JPH04336301A - Feedforward controller by reverse system - Google Patents

Abstract

PURPOSE:To enable excellent control even when the characteristic of an object part of process control is nonlinear. CONSTITUTION:A compensator 100 is incorporated in a system which feeds the output of the object part of process control back by a proportional plus integral operation controller so as to prevent control disorder due to known disturbance, and the sum of the output u'' of the proportional plus integral operation controller and the output u deg. of the compensator 100 is a manipulated variable (u) to the object part of process control. This compensator 100 has a mathematical expression model 101 for estimating the output of the object part of process control, a primary delay element 104 which is arranged in parallel to the mathematical expression model 101 and stabilizes the model, and proportional plus differential operation controllers 107, 108, 109, and 110 which put in an error delta in proportional plus differential operation.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a feedforward control device using an inverse system that is useful for controlling steam temperature of process products.

0002

2. Description of the Related Art A typical example of a process control target part and a control system is shown in FIG. In the figure, the control amount y, which is the output of the process control target section 1, is compared (subtraction operation) with the target value r in the subtracter 2, and the control deviation ε (this value is y
and r) is output. The control deviation ε is input to a proportional-integral action regulator (hereinafter abbreviated as "PI regulator") 3, and the output u'' of the PI regulator 3 and the output u coming from the compensator 10 are
° are added by an adder 4. The output of the adder 4 becomes the manipulated variable u and is output to the process control target section 1. Here, we will discuss the case where there are x and w as known disturbances to the process control target section 1.

Conventionally, the control amount y due to known disturbances x and w
In order to minimize the influence on the output, the output u° is determined by a compensator 10 as shown in FIG. That is, the signal obtained by compensating the known disturbance w in the compensation circuit 5 and the signal obtained by compensating the known disturbance It is added to one end of 4.

The compensation characteristics of the compensation circuits 5 and 6 are linear characteristics, and are determined from the following relationship. The control amount y is expressed by the following equation. y=Gwy(s)w+Gxy(s)x+Guy
(s)u...(1) Next, when the known disturbances w and x are input, u at which y becomes zero is found from equation (1), resulting in the following equation. u=-Guy(s)-1{Gwy(s)w+G
xy(s)x}...(2) Here, Gwy(s), Gxy(s) and Guy(s) are transfer functions, and s is a Laplace operator. That is, using u obtained by equation (2) as the output u°, the transfer function G1 (s) of the compensation circuit 5 and the transfer function G2 (s) of the compensation circuit 6 can be expressed by the following equations. G1 (s)=-Guy(s)-1Gwy(s
)
...(3) G2 (s) = -Guy(s) - 1G
xy(s)
…(4)

[0005]

However, since the conventional compensator 10 has linear characteristics, it has the disadvantage that sufficient control characteristics cannot be obtained for a nonlinear process control object. The present invention provides a feedforward control device using an inverse system that can obtain good control performance even if the characteristics of a process control target part are nonlinear.

[0006]

[Means for Solving the Problems] The present invention, which explains the above problems, has the following features. b) Compensation using an inverse function is effective when improving the control performance of a nonlinear process control target, but here the compensator configuration is nonlinear. (b) Normally, an inverse function of the nonlinear transfer characteristic would be required, but here we did not use an inverse function, but instead obtained an inverse function-like value as a result. c) The inverse functional value is obtained by a method of feeding back the function, and at this time, a simple characteristic is added in parallel to the function in order to stabilize the feedback system. d) The simple characteristic arranged in parallel is 1, which corresponds to the inverse function of the proportional differential action regulator provided in the feedback system.
The second lag characteristic was used.

[0007]

[Operation] In the present invention, a nonlinear process control target part is expressed by a nonlinear mathematical model, and by making full use of the nonlinear mathematical model, the manipulated variable is expressed as a mathematical formula to minimize fluctuations in the controlled variable of the process control target part. It is determined on the model and supplied to the process control target part as part of the manipulated variable signal. Specifically, first, the control amount y is estimated using a nonlinear mathematical model, and
A closed-loop system is constructed in which the estimated value y° of the controlled variable can be constantly brought closer to the target value r. A proportional differential action regulator is then placed within that loop. By doing so, a manipulated variable u is obtained that can suppress fluctuations in the controlled variable y when known disturbances x and w are applied to the process control target section. At this time, since the first-order lag characteristic is placed in parallel with the nonlinear mathematical model, the feedback system acts as a more stable system.

