JPS60263207A - Process control device - Google Patents

Abstract

PURPOSE:To reduce remarkably an operator, and also to improve the quick responsiveness of a control system by calculating a control constant of a feedback/ feedforward control system by using a parameter of a controlled system and a new reference model. CONSTITUTION:A feedback/feedforward controller 1 generates a final operating signal u(t) by providing a main controller 3 for inputting an output y(t) of a controlled system 2, and inputting a deviation e(t), a feedback compensating operator 4 of the output y(t), and a feedforward compensating operator 5 of a target value r(t). A control constant operating device 6 is constituted of a transfer function parameter input operator 7 for executing a transfer of a parameter of a transfer function Gp(s) of the controlled system 2, a design parameter input operator 8 for inputting a parameter of a reference model and a degree of the compensating operator, and a control constant operator 9 for calculating a control constant based on a partial model matching, inputs an output of the operator 7 and 8 to the operator 9, and outputs a control constant of each compensator of the controller 1 and the main controller.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a process control device that controls a process to be controlled using a feedback controller and a feedforward controller.

[Technical background of the invention and its problems]

Until now, the calculation formula for the control constants of a feedback/feedforward servo system that suppresses disturbances and causes the output to follow the target value signal has been developed using the method of Mori et al. Design of multivariable linear optimal tracking system, Proceedings of the Society of Instrument and Control Engineers, Vol. 1
0. NQl, 1972) is known. In other words, this defines the dynamic characteristics of the controlled object as x (t) = A・
x (t) + B−u (t) −−−−・−(1
)y (t)=C-x (t)...
(2), and by using an expanded system including an integral controller and a feedforward compensator, the operation signal that minimizes the evaluation function is set to 1(i
The calculation is performed using the solution matrix of the ccati equation. Here, V(t) is the output, X(t) is the state quantity, u
B) is the input, A and BXC are nxn and nx respectively
m, Pxn matrix.

However, it is difficult to accurately measure the dynamic characteristics of an actual controlled object in the form of an equation of state, and R
This method has practical problems because it is not easy to solve the 1ccati equation efficiently.

On the other hand, Mr. Kitamori (Design method of control system based on partial knowledge of controlled object, Proceedings of the Society of Instrument and Control Engineers, Vol.
, 15. No. 4.1979), a control system model Gm(S).

(3) We have proposed a method of calculating the control constants of the I-PD control system and the PID control system so as to match this model. This method is highly practical because it is based on low-frequency parameters of the controlled object that are relatively easy to identify, and the compensator can be constructed based on usable measurement signals. However, this is basically a proposal for a calculation method for the control constants of the feedback compensation calculator, and the design of the zero point that contributes to improving the correspondence with the target value of the control system, that is, the control constants of the feedforward compensation calculator. does not indicate any specific calculation method.

[Purpose of the invention]

An object of the present invention is to provide a process control device that improves coordination with a target value based on partial knowledge of a controlled object and has means for calculating control constants for a feedback control system and a feedforward control system. There is.

[Summary of the invention]

A process control device according to the present invention is a means for calculating control constants of a feedback/feedforward control system based on the concept of partial model matching using parameters of a controlled object and a new reference model having poles and zeros. It is equipped with the following.

〔Effect of the invention〕

In the feedback/feedforward control system,
Since the control constants of not only the feedback compensation calculator but also the feedforward compensation calculator can be calculated using only a few simple steps of four arithmetic operations based on partial model matching, the number of calculators can be significantly reduced. In addition, due to feedforward compensation, there is a component that acts directly on the operating end in response to a change in the target value, so that the coordination of the control system can be improved. Furthermore, since feedforward compensation operates outside of feedback compensation, it has the advantage that the ability to suppress disturbances does not deteriorate in any way.

[Embodiments of the invention]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of a process control device according to an embodiment of the present invention.

In the figure, the part 1 surrounded by the two-dot chain line indicates the feedback feedforward controller, and 2 indicates the controlled object.The feedback feedforward controller 1 is the output of the controlled object 2. a main controller 3 that feeds back y(t) and inputs the target value r(t) to input the deviation e(t) in order to generate the final operation signal uB);
It includes a feedback compensation calculator 4 for output y(t) and a feedforward compensation calculator 5 for target value r(t). A portion 6 surrounded by a two-dot chain line is a control constant calculation device. This control constant calculation device 6 inputs the transfer function parameter input calculation unit (GPIN) 7 which inputs the parameters of the transfer function G11(S) of the controlled object 2, the parameters of the reference model, and the order of the compensation calculation unit. A design parameter input calculator (DPI) 8 and a control constant calculator (C) that calculates control constants based on partial model matching.
PD)9, GPIN7 output and DPI8
is input to the CP[) 9 to output the control constants of each compensation calculator and the main controller of the feedback/feedforward controller 1.

Next, the calculation of each control constant by the control constant calculation device 6 will be explained.

Now, suppose that the transfer function Gl)'(S) of the controlled object 2 has been identified as ......(4) by some identification means, and GPIN7 is 00,
gl, (12, lJ3).Next, DPI8 inputs the parameters of the reference model Mr(S) expressing the control specifications of the servo system, ......(5), α2.α3. α4..., β1
As well as inputting ゜β2..., the transfer functions μΣ, Fe(S), FF(S),
However, input the orders nc, nB, and nF.

