JPH04370804A - Controller - Google Patents

Abstract

PURPOSE:To remove idle time elements by predicting the control motion of a controlled system which has the idle time elements. CONSTITUTION:This controller is equipped with a time identifying arithmetic part 24 which continuously measures a manipulated variable inputted to the controlled system varying in idle time and a controlled variable outputted by the controlled system to identify the idle time up to the appearance of the manipulated variable in the controlled variable and a controlled variable predictive arithmetic part 26 which predicts and calculates a controlled variable after the idle time by using the calculated idle time and the relation model between the controlled variable and manipulated variable.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention relates to a control device that automatically controls liquid level, temperature, shape, etc. This invention relates to a control device suitable for control.

0002

[Prior Art] Molten metal level control in a mold in a continuous casting process (for example, Japanese Patent Publication No. 16218/1983);
Room temperature control and plate thickness control in the rolling process are generally based on fixed value control using PID calculations, especially proportional action (P) and integral action (I).
) is mainly controlled.

However, if there is a non-negligible dead time in the controlled object, in conventional PID control, each time a disturbance occurs, the target value suddenly increases or decreases, resulting in sustained vibration. Therefore, it is not suitable for control that requires the amount of variation to be minimized. The Smith method is known as a method for stably controlling systems that include dead time elements. This method is a method of equivalently removing dead time elements by adding a control element for compensating for dead time in addition to control elements such as PID control.

0004

[Problem to be Solved by the Invention] However, even if this method is adopted, it is necessary to determine an appropriate value of dead time for each system, and it is necessary to determine an appropriate value of dead time for each system. Therefore, it is necessary to determine the dead time in real time. Therefore, it is an object of the present invention that a controlled object has a non-negligible dead time, and
It is an object of the present invention to provide a control device capable of performing stable control even when it changes over time.

[0005]

[Means for Solving the Problems] FIG. 1 is a diagram showing the basic configuration of the present invention. In the figure, the control device of the present invention is a control device equipped with a manipulated variable calculating means 12 that calculates a manipulated variable for making the observed value match the target value from the deviation between the target value and the observed value and provides it to the controlled object 10. , a dead time sequential identification means 14 sequentially and continuously identifies the dead time of the controlled object 10 from the manipulated variable given to the controlled object 10 and the controlled variable observed in the controlled object 10; Controlled variable prediction calculation means 16 predicts and calculates the controlled variable after the dead time has elapsed using a model of the controlled object 10 from the value of the dead time identified by the time-sequential identification means 14, the control amount, and the manipulated variable.
It is characterized in that the difference between the target control amount and the control amount predicted by the control amount prediction calculation means 16 is given to the operation amount calculation means 12 as the deviation.

[0006]

[Operation] By sequentially performing calculations for dead time identification by the dead time sequential identification means 14, high-speed dead time calculation is achieved, and the dead time that changes from moment to moment is identified in real time. Dead time compensation is realized in real time.

[0007]

[Embodiment] FIG. 2 shows a control device 20 according to an embodiment of the present invention.
A block diagram of a control system using the is shown. 22 represents a controlled object, and it takes τ time as a dead time for the controlled variable y(t) to change in response to the applied manipulated variable u(t), and the dead time τ varies depending on time t. , τ(t
).

The dead time sequential type identification calculation section 24 identifies (estimates) this dead time τ(t).
The cross-correlation coefficient between (t-d) and y(t) is obtained by varying d, and the maximum d is set as the delay time τ'(t). The greatest feature of this control device is that the cross-correlation coefficient is found sequentially using the forgetting coefficient.

The control amount prediction calculation section 26 uses the estimated dead time τ'(t) obtained above to calculate τ by model calculation.
'(t) The control amount y(t+τ'(t)) after time is predictively calculated. The manipulated variable calculating section 28 calculates the manipulated variable based on the difference between the target controlled variable and the predicted controlled variable y (t+τ'(t)) after τ'(t) calculated by the controlled variable prediction calculating section 26. . This uses, for example, a well-known PID control calculation logic.

Reference numeral 30 denotes an operating end that acts on the controlled object 22 by the operating amount calculated by the operating amount calculating section 28;
32 is a detection end for detecting the control amount. 3 to 5 show examples of possible uses of the control device of the invention. FIG. 3 is an example of plate thickness control of a steel plate, where the manipulated variable u(t) is the roll reduction amount by the hydraulic pressure reduction device 40, and the controlled variable y(t) is the thickness total 4
This is the steel plate thickness measured in No. 2. In this case, since the thickness gauge 42 is installed at a position several meters away from the rolling down device 40, there is a dead time.
slips, so the steel plate speed does not match the roll circumferential speed, and the dead time changes.

FIG. 4 shows an example of mold level control in a continuous casting machine, where the manipulated variable u(t) is the opening degree of the sliding nozzle 50, and the controlled variable y(t) is the mold level measured by the level meter 52. In this case, the immersion nozzle 54 is not filled with molten steel, and there is a time delay before the molten steel in the tundish 56 flows into the mold 58, and the dead time changes depending on the situation inside the nozzle.

FIG. 5 shows an example of indoor temperature control in which the manipulated variable u(t
) is the opening degree of the control valve 62 that controls the amount of steam from the boiler 60, and the control amount y(t) is the room temperature measured by the thermometer 64. Even in this process, there is dead time since the indoor temperature changes after the valve 62 opens and closes. The calculation process in the dead time sequential identification calculation unit 24 in FIG. 2 is simplified by sequentially calculating the cross-correlation coefficient using the previous calculation results, so that calculation time and memory can be saved. This can be realized using software processing on a computer with the same power as a personal computer.

