JP2839679B2 - Automatic tuning device for control parameters - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to an auto-tuning apparatus for automatically adjusting a plurality of parameters in components necessary for controlling a control target. About.

2. Description of the Related Art In a control device having an automatic adjustment function, an automatic adjustment method based on trial and error has conventionally been used for adjusting a parameter of a component that needs to be adjusted among components necessary for controlling a control target. Has been used. FIG. 3 shows an example of a tuning apparatus for adjusting the trial and error method.

In FIG. 3, the basic configuration of the control system includes a control device 2 including a controller 3 for calculating an operation amount for making a deviation between a target value of a control target and a control amount zero, and the control device 2
And a control target 4 controlled by the control unit. Since the control amount of the control target 4 is compared with the target value, the control device 2
Will be fed back. In order to adjust a plurality of parameters included in the controller 3, a control characteristic analysis unit (first unit) 11, a parameter estimating unit (second unit) 12, which checks characteristics of a control system including the control target 4; And an adjustment parameter setting means (third means) 13 are provided. The first means 11 includes a test signal generating means 11a, a waveform recognizing means 11b, and an adjustment end determining means 11c. A test run is performed by the test signal generating means 11a including the control target 4, a characteristic value of a waveform indicating the control characteristic is obtained by the waveform recognizing means 11b, and the result is adjusted by the adjustment end determining means 11c. As long as it is not determined, it is input to the parameter estimating means 12 to calculate an estimated value. Using the value estimated here, new setting for the controller 3 is performed again through the adjustment parameter setting means 13, and adjustment is performed again. The above operation is repeated until the adjustment end determining means 11c determines that the adjustment is completed, and the desired parameter adjustment is performed.

(Problems to be Solved by the Invention) The conventional auto-tuning device described above has the following problems.

(1) At the end of parameter adjustment by the auto-tuning device, the estimation result does not easily converge while oscillating around the optimal adjustment value, so that it takes time to adjust.

(2) In the case where there are a plurality of adjustment parameters and the values of the parameters are different, when the same characteristic value is obtained, it is not possible to make an accurate determination as to which one to select.

(3) In the initial stage in which the auto-tuning device adjusts the parameters, even if the adjustment target is adjusted to be similar, it takes a long time for the adjustment every time since the adjustment is performed again as a completely new control target.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an automatic tuning device for control parameters that can achieve a reduction in adjustment time and an improvement in accuracy.

[Configuration of the Invention] (Means for Solving the Problems) In order to achieve the above object, the present invention provides a method for controlling a plurality of control parameters in components required for controlling a control target. First means for examining characteristics of the control system including the object, second means for estimating more appropriate parameter values of the control system from the examined characteristics in accordance with a predetermined adjustment rule, and values of the parameters A control parameter auto-tuning device comprising a third means for newly setting the parameter to the component being adjusted, a fourth means for storing an estimation process of the second means and outputting a locus of parameter transition. Means, a fifth means for judging the adjustment state from the locus of the transition of the parameter, and a selection of the adjustment parameter when it is determined that the adjustment is necessary as a result of the judgment by the fifth means. And sixth means for correcting the estimation result of the second means based on each output of the fourth means, the fifth means and the sixth means. I do.

Further, the automatic tuning apparatus according to the present invention further comprises an eighth means for storing data indicating a result of adjustment of the control parameter in combination with data indicating a known operating condition based on the output of the seventh means; A ninth method of searching data stored in the means for data similar to the current operating conditions
And a tenth means for estimating an adjustment parameter based on the similar data retrieved by the ninth means.

(Operation) According to the above configuration, the convergence value of each parameter can be satisfactorily estimated from the adjustment trajectory up to the previous time by the seventh means, whereby appropriate control parameter adjustment can be quickly and accurately performed. .

Further, by using the above-mentioned adjustment data and data obtained before the adjustment, the auto-tuning device can be provided with a learning function, whereby the control parameters can be adjusted more quickly and appropriately.

(Embodiment) FIG. 1 shows a first embodiment of an auto tuning apparatus according to the present invention.
FIG. Control device 2 including controller 3 which needs to adjust parameters for controlling control target 4
On the other hand, the control characteristic analysis means (first means) 11 analyzes the control characteristic of the system including the control target 4 of the currently set parameter. This characteristic analysis is performed by giving a test signal such as a step waveform to the control device 4. This test result is input to the control characteristic analysis means (first means) 11 as a feedback amount of the control amount.
From the input control amount waveform, for example, the rise time,
Characteristic values such as settling time, overshoot, and offset are calculated, and the calculation results are provided to parameter estimating means (second means) 12. The parameter estimating means (second means) 12 estimates an adjustment parameter for adjustment by the control device 2 in accordance with a predetermined rule as described below. Here, two parameters of parameter 1 and parameter 2 are considered as adjustment parameters. The parameters 1 and 2 are, for example, parameters of a proportional (P) element and parameters of an integral (I) element in the PI controller.

(Rule 1a): Parameter 1 if there is overshoot
The value of.

(Rule 1b): If there is no overshoot, the value of parameter 1 is increased.

(Rule 1c): If the settling time is late, the value of parameter 2 is increased.

(Rule 1d): If the settling time is short, the value of parameter 2 is reduced.

