JPH0610763B2 - Auto selector adjustment device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial field of application) The present invention uses two or more controllers, and automatically selects the output signal of any one of the controllers according to the magnitude relationship. The present invention relates to an automatic selector adjusting device that outputs to a process. More specifically, the present invention relates to an automatic selector adjusting device using a controller having a self-tuning function for automatically adjusting at least proportional (P) and integral (I) operation constants to optimum values.

(Prior Art) FIG. 6 is a configuration block diagram of an automatic selector adjustment device. This device includes a first controller C ₁ and a second controller C ₂
And a selection circuit SEL for selecting the outputs of the controllers C ₁ and C ₂ . In the pipeline ₁ , for example, the first controller C ₁ inputs the process amount PV1 related to the discharge pressure m ₁ and outputs the control amount MV1 of the valve 2, and the second controller C ₂ outputs the suction pressure m ₂ The process amount PV2 related to the above is input and the control amount MV2 of the valve 2 is output.
Selection circuit SEL, the first, of the second controllers C _1, C ₂ of the output signal MV1, MV2, selects the output signal of the direction for example greater, controls the valve 2 by the output signal.

Now, the control deviation of the suction pressure m ₂ is large and the valve 2
When There being operated by the second controllers C ₂ of the output signal MV2, for example, of the control deviation increases discharge pressure m _1, the first controllers C ₁ of the output signal MV1 is the second of the controller If it becomes larger than the output signal MV2, the selection circuit SEL
By this, the output signal MV1 of the first controller C ₁ is automatically selected, and the control loop is switched by the discharge pressure m ₁ .

As described above, the automatic selector adjustment device performs the control operation for the process amount having the larger control deviation among the two process amounts, and can control both process amounts together. Therefore, the automatic selector adjusting device is often used, for example, in order to control the process control amount so as not to be above a certain upper limit value or below a certain lower limit value in terms of product quality and plant safety maintenance. .

(Problems to be Solved by the Invention) By the way, in the automatic selector adjusting device having such a configuration, the PI operation constants of the _first and _second controllers C ₁ and C ₂ are automatically adjusted to optimum values. If this is done, it is necessary to provide circuit means for performing self-tuning to each of the controllers C ₁ and C ₂ , which causes a problem that the configuration becomes complicated.

The present invention has been made in view of such problems,
The purpose thereof is to realize an auto selector adjusting device having a simple structure and stably operating.

(Means for Solving Problems) According to the present invention for solving the above problems, a plurality of controllers having a plurality of PI control blocks for performing at least proportional and integral calculations, and output signals from the plurality of controllers. Maximum, minimum,
A selection circuit that selects one output signal such as an intermediate value, and first and second drive circuits that are driven according to the signal selection of this selection circuit.
And a deviation signal obtained from any of the plurality of controllers via the first changeover switch, and the second changeover switch so that the process amount has an optimum response. PI of PI control block of any of the plurality of controllers
It is equipped with a self-tuning circuit that adjusts the arithmetic constants, and this self-tuning circuit observes the waveform of the deviation signal between the process amount or the process amount and the set value, and when the signal exceeds a predetermined value, the waveform pattern Of the overshoot amount OVS (= -E2 / E1) and the damping value DMP {= (E3-E2) / based on the peak information (E1, E2, E3) and the time information (t1, t2, t3) when the peak occurs. (E1-E
2)} and a vibration period Tp (= t3−t1), a waveform observing unit, a target setting unit that sets at least an overshoot amount and a damping value that are targets of controllability of the process, and the waveform observing unit. The amount of overshoot (O
VS), damping value (DMP) and vibration period (TP)
And a target value set by the target setting means, respectively, and parameter calculating means for calculating a parameter to be set in the PI control block, wherein the parameter calculating means is obtained by the waveform observing means. P is divided into two sections according to whether the damping value (DMP) is larger than a predetermined value or not, and one of the sections is currently adjusted.
The ratio R (= T) between the integration constant (TI) set in the I control block and the vibration period (TP) obtained by the waveform observation means.
I / TP) is divided into a plurality of calculation sections, and the difference between the target value of the overshoot given by the target setting section and the overshoot amount obtained by the waveform observing section is divided for each section. Error overshoot amount (Eovr), an error damping value (Edmp) represented by the difference between the target value of the damping value given by the target setting means and the damping value obtained by the waveform observing means,
Ratio (R) of the integration time (TI) and the vibration period (TP)
By applying different arithmetic expressions using at least one of the above, at least the proportional and integral constants are calculated, and the obtained proportional constant,
It is an automatic selector adjusting device characterized by setting an integration constant in a PI control block.

(Example) FIG. 1 is a configuration block diagram showing an example of an apparatus according to the present invention. In the figure, C ₁ is the first controller and C ₂ is the second controller.
Each of these controllers has addition means 11 and 21, and PI control blocks 12 and 22 that perform PI calculation on the deviation signal from the addition means.

STC is a self-tuning circuit, which inputs the deviation signal ε ₁ or ε ₂ obtained from each adding means 11, 21 of the first controller C ₁ or the second controller C ₂ via the first changeover switch SW1. Then, the control signal for setting the PI calculation constant is output to the PI control block 12 or 22 of the first controller C ₁ or the second controller C ₂ via the second changeover switch SW2. Each of the first and second changeover switches SW1 and SW2 is in contact with the contact a or b in synchronization with the drive signal from the selection circuit SEL.
Connected to.

FIG. 2 is a configuration block diagram of the self-tuning circuit STC used in the apparatus of FIG. Reference numeral 31 is a waveform observing means for observing the waveform of the deviation signal ε applied via the first changeover switch SW1, which is an overshoot amount OVS of the deviation signal ε, a damping value DMP, and a vibration cycle T.
It has the function of finding p. Reference numeral 32 is a target setting means for setting a target value which is a target of controllability, and here, at least an ideal overshoot amount OVS in the process 2 is set.
And the damping value DMP are set. Reference numeral 33 is a parameter calculation means for calculating the PI constant by calculation, and calculates the PI calculation constant so that the overshoot amount and the damping value obtained from the waveform observation means 31 approach the target value set by the target setting means 32. Each of these means 31, 32, 33
Are all realized by a program by a microcomputer installed in the controller.

FIG. 3 is an explanatory diagram of a waveform observation method in the waveform observation means 31. The waveform observing means 31 first obtains a deviation ε between the set value SV and the process amount PV, and when the deviation ε becomes larger than a predetermined value ΔE set in the target setting means 32 in advance, starts the waveform observation. Assuming that this deviation ε changes with time as shown in the figure, this deviation ε
When the waveform becomes larger than ΔE, the waveform observation is started, the peak values E1, E2, E3 of the waveform are detected, and the times t1, t2, t3 until the peak values are reached are measured. Using these values, the overshoot amount OVS and damping value DM
P and the vibration cycle Tp are obtained by the following arithmetic expression, and the arithmetic result is given to the parameter arithmetic means 33.

Although the example of FIG. 3 shows a case where the waveform pattern observed by the waveform observing means 31 oscillates around the 0 level, it may fluctuate around a certain value (for example, a set value). , Peak information E1 to E3 of the observed waveform
Will use its absolute value in the following calculations. Further, the waveform observing means 31 detects each peak value of the observed waveform, but if no peak is detected within a predetermined time after starting the observation of the waveform (for example, in the case of a non-oscillating system), Is the peak value E1
It is assumed that the peak value E2, E3 corresponding to the second and third peak values is processed as 0, and so on.

Here, the overshoot amount OVS and the damping value DM
The significance of defining P by the following arithmetic expression is that those values can be obtained only by the peak value information E1, E2, E3 of the waveform.

The Tp = t3-t1 parameter calculation means 33 is set in the PI control block 12 or 22 of the first controller C ₁ or the second controller C ₂ based on the calculation result obtained by the waveform observing means 31. What is the relationship between the current proportional constant and integral constant with respect to the target proportional constant and integral constant for realizing the overshoot amount and damping value set by the target setting means 32? Recognize and apply different calculation formulas according to some divisions determined by the respective relationships, and calculate at least new proportional constants and integration constants (changes from the current value). The new proportional and integral constants obtained in this way are used in the second changeover switch SW2.
Is reset to the PI control block 12 or 22 via the
The control block 12 or 22 then performs a PI operation based on this newly set constant and outputs an operation signal to the process 1.

FIG. 4 is a flowchart showing an example of the operation of the parameter calculation means 33. Here, an example in which four divisions are determined is illustrated.

First, the damping value DM obtained by the waveform observation means 31
It is determined whether P is smaller than "0" (step 1). As a result, it is determined whether the P and I constants currently set in the PI control block 2 are in the A section described later.
That is, when the damping value DMP is smaller than “0”, it shows a non-oscillating response characteristic, and it is determined that the current controllability by the P and I operation constants is in the A category, and is set in the target setting means 32. A calculation for changing the P and I calculation constants is performed so as to approach the target value (step 2). If the damping value DMP ≧ 0 (“No” in step 1), the value R of (accumulation calculation constant T _I ) / (vibration period T _P ).
Is calculated (step 3), and the magnitude of this value R is judged (step 4), that is, if R <0.2, it is judged that the currently set PI calculation constant is in the B category, Go to step 5. If 0.2 ≦ R ≦ 0.4,
It is judged to be in the C section, and in step 6, if R> 0.4, it is judged to be in the D section, and the process proceeds to step 7. In this way, according to the damping value, the vibration period, and the size of R, the four divisions of A, B, C, and D are based on the empirical rule so far. The arithmetic expression in is shown in FIG. In FIG. 5, the horizontal axis is the value of R and the vertical axis is the damping value DMP.
Has been taken. In the area where the target values are in the hatched area (near the damping value DMP = 0.2, R = 0.2) in this figure, the predetermined values for obtaining the proportional operation constant PB and the integral operation constant Ti shown for each section are set. By performing the calculation, the P calculation constant and the I calculation constant are obtained so that the controllability approaches the target value from any of the sections. The self-tuning circuit STC that performs such an operation is
The present invention can be similarly applied to a controller including a differential (D) operation, and FIG. 5 also shows, for reference, an arithmetic expression for obtaining the differential operation constant Td.

In each arithmetic expression shown in FIG. 5, PB1 and PB2 are present and next proportional arithmetic constants, Ti1 and Ti2 are present and next integral arithmetic constants (integral time), Eovr is error overshoot, and Edmp is error damping. .

Here, the error overshoot Eovr is the difference between the overshoot amount OVS given from the waveform observing means 31 and the ideal overshoot amount OVS set in the target setting means 32, and the error damping Edmp is the waveform. It is the difference between the damping value DMP given from the observing means 31 and the ideal damping value DMP set in the target setting means 32. If the observed waveform pattern shows an ideal response, these are the differences. Each error value Eov
Both r and Edmp are zero.

The formulas (1A) and (2A) in the category A are the same as the step 2
Applied in, proportional calculation constant PB2, integral calculation constant Ti2
Is required to be a value that increases from the current constant according to the values of the error damping Edmp and the error overshoot Eovr. If Edmp and Eovr are 0, PB2 and Ti
2 is the same value as PB1 and Ti1 this time.

The equations (1B) and (2B) in the section B are the same as those in step 5
Applied in, proportional calculation constant PB2, integral calculation constant Ti2
Are required to be values that decrease from the current constant at a ratio of R / 0.2 (here, R is smaller than 0.2).

Hereinafter, in the same manner, (1C expression), (2
The expression (C) is applied in step 6, and the expressions (1D) and (2D) in the section D are applied in step 7.

In step 8, the proportional operation constant and the integral operation constant obtained in any of steps 2, 5, 6, and 7 are reset in the PI control block.

By the above operation, the PI control block 2
Whatever value the proportional calculation constant and integral calculation constant are set to, the controllability will be automatically adjusted so as to finally reach the optimum target value.

In the device configured as described above, the first controller C
_If the output e _{1 of 1} is larger than the output e ₂ of the _second controller C ₂ , the selection circuit SEL selects the output e ₁ of the _first controller C ₁ and outputs it. To the operation output signal MV. Further, in this state, the selection circuit SEL is
The first and second changeover switches SW1 and SW2 are both contact a
Connected to the self-tuning circuit STC
The first inputs a deviation signal epsilon ₁ obtained from Controller C _1, the controllers PI calculation constants that are set in the PI control block 12 of the C ₁ is controlled so that each becomes optimal target value To do.

