JPH03253905A - Identifying method for control parameter - Google Patents

Abstract

PURPOSE:To identify the control parameter in a short time even in the case the process quantity is scarcely varied by using the sum of an overshoot period at the time when a manipulated variable is varied from '1' to '0' and a period in which the process quantity ascends from a prescribed value. CONSTITUTION:A manipulated variable of a level S1 is given to a process, and at the time point when the process quantity exceeds P1 and reaches P2, the manipulated variable is varied to a level S2, and thereafter, when the process quantity returns to P2, the manipulated variable is varied to the level S1, and the manipulated variable is maintained in the level S1 until the process quantity reaches P3. In this case, the time t1' until amplitude a' of the process quantity, and the process quantity PV become the level P1, and thereafter, reach a second level P2, and the time t2' until they reach second P2, and thereafter, reach P3 are measured, an equivalent waste time L, etc., are derived, and by referring to these values, control parameters of a proportional constant, a differential constant, and an integration constant are derived. In such a way, even in the case the process quantity is scarcely varied, the control parameter can be identified.

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial Application Field> The present invention relates to an improvement in a method of applying vibration to a process and identifying its #JIXI parameter.

<Prior Art> To control a process, a manipulated variable is determined by performing proportional, differential, and integral calculations on the deviation between a target value and a process amount, and the process is controlled using this manipulated variable. In order to properly control the process, it is necessary to accurately determine control parameters such as proportional constants, differential constants, and integral constants for performing proportional, differential, and integral operations.

Fig. 3 shows the configuration of a device for determining control parameters. In Fig. 3, a set value SO specifying the position of vibration and a process amount PV are inputted to the identification section 1, and a manipulated variable for oscillating the process variable. is output to process 2. When the command signal is input, the identification unit 1 alternately applies two levels of manipulated variables to the process 2, and identifies control parameters from the state of change of the process variable at that time. This control parameter is set in a control section (not shown), and the control section executes a PID calculation based on this control parameter to obtain an operation amount, and outputs it to process 2 to control this process 2.

FIG. 4 shows changes in manipulated variables and process variables for identifying control parameters to be applied to the process.

When a command is issued by an operator's operation such as pressing a push button switch at time T1, the manipulated variable is set to 1 (full scale), and after a slight delay, the process amount increases to a predetermined value at 0 time T2. When the value SO is reached, the manipulated variable is set to zero. This SO specifies the position where vibration is to occur. The process amount continues to increase for a while, but eventually starts to decrease at time 0 T3, when the process amount (
When SO-b) is reached, the manipulated variable is set to 1 again. The process amount begins to increase and reaches SO at time T4.

During this time, the process quantity undergoes one cycle of oscillation.

If the amplitude of the process amount at this time is a, and the times t1 and t2 are defined by the following formula, tl = (T3-72) t2 = (T4-73) The equivalent dead time of the process, the process gain K, and the time constant T The relationship can be determined as shown in the following equation using empirical rules based on waveform analysis.

L= (1-b/a) book (tl *
t2 / (tl + t2)) ・・・・・・
・・・・・・・・・・・・・・・・・・（１〉L＊に／
T = (a-b) - (2) This formula (1) is the value referenced when calculating the differential constant and integral constant, and (2
) formula serves as a reference value when calculating the proportionality constant.

<Problem to be solved by the invention> However, even if the manipulated variable is changed to O at T2 in Fig. 4,
A case may occur in which the process amount does not change much from the point of view of SO. In this case, there is a problem that the control parameters cannot be identified because the process l does not become (so-b).

<Object of the Invention> An object of the present invention is to provide a control parameter identification method that can identify control parameters even when the change in process amount is small.

Means for Solving the Problem> In order to solve the above problem, the present invention provides a process with an operation amount of level S1, and when the process amount exceeds 21 and Pl
When the process amount reaches P3, the manipulated variable is changed to the level S2, and when the process amount returns to the P2, the manipulated variable is changed to the level S1, and the manipulated variable is kept at the level S1 until the process amount reaches P3. control parameters are identified from the time from when the process quantity exceeds P1 until it returns to P2, the time from P2 to P3, and the vibration amplitude of the process quantity. This is what I did.

Effect> Control parameters can be identified even if the changes in process quantities are small.