The output of the second means 12 is input to a transition output means (fourth means) 14 of the adjustment parameter. This fourth means 14
The trajectory of the transition of the parameter in the adjustment process is obtained by the above, and the adjustment state is determined by the adjustment state determination means (fifth means) 15 based on the result. Adjustment state determination means (fifth
Means 15 determines the following contents based on the locus of transition of the value of each parameter.

(Decision a): The way of changing each parameter is the same many times.

(Decision b): The way of changing each parameter oscillates around a certain value.

(Decision c): The way of changing each parameter is small.

(Determination d): Fine adjustment is required for each parameter individually.

(Decision e): The control target is special, and adjustment cannot be performed by a predetermined rule.

Of the above determinations, in the case of (determination e), the adjustment parameter selecting means (sixth means) 16 selects to perform adjustment from adjustment parameters having higher importance for the control system. In the case of other judgment results of the adjustment state judging means 15, the adjustment correcting means (seventh means) 17 performs correction according to the following rules.

(Rule 2a): In the case of judgment a, it is determined that the rate of change of the parameter is small, and the rate of change is increased.

(Rule 2b): In the case of determination b, it is determined that the rate of change of the parameter is large, and the rate of change is reduced to perform fine adjustment. It also estimates the convergence value at the center of the vibration and outputs that value.

(Rule 2c): In the case of judgment c, the change is reduced to make final adjustment of the parameters.

(Rule 2d): In the case of the judgment d, the change of each parameter is reduced in order to finely adjust each parameter according to the judgment result.

According to the auto-tuning device configured as described above, it is possible to quickly and accurately adjust a parameter according to the state of adjustment.

FIG. 2 shows a second embodiment of the present invention. In this embodiment, an information storage means (eighth means) 18 for storing information on the control target 4 is provided.
The information storage means 18 stores the rating and control specifications of the control target 4,
It is assumed that the adjustment result is stored. It is assumed that the control target adjuster manually inputs the rating and the control specification, and the other adjustment results are automatically stored.
Searching means (ninth means) 19 searches the data stored in the information storage means 18 for data similar to the rating or control specification of the control target 4 currently being adjusted, and outputs the data. If there is a difference in the adjustment specification or rating currently being adjusted, parameter estimation means (tenth means) 20 performs estimation according to the following rules.

(Rule 3a): If the detected control specification is faster than that currently being adjusted, the value of parameter 1 is increased.

(Rule 3b): If the detected control specification is later than the one currently being adjusted, the value of parameter 1 is reduced.

(Rule 3c): If the detected rating is larger than that currently being adjusted, the value of parameter 2 is increased.

(Rule 3d): If the detected rating is smaller than that currently being adjusted, the value of parameter 2 is reduced.

Note that these rules are not uniform, and are appropriately changed depending on the type of the adjustment element to be adjusted. The value estimated by the parameter estimating means 20 is adjusted by the adjusting and correcting means (the seventh
Means 17). The adjustment correction unit 17 outputs the estimated value of the parameter estimation unit 10 as a setting value of the adjustment parameter only when a test run has not been performed for adjustment.

With the above configuration, the initial value of the adjustment parameter can be set to an appropriate value. In addition, to store adjustment record data,
A more appropriate value can be set. Therefore, it is possible to make settings close to the final adjustment result to some extent,
Adjustment time can be further reduced. Further, since a value that is not so appropriate is not set, a situation in which the control system becomes unstable can be eliminated.

〔The invention's effect〕

As described above, according to the present invention, a determination unit that determines data in an estimation process is provided, and correction is performed according to the determination, thereby reducing time for parameter adjustment,
High-precision adjustment can be performed. Also, by storing the adjustment result in advance in association with data such as control specifications and ratings known before adjustment, an appropriate initial value can be set.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of an auto-tuning device according to the present invention, FIG. 2 is a block diagram showing a second embodiment of the present invention, and FIG. 3 is a block diagram of a conventional auto-tuning device. It is. 2 ... control device, 3 ... controller, 4 ... controlled object, 11 ... control characteristic analysis means (first means), 12 ... parameter estimation means (second means), 13 ... adjustment parameter setting Means (third means), 14 adjustment parameter transition output means (fourth means), 15 adjustment state determination means (fifth means), 16 adjustment parameter selection means (sixth means) , 17 ... Adjustment correction means (seventh means), 18 ... information storage means (eighth means), 19 ... similar data search means (ninth means), 10 ... parameter estimation means (tenth means) Means).

Claims (2)

(57) [Claims] A first means for examining the characteristics of a control system including a control object in order to adjust a plurality of control parameters in components necessary for controlling the control object, and A control device comprising: a second means for estimating a more appropriate parameter value of a control system according to a predetermined adjustment rule; and a third means for newly setting the parameter value to a component being adjusted. The fourth means for storing the estimation process of the second means and outputting the locus of the parameter transition, and the fifth means for judging the adjustment state from the locus of the parameter transition. Means, a sixth means for selecting an adjustment parameter when it is determined that adjustment is necessary as a result of the determination by the fifth means, and the fourth means, the fifth means and the sixth means. Auto tuning device of the control parameter, characterized in that a and seventh means for correcting the estimation result of said second means based on the output of. 2. An eighth means for storing data indicating a result of adjusting a control parameter based on an output of said seventh means in combination with data indicating a known operating condition, and storing the data in the eighth means. Ninth means for retrieving data similar to the current operating conditions from the data, and tenth means for estimating the adjustment parameter based on the similar data retrieved by the ninth means are further provided. The control parameter auto-tuning device according to claim 1, wherein: