JP3267841B2 - Controller with phase compensation function - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a control device having a phase compensation function. For more details, shorten the settling time,
The present invention relates to a control device with a phase compensation function capable of improving the stability of a process.

[0002]

2. Description of the Related Art Conventionally, a Smith compensator has been used for phase compensation of a process including a dead time in a system including a dead time in a process, for example, control of a furnace temperature, control of a plate thickness of a rolling mill, and the like. FIG. 6 shows a block diagram of an example of a control device using a conventional Smith compensator. In FIG.
11 is a Smith compensator which is a phase compensator, 13 is a control target, and 14 is a PI controller.

A control device using a Smith compensator to control the thickness of a rolling mill will be described below. First, the thickness deviation on the delivery side of the rolled material is detected by a thickness gauge installed on the delivery side of the rolling mill. The detected value is the PI controller 14
The PI controller 14 obtains a control amount by performing a proportional operation process, an integral operation process, or a proportional integral operation process. The control amount obtained by the PI controller 14 is converted into a Smith compensator 1 as a phase compensator.
After that, the compensation signal obtained by the Smith compensator 11 is fed back to the PI controller 14. A series of steps up to feedback is repeated until the sheet thickness deviation becomes the preset sheet thickness deviation setting “0”. By the way, in a system including a dead time, the Smith compensator 11 has a dead time estimation transfer function e estimated in advance according to the control target as shown in FIG.
^-Ts, it is a process estimating the transfer function G (s) ps, and process estimator gain K _s components. That is,
The conventional Smith compensator 11 predicts fluctuation after a dead time with respect to the control amount, and cancels the actual measurement value Y to prevent overcontrol (an unstable phenomenon such as hunting) due to an excessive control gain.

[0004] In the control device using the Smith compensator, the transfer function G (s) c of the PI controller 14 using the integral time constant T _i and total gain K _o, the following formula (1)
It is represented by Further, the process transfer function of the controlled system 13 can be approximated by a cubic delay function using the constants T _p time vary depending controlled object (see the following equation (2)).

[0005]

(Equation 1)

[0006]

When the conventional control device using the Smith compensator is applied to a control device having a long dead time and a small time constant, as in the control of the thickness of a rolling mill, the rolling device is not suitable for rolling. Although the thickness accuracy is improved in the low-speed region of the machine, the accuracy is not improved in the high-speed region. In addition, the conventional control device needs to readjust the control constants (total gain Ko and integration constant) to add a Smith compensator in the existing PI control loop. Is difficult, and the stability against model errors is poor.

When the dead time T approaches 0 in the equation (1) when the dead time is short and the time constant is large as in the furnace temperature control, PID (proportional differentiation and integration) ) Only control equivalent to control can be performed, and controllability is poor.

An object of the present invention is to solve such a problem and to provide a control device having a phase compensation function which is easy to adjust and has improved stability.

[0009]

A control device with a phase compensation function according to the present invention inputs a process estimation gain estimated in advance for a control signal from a PI controller to a phase compensator and outputs the obtained output to the phase compensator. A phase compensating means for feeding back to a PI controller and compensating a phase of a measured value with respect to a control target value; and a stability compensating means for stably compensating a system by a stability compensator included in the phase compensator. A control with a phase compensation function used for controlling a process having a dead time, wherein a transfer function of a controller includes a product of a quotient of a stability compensator time constant and an integration time constant and a reciprocal of the process estimation gain. Device.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, an embodiment of a control device with a phase compensation function of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of a control device having a phase compensation function according to the present invention. In FIG. 1, 1 is a phase compensator, 2 is a stability compensator, 3 is a control target, and 4 is a PI controller. Here, the stability compensator 2 is represented by a first-order lag transfer function to ensure stability, and the transfer function G (s) h is represented by a stability compensator time constant T
with _h is represented by the following formula (3), by adjusting a stable compensation time constant T _h, it is ensured the stability of the control system.

[0012]

(Equation 2)

The phase compensator 1 has a process estimation transfer function G
(S) Ps (in FIG. 1, the process transfer function G (s) ps
And expressed by the product of dead time estimated transfer function e ^-Ts), has a stable compensator 2 transfer function G (s) h, and the process estimated gain K _s components.

A control method when the controller with a phase compensation function of the present invention is used for a process having a long dead time and a small time constant, for example, for controlling the thickness of a rolling mill will be described below.

First, the thickness is detected by a thickness gauge (the detection process is not shown), and the detected value is used as the PI controller 4 in FIG.
Is input to When detecting the plate thickness, tracking is performed by a plate speedometer for measuring the speed of the plate used for calculating the dead time. Based on the input value, the PI controller 4 performs a proportional operation, an integral operation, or a proportional integral operation. After the arithmetic processing is performed, the output obtained by inputting to the phase compensator 1 (including the stability compensator 2) is represented by P
This is fed back to the I controller 4. These series of operations are repeated until a desired plate thickness is obtained.

If there is a disturbance ΔG after the closed loop composed of the PI controller 4 and the phase compensator 1, a fluctuation of the actual measured value Y by the disturbance ΔG is detected, and the delay after the dead time with respect to the control amount is detected. An error occurs in offsetting the fluctuation with the predicted value. PI controller 4 with the error
Feedback to These series of operations are repeated until a desired plate thickness is obtained.