FIG. 6 shows this dead time sequential type identification calculation section 24.
3 is a flowchart of software processing that implements the control amount prediction calculation unit 26 of FIG. 2. FIG. Both the manipulated variable u(t) and the controlled variable y(t) are A/D converted and captured and stored in each calculation cycle, and u(t) is stored together with past values as u(t-d),...u (t-1), D samples of u(t) are held in memory (step a)
. Note that when the controlled object has an integral element as in the example in Fig. 4, the difference value Δy(t) =
Use y(t) -y(t-1). Next, using the forgetting coefficient ρ (0<ρ<1) which is a preset constant, the recurrence formula (
1) P(t) =1+ρ・P(t-1), P(0) =0
...From (1), calculate the parameter P(t), and use the value to calculate recurrence formulas (2) and (3), ym (t) = ym (t-1) + {y(t) -y
m (t-1)}/P(t), ym (0) =0
...(2) um (t, d) = um (t-1, d) + {u(t
-d) -um (t-1, d)}/P(t), um (0, d) = 0; however, d = 0, 1...D
...(3), calculate the weighted average value ym (t) of y(t) and the weighted average value um (t, d) of u(t-d), and use the recurrence formula (4 ) R(t, d) = R(t-1, d) +1/P(t) [ρ・P(t-1)/P(t)2・{
y(t) -ym (t-1)}{u(t-d)-um
(t-1, d)}+{y(t) -ym t)}{u
(t-d)-um (t,d) }-R(t-1,d)
], R (0, d) = 0
...(4) u(t-d) with the manipulated variable u(t) delayed by d
The cross-correlation coefficient R(t, d) between the control amount y(t) and the control amount y(t) is calculated (step b). Then, the calculated R(t,
d); d=0,1...maximum R(t,d) among D
Let d giving the estimated dead time τ'(t) be the estimated dead time τ'(t) (step c).

[0014] If the controlled object model is ym (t) = f (u (t - τ (t)), then τ
'(t) The predicted control amount after time is ym (t+τ'(t
))=y(t) +(f(u(t))-f(u(t-τ
'(t))) (step d). However, y(t) is an actual value.

[0015] The calculated predicted control amount y(t+τ'(t)
) is output (step e), and the deviation from the target control amount is supplied to the manipulated variable calculation unit 28 (FIG. 2). FIG. 7 is an example of control using this control device, and is a mold level control chart of the continuous casting machine of FIG. 4. The calculation contents of this control device will be explained assuming that the current time t is 34 minutes and 40 seconds. First, the dead time τ is calculated from the manipulated variable SN (sliding nozzle opening degree) u(t) and the controlled variable mold level y(t).
′(t). (However, in this case, the control target is the mold 58 (Fig. 2)
Since there is an integral element, u(t) and Δy(t)=
Find τ'(t) from y(t) -y(t-1). ) In this case, the identification result is τ'(t) = 2 seconds, so 2
u(k) from 34 minutes 38 seconds ago to 34 minutes 40 seconds
Since the effect of (t)). Feedback control is performed based on this predicted level.

[0016]

[Effects of the Invention] As described above, according to the present invention,
A control device that performs stable control even when the controlled object includes non-negligible dead time elements and the value of the dead time changes over time can be realized using a relatively inexpensive computer. .

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the principle configuration of the present invention.

FIG. 2 is a diagram representing an embodiment of the present invention.

FIG. 3 is a diagram showing an example in which the present invention is applied to plate thickness control in a steel plate rolling process.

FIG. 4 is a diagram showing an example in which the present invention is applied to mold level control of a continuous casting machine.

FIG. 5 is a diagram showing an example in which the present invention is applied to room temperature control.

FIG. 6 is a flowchart of processing for dead time identification calculation and predicted control amount calculation.

FIG. 7 is a diagram showing control results of mold level control of a continuous casting machine by the control device of the present invention.

[Explanation of symbols]

40...Reducing device 42...Thickness gauge 44...Roll 46...Steel plate 50...Sliding nozzle 52...Mold level meter 54...Immersion nozzle 56...Tundish 58...Mold 64...Thermometer

Claims (3)

[Claims] [Claim 1] The control target (
In a control device equipped with a manipulated variable calculating means (12) given to a controlled object (10), the controlled amount of the controlled object (10) is calculated from the manipulated variable given to the controlled object (10) and the controlled variable observed in the controlled object (10). A dead time sequential identification means (14) identifies the dead time sequentially and continuously, and the control is performed based on the value of the dead time identified by the dead time sequential identification means (14), the control amount, and the manipulated variable. Target (10
) for predicting and calculating the controlled variable after the dead time has elapsed using a model of The difference is used as the deviation by the operation amount calculation means.
(12) A control device characterized by: [Claim 2] The dead time sequential identification means (10)
2. The control device according to claim 1, wherein: calculates a cross-correlation coefficient between the manipulated variable and the controlled variable delayed by a plurality of delay times, and sets the delay time at which the cross-correlation coefficient is maximum as the dead time. 3. The dead time sequential identification means (10)
3. The control device according to claim 2, wherein: calculates the current cross-correlation coefficient by a recurrence formula using the cross-correlation coefficient in the previous calculation.