (Step 2). When DV>DB, the process moves to the next Step 3. In step 3, according to the system program that functions as the waveform observation means 31 in FIG.
Process amount PV or process fiPV and control target value Sv
Observe the waveform of the deviation nv from the observed waveform and analyze the pattern of the observed waveform. Furthermore, when the pattern analysis is completed, the evaluation index of the observed waveform is calculated (steps 4 and 5).
).

FIG. 4 is an explanatory diagram of a waveform observation method in this waveform observation means 31. Here, the response waveform of the deviation signal DV when the load fluctuates is shown. Although it changes greatly when the load fluctuates, the change gradually becomes smaller due to the operation of the PI control means 2 and eventually converges to zero.

The waveform observation means 3I observes the behavior of this deviation signal DV, and continues from 1+ (,
Time point at which the second peak (nv2) is observed (time point at which the second peak occurs) Area related to the deviation signal DV up to 11 ^1
, ts + (
ts-Lt) (the time when Ls-Lt has elapsed from ts), the area A related to the deviation signal DV is calculated. Then, using these areas A and M, AI+A
Calculate the ratio (AR) of . Examples of formulas for calculating area A, A, and area ratio ^R are formulas (1), (2),
It is shown in equation (3).

A, = 7″ (DV-DVs) dt (
2)t! Here, the calculation formula (1) is used to calculate the area A, which is shaded in FIG. 4, and the calculation formula (2) is used to calculate the area A, which is shaded in FIG. It is something to seek.

In addition, the maximum overshoot amount (overshoot) OVS (
4) Obtain by the calculation formula of Eq.

Further, the vibration period TP is determined by the calculation formula (5).

Tp= 2・(t2−t+) (
5) The area ratio AR, overshort OVS, and vibration period TP thus obtained are given to the parameter calculation means 3 as evaluation indicators representative of the observed waveform.

In the apparatus of the present invention, the waveform observation means 31
is the first observed waveform. Area ratio AR using the area Al#^ of the observed waveform that continues from the point of occurrence of the second peak
is given to the parameter calculation means Go2 as one of the evaluation indicators of the observed waveform, and there is no need to detect the third peak, thus providing the effect of not being affected by noise.

FIG. 5 is a waveform diagram for explaining the influence of noise included in observed waveforms. In a feedback-controlled control system, the response waveform of the process variable PV or deviation v that generally occurs due to load fluctuations, changes in the control target value, or other disturbances initially changes significantly.
2nd Peak Scout DV.

Although Dv is relatively easy to detect, its amplitude becomes small near the third peak. For this reason, when detecting the third peak, the amplitude of the mixed nozzle NS may be detected as the third peak, giving an incorrect evaluation index.

As in the present invention, first. Time of occurrence of second peak 1. ,
1. If the area ratio ﾾ藺 calculated using the area AI and A2 of the observed waveform following from is used as an evaluation index, even if noise MS is mixed into the observed waveform DV as shown by the broken line in Figure 5, the signal waveform The area of the area, for example, the shaded surface 1lls, is not affected by noise contamination because the portion that is larger and the portion that is smaller than the signal waveform when no noise is mixed cancel each other out.

In FIG. 3, after completing the calculation of the evaluation indicators (AR, OVS, TP) of the observed waveform as described above in step 5, the parameter calculation means 33 calculates each evaluation index given from the waveform observation means 31 and the target setting. P1 is compared with each data of the optimal response model stored in advance in the means 32.
Calculate the calculation parameters and convert this PI calculation parameter to P
1 Setting (tuning) to control means 2 (step 6
). Details of this step 6 will be described later.

The PI control means 2 performs a P1 calculation using the PI calculation parameters set and changed by the parameter calculation means 33 (step 7), and outputs the calculation result to the controlled object 1 as an operation signal Mv (step 8).

In addition, in step 2, when the deviation DV is within the predetermined set value range (deviation set value for self-tuning activation) DB (DV< DB), control is considered to be good.
Proceeding to step 7, P1 control calculation is performed using the already set P1 calculation parameters. In step 4, if the pattern analysis of the observed waveform is not completed (
1st. If a certain amount of time is required until the second peak is observed, and the analysis is still in progress, the process similarly proceeds to step 7.

FIG. 6 is a flowchart showing details of step 6 in FIG.

First, the parameter calculation means 33 starts with the waveform observation means 3.
It is determined whether the area ratio value AI obtained in step 1 is smaller than "0" (step $1). Thereby, it is determined whether the P and I calculation parameters currently set in the P1 control means 2 are in the category described below. That is, if the area ratio ratio is smaller than "0", it indicates a non-oscillatory response characteristic, and it is determined that the controllability based on the current P and I calculation parameters is in category A, and the target setting means 32 sets A calculation is performed to change the P and I calculation parameters so that they approach the set target values (step 62). Area ratio AR≧
0 (“NO” in step 61), calculate the value R of (integral calculation constant T1)/(vibration period Tp) (step G3), and judge the magnitude of this value R (step 6).
4), that is, if R<8.2, it is determined that the currently set P1 calculation parameters are in 8 categories, and the process moves to step 65. Further, if 0.2≦R≦0.4, it is determined that it is in the C category, and in step 66, R
If >0.4, it is determined that it is in the 0 category, and the process moves to step 67. In this way, the area ratio AR,
la motion period Tp, depending on the size of R, A, B, C,
The division into four divisions D is based on past empirical rules, and the concept of each division and an example of the calculation formula for each division are shown in FIG.

In FIG. 7, the horizontal axis is the value of R, and the vertical axis is the area ratio AR. In this figure, the hatched area (near area ratio AR = 11.2, R = 1.2) is the target value area, and the proportional calculation parameter PB and integral calculation parameter Ti are determined for each section. By performing predetermined calculations for
■Calculation parameters are calculated. Note that the present invention can be similarly applied to a controller that also includes differential (D) calculation, and FIG. 7 also shows the calculation formula for obtaining the differential calculation parameter Td for reference. be.

In each calculation formula shown in Figure 7, FBI and PIN are the proportional calculation parameters for the first time this time, and Til and Ti2 are the current proportional calculation parameters.

Next integration calculation parameter (integration time), Eovs is error overshoot, and EAR is error area ratio.

Equations (1^) and (2A) in section A are step ε
2, the proportional calculation parameter PB2.

The integral calculation parameter Ti2 is the error area ratio EAR.

Depending on the value of the error overshoot Eovg, values are determined that are each greater than the current parameters. If EAR, Rows is O, PI3. Ti
n is the same value as the current FBI, Til.

Equations (IB) and (2B) in section B are applied in step 65, and the proportional calculation parameter PB2゜integral calculation parameter Ti2 is set to R10,2 (here R
is smaller than 0.2). Values are calculated that each decrease from the current parameter.

Hereinafter, the (IC) formula for category C will be explained in the same way.

Equation (2C) is applied in step 66, Equation (ID) in the partition, Equation (2D) is applied in step 67
applicable.

In step 68, the proportional calculation parameters and integral calculation parameters obtained in any of steps 62, 65, 65, and 67 are reset to the P1 control means 2.

Through the above-described operation, the PI control means 2 is able to finally reach the target f-mount with optimal controllability, no matter what the proportional calculation parameters and integral calculation parameters set therein. It will be automatically adjusted.

In the above explanation, the waveform observation means 3I is configured to observe the waveform of the deviation signal DV between the process quantity PV and the control target value sv. However, assuming that the control target value is constant, , the waveform of the process amount PV may be observed. Further, in the above description, the adjustment means having a PI control means was taken as an example, but the present invention may also be applied to an adjustment means having a PID control means.

(Effects of the Invention) As explained above in detail, the present invention provides the following as one of the indicators for grasping the response pattern of the process quantity or deviation signal accompanying changes in control target values, load fluctuations in processes, etc.
The first response waveform. Process amount or deviation, area related to the signal following the occurrence of the second piston ^1. Introducing the ratio of A, and based on the evaluation index including this area ratio, PI
The calculation parameters are changed to the optimum values. Therefore, according to the present invention, it is possible to realize an auto-tuning adjustment device that is not affected by noise and can set optimal P1 calculation parameters in a short time.

[Brief explanation of the drawing]

Figure 1 is a basic functional block diagram of the device of the present invention, Figure 2 is a block diagram of the configuration of the device of the present invention, Figure 3 is a flowchart showing an example of the operation of the device in Figure 2, and Figure 4 is a waveform. An explanatory diagram of the waveform observation method in the observation means, FIG. 5 is a waveform diagram for explaining the influence of noise included in the observed waveform, FIG. 6 is a flowchart showing details of step 6 in FIG. 3, and FIG. FIG. 8 is an explanatory diagram showing the concept of each division determined by the parameter calculation means and the calculation formula for each division. FIG. 8 is an explanatory diagram showing an example of a response pattern recognition method in a conventional device. 1... Controlled object (process), 2... PI control means. 3... Parameter tuning means, 31... Waveform observation means, 32... Target setting means, 33... Parameter calculation means. Figure 6

Claims (2)

[Claims] (1) PI control means for performing at least proportional (P) and integral (I) calculations on the deviation signal between the process amount from the controlled object and the control target value and outputting the obtained manipulated variable to the controlled object; The above-mentioned PI
A control device comprising: parameter tuning means for tuning P and I calculation parameters of the control means, wherein the parameter tuning means is configured to adjust the timing at which the first peak (DV_1) of the process quantity or deviation signal (DV) occurs ( The area (A_1) related to the process amount or deviation signal (DV) from t_1) to the time of occurrence (t_2) of the second peak (DV_2), and the time of t_2+(t_2-t_1) following from the time of occurrence (t_2). The process amount or the area (A_2) related to the deviation signal (DV) is calculated, and the area ratio (AR) of the obtained areas (A_1) and (A_2) is used as one evaluation index representative of the waveform pattern. 1. An adjustment device comprising: a waveform observation means for outputting a waveform observation means; and a parameter calculation means for calculating P and I calculation parameters based on an evaluation index including an area ratio given from the waveform observation means. (2) In the waveform observation means, area (A_1), area (
A_2) and the area ratio (AR) are calculated according to equations (1), (2), and (3). ▲There are mathematical formulas, chemical formulas, tables, etc.▼(1) ▲There are mathematical formulas, chemical formulas, tables, etc.▼(2) AR=A_2/A_1(3)