JPH0535306A - Dead time compensation control device - Google Patents

Abstract

PURPOSE:To realize much higher controllability by correcting the gain of a controlled system model and effectively utilizing the function of a Smith dead time compensation method while using a suitable gain ratio signal even to the gain change of a controlled system. CONSTITUTION:A dead time compensation part 5 equipped with a controlled system model 7 and a first-order delay means 6 is added to controlled systems (1-3) and although control is executed by outputting dead time elements to the outside of a control loop on the condition of the Smith method, the controllability is degraded by the deviation of the relevant condition due to the gain change of the controlled system. Then, the ratio is calculated according to a controlled variable and the output of the controlled system model by a gain ratio arithmetic means 11, stored in a storing means 12, afterwards transmitted to a gain correcting means 13 and multiplied to the output of the dead time compensation part 5, and the gain of the dead time compensation part 5 is corrected. On the other hand, the zero output of the dead time compensation part 5 is discriminated by updating and judging means (14 and 15) and the gain ratio is updated to the storing means 12.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a dead time compensation control device using a Smith method for compensating for dead time, and particularly to a dead time compensation control with an improved gain correction function of a controlled object model with respect to a gain change of a controlled object. Regarding the device.

[0002]

2. Description of the Related Art Controllers having PI or PID adjustment functions have been widely used in all industrial fields since the history of process control, and have become indispensable for plant operation.

If the controlled object is approximated by the dead time L and the first-order transfer function T (time constant), the dead time L can be easily controlled by the PID control when the controlled object has only the first-order lag. If it contains PID as the dead time L increases, that is, as L / T increases.
Control becomes difficult only with control.

Therefore, as a method for improving the controllability of the control system including the dead time, the O.D. J. MSmith (person name)
Is a so-called Smith, in which a dead time compensator using a controlled object model is added to PID control and the dead time is output to the outside of the control loop so that only the apparently first-order delayed controlled object is controlled. Method or Smith dead time compensation method has been proposed and is now widely used.

FIG. 3 (a) is a functional block diagram of a control device using the Smith method. This control device performs a deviation calculation means 1 for obtaining a deviation En between a target value SV _n and a control amount PV _n, and executes a PI or PID adjustment calculation based on this deviation En to control the obtained operation signal MV _n . The control system including the PID adjusting means 3 applied to the target 2 is provided with a dead time compensating unit 5 that allows the dead time to be controlled outside the control loop under a certain condition.

The dead time compensation unit 5 is a PID adjusting means.
Operation signal MV from 3_nThrough the first-order lag transfer function
First-order model means 6 to be controlled, similarly PID adjustment
Operation signal MV from the means 3_n1st delay and dead time
Object model means for outputting through transfer function with
(Hereinafter, this controlled object model means is called a controlled object model.
7) The output of the primary model means 6 is used as a control pair.
The subtraction means 8 for subtracting the output of the elephant model means 7
The output of this subtraction means 8 is the output side of the deviation calculation means 1
The deviation E of the deviation calculation means 1 is introduced into the subtraction means 4 provided in
The output of the dead time compensation unit 5 is subtracted from n.
There is. However, in FIG. G_Pe^{-Lp s} : Transfer function of controlled object G_P= K_p/ (1 + T_Ps): waste from the characteristics of the controlled object
Transfer function excluding time L_p: Dead time of controlled object K_p: Gain to be controlled T_p: Time constant of controlled object s: Laplace operator Also, G_M・ E^-LMs: Transfer function of controlled model G_M= K_M/ (1 + T_Ms): waste from controlled model
Transfer function excluding time L_M: Dead time of controlled model K_M: Gain of controlled model T_M: Time constant of controlled model , And the equivalent conversion of FIG.
Then, the configuration is as shown in FIG. By the way,
In the device of (a), the disturbance D is small and can be ignored.
The characteristics of the controlled object and the characteristics of the controlled object model 7 match
Condition, that is,   Disturbance = small, T_p= T_M, L_p= L_M                          …… (1) Assuming that the relationship is_P= G_MNext to this
SV in case_n→ PV_nThe transfer function between PV_n/ SV_n= {(G_c・ G_M) / (1 + G_c・ G_M)} E^{-LP s}                                                             …… (2)

Therefore, it can be converted into a simple structure as shown in FIG. Therefore, this control device can output the dead time element 9 to the outside of the control loop, and means that the PID adjusting means 3 may feedback-control the primary model means 6 excluding the dead time. Therefore, even if there is dead time in the controlled object, it can be easily controlled by the PID adjusting means 3, and good controllability can be realized.