And CPD9 is gi (i = 0.1, 2...
), αi (i = 2', 3...), ri (
i=1.2. ), no2port B, nF, the following control constants, Co, C1, C2, ・= −
, Cn.

fso, fBs,..., fBnBfFl, f
F2 l..., calculate fFnF. For example, ri
. == o, nB=1, np=0, α2, α3, α4
,β! , then from the target value r(t) to the output y(
The transfer function up to t-) is y (t) C(S) 19F
F(S)...(7) Therefore, when the denominator and numerator are grouped into powers of S and matched with the reference model of equation (5), the following is obtained.

1+β1σS+α2I2(σS)2 (8) Therefore, by assuming that the denominators and numerators on both sides are equal, the following equation is obtained.

Here, in equation (6), n=o, nB=i, nF
= O, the following relationship can be obtained from equation (9).

From this equation (10), the following design rule is derived.

That is, when the order of each compensation calculator is set, (11)
Control constants for not only the feedback system but also the feedforward system can be calculated using simple arithmetic operations such as the following.

Next, the results of controlling a process using this control device will be explained.

In this control device, no-0, nB=1, n, w
By setting O, the feedforward I-PD
A control system is constructed, and FIG. 2 is a block diagram thereof. Each control constant is calculated by equation (11) using the process parameters and set in each compensation I calculation unit. Third
The figure shows the response waveform of this control system. Then, in FIG. 3(a), the scattering is set to d(t)=C, and the eye 11
It shows the response of the output V(t) when 111r(t) is changed by a unit step with 1=0. Moreover, the dotted line is the output response when there is no feedforward control system, and the solid line is the output response when the designed feedforward control system is used. Thus, it can be seen that the interlockability is improved. Figure 3(b) shows the target value r(
t)=C and the disturbance d(t) is changed by a unit step at time 1=0. Similar good regulation waveforms are obtained regardless of whether a feedforward control system is used or not.

Note that the present invention is not limited to the embodiments described above. In other words, the control system in Fig. 2 is equivalently transformed to:
It can be configured with a feedforward PID control system having a target value filter as shown in FIG. Here, 3-2 is the target value filter, and 3-1 is the PID II
It is a JII device.

Kc -T; , Td is from Co, Cs, C2 (
12) It can be determined from the relationship of Eq.

Furthermore, the control systems shown in Figures 1 and 2 have a structure in which no steady-state deviation remains with respect to the step target value and step disturbance, but a steady-state deviation remains with respect to the ramp target value and lamp disturbance. Put it away. In such a case, n=2 may be used in the control system shown in FIG. However, in this system, each compensation calculation polynomial 〇(S), F
F (S ) and Fe (S ) are respectively C(S
) =C(1+Ct S +-...”+cn, 5nc
FF (S)=fp(1+fFl 3+...
・+L□fSnfFB (S) = fBtr +
fat S +−−−・−+f snbnb.

Generally, if the order of the larger of the external signal, the target value, and the disturbance is ne, then by setting n=n + 1 (13), the steady-state deviation can be made zero. can. However, the order of the step signal is ne = 0, and the order of the ramp signal is ne-i
It is. By configuring as shown in FIG. 5 even for such high-order external signals, it is possible to obtain a control device in which the steady-state deviation is zero. Moreover, it can be realized not only for control systems using continuous arithmetic elements but also for discrete value systems such as digital process controllers with built-in microcomputers.

[Brief explanation of drawings]

FIG. 1 is a block diagram of a control device according to an embodiment of the present invention, and FIG. 2 is a feedforward/IPD diagram of the same device.
Figure 3 is a block diagram of the control system; Figure 3 is a response waveform diagram of the control system; Figure 4 is a block diagram of the target value filter and feedforward PID control system equivalent to Figure 2; Figure 5 is a diagram of the response waveform of the control system; FIG. 3 is a block configuration diagram of a feedforward (PD) control system in which no steady-state deviation remains for the next external signal. 3... Main controller, 4... Feedback compensation calculator, 5... Feedforward compensation calculator, 6... Control constant calculator, 7... Transfer function parameter input calculator 8... Design parameter input calculator, 9... control constant calculator. Applicant's agent Patent attorney Takehiko Suzue Figure 1 8

Claims (1)

[Claims] (1) A feedback control calculator that inputs an output signal and a target value signal of a process to be controlled and outputs a first operation signal, and a second feedback control calculator that inputs the target value signal and directly acts on the control target. a feedforward control calculator that outputs an operation signal of the control object, and adds the first operation signal and the second operation signal to obtain an operation signal of the control object, By inputting and setting the dynamic characteristic parameters of the reference model having zero points that define the control characteristics of the control, and the orders of the feedback control calculator and the feedforward control calculator, the partial model can be created. A process control device comprising means for calculating control constants of the feedback control calculator and the feedforward control calculator based on a matching method. (2) The feedback control calculator performs proportional, integral, and differential calculations on the target value filter, the deviation signal between the signal obtained by the target value filter, and the output of the process to be controlled, and obtains the operation signal for the process. The process control device according to claim 1, characterized in that it is constituted by a main controller. A subtracter that calculates the deviation between A process control device according to claim 1, characterized in that: