JPH0283702A - Method and device for non-interference control - Google Patents

Abstract

PURPOSE:To increase the control output even in case a non-linear element is included in an interference element, and improve the response of a non-interference control by controlling the actual state value of a control subject in response to the estimated model state value. CONSTITUTION:The control output is produced based on the deviation between the output feedback Y1 and a control instruction Y1ref when a control subject 101 is controlled based on a manipulated variable U1. At the same time, the optimum control output is produced based on the control output and the estimated model state value -X1. In addition, the variable U1 is successively updated based on the model standard form control output U1M, the non-interference control output UC21 given from a 2nd control system, and the optimum control output UO1. Then the state value X1 is controlled based on the value -X1. In this case, the non-interference control is performed via the feedforward control based on the state -X1. Therefore it is possible to instantaneously eliminate the interference caused by the change of the variable U1.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a non-interference control method and apparatus, and in particular,
A non-interference control method and device suitable for performing non-interference control of each control object to a robust system 1 using a plurality of control systems even when a plurality of control objects interfere with each other and include non-linear elements as interfering elements. Regarding.

[Conventional technology]

In cases where multiple control systems interfere with each other's control systems, the method described in JP-A-61-131103 is known as a non-interfering control that cancels out the interfering elements of each control system. It is being

[Problem to be solved by the invention]

In the prior art, a dynamic characteristic key having a function that is an inverse matrix to the dynamic characteristic gain element of the controlled object is added as a non-(1 interference element) to the feedback system of each control system,
The dynamic characteristic gain and the dynamic characteristic time constant are organically combined as a dynamic characteristic gain to improve the response characteristic, but if the interference element includes a nonlinear element, it is not enough to improve the response characteristic of the control system. It wasn't.

A first object of the present invention is to provide a non-interference control method and apparatus that can control a plurality of control objects in a non-interfering manner by matching the actual state amount of the control object to the estimated model state amount. .

[Means to solve the problem]

In order to achieve the above object, the present invention provides a system in which a plurality of control objects have interference elements that interfere with each other, and each control object is controlled by a plurality of control systems, and each control system has control commands and control elements. A control output corresponding to the deviation from the actual state quantity of the target is generated, an estimated model state quantity according to the control output is calculated from a controlled target model in which the controlled target is modeled according to specific parameters, and the estimated model state quantity is calculated according to the control output. A model reference form control output is generated from the deviation between the quantity and the actual state quantity of the controlled object, and a non-interfering control output for the other control system is generated based on the estimated model state quantity and the actual state quantity of the controlled object. , a non-interference control method is adopted in which a manipulated variable is determined from the control output, a model standard control output, and a non-interfering control output from the other control system, and the controlled object is controlled by this manipulated variable.

In systems in which multiple control targets have interfering elements that interfere with each other, and each control target is controlled by multiple control systems, each control system is controlled according to the deviation between the control command and the actual state quantity of the control target. Generate a control output, generate an optimal control output from the deviation between the control output and an estimated model state quantity of a controlled object model in which the controlled object is modeled according to specific parameters, and generate an estimated model state quantity according to this optimal control output. is calculated from the controlled object model, a model reference form control output is generated from the deviation between this estimated model state amount and the controlled object's actual state amount, and the model reference form control output is calculated based on the estimated model state amount and the controlled object's actual state amount. to generate a non-interfering control output for the other control system, and determining a manipulated variable from the optimal control output, the model reference control output, and the non-interfering control output from the other control system,
A non-interference control method is adopted in which the controlled object is controlled by this manipulated variable.

The optimal control output is generated from the deviation between the control output and the state quantity obtained by multiplying the estimated model state quantity by the optimal gain, and the deviation between the estimated model state quantity and the actual state quantity of the controlled object is equivalent to the optimum gain. This method employs the Benshi Wataru control method, which generates a model reference form control output by multiplying it by an optimal gain.

A compensation element for adding a detection delay element of a means for detecting the actual state quantity of the controlled object to the actual state quantity of the controlled object, and compensating for the delay in the manipulated variable due to the detection delay element to the estimated model state quantity. A non-interfering control method is adopted in which the difference between the state quantity of the controlled object to which the detection delay element has been added and the estimated model state quantity to which the compensation element has been added is multiplied by an optimal gain to generate a model reference form control output. This is what I did.