[0008]

However, as is clear from the above description, the control device using the Smith method as described above is shown in FIG. 3 (c) unless the condition of the expression (1) is satisfied.
Cannot be configured as

However, in actual plant control,
It is difficult to always satisfy the condition of the formula (1), and the condition of the formula (1) shifts with time due to the characteristic change of the controlled object 2 such as the ambient temperature, the catalyst concentration, the raw material condition, and the load. It is a thing. As a result, the controllability deteriorates as the condition of equation (1) shifts, and the function of the Smith method is no longer fulfilled, which has a problem of greatly affecting the controllability of the plant.

The present invention has been made in view of the above situation, and it is possible to sufficiently exhibit the function of the Smith method even when the gain of the controlled object is changed due to the characteristic change of the controlled object or the environmental change. An object of the present invention is to provide a dead time compensation control device capable of executing highly efficient control.

[0011]

In order to solve the above problems, the present invention provides a control system for feedback controlling a controlled object with a dead time compensation having controlled object model means, first order model means and subtraction means. In the dead time compensation control device provided with a section,

Gain ratio calculating means for obtaining a gain ratio signal from the controlled variable from the controlled object and the output of the controlled object model means, storage means for storing the gain ratio signal obtained by the gain ratio calculating means, and this storage Gain correction means for multiplying the output of the dead time compensating section by the gain ratio signal stored in the means to correct the gain of the dead time compensating section due to the gain change of the controlled object,

Further, a signal discriminating means for discriminating a substantially zero signal output from the dead time compensating portion and a zero signal discriminated by the signal discriminating means are discriminated, or the discrimination time of the zero signal continues for a predetermined time. The gain ratio signal is updated to the storage means at this time, and the time ratio means does not update the gain ratio signal at other times.

[0014]

Therefore, according to the present invention, by taking the above means, when the gain of the controlled object changes so as to increase, the gain ratio signal obtained from the gain ratio calculating means increases, and this gain ratio signal is By multiplying the output of the dead time compensating unit, the gain of the dead time compensating unit can be automatically corrected in the direction of increasing the gain, and thus the appropriate operating state of the Smith method can be maintained.

Moreover, since the gain of the dead time compensating unit is corrected by using the gain ratio signal when the controlled object reaches the time corresponding to the complete response after the change of the operation signal, the Smith method function is optimally operated. It can be in a state, and higher controllability can be realized.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the device of the present invention will be described below with reference to FIG. In the figure, the same functions or parts as those of FIG. 3 are designated by the same reference numerals, and detailed description thereof will be omitted.

In this control device, a model gain correction unit 10 that automatically corrects the gain of the dead time compensation unit 5 in response to the gain change of the controlled object 2 is provided in the dead time compensation unit 5 to which the Smith method is applied. It is provided.

The model gain correction unit 10 takes in the controlled variable PV _n from the controlled object 2 and the output signal PV _Mn of the controlled object model means 7 forming a part of the dead time compensation unit 5, and controls the controlled object. Gain ratio calculating means 1 for obtaining a gain ratio signal k _n = PV _n / PV _Mn according to the gain change of 2
1 is provided, and the gain ratio signal k _n obtained here is stored in the storage means 12. Reference numeral 13 is a gain correction means for correcting the gain of the controlled object model means 7 according to the change in the gain of the controlled object 2. Here, the output of the dead time compensation section 5 is multiplied by the gain ratio signal read from the storage means 12. The gain-corrected signal thus obtained is supplied to the subtracting means 4.

The model gain correction section 10 also has a signal discriminating means 14 for discriminating an almost zero signal output from the dead time compensating section 5 and outputting a zero discrimination signal. The signal is introduced into the timing means 15. The time limiter 15 has a time counting function, and when the zero determination signal, that is, the output zero of the dead time compensation unit 5 continues for a predetermined time Ts or more, gives an update command signal to the memory 12 to give a gain ratio signal K _n. Of the controlled object model 7 is updated, and the update storage is stopped at other times, and the gain of the controlled object model 7 is corrected while holding the immediately preceding gain ratio signal K _n .

