JP4019350B2 - CONTROLLER WITH INTEGRATION UNIT AND ITS INTEGRATION METHOD - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an improvement of a PID controller used for control of a physical quantity such as temperature, flow rate or pressure in a chemical plant or a steel plant, or for positioning or speed control in an industrial robot, machine tool or information processing related equipment. The present invention relates to a PID controller having a function of changing the integral constant according to the magnitude of a deviation between a command value and a control amount.
[0002]
[Prior art]
In general, the control by the integral operation of the PID controller is effective in improving the deviation in the steady state that cannot be eliminated by the proportional operation. On the other hand, depending on the value of the integral constant, an undesired phenomenon such as overshoot, undershoot, or windup is caused, and the control characteristic is deteriorated.
[0003]
Also, when the manipulated variable reaches saturation due to the integral operation, a problem is similarly caused in the control characteristics.
[0004]
Therefore, in the past, to deal with this, even if the responsiveness was sacrificed, if the integral constant of the controller was set large, or if the operation amount was saturated, the integration operation was temporarily stopped, and the saturation state was resolved. At this stage, a method such as restarting the integration operation has been adopted.
[0005]
Further, in the invention described in Japanese Patent Application Laid-Open No. 9-282005, as shown in FIG. 3, the constant related to the integration operation is not constant, and the constant may be changed or the deviation value may be changed depending on the deviation value. A method has been adopted in which a certain threshold value is provided, and when the actual deviation exceeds the threshold value, a constant related to the integration operation is changed stepwise.
[0006]
[Problems to be solved by the invention]
However, in these methods, the deviation cannot be reduced in a short time, or even if the constant relating to the integration operation is changed, it is not specifically shown how to change the constant. There were problems such as.
[0007]
The present invention has been made to solve the above-described problems. The object of the present invention is to calculate the integral constant T _I of the absolute value of deviation | e in a PI controller, PID controller, or an equivalent controller. As a function of |, it is large when the deviation is large, and conversely small when the deviation is small, and it can be changed smoothly and continuously. It is to realize a controller capable of precise response.
[0008]
[Means for Solving the Problems]
In order to solve the above-mentioned problem, the invention according to claim 1 is a means for calculating a deviation which is a difference between an input signal and a feedback signal; And a means for calculating and outputting the absolute value of the deviation, a means for outputting the absolute value by a proportionality constant, and a value obtained by multiplying the absolute value by the proportionality constant. Means for calculating and outputting the reciprocal of a value obtained by multiplying the absolute value by a proportionality constant and adding 1; and means for multiplying the deviation by the reciprocal and outputting, and the integrating means comprises: The integration time constant is calculated based on a value obtained by multiplying the deviation by the reciprocal and a preset initial integration constant.
According to a second aspect of the present invention, there is provided an integration of a controller comprising: means for calculating a deviation which is a difference between an input signal and a feedback signal; In the processing method, the absolute value of the deviation is calculated, the absolute deviation process is performed, the output of the absolute deviation process is multiplied by a proportionality constant, 1 is further added, the multiplication and addition process is performed, and the multiplication and addition process is performed. Calculate the reciprocal of the output, perform reciprocal processing, multiply the output of the reciprocal processing by the deviation, perform multiplication processing, and calculate the integration time constant based on the output of the multiplication processing and a preset initial integration constant Then, integration processing is performed.
[0009]
Specifically, the value T _I (integration time constant) corresponding to the integration constant is adjusted as a function of the absolute value | e | of the deviation as follows.
T _I = T _{I 0} (K _I | e | +1)
However, T _{I 0} is an initial setting integral constant, and the maximum time constant T _{L of the} control target _Less than _M (ie, T _{I 0} < _{TL M} ). K _I is, proportional gain.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
A preferred embodiment according to the present invention will be described below with reference to FIGS. 1 to 2.
[0011]
FIG. 1 is a block diagram of a control system in which a calculation function for monotonously changing an integral constant as a function of deviation is added to the control system of a conventional PID controller.
[0012]
The deviation is given as a difference between the command value r and the control amount x. In this embodiment, the deviation is input to the block 1 that performs the absolute value calculation, and the absolute value | e | of the deviation is output.
[0013]
Then, the output signal is input to a block 2 for K _I times multiplication operation, further, after the process of adding 1 is performed, in block 3 1 / arithmetical operations (K I _| +1 | _e) Done.
[0014]
This output signal is input to a block 4 for performing a multiplication operation, and after a product with the deviation e is obtained, it is input to a block 5 as an integrator together with an initial setting constant T _I0 (1 / ( T _I0 (K _I | e | +1))) edt is calculated.
[0015]
Finally, the manipulated variable u is obtained by adding the outputs of the block 5, the block 6 and the block 7 and then multiplying them by K _P.
[0016]
That is, in this embodiment, a conventional PID controller is used as a basic configuration, and only the integration operation of the conventional PID controller is replaced with a value that is integrated with a monotonically changing deviation function, and the manipulated variable u is replaced by K _P (E + (1 / (T _I0 (K _I | e | +1))) ∫edt + K _D de / dt)
[0017]
Note that all the operations in the above-described embodiments can be executed by a microprocessor.
[0018]
Next, the reason why the integration constant is monotonously changed as a function of deviation in the above-described embodiment will be described below.
[0019]
In the conventional PID controller, as described above, a method has been used in which the gains K _P , K _I , and K _D of the PID calculation are made constant or constants related to the integration operation are switched stepwise.
[0020]
However, the constant gain method causes problems such as overshoot due to the integral operation, so it is common to set the controller's integral constant to a large value, and control specifications that require high-speed response and small deviation. Could not be satisfied.
[0021]
In addition, in the method of switching the constant related to the integral operation, the switching time must be determined in advance, and the control target is subjected to vibration or sudden shock due to a sudden change in the operation amount resulting from the switching of the constant related to the integral operation. There was a fear of generating.
[0022]
On the other hand, in the above embodiment, by setting the integral constant to T _I0 (K _I | e | +1) as described above, the integral constant is large when the deviation is large, and conversely small when the deviation is small. Moreover, it can be changed smoothly and continuously.
[0023]
Therefore, it is not necessary to set the integral constant as large as in the prior art, and it can be set to a fraction of the conventional value, and a PID controller capable of high-speed and precise response can be realized.
[0024]
Figure 2 is a conventional PID controller in which the integral constant _{T I} and a constant value 50 × ^{10 -6,} and a _{K P} to _{100, T I0} to ¹⁰ × 10 -6, _{K I} and 50, _{K D} 0 above It is the result of having compared the response with the thing of an Example, and the operation amount.
[0025]
In the case of the above embodiment, as is apparent from the drawing, there is no problem of overshoot or undershoot as in the conventional PID controller, and it can be confirmed that the settling time is greatly improved.
[0026]
【The invention's effect】
As described above, according to the present invention, PI controller, the PID controller or any equivalent controllers, the integral constant T _I, the absolute value of the deviation | e | as a function of, when the deviation is large is large, reverse When the deviation is small, it can be changed smoothly and continuously, and it is not necessary to set a large integral constant as in the conventional case, and a controller capable of high-speed and precise response can be realized. .
[Brief description of the drawings]
FIG. 1 is a block diagram showing the configuration of a PID controller for implementing the method of the present invention. FIG. 2 is a step response comparing the method of the present invention with a conventional method. FIG. 3 is a conventional position control routine for changing an integral gain. Flowchart