In systems where multiple control objects have interfering elements that interfere with each other and each control object is controlled by multiple control systems, a control output is generated according to the deviation between the control command and the actual state quantity of the control object. a control output generation means; an estimated model state amount calculation means for calculating an estimated model state amount according to the control output from a controlled object model in which the controlled object is modeled according to specific parameters; and an estimated model state calculated by the calculation means. a model reference form control output generation means for generating a model reference form control output from a deviation between the quantity and the actual state quantity of the controlled object; and another control system based on the estimated model state quantity and the actual state quantity of the controlled object. a non-interfering control output generation means for generating a non-interfering control output for the control system; a manipulated variable is determined from the control output, a model reference control output, and a non-interfering control output from the other control system, and the controlled object is controlled by this manipulated variable; This configuration constitutes a non-interference control device in which each control system has a control means for controlling the control system.

In systems where multiple control objects have interfering elements that interfere with each other and each control object is controlled by multiple control systems, a control output is generated according to the deviation between the control command and the actual state quantity of the control object. a control output generation means; an optimum control output generation means for generating an optimum control output from a deviation between the control output and an estimated model state quantity of a controlled object model in which the controlled object is modeled according to specific parameters; An estimated model state amount calculating means for calculating an estimated model state amount from the controlled object model according to the output of a model reference type control output generation means for generating an output; a non-interference control output generation means for generating a non-interference control output for the other control system based on the estimated model state quantity and the actual state quantity of the controlled object; A non-interference control device is constructed in which each control system has a control means that determines the manipulated variable from the optimal control output, the model reference control output, and the non-interfering control output from the other control system, and controls the controlled object using this manipulated variable. This is what I did.

[Effect]

A non-interfering output is generated for the other control system based on the estimated model state quantity and the actual state quantity of the controlled object, and the manipulated variable is generated using the control output, model reference form control output, and non-interfering control output from the other control system. Since the controlled object is controlled using this manipulated variable, it is possible to control the state quantity of the controlled object in accordance with the state quantity of the estimated model. In addition, an optimal control output is generated from the deviation between the estimated model state quantity and the control output, an estimated model state quantity corresponding to this optimal control output is calculated from the controlled object model, and this optimal control output, model reference form control output, and the other The manipulated variable is obtained from the non-interfering control output of the control system, and the controlled object is controlled by this manipulated variable, so even if the state quantity of the controlled object deviates from the state quantity of the estimated model, this deviation can be reliably corrected. be able to. In this case, by setting the optimal gain when generating the optimal control output and the optimal gain when generating the model reference form control output as equivalent gains, the discrepancy between the state quantity of the controlled object and the state quantity of the estimated model can be reduced. It can be suppressed reliably. Furthermore, a detection delay element is added to the actual state quantity of the controlled object, a compensation element is added to the state quantity of the estimated model, and a model reference form control output is generated by considering the delay element and the compensation element. It becomes possible to increase the optimum gain of the optimum control output, which can contribute to improving the responsiveness.

〔Example〕

Hereinafter, one embodiment of the present invention will be described based on FIGS. 1 and 2.

In FIG. 1, the first . The second control system goes to main controller 1 from 0 to 12.
1, 122, controlled object model 111° 112, non-interference control block 113, 1i, 4, computing unit 105, 106,
Adders 107, 108, 109.1.10, model reference form control gain M1. M2, gain α, 1-α, β, 1
-β, and interference terms 103 and 104 as interference elements are connected to the controlled objects 101 and 102.

As shown in FIG. 2, the main controller 1-○-ra] 21 has an arithmetic unit 123, an optimum integral gain K/S, 1 for the optimum gain, and an adder 125, and inputs control commands Y, rcf and Output feedback Y of the controlled object 101, (state quantity X
□) is calculated by the calculator 123, and this deviation is multiplied by the optimal integral gain by ■/Sk to generate the control output, and the output of the controlled object model]11, in which the controlled object 101 is modeled according to specific parameters. Adder 12 adds the control output to the estimated model-like g quantity X1 and the state quantity obtained by multiplying the optimal gain by 1.
5 and the optimum control output U. , generates.

That is, the main control roller 121 is configured as a control output generation means and an optimum control output generation means.

Moreover, the controlled object model 111 has an optimal control output U. , is configured as an estimated model state amount calculation means for calculating the estimated model state amount X1 from .