Next, the circumstances leading to the adoption of the configuration shown in FIG. 1, and the results of experiments and examinations will be described. Now, assuming that the complete response required time of the controlled object 2 is T _R , the complete response required time T _R is: T _R = almost dead time L _P + (3 to 5) · T _P

Can be expressed as Therefore, if the characteristics of the controlled object and the characteristics of the controlled object model match, if the characteristics of FIG.
After the change of the manipulated variable MV _n , as shown in (4), the controlled variable PV _n and the output PV _Mn of the controlled object model 7 match when the complete response required time T _R of the above equation has elapsed. Therefore, the gain ratio K _{n at} this time must be K _n = PV _n / PV _Mn = 1.

However, in practice, there may be cases where K _n = 1. This is because the gain of the controlled object characteristic changes due to changes in the characteristic of the controlled object 2 and environmental conditions such as ambient temperature, catalyst concentration, raw material conditions, and the size of the load, and the gains of the controlled object characteristic and the controlled object model do not match. Is. Therefore,
If the gain ratio K _n is obtained so that the gain of the controlled object model matches the gain of the controlled object characteristic, the condition of the Smith method is satisfied.

Therefore, it is necessary to measure the complete response time T _R , but since the dead time compensator 5 originally removes the dead time, the dead time L _M is removed from the complete response time T _R. Remaining time T _S , that is, T _S = (3-5) · T _M

By using the gain ratio K _n obtained after the elapse of this time T _S , a value that properly reflects the gain change of the controlled object characteristic and this gain is obtained. If the gain of the controlled object model is compensated by using the ratio K _n , the condition of the Smith method can be satisfied.

Therefore, the present apparatus has been tested and studied as described above.
It was realized while stepping on the results.
For modifying the gain of the controlled model for changes in gain.
I will explain with a mathematical formula. First, the PID adjustment means 3
Operation signal MV given by_nControlled variable PV_nAnd
And output PV of controlled object model 7_MnIs   PV_n= MV_n・ G_P・ E^{-LP ・ S}         = MV_n・ {K_P/ (1 + T_P・ S)} ・ e^{-LP ・ S}    …… (3)   PV_Mn= MV_n・ G_M・ E^{-LM ・ S}         = MV_n・ {K_M/ (1 + T_M・ S)} ・ e^{-LM ・ S}    …… (4) Can be expressed as On the other hand, the gain ratio calculation means 11
Output gain ratio signal K_nIs     K_n= PV_n/ PV_Mn                                    …… (5) However, the above equations (3) and (4) are substituted for this equation (5).
When entered, the following equation (6) is obtained.     K_n= (K_P/ K_M) ・ {(1 + T_M・ S) / (1 + T_P・ S)}                                           ・ E^{- (LP-LM) ・ s}…… (6)

Here, the output of the dead time compensator 5 becomes zero, the signal discriminating means 14 discriminates the zero signal, and the time limiting means 15 is reached.
When the update command signal is output from, since the complete response required time T _R has been exceeded after the change of the operation signal MV _n , (1 + T _M · s) / (1 + T _P · s) = 1 in the equation (6) above. , e ^{- Since (LP-LM) · s} = 1, the equation (6) eventually becomes K _n = (K _P / K _M ) ... (7).

Therefore, the gain ratio obtained by this equation
A dead time compensator for introducing the rate signal into the gain correction means 13.
Output signal X of 5_nGain correction by multiplying
Output Y from stage 13 as_nIs obtained.   Y_n= K_n/ X_n                                          …… (8)       = (K_P/ K_M) ・ {K_M/ (1 + T_M・ S)}                         ・ (1-e^{-LM ・ S}) ・ MV_n       = {K_P/ (1 + T_M・ S)} ・ (1-e^{-LM ・ S}) ・ MV_n… (9)

Therefore, as is apparent from the equation (9), if the output X _n of the dead time compensation unit 5 is multiplied by the gain ratio signal K _n of the controlled object model gain correction unit 10, the dead time compensation unit 5's It is shown that the controlled object model gain K _M is always automatically corrected by the controlled object gain K _P.

Therefore, in general, the Smith dead time compensation method is applied to a controlled object including dead time when the disturbance is small and the controlled object characteristic and the controlled object model match. The gain of the controlled object characteristic changes due to the change in the characteristic of the controlled object 2 or the environmental conditions, and the gains of the controlled object characteristic and the controlled object model do not match, which greatly affects the controllability.

Therefore, the present apparatus obtains the gain ratio signal from the output of the controlled variable and the controlled object model, and stores the gain ratio at the time corresponding to the complete response time after the change of the operation signal. Since the gain of the controlled object model is modified using the stored gain ratio signal, the controlled object model gain of the dead time compensating unit 5 is modified based on the change gain of the controlled object. It is possible to realize a dead time compensation control device having a high controllability by fully exerting the function of the time compensation method.

Particularly, in an actual plant, frequently and
The gain of the controlled object changes greatly, but in such a case, by automatically correcting the controlled object model gain, the applicability to the actual plant is significantly improved, and if it is often used for control devices scattered throughout the plant. In addition, it can greatly contribute to the improvement of controllability of the entire plant, and can realize full-scale flexibility of plant operation, super automation and high quality products.

The present invention is not limited to the above embodiment. That is, in the above embodiment, after the output of the gain ratio calculation means 11 is stored in the storage means 12 as it is,
Although it is sent to the gain correction means 13, the ratio K _{n is} output to the output side of the ratio calculation means 11 or the storage means 12, for example.
Alternatively, the smoothing means for removing the instantaneous fluctuation of Also, the time limit means 15 is removed and updating is performed with the output of the dead time compensation unit 5 being zero, or the time limit means 15 is used.
Of may be zero, for example, T (3~5) · T _M of the time set value. In this case, the region due to the time constant T _M is simply ignored and the gain change can be approximately applied. In addition, the present invention can be modified in various ways without departing from the scope of the invention.

[0033]

As described above, according to the present invention, the function of the Smith dead time compensation method can be fully utilized even when the gain of the controlled object is changed due to the characteristic change of the controlled object or the environmental change. In addition, it is possible to provide a dead time compensation control device capable of performing gain correction using an appropriate gain ratio signal and maintaining higher controllability.

[Brief description of drawings]

FIG. 1 is a functional block diagram showing an embodiment of a dead time compensation control device according to the present invention.

FIG. 2 is a diagram for explaining the progress until almost complete response when a system having a dead time L _M and a time constant T _M is changed.

FIG. 3 is a functional block diagram of a conventional dead time compensation control device.

[Explanation of symbols]

1 ... Deviation calculation means, 2 ... Control object, 3 ... PI or PI
D adjusting means, 4 ... subtracting means, 5 ... dead time compensation section, 6 ...
Primary model means, 7 ... Controlled model means, 8 ... Subtraction means, 10 ... Model gain correction section, 11 ... Gain ratio calculation means, 12 ... Storage means, 13 ... Gain correction means, 14 ...
Signal discriminating means, 15 ... Timer means.

Claims (2)

[Claims] 1. A control system for executing a PI or PID adjustment calculation in an adjusting means on the basis of a deviation between a target value and a controlled variable from a controlled object and applying a resulting operation signal to the controlled object is wasteful. A controlled object model means approximated by time and a first-order transfer function, a first-order model means excluding dead time elements, and a subtraction means for subtracting the output of the controlled object model means from the output of the first-order model means. In the dead time compensation control device having a dead time compensating section by the Smith method for providing the dead time compensation signal obtained from the subtracting means to the input side of the adjusting means, the control amount from the control target and the control Gain ratio calculation means for obtaining a gain ratio signal from the output of the target model means, storage means for storing the gain ratio signal obtained by the gain ratio calculation means, and this storage means Gain correction means for multiplying the stored gain ratio signal by the output of the dead time compensating section to correct the gain of the dead time compensating section due to the gain change of the controlled object, and almost zero output from the dead time compensating section. And a determination unit for updating the gain ratio signal with respect to the storage unit when the above signal is discriminated. 2. The update determining means includes a signal determining means for determining a substantially zero signal output from the dead time compensating section,
When the zero signal determination time determined by the signal determination means continues for a predetermined time, the gain ratio signal is updated in the storage means, and the gain ratio signal is not updated at other times. And a dead time compensation control device.