[0008] The above-mentioned action will be explained below using mathematical formulas. It is assumed that the estimated value y° of the control amount can be expressed by the following equation. y°=f°(u, x, w)

...(5) Here, u is the manipulated variable of the process control target part, and x and w mean known disturbances. The manipulated variable u, the output u° of the compensator, and P for controlling the process control target part
The function of the output u″ of the I regulator is given by the following formula: u=u°+u″

...(6) Also, by constructing a feedback system on the mathematical model, the following equation is obtained. δ=r−[y°+{1/K(1+TS)}u°]
…(7) u°
=K(1+TS)δ
…
(8) Here, δ means error, 1/K(1+TS)
means a first-order delay, and K(1+TS) means a proportional differential operation. Note that S is a Laplace operator. next,(
From equations 7) and 8, we obtain the following functional equation. 2δ=ry°

(9) Therefore, if the feedback system parameters K and T are adjusted to reduce the error δ, the estimated value y° of the control amount y approaches the target value r. That is, if the output u° thus obtained is applied as the output of the compensator to the process control target, controllability can be improved.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of the present invention will be described below in detail with reference to the drawings. Note that parts that perform the same functions as those in the prior art are given the same reference numerals, and redundant explanations will be omitted. FIG. 1 is a block diagram showing the entire feedforward control device according to an embodiment of the present invention, and FIG. 2 is a block diagram showing an extracted compensator 100. In FIG. 1, the process control target section 20 has nonlinear characteristics, and good control can be obtained by adjusting the output u° output from the compensator 100.

In FIG. 2, a mathematical model 101 simulates a process control target section 20, and a known disturbance x
, w and the manipulated variable u as inputs, and outputs an estimated controlled variable y° obtained by estimating the controlled variable y. Note that the manipulated variable u is the compensator 100
The output u° and the output u″ of the PI controller 3 are added to the adder 102.
It can be obtained by adding .

Next, a circuit in which a coefficient unit 103 and a first-order lag element 104 are connected in series is provided in parallel with the mathematical model 101 simulating the process control target section 20, and the output u of the compensator 100 is used as an input of the coefficient unit 103. °, and its output is 1
Input to the next lag element 104 and obtain the first lag element 10
4 and the output of the mathematical model 101, that is, the estimated value y° of the control amount, are added by an adder 105. The output of the adder 105 is subtracted from the target value r by a subtracter 106, resulting in an error δ.

The error δ is sent to the adder 10 via the coefficient unit 107.
At the same time, the output of the coefficient unit 107 is supplied to the differentiator 110 via the coefficient unit 109, and its output is connected to the other side of the adder 108. Then, the output of the adder 108 is input to the first-order lag element 111, and its output becomes the output u° of the compensator 100. Here, the first-order lag element 111 is provided to facilitate numerical calculations, and its time constant τ is set to a small value.

Note that K and T are adjustment parameters, which are determined by trial and error. In other words, K and T are selected so that controllability can be improved even if the characteristics of the process control target section 20 are nonlinear.

[0014]

[Effects of the Invention] As specifically explained above in conjunction with the embodiments, the present invention provides the following effects. (1) Since the compensator was created using a nonlinear mathematical model, the controllability is better than the conventional linear mathematical model. (2) Since there is no need to find an inverse function, it is easy to process when finding the output of the compensator. (3) The value corresponding to the output of the inverse function is the normal function (
(not an inverse function) is calculated by feeding back, and at this time, a first-order delay element is placed in parallel with a mathematical model that simulates the process control handling section, so the stability of the closed-loop system is improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the entire embodiment of the present invention.

FIG. 2 is a block diagram showing a compensator extracted from the embodiment.

FIG. 3 is a block diagram showing the entire conventional control device.

FIG. 4 is a block diagram showing an extracted conventional compensator.

[Explanation of symbols]

1, 20 Process control target part 3. Proportional-integral action regulator 10,100 Compensator 101 Mathematical model 103, 107, 109 Coefficient unit 104,111 1st order lag element 110 Differentiator

Claims (1)

[Claims] [Claim 1] A control deviation ε, which is the difference between a controlled variable y outputted from a process control target part to which the manipulated variable u is input, and a target value r, is determined, and this control deviation ε is calculated using a proportional-integral operation regulator. The output u'' obtained by the proportional integral calculation and the output u output from the compensator in order to reduce the influence of known disturbances on the control amount y.
In the control device which adds the operation amount u to the operation amount u,
The process control target part has non-linear input/output characteristics, and the compensator calculates a control amount y output from the process control target part according to the manipulated variable u from the manipulated variable u and a known disturbance.
A mathematical model that estimates and outputs the estimated control amount y°, a first-order lag element that processes the output u°, which is the output of the compensator itself, with a first-order lag, and the output of this first-order lag element and the mathematical model. The compensator is equipped with a proportional differential operation regulator that performs proportional differential calculation on the error δ, which is the difference between the sum of the outputs and the target value r, to obtain the output u°, and uses this as the external output of the compensator. A feedforward control device with a characteristic inverse system.