The output X of the controlled object model 1]1 is supplied to the arithmetic unit 104 via the phase matcher 152. The calculator 105 also contains the actual state quantity X of the controlled object 101.
1 is supplied with a state quantity obtained by adding a temporary delay element 151, which is a detection delay element of the means for detecting the actual state quantity X]- of the controlled object 101, and the output of the phase matcher 152 and the temporary delay element 151 The deviation between the two and
It is designed to output to. The phase matcher] 52 is provided as a compensation element for compensating for the delay in the manipulated variable U1 caused by the temporary delay element 151, and - the time delay element 15
1, an arithmetic unit 105, an optimum gain of 1, and a model reference form control output generating means.

A state quantity obtained by multiplying the state fiX, 1 by Keyne α and a state quantity obtained by multiplying the output of the − time delay element 151 by a gain (1−α) are respectively supplied to the adder 109, and the output of the adder 109 is It is designed to be generated as a non-interference control output UC12 via an interference control block 113. That is, the gains α, 1-α, the adder 109, and the non-interference control block 113 are configured as non-interference control output generation means, and the non-interference control output UC12 is supplied to the second control system, and this control output The second control system is controlled in a feedforward manner.

Non-interference control output UC21 generated by the second control system
is supplied to the adder 107, and this control output U21
The manipulated variable U1 is generated by adding the optimal control output UOI and the model standard control output UIM. That is, the adder 107 is configured as a control means that generates the manipulated variable U1 and controls the controlled object 101. still,
Since the second control system also has the same configuration as the first control system,
The specific details of the second control system will be omitted.

In the above configuration, the controlled object 10
When control 1 is executed, a control output is generated based on the deviation between the output feedback Y1 and the control command Y1ref, and an optimal control output is generated based on this control output and the estimated model state quantity X1. . Further, the manipulated variable U1 is sequentially updated from the model standard control output UIM that suppresses the deviation between the estimated model state quantity X1 and the state quantity x1, the non-interference control output UC21 from the second control system, and the optimal control output UOI, Control is performed to match the state quantity X1 with the estimated model state quantity X1. At this time, the estimated model state quantity X
Since the non-interference control based on 1 is performed by feedforward control, interference caused by a change in the manipulated variable U] can be instantly removed. Furthermore, since non-interference control based on the state quantity X1 is performed on the second control system, interference due to an error in the estimated model state quantity X1 can be alleviated.

Further, according to this embodiment, since the model reference form control output is generated based on the deviation between the estimated model state quantity X1 and the state quantity x1, control due to response delay and detection accuracy of the state quantity X1 detection means Main controller 12
There is no need to consider 1. That is, the feedback of the main controller 121 is the estimated model state quantity X
This is because 1 is used. Therefore, the gain of the main controller 121 can be increased, and even if there is a delay in response, non-interference control can be performed, contributing to improved responsiveness.

Next, an embodiment in which the present invention is applied to a looper control system installed between stands of a hot tandem rolling mill will be described with reference to FIG.

In Fig. 3, the control system is composed of an angle control system and a tension control system, and the control target of each control system is the rolling phenomenon 201.
The object to be controlled by the tension control system is the rolling phenomenon 20
It is represented as 2.

The angle control system includes an optimal control block 221, a model reference control block 211, and a looper motor transfer function 20.
5, the tension control system includes an optimal control block 222, a model reference form control block 212, and a main material motor transfer function 20.
6, and filters 1 and 2 are provided as temporary delay elements. Also, tension → interference term 204 and angle → tension interference term 2
03 is provided. Furthermore, a non-interference control block 213 is provided between each control system.

The control system with the above configuration is expressed by the following physical equation.

Ta・a+c=-Ka・(1+f)Vn+K
a・nr・b −K((J) −(1)J'6=
ζφ・Ia-craAaQr-flK(θ)+D(J-
Td ・(2) Tc4a +Ia = Iref
−(3) TrrVn +V
R: VRref - (4)
Here, σ: Tension vR: Roll speed θ Nilooper angle Ia Nilupar current f: Advance rate Ka,
Tc: Influence coefficient A: Material cross-sectional area Qr Nil-Parr arm length K (θ): Angle coefficient J Nil-Parr system moment of inertia 1-D Nil-Parr system damping element Td Nil-Parr own weight and material weight 1-Lux ζφ Nirupar motor 1 herk coefficient Tc Nirupar motor actual flow time constant Tn: main engine speed time constant suffix ref: command The physical model after non-interference according to the above physical equation is expressed by the following equation.

Angle control system: J゛6:ζφIa +D (+-Td -(
2) 'TcIa +Ia = Iref
−(3)′ Tension control system: Tσ・mi+σ=−toσ(I−If)VR−・(1)′T
r-VR+Vo ”Vn ref −(4
)'''1 When the notation is converted into state space form, it is expressed by the following equation.

Z 1 ” A sχ, +B,・I ref
-(5) A, 2o, e, ・
・(6) Z 2 = A 2 Z 2 ” B 2Δ■
ref・(7)y2”C2χ7
...(8) Here, χ, = [klo, 1b, +°
6) Tχ2”(kzσl k2σ]TC,,=I:.

0, 0] C2-hit, 0] a, = 1-/Ta, b, =], /Tn k, = 1-/ (Vn (Hf) (σ to -) a2b2:1
When each control system is configured based on the above equations, the control output of the model standard control block is the actual detected value of the state estimates ix, , x2 obtained from equations (5) and (7) in section 1. It can be obtained by multiplying the deviation from the state quantity obtained through filters 1 and 2 by 2. Here, Kl, 2 is the optimal control block 221
Make it the same as the state feed bank gain. this is,
Optimal Control Flock 22] and Model Normative Control Flock 2
11 - Which interference to avoid, optimal control block 221
The state feedback value of the optimal control block 22 is calculated using the estimated value of the model, and in order to satisfy the predetermined performance of the optimal control block 221, the apparent control output of the optimal control block 22 and the actual state feedback are made to match. It is. This is clear from the following equation.

Output of optimal control block 221 = KX+ + ('/r
ef+Y) Output of model reference form control block 211=K
(X-X)” - Output of each control block =
:KX+ becomes +(Yref-Y), and the apparent output of the optimal control block 221 becomes the same as the actual state quantity feedback.

Next, when configuring the non-interference control block 213, it is configured using actual state quantities and values estimated by the model in order to perform feedforward control that takes model deviations into consideration. Furthermore, in the case of looper control, since T, c, and 'I'r are sufficiently short compared to the control system, the dynamic characteristics of the control system are expressed by the following equation.

Gc, (Z)=(Q r−K (0)/(VRc+
)) <a b )+(1+ ct )+ b )G
c2(z)=(Qr-K(0)/Sφ)(B σ+(1
−β)σ) where 0. σ is the estimated value by the model, σ is the actual detected value α, β<1 Next, the optimal control block 221 performs the above (5) and (6).
Based on the formula, an integral type optimal regulator configuration is adopted.

Introducing the disconnection state quantity 'l, = 0-Oref, X1 = (
k, 0. −.

k10□+k30xr Zl)", Xl is expressed by the following equation.

Number J 1”f o (Xs QI XI” I ref
A steady-state solution that minimizes R4I ref) dt is found using the Ricasoti equation.

Further, the optimal control block 222 is configured as an integral type optimal regulator based on equations (7) and (8). In this case, we introduce the disconnection state quantity Z,=a-aref, and X2=(k2a
.

k2σ, Z2)T, X2 is expressed by the following equation.

1 pa [combination: :]
1ml・3・mu°1゛1pa°゛°1 ″′3″C″・”・”[8: :1・8pa[g・1 From the above formula, the optimal feedback gain is the evaluation function Jz”fa ( XzQ, X2+Vref R2Vref
) is determined by finding a steady-state solution that minimizes dt using the Rikatti equation.

In this embodiment, when the current command I ref changes, the torque changes and the speed θ changes. and the looper angle by integrating the angular velocity O.

changes. Then, this looper angle θ is fed back as a state quantity, and control is executed to match this looper angle θ to the model. Also, the speed V of the main motor
When ref changes, the speed of the main material motor changes, and the tension changes in accordance with this speed. Changes in this tension are also controlled according to the model.

〔Effect of the invention〕

As explained above, according to the present invention, the actual state quantity of the controlled object can be controlled in accordance with the estimated model state quantity, so even when the interference element includes a nonlinear element, the gain of the control output can be increased. This can contribute to improving responsiveness.

Furthermore, even if there is a response delay in the control system, non-interference control can be reliably executed. Further, by setting the optimum gain of the optimum control output and the optimum gain of the model reference control output to be equivalent gains, it is possible to suppress the error between the actual state quantity of the controlled object and the estimated model state quantity. In addition, since a detection delay element is added to the actual state quantity of the controlled object and a compensation element is added to the estimated model state quantity, the response delay and detection accuracy of the means for detecting the actual state quantity of the controlled object are taken into account. Therefore, the gain of the optimal control output can be increased without causing any problems, and this can contribute to improving responsiveness.

[Brief explanation of the drawing]

FIG. 1 is an overall configuration diagram showing an embodiment of the present invention, FIG. 2 is a specific configuration diagram of a first control system, and FIG. 3 is a configuration diagram of a looper control system. 101.102...Controlled object, 105,106...Arithmetic unit, 107,108,109.・110...adder,
111, 112...Controlled object model, 121°122
...main controller, X,,X2...state quantity,
X engineering, X2... Estimated model state quantity, Ml, M2...
・Model reference form control gain.

Claims (1)

[Claims] 1. A plurality of controlled objects have interference elements that interfere with each other,
In systems where each controlled object is controlled by multiple control systems, each control system generates a control output according to the deviation between the control command and the actual state quantity of the controlled object, and models the controlled object according to specific parameters. An estimated model state amount corresponding to the control output is calculated from the controlled object model, a model reference form control output is generated from the deviation between the estimated model state amount and the actual state amount of the controlled object, and the estimated model state amount is and generating a non-interfering control output for the other control system based on the actual state quantity of the controlled object, and determining a manipulated variable from the control output, the model reference form control output, and the non-interfering control output from the other control system, A non-interference control method characterized by controlling a controlled object using this manipulated variable. 2. A plurality of controlled objects have interference elements that interfere with each other,
In systems where each controlled object is controlled by multiple control systems, each control system generates a control output according to the deviation between the control command and the actual state quantity of the controlled object, and models the controlled object according to specific parameters. An optimal control output is generated from the deviation between the estimated model state quantity of the controlled target model and the control output, an estimated model state quantity corresponding to this optimal control output is calculated from the controlled target model, and this estimated model state quantity and control Generate a model reference form control output from the deviation from the actual state quantity of the target, generate a non-interfering control output for the other control system based on the estimated model state quantity and the actual state quantity of the controlled target, and A non-interference control method characterized in that a manipulated variable is determined from an optimal control output, a model reference control output, and a non-interfering control output from the other control system, and a controlled object is controlled using this manipulated variable. 3. Generate the optimal control output from the deviation between the control output and the state amount obtained by multiplying the estimated model state amount by the optimal gain, and apply the optimal gain to the deviation between the estimated model state amount and the actual state amount of the controlled object. 3. The non-interference control method according to claim 2, wherein the model reference form control output is generated by multiplying the model reference form control output by an optimum gain equivalent to . 4. Add the detection delay element of the means for detecting the actual state quantity of the controlled object to the actual state quantity of the controlled object, and add it to the estimated model state quantity to compensate for the delay in the manipulated variable due to the detection delay element. The system is characterized in that a compensation element is added, and the deviation between the state quantity of the controlled object to which the detection delay element is added and the estimated model state quantity to which the compensation element is added is multiplied by an optimal gain to generate a model reference form control output. The non-interference control method according to claim 2 or 3. 5. A plurality of controlled objects have interference elements that interfere with each other,
In systems where each controlled object is controlled by multiple control systems, a control output generation means that generates a control output according to the deviation between the control command and the actual state quantity of the controlled object, and a control output generation means that models the controlled object according to specific parameters. an estimated model state amount calculation means for calculating an estimated model state amount according to the control output from the controlled object model, and a model norm based on the deviation between the estimated model state amount calculated by the calculation means and the actual state amount of the controlled object model reference control output generation means for generating a control output, and non-interference control output generation means for generating a non-interference control output for the other control system based on the estimated model state quantity and the actual state quantity of the controlled object. , each control system has a control means for determining a manipulated variable from the control output, a model reference control output, and a non-interfering control output from the other control system, and controlling a controlled object using the manipulated variable. Non-interference control device. 6. A plurality of controlled objects have interference elements that interfere with each other,
In systems where each controlled object is controlled by multiple control systems, a control output generation means that generates a control output according to the deviation between the control command and the actual state quantity of the controlled object, and a control output generation means that models the controlled object according to specific parameters. an optimal control output generation means for generating an optimal control output from the deviation between the estimated model state quantity of the controlled object model and the control output; and an estimated model state quantity according to the output of the control output generation means from the control pair model an estimated model state quantity calculation means for calculating, a model reference form control output generation means for generating a model reference form control output from the deviation between the estimated model state quantity calculated by this calculation means and the actual state quantity of the controlled object; non-interfering control output generation means for generating a non-interfering control output for the other control system based on the estimated model state quantity and the actual state quantity of the controlled object; the optimum control output, the model reference form control output, and the other control system; A non-interference control device characterized in that each control system has a control means for determining a manipulated variable from a non-interfering control output from a non-interfering control output, and controlling a controlled object using this manipulated variable.