Next, when the output e _{2 of} the second controller C ₂ increases from the above state to become larger than the output e _{1 of} the first controller C, the selection circuit SEL causes the second controller C this time. ₂ select the output e _2, and outputs this as an operation output signal MV. In this state, the selection circuit SEL has the first and second changeover switches.
Both SW1 and SW2 are connected to the contact b side, the self-tuning circuit STC inputs the deviation signal ε ₂ obtained from the second controller C _2, and the PI control block of this controller C _1.
The PI operation constants set in 22 are controlled so that they are optimal target values.

In the above explanation, as a self-tuning circuit,
Although the one shown in the block diagram of FIG. 2 is used, the waveform observing means 31 may observe the waveform of the signal from the process here. The self-tuning circuit may have another circuit configuration. Also, the first
Each of the second controllers may use the PID control block. Further, a plurality of controllers may be used, and the selection circuit may be a circuit for selecting any one of the maximum, minimum, intermediate value and the like among these controller outputs.

(Effects of the Invention) As described above, according to the present invention, since only one self-tuning circuit needs to be provided for a plurality of controllers, an automatic selector adjusting device having a simple structure and operating stably. Can be realized.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of an apparatus according to the present invention, FIG. 2 is a block diagram showing a self-tuning circuit, FIG. 3 is an explanatory view of a waveform observation method in a waveform observation means, and FIG. 4 is a parameter calculation. FIG. 5 is a flowchart showing an example of the operation of the means 33, FIG. 5 is an explanatory view showing the concept of each section defined by the parameter calculating means and an arithmetic expression in each section, and FIG. 6 is a block diagram of a conventional auto selector adjusting device. It is a figure. C _1, C ₂ ... Controller, SW1, SW2 ... changeover switch, SEL ...
Selection circuit, STC ... Self tuning circuit, 12,22 ...
PI control block.

Claims (1)

[Claims] 1. A plurality of controllers having a plurality of PI control blocks for performing at least proportional and integral calculations, and maximum, minimum, and maximum output signals from the plurality of controllers.
A selection circuit that selects one output signal such as an intermediate value, and first and second drive circuits that are driven according to the signal selection of this selection circuit.
And a deviation signal obtained from any of the plurality of controllers via the first changeover switch, and the second changeover switch so that the process amount has an optimum response. PI of PI control block of any of the plurality of controllers
It is equipped with a self-tuning circuit that adjusts the arithmetic constants, and this self-tuning circuit observes the waveform of the deviation signal between the process amount or the process amount and the set value, and when the signal exceeds a predetermined value, the waveform pattern Of the overshoot amount OVS (= -E2 / E1), based on the peak information (E1, E2, E3) and the time information (t1, t2, t3) at which the peak occurs.
Damping value DMP {= (E3-E2) / (E1-E
2)} and a vibration period Tp (= t3−t1), a waveform observing unit, a target setting unit that sets at least an overshoot amount and a damping value that are targets of controllability of the process, and the waveform observing unit. The amount of overshoot (O
VS), damping value (DMP) and vibration period (TP)
And a target value set by the target setting means, respectively, and parameter calculating means for calculating a parameter to be set in the PI control block, wherein the parameter calculating means is obtained by the waveform observing means. P is divided into two sections according to whether the damping value (DMP) is larger than a predetermined value or not, and one of the sections is currently adjusted.
The ratio R (= T) between the integration constant (TI) set in the I control block and the vibration period (TP) obtained by the waveform observation means.
I / TP) is divided into a plurality of calculation sections, and the difference between the target value of the overshoot given by the target setting section and the overshoot amount obtained by the waveform observing section is divided for each section. Error overshoot amount (Eovr), an error damping value (Edmp) represented by the difference between the target value of the damping value given by the target setting means and the damping value obtained by the waveform observing means,
Ratio (R) of the integration time (TI) and the vibration period (TP)
By applying different arithmetic expressions using at least one of the above, at least the proportional and integral constants are calculated, and the obtained proportional constant,
An auto selector adjusting device characterized by setting an integration constant in a PI control block.