Claims (2)

In a controller comprising: means for calculating a deviation which is a difference between an input signal and a feedback signal; and an integrating means for performing an integration operation by multiplying the deviation by an inverse of an integration time constant.
Means for calculating and outputting an absolute value of the deviation;
Means for multiplying the absolute value by a proportional constant and outputting,
Means for adding 1 to a value obtained by multiplying the absolute value by a proportionality constant;
Means for calculating and outputting the reciprocal of a value obtained by multiplying the absolute value by a proportionality constant and adding 1;
Means for multiplying the deviation by the reciprocal and outputting,
The controller, wherein the integration means calculates the integration time constant based on a value obtained by multiplying the deviation by the reciprocal and a preset initial integration constant. In an integration processing method of a controller, comprising: means for calculating a deviation which is a difference between an input signal and a feedback signal; and integration means for performing an integration operation by multiplying the deviation by an inverse of an integration time constant.
Deviation absolute value processing for calculating the absolute value of the deviation,
Multiplying the output of the deviation absolute value processing by a proportionality constant, adding 1 further, and multiplying and adding,
Reciprocal processing for calculating the reciprocal of the output of the multiplication and addition processing;
Multiply the output of the reciprocal process by the deviation,
An integration processing method for a controller, wherein the integration time constant is calculated based on an output of the multiplication process and a preset initial integration constant, and the integration process is performed.

Images