JPS58168106A - System for discriminating instability of process control - Google Patents

Abstract

PURPOSE:To discriminate the instability of a process control applicable to the adaptive control, by utilizing poles of the oscillation of a process variable and an operation variable, the oscillating period and the attenuation factor of the amplitude of oscillation. CONSTITUTION:The three poles X1-X3 and their generated times tX1-tX3 are obtained from the timewise change of the process variable X, the three poles Y1-Y3 and their generated times tY1-tY3 are obtained from the operating variable similarly, the period (tX3-tX1) of the waveform X and the period (tY3- tY1) of the waveform Y are compared and whether the ratio of both is close to the unity in a certain range or not is discriminated. Further, whether the attenuation factor of the amplitude of the waveform X is larger than a prescribed attenuation factor or not is discriminated according to the formula of discrimination. Similarly, the same is performed for the amplitude of the waveform Y. When the three conditions are established, it is discriminated that the control operation is instable and hunting takes place.

Description

Detailed Description of the Invention The present invention provides a method for automatically determining whether or not a control operation has become unstable in a closed loop control system that performs feedback control on a process. Adaptive control is known, which achieves good control by adapting the control parameters of a joint needle to the characteristics of the process. One example of such adaptive control is, for example, in feedback control of a process that has been operating stably until then, when the control operation becomes unstable due to a change in the characteristics of the process and hunting occurs, the hunting is stopped. 1 As a result, it is possible to change the control parameters of the controller to a weaker one and return to stable control operation again, but the present invention relates to control operation/hunting detection means that can be used for the above-mentioned applications. Now, this type of control operation instability determination method, in other words, hunting detection method, can be thought of as a color in principle, but no concrete means are known yet. Therefore, the present invention was made to meet the above-mentioned technical needs, and therefore, the purpose of the present invention is to destabilize process control that can be applied to the above-mentioned adaptive control. The principle of the present invention is to provide a discrimination method.The principle of the present invention is as follows. In other words, in process feedback control, the process variable changes oscillatedly with respect to time, and the oscillation decays slowly, and the manipulated variable remains the same (oscillates and decays slowly, and the process variable It is based on the principle that if the vibration period and the vibration period of the process variable are equal, it is determined that hunting has occurred.Unless the above three conditions are met, for example, just because the process variable vibrates, This is not determined to be hunting, as it may simply be due to noise.Note that damping means the amplitude of each half cycle of vibration.

Therefore, the main points of the configuration of the present invention are: a first means for measuring a process variable and detecting a predetermined number of extreme points in its vibration;
The means of s2 for detecting a predetermined number of extreme points in the vibrations of the operating variables from the control unit that are lost in the process; 1. The noodle and value at the time of detection of the extreme points detected by the means of @1; and the second means. a third means for determining whether or not the deviation between the vibration period of the process variable and the vibration period of the manipulated variable is within a predetermined limit using the detection value and the value of the extreme point detected by the third means; A fourth means for determining whether the damping rate of the vibration amplitude of the process variable is larger than a predetermined damping rate using the pole point, , QIII & output value detected by the means of 1, and the means of II2 above. and a fifth means for determining whether or not the damping rate of the vibration amplitude of the manipulated variable is larger than a predetermined damping rate using the detected value at one point detected by the third to third means. When each of the fifth means determines that the control operation has become unstable, it is determined that the control operation has become unstable.

Next, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an example of the application of the destabilization determination method according to the present invention. In the figure, 1 is the process, 2 is the investigation section, and 3 is the instability determining means (hunting detection means).
, 4 is a control parameter adjustment unit, and 5 is a discrimination parameter setting unit. In FIG. 1, the adjustment unit 2 detects a process variable 1, and so-called feedback control is performed. The destabilization determining means 3 monitors the process variable Send it to the adjustment section 4.

In the adjustment section 4, assuming that the adjustment section 2 is performing PID control, for example, the adjustment section 2 sends a command W to change the control parameters (proportional gain, integral time constant, derivative time constant II).
The control parameters are changed in order to prevent hunting from occurring. Note that the determination parameter setting section 5 is for setting a parameter U necessary for detecting hunting in the destabilization determining means 3. As mentioned above, it can be used as the discrimination means (hunting detection means) 3.

Next, the operating principle of the present invention will be specifically explained with reference to $211.

No. 2110) is a graph showing an example of a temporal change in the process variable X, and No. 2 (1) is a graph showing an example of a temporal change in the manipulated variable Y at the same time.

The values of the three poles (peaks of peaks or valleys) at the second wJ pi) and the time of occurrence N txt
# txg p lx3 is found, and fil Ys # Y2 e of similar three poles in 2nd g-)
It is assumed that Ya and the occurrence time tyte jY2 #IY3 have been found.

At this time, the period of the waveform of X (1'xs 1 work 1) and Y
Compare the period of the waveform (tya-tyx) and determine whether the ratio of the two is close to 1 within the tolerance range &1.0, that is, determine whether the following formula is satisfied.Furthermore, It is determined whether the attenuation rate of the amplitude of the waveform X is greater than a predetermined attenuation rate α. In other words, it is determined whether the following equation is satisfied.

Waveform Y
It is determined whether the attenuation rate of the amplitude of is greater than a predetermined attenuation rate α. That is, it is determined whether the following equation is satisfied.

l Ys-Yz I >α·1Yz-Yll □ When the above three conditions are satisfied, in the present invention, it is determined that the control operation has become unstable, that is, hunting has occurred.

From the above, in the present invention, it is necessary to monitor the temporal changes of the process variable It will be understood that it is essential to perform predetermined calculations on .

Jli31iU is a block diagram showing an embodiment of the present invention. In the figure, the block S surrounded by a broken line can be used as it is as the destabilization determining means 3 in the 1611th block. DXl is a detection unit for the first extreme point in process change #X (for example, Xl in FIG. 2(a)), and sends the detected value (Xl and t-work 1) to the discrimination calculation unit E. Similarly, DX2 is the IIk part of the second pole (for example, X2 in Figure 2 (a)), and DX3 is the third pole (41
i1 is, for example, the detection section of Xa) in FIG.

DYI, L)Y2, DY3 are the first . Second. Yl at the third pole (Fig. 2←).

Y2 e Ya ), and sends each detected value (Yl and 'YlpY2 and lY2 p Ya and tya) to the discrimination calculation section E.

The discrimination calculation unit E is one pole detection unit DXI to DX3, DY
, ~DYa, a predetermined calculation is performed on the detection value of each pole point, and a determination result V of the presence or absence of hunting is thereby output. The hm calculation unit E is set by the setting unit 5 with a determination parameter U necessary for performing a predetermined calculation to determine the presence or absence of hunting. Note that some of the parameters U' are also variables and meters necessary for detecting the extreme points, so they are further transmitted from the discrimination calculation unit E to the extreme point detection units DX1 to DX1.
The OR sent to DX3 and DY1 to DY3 is a reset signal.Whether it is the process variable Therefore, a place that is not an extreme point may be mistakenly detected as an extreme point, and in order to avoid this, special measures are required for the detection means, which will be described later.

Next, the above-mentioned discrimination parameter U will be explained.

Once the process to be controlled is determined, the oscillation period Ti of the variable X (or manipulated variable Y) when hunting occurs can be roughly predicted, so this *motion jllllTh can be predicted and a single parameter Toshite decided. The genus of the parameter U is listed below. ... Assumed vibration period DX ... Minimum amplitude for determining that X is vibrating (If the amplitude of X is smaller than DX, it is considered to be due to noise and is not considered as vibration. DY...Minimum amplitude for determining that Y is vibrating (If the amplitude of Y is smaller than DY, it is considered to be due to noise and is not considered as vibration) C1...・Tolerance of deviation between X and Y vibration cycles ξ2...Tolerance between actual vibration cycle and expected cycle α...Minimum damping rate of amplitude to be determined as unstable (If the amplitude attenuation rate is α or more, it is determined to be unstable.) Among these, Tl p Tl is also used as U.

If the fiww degree of Th is poor, the allowable wA difference -2 is increased or the use is discontinued. It is good to assume TI and set it in a pot. These Tl e Tl t Th # D
An example of XeDY is shown in FIG.

Next, regarding the operation of detecting the extreme point and its subsequent discrimination operation, 1.
This will be explained with reference to FIGS. 12 and 3.

■When the polar point detection unit DX1 starts the polar point determination operation,
Using the time span of T1 as a unit, we examine the change in the pulse variable is not updated in the remaining period of the time width Tl after the time of its maximum value or minimum value, the said time is lxl%, the amplitude of the change WkX at that time is Xl, and these values are the first It is sent to the discriminator calculation unit E as a representation of the extreme point and stored in memory. From the graph of 1m (No. 2-0) where the detection of the extreme point of Happy I was #I, 13), it is the third point in the same plane as the memory. If the second point is not detected even after one time span (0 @111ell#) detection is 1ltTlt scissors, TlF), then it is assumed that Bootes Demon 1lkx is not fibrillating! Nm and return to above (■).

■When the third point is detected by the inspection tooth DXs, calculate that value and -m (xl and @XZ in the second cabinet).
m to be memorized.

■ Poem ml for which detection of the second point has been completed (gH&!I's t
8), similar to 0 above, issue 113 (Xistx
Even if s) is 1Ill, if * is not raised, the book returns to ■.

O In this way, if the first point is the top of the mountain, then the three points are (mountain, valley, mountain), and if the first point is the bottom of the valley, then the three points are (valley, mountain, pot). Seek.

ΦOnce the three points are found in this way, the discriminator calculation layer compares the values of the adjacent points, and if the difference is not greater than the distance between the two points, the first point is drawn and the second point is determined. , II3 as the first . Move to the second step and find a new pole of $3. For example lXl-X21>DX or lX2-Xa I>D
X... If (1) holds true, the discriminant calculation unit E proceeds to the next 1ill-specific calculation; however, if the above (1) does not hold, the obtained three poles (Xi #X2 t Xa ) to CX2 e Xl
#*] and replace 1 Find the third pole *.For the manipulated variable Y, find the three poles in exactly the same way: IYI-Y21≧DY or 1Y2-Yal≧DY...
(Determine ω, and if it does not hold, remove the first extreme point, move the second and third points to 1.$2, and find the third extreme point again.

(2) If three extreme points are detected for both the push variable X and the manipulated variable Y, use them to check whether the following discriminant holds true.

1Xa-Xzl>α・lX2 Xll...
... (4) IYs-Yzl>α・1Y2-Yll
...... (5) If all the above discriminants hold true, it is determined that hunting has occurred, and the discriminant calculation sE outputs the discriminant result ■.0However, among the above three discriminants, If any one of the conclusions is not met, the next extreme point is searched uniformly with ■.

■If a single point is not detected during the search operations described in ■ and ■ above, return to ■ with the same aircraft as in ■ and ■ above.

[Phase] Discrimination condition (3) # (4) t (5
In addition to ), if the period Th of unstable vibration is given with high accuracy for some reason, the following conditions can be added to improve the reliability of the discrimination.

In addition, T1. The method of determining Tl has been described above, but it is preferable to determine it as follows in relation to the accuracy of the predicted vibration period Th. Regarding Tl, set it as close as possible to the noise.

The role of the time span TI in the search for the peak point will be explained with reference to FIG.

1114- is a graph showing a change over time in the process variable X, and it is assumed that a peak M, which is 5 IkA and is extremely high, has occurred due to noise. However, by appropriately determining the time width T1 of the pole search, this M point will not be connected to the pole. In this case, since the Iidairen of the process variable X is updated at the point when the point M is passed within the tortoise time width Tl, the point M is not determined to be the extreme point.

Figure 85 is a graph showing how the process variable Since it is determined that the process variable to terminate the judgment based on the previous data. the result,
The pole searching operation of II2 does not continue indefinitely, and the adaptability of w*m operation of unstable vibration is guaranteed.

111611 no 11i 101fil+hfh7'a
This is a graph showing an example of the temporal change of the seven X variables x and the manipulated variable Y. With reference to these graphs, a method of determining the minimum attenuation rate α, which is one of the discrimination parameters, will be explained.

In general, the response of the Puwara Setsu variable WkX and the repetition variable @Y is the 65th
The damping shown in is considered to be slI-like, and in that case, the damping rate aν/al is said to be approximately 0.25, and the 0 damping rate, which refers to the damping of such vibrations as 25-damping, is equal to this 0 damping rate. If the value is less than .25, the vibration will be on the stable side, and if it is very small, it is said to be overkill, and the control operation will lack adaptability. Conversely, if the damping factor is greater than 0.25, the vibrations become unstable. When the vibration oscillates as shown in FIG. 7E, the damping factor is greater than 1.0, and when the damping is poor as shown in FIG. 8, the damping factor is close to 1.0.

Taking this into consideration, the minimum attenuation rate α is set to 1
．． A value slightly smaller than 0 (for example, f14 is 0.8 or 0.
7)! It is appropriate to set this value, and if you want to avoid vibration to an extreme extent, you can set it to a small value.0 In an actual control system, when the vibration of the manipulated variable Y accompanies the vibration of the process variable There are both cases where you proceed with
Since the polarity of the vibration may be reversed, the relationship between the poles during vibration may be as shown in the 9th iEl and the 10th g.

In other words, FIG. 9 is a graph when the vibration phases of process variable X and manipulated variable Y are out of phase by 180 degrees, and the 1θth
Although the figure shows a graph where the phase relationship is unclear, it is needless to explain that hunting can be detected by the present invention even in these cases. This point is also an advantage of the present invention.

According to the destabilization determination method of the present invention, the determination conditions are simple and the capacity required for the determination is small, so it can be implemented in a small-capacity digital control system at the microcomputer level. In addition, since the judgment is based on the vibration status of both the process variable and the manipulated variable, it is within two cycles of vibration (in detail, LL1,5
It has the advantage that hunting can be determined stably (without malfunctions due to noise) in a short period of time (the scrap period).

If the destabilization determination method according to the present invention is incorporated into a process feedback control method, it is possible to strengthen the ll1lIllllII operation as much as possible to the point just before hunting occurs, and to improve coordination. This is because even if hunting force is generated due to subtle changes in process variables, it can be quickly detected and the control parameters of the controller can be changed to prevent hunting.

In the application example shown in FIG. 1, the control parameter gas4 is configured to output a command to change the control operation to the first adjustment section 2. The present invention can be similarly utilized as a configuration in which this function is enhanced to provide a function to provide a setting value to be changed for the control parameter used in the adjustment section 2, and in that case,
It becomes possible to make the joint part 2 simple.

Furthermore, this invention can be applied not only to the commonly used PID feedback control system but also to a multi-St* feedback control system.

[Brief explanation of the drawing]

Figure 1 is an example of the application of the destabilization determination method according to the present invention (a block diagram shown, Figure 2 ←) is a graph showing an example of a temporal change in the process variable
3 is a block diagram showing an example of the present invention. FIG. 4 is a graph of process variation WkX to explain the role of time span T1 in pole search. FIG. 5 is a graph of the process variable X for explaining the role of the set time width T2, and FIGS. 6 to 10 are graphs showing examples of vibrations of the process variable X and the manipulated variable Y, respectively. Description of symbols 1...Process, 2...Adjustment section, 3.
...Instabilization determination means, 4...Control parameter adjustment section, 5...[II parameter setting section, X...process change il, Y. ...l-operated variable, U...1-discrimination parameter, ■...
・Discrimination result, W... Control parameter change command, DX1 to DX3... Extreme point detection unit of process variation WkX, DYI to DY3... Extreme point detection of operation error Y Section, E...discrimination calculation section. Agent Patent attorney Akio Namiki Agent Patent attorney Kiyoshi Matsuzaki 11I11
;X/IJI force sequence 5 Iko 7 Figure 6

Claims (1)

[Claims] 1) A method for determining when a control operation in process feedback control has become unstable, comprising a first means for measuring process control and detecting a predetermined number of extreme points in the vibration, and a second means for the process. 182 means for measuring the operating variables and detecting a predetermined number of extreme points at 1,000ii*; 1) the detection time and value of the extreme points detected by the first means; iI3 means for determining whether the deviation between the llwII circumference layer of the prism change and the oscillation period of the repetition variable is within a predetermined degree using the detection time and value of the pole; (114) to determine whether the attenuation rate of the process control 1118114m is greater than a predetermined attenuation rate using the detection value of the pole detected by the first means;
A fifth method for determining whether or not the attenuation rate of the SRS convergence on the operation surface is larger than a predetermined attenuation rate by layering the detected values of the extreme points detected by the second means (I5). A method of process control characterized in that when it is determined that each of the third to II5 means is present, it is determined that the control operation has become unstable. Destabilization discrimination method.