Here, in order to cooperate with the compensation signal from the phase compensator 1 (including the stability compensator 2), the transfer function G (s) c of the PI controller 4 includes a process estimation gain K _s and an integral using the time constant T _i and stable compensator time constants T _h, it is uniquely determined by the following equation (4).

[0018]

(Equation 3)

That is, the transfer function G (s) c of the PI controller 4 is obtained by dividing the stability compensation time constant _Th by the integral time constant T _i (T _h / T _i ) and the reciprocal of the process estimation gain K _s (T _h / T _i ). 1 / K _s ).

[0020] However, when describing the process estimator gain K _s In one example, the motor terminal voltage (input) and either pre-computed gain, such as the relationship between the rotational speed output of the motor, or measured by real equipment control It is set in the device.

When the stability compensator of the present invention is used for controlling the thickness of a rolling mill or controlling the furnace temperature, the above-mentioned stability compensator time constant _Th is (１ ／) T _i < _Th <. It is set to a range of about 2T _i .

FIG. 1 can be modified as shown in FIG. 2 based on the equations (3) and (4). That is, FIG. 2 is an equivalent block diagram of FIG.

Since the configuration can be modified as shown in FIG. 2, open-loop control can be equivalently performed. For this reason, adjustment is easier than in the conventional Smith method. Moreover, since the control is equivalent to the open loop control, the amount of phase delay is smaller than that in the closed loop system of the Smith compensator in the Smith method, and the process is more stable.

FIG. 3 shows K = Ks = 0.5, t = T = 1, T _i
= 0.1, in case of the T _h = 0.2, the response curve in the absence of a model estimation error, in comparison to the response curve when using a conventional Smith compensator. Curve A in FIG. 3 is a response curve when the control device of the present invention is used, and curve B is a response curve when a conventional Smith compensator is used. The vertical axis and the horizontal axis indicate the step amount and the time, respectively. As shown in FIG. 3, when the settling time was about 2.2 seconds when the Smith compensator was used,
When the control device of the present invention is used, the settling time is 1.
It was reduced to about 8 seconds. That is, the settling time is about 20%
It can be seen that the degree has been improved to a more stable level.

FIG. 4 shows that K = Ks = 0.5, t = T = 1,
A response curve when a model estimation error occurs is shown in comparison with a response curve when a conventional Smith compensator is used, where T _i = 0.1 and _Th = 0.2. The vertical axis and the horizontal axis indicate the step amount and the time, respectively. The estimated constant is increased by 20% compared to the estimated constant applied in FIG. FIG.
Similarly to the case, the curve A is a response curve when using the control device of the present invention, and the curve B is a response curve when using the conventional Smith compensator. As shown in FIG. 4, when the Smith compensator is used, the swing width is large and diverges, whereas when the control device of the present invention is used, the swing width is small and tends to converge. It can be seen that the stability has improved.

The control system shown in FIG. 1 is also used when the control device with a phase compensation function of the present invention is used for a device having a short dead time and a large time constant, for example, for controlling furnace temperature. Also in this case, since the control is equivalent to the open loop control, the amount of phase delay is smaller than that of the normal PI control, and the controllability is improved.

FIG. 5 shows K = Ks = 0.5, t = T = 1, T _i
= 10 and T _h = 22 are shown in comparison with the response curve using a conventional Smith compensator. The vertical axis and the horizontal axis indicate the step amount and the time, respectively. Curve A in FIG. 5 is a response curve when using the control device of the present invention, and curve B is a response curve when using the conventional Smith compensator. As shown in FIG. 5, it can be seen that the settling time is improved by about 20% and the controllability is improved.

[0028]

According to the present invention, since the control device can be considered as an open loop equivalently, if the control device with the phase compensation function of the present invention is used for a device having a long dead time and a small time constant, Adjustment of control constants is easy, and stability is good even for models with model errors.

If the control device with the phase compensation function of the present invention is used for a device having a short dead time and a large time constant, the controllability is improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a control device with a phase compensation function according to an embodiment of the present invention.

FIG. 2 is an equivalent block diagram of the control device with a phase compensation function of FIG. 1;

FIG. 3 shows a response curve when a controller with a phase compensation function according to an embodiment of the present invention is used for controlling the thickness of a rolling mill and a conventional Smith compensator when there is no model estimation error. It is a figure which shows and compares the response curve of A.

FIG. 4 uses a response curve and a conventional Smith compensator when a control device with a phase compensation function according to one embodiment of the present invention is used for controlling the thickness of a rolling mill when a model estimation error occurs. FIG. 7 is a diagram showing a comparison of response curves in the case of the above.

FIG. 5 is a diagram showing a comparison between a response curve when a control device with a phase compensation function according to one embodiment of the present invention is used for controlling a furnace temperature and a response curve when a conventional Smith compensator is used.

FIG. 6 is a block diagram of a control device using a conventional Smith compensator.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Phase compensator 2 Stability compensator 3 Control object 4 PI controller

Claims (1)

(57) [Claims] 1. A process estimation gain estimated in advance for a control signal from a PI controller is input to a phase compensator, and an obtained output is fed back to the PI controller. The PI compensator comprises a phase compensator for compensating, and a stability compensator for stably compensating the system with a stability compensator included in the phase compensator. A control device with a phase compensation function used for controlling a process having a dead time, comprising a product of a quotient obtained by division and a reciprocal of the process estimation gain.