JPS6142003A - Automatic adjusting method of control constant - Google Patents

Abstract

PURPOSE:To adjust a control constant to an optimum value and to allow process variables on the lower side of a control system to respond with excellent stability and promptness by identifying the transmission function of a controlled system in the control system and that of the controlled system on the lower side of the control system and allowing the transmission function from a target value of a controlled variable to a process variable, whose response control is required, on the lower side of the control system to coincide with the transmission function of a reference model showing an ideal response characteristic. CONSTITUTION:A control operation device is provided with the second controlled system (process) 6 on the lower side of the control system, the second identifying operation part 7 which identifies the transmission function of the second controlled system in accordance with a controlled variable signal ys(t) and a process variable signal xs(t), and the second detector 8 which detects a process variable x(t) to output the process variable signal xs(t). The first and the second controlled systems are approximated by self-regression moving average models, and G1(s) and G2(s) are identified in accordance with time series signals vs(t),ys(t), and xs(t) before and after them. A control constant operation part 4 obtains the transmission function from a target value rs(t) or y(t) to the process variable x(t) and obtains the optimum control constant on a basis of an operation formula so that this transmission function coincides with a transmission function Gm(sigma, s) of the reference model showing the ideal response characteristic, and the control constant of a control operation device 2 is corrected. Thus, the control constant by which the process variable in the lower side of the control system responds ideally is obtained easily without feeding back this process variable to the control system.

Description

Detailed Description of the Invention [Field of Application of the Invention] The present invention relates to a control constant automatic adjustment method for automatically adjusting the control constant of a controller that controls a process to an optimal value with good coordination and stability. [Background] This type of method is described in ``Control system design method based on partial knowledge of the controlled object'' (Proceedings of the Society of Instrument and Control Engineers, Vol. 5,
The (partial) model matching method described in No. 4, 5491555, 1984-8) is known. A brief overview of this is given below. Figure 1 shows an overview of the model matching method.

l is the controlled object (process), 2 is the control calculation section that performs control calculations such as PID, 3 is the identification calculation section that identifies the transfer function of the controlled object, 4 is the transfer function G of the control calculation section 2, (8)
5 is a control constant calculation unit that calculates the optimal value of the control constant of the controlled object, and 5 is a detector that detects the controlled variable y(t) and outputs the controlled variable signal y,(t). The transfer function ()+(1m) for the controlled variable is identified, and the closed-loop transfer function W(I) from the target value signal ``,(1) to the controlled variable y(I) is the reference model representing the ideal response. This is a method of determining the optimum value of the control constant of the control calculation device 2 so as to match the transfer function G, (σ+').

First, the controlled object is approximated by an autoregressive moving average model, and its transfer function G + (S) is identified from the operation signal v, (ri and control amount signal ys (l) using the Kalman filter algorithm, and is expressed as Deform @Next, for example, when the control calculation device 2 is in PI operation, its transfer function G, (
S) is K as follows, where K is the proportional gain and T
x is a control constant called an integral time constant.

Using G, (S), G, (S), the closed loop transfer function W(S) becomes as follows.

Next, the transfer function G of the reference model representing an ideal response,
Put (σ, s) as follows◎G, (σ, s) −(
4) α豐＋屯σS十α嘗σ! S3+α穆σsu8+…(α1=
α+=1) A reference model that responds to the step increase (decrease) K of the target value signal r, (t) without overshoot (undershoot) of the control amount y(t) is α+=0.
．． 375, αs=o, o625, α4=0.0039.

If equations (3) and (4) match, the following equation holds true.

Gl(8) G,(s) = (
s)■ =-[la+ (gr Fhthσ)3σ3 + (gr-g+α snow σ0g・(αI-αl)σ1)+
...) (7) From the condition that equations (2) and (7) match, the following equation is derived by comparing each term of the numerator.

(0th order term of S in numerator) (1st order term in 3 in numerator) [(p=z-(g wealth - g・αkuraσ)
(9) σ (quadratic term of 3 in the numerator)
(10) Although there are actually terms higher than the third order of 3 in the numerator, ideal control characteristics can be obtained by matching the lower order terms of S, which are important in terms of characteristics. In this case, there are only three unknowns: the control constant of □Tx and σ corresponding to the rise time, so find σ from equation (10) and convert it into (9
) to obtain K, and substitute these into equation (8) to obtain Tx.

This method is a simple method in which the optimal values of control constants, which were conventionally determined by many parameter surveys, are determined by solving algebraic equations. There is a problem in that it cannot be applied when there is a process having the characteristics of Kal and others and it is necessary to give an ideal response to the process quantity. The purpose is to create a control constant a that can provide a stable and highly responsive response to the downstream process variables of the control system.
[Summary of the Invention] The present invention identifies a transfer function of a controlled object in a control system and a transfer function of a controlled object on the downstream side of the control system, and calculates a response from a target value of a controlled variable. The transfer function to the process variable on the downstream side of the control system that needs to be adjusted is made to match the transfer function of the reference model representing the ideal response characteristics, and the control constant is set to the optimum value Vcv! 4
It was designed to be adjusted.

[Embodiments of the invention]

FIG. 2 shows an embodiment of the present invention. 6 is the second controlled object (process) on the downstream side of the control system, 7 is the controlled variable signal y, (t)
and process quantity signal x, (2) a second identification calculation unit that identifies the transfer function of the second controlled object from It is a vessel.

First, the first controlled object and the second controlled object are approximated by an autoregressive moving average model, and the time series signals before and after that, v, (t),
From y, (t), x, (t), Gl(8), am(s)
is identified and transformed as follows.

Next, using a, (s), G+('), Gj (!I), [from the target value r of y(t), (t), the process amount x(t)
The transfer function w*(s) up to
G, (!l) G word (S) Now, assuming that the control calculation unit 2 is in PI operation, and substituting equations (14) 2K (2), (4), (11), and (12), we get the following. 0 The formula holds true Next, the coefficients of each term are compared and the following formula holds true.

(0th order coefficient of S) Ts c1+a Fhs (1st order coefficient of S) (2nd order coefficient of S) Find the minimum positive σ from this formula (1B) and convert it to (17
), calculate , and substitute these into equation (16) to obtain T! seek. The control constant calculation unit 4 calculates K $ and T s based on these calculation formulas, and corrects the control constants in the control calculation unit 20. As described above, in one embodiment of the present invention, the This embodiment has the advantage that control constants that can provide an ideal response to the process variables downstream of the control system can be easily determined without feeding the process variables downstream to the control system. In this case, the control system may be an I-PD control system.

In the embodiment described above, a compensation circuit may be included in the control system.

〔Effect of the invention〕

According to the present invention, when there is a process on the downstream side of the feedback control system and it is necessary to control the process quantity, it is possible to easily obtain control constants that can give an ideal response to the process quantity. can. Furthermore, there is no need to feed back this process amount, that is, there is no need to improve the existing control system, so it is highly economical.

[Brief explanation of the drawing]

This is a block diagram.

Claims (1)

[Claims] 1. Calculate the control constant from the identified control target transfer function,
In a method of adjusting control constants of a control calculation device, the control constants of the control calculation device are adjusted so that the transfer function from the target value of the controlled variable to be fed back to the most important amount of change in the process matches the transfer function of the reference model. A control constant automatic adjustment method characterized by automatic adjustment.