<Embodiment> FIG. 1 shows a flowchart illustrating an embodiment of the control parameter identification method according to the present invention. When a command signal for identifying a control parameter is input, this flow is started. First, the operating lid is brought to level S1. Sl is usually 1
(full school) is 0 Next, wait until the process amount passes through Pl and becomes larger than Pl, and when it becomes larger, set the manipulated variable to 82 (normally 0).The relationship between Pl and Pl is b
Set the bias amount to Pl = (Pl - b).Furthermore, wait until the process amount P becomes smaller than P2 again, and change the manipulated variable to 81 (=1
) and maintain that value until the process amount Pv becomes larger than P3. P3 is set to P3 = (Pl +b) using the bias amount. Then, the time t1° from when the process amount amplitude a process amount PV reaches level P1 to the second level P2, and the time t2- from when it reaches Pl for the second time until it reaches P3 are measured. , a = a - + b t1 = t1 t2 = t2, substitute them into the above formulas (1) and (2> to find the dead time etc., and refer to these values to calculate the proportional constant, differential constant, integral Try to find constant control parameters.

FIG. 2 shows changes in the process amount and manipulated variable according to the embodiment shown in the flowchart of FIG. 1, where (A) shows the process amount and (B) shows the change in the manipulated variable. Time T10
When a control parameter identification command is issued, the manipulated variable S1
(=1> is output to the process. The process amount increases with a delay of time related to the dead time of the process. At time 0 T11, the process amount exceeds Pl, and at time T1
When it reaches 22 at 2, the manipulated variable is set to 32 (-0), and the process variable begins to decrease after a slight overshoot.

When the process amount returns to Pl at time 713, the manipulated variable is again set to Sl (=1>. After a slight delay, the process amount increases again until the process amount exceeds P3 at time 0 T14. is held at 81.

From the change in process amount at this time, a-=amplitude of process amount t1=(713-T11nt2-=
(T14-713), and using a = a + b tl = tl tl = t2, use the above equations (1) and (2>) to calculate the value of the equivalent dead time of the process, and the equivalent waste time and process gain. The relationship between K and time constant T is determined, and from these relationships, the proportional constant, differential constant, and integral constant are determined using known methods. tl in Fig. 4 indicates the process when the manipulated variable changes from 1 to O. On the other hand, tl - in Figure 2 is given by the sum of the overshoot period when the manipulated variable changes from 1 to O and the period during which the process variable falls below a constant value. is the sum of the periods during which the
Since equation (2) is an empirical equation, using tl - instead of tl will not result in a large error.In addition, in actual identification, the change in process amount itself changes smoothly as shown in Figure 2. Since there are few numbers, there is always an error in identification. Furthermore, in the conventional example shown in FIG. 4, the process amount is s.

The manipulated variable was changed when the process amount became (So-b), whereas in FIG. 2 according to the present invention, the manipulated variable was changed when the process amount reached 81, so a=a +b. Further, tl and tl - are equal because they are both the sum of the overshoot period of the process amount when the manipulated variable changes and the period until the process l reaches a constant value.

In this embodiment, a case has been described in which the process amount increases during the identification process, but identification can be performed in a similar manner even when it is desired to reduce the process amount in a short time.

<Effects of the Invention> As explained above in detail based on the examples, in this invention, when the process amount reaches 22 in the rising process, the manipulated variable is set to 82, and when it becomes Pl in the downward process, 81, the control parameters are determined from the time from when the process amount reaches Pl until it reaches Pl in the descending process, the time from then until it reaches P3, and the amplitude of the process amount. Therefore, there is an effect that the control parameters can be identified in a short time even when the change in the process amount is small.

Also, if you want to raise the temperature in a short time using a heating furnace, etc.
The effect is large because the amount of decrease in the process amount during identification is small.

[Brief explanation of drawings]

FIG. 1 is a flowchart showing an embodiment of the control parameter identification method according to the present invention, FIG. 2 is a characteristic curve diagram for explaining its operation, and FIG. 3 is a configuration diagram of a device for identifying control parameters. FIG. 4 is a characteristic curve diagram for explaining a conventional control parameter identification method. 1...Identification part, 2...Process.

Claims (1)

[Claims] In a control parameter identification method in which a manipulated variable of level S1 is given to a process and the process variable exceeds P1, When P2 is reached, the manipulated variable is changed to level S2, and then when the process amount returns to P2, the manipulated variable is changed to level S1, and the manipulated variable is kept at the level S1 until the process amount reaches P3. S1, and control parameters are determined from the time from when the process amount exceeds P1 until it returns to P2, the time from when the process amount returns to P2 until it reaches P3, and the vibration amplitude of the process amount. A control parameter identification method characterized in that the control parameters are identified.