JPH0442303A - Double freedom degree controller - Google Patents

Abstract

PURPOSE:To specify the optimum value of a double freedom degree coefficient for the integral time by using a target value filter where a primary delay element defining the integral time as a time constant is combined with a controlled system model means. CONSTITUTION:A target value filter means 10 is provided at least with a primary delay element 15 defining the integral time of a PI (proportion and integration) control means 4 as a time constant and a controlled system model means 14 which has a transmission function including the unit transmission function of a controlled system model out of the transmission functions of a controlled system 2. Therefore the means 10 can have a transmission function accordant with the characteristic of the system 2. Thus the optimum value of the double freedom degree can be optimized for the integral time.

Description

Detailed Description of the Invention [Objective of the Invention] (Industrial Application Field) The present invention relates to a two-degree-of-freedom PID adjustment device having a target value filter, and in particular a first-order delay element having an integral time as a time constant. The present invention relates to a two-degree-of-freedom control device using a target value filter combined with a controlled object model means.

(Prior Art) PI or PID control devices have been widely used in all industrial fields since the beginning of process control, and it is no longer possible to do without PID control devices in the instrumentation field.

By the way, various control calculation methods have been proposed for PID control devices, and even now that the analog control system has been replaced by a digital control calculation method, the throne of the PID control device has not changed in the slightest, and it is important for plant operation. It forms the basis of control.

That throne is unlikely to change in the future.

The basic formula for such a PID adjustment calculation method is... (
1) It is expressed as. However, MV (s) is the operation signal, E (s)
is the deviation, K is the proportional gain, T1 is the integral time, TD is the differential time, S is the Laplace operator, and (1/η) is the differential gain.

Such a PID adjustment calculation method is called a one-degree-of-freedom PID adjustment calculation method, and is a so-called one-degree-of-freedom method in which only one set of PID parameters can be set. As a result, in an actual control system, the values of the disturbance suppression optimum PID parameter and the target value tracking optimum PID parameter are significantly different, and it is not possible to simultaneously optimize both the disturbance suppression characteristic and the target value tracking characteristic. It has contradictory properties.

In other words, the PID
When parameter values are set, the target value tracking characteristics become oscillatory, and conversely, the PI
Setting the D parameter value will improve the disturbance suppression characteristics.
It becomes.

Therefore, in this type of PID control device, there is a desire for a technology to simultaneously optimize the disturbance suppression characteristics and the target value tracking characteristics.

In contrast, in 1963, I 5sac, M +H
Two degrees of freedom PI or PID algorithm (T vOD degrees of 'F r
eedom P I or P I D A Igo
The basic concept of ritl+n+ (hereinafter collectively referred to as 2DOF PID) was announced, and in recent years, 2DOF PID has been put into practical use in Japan based on this basic concept, and is greatly contributing to the advancement of plant operation control.

Figure 7 shows a conventional 2DO with two degrees of freedom for P and D motions.
FIG. 2 is a diagram showing the configuration of an FPID control device. This device is provided with a target value filter 1 that performs arithmetic processing to make the proportional gain KP and the integral time TI two degrees of freedom, and the target value S
2, the control target value SVo obtained by the target value filter 1 and the control amount Pv' from the controlled object 2 are led to the deviation calculation means 3, where the calculation (SVo - PV') is performed to calculate the deviation E. and introduces it into the PI adjustment means 4. The PI adjustment means 4 receives the deviation E' and performs a PI adjustment calculation to obtain a manipulation signal MV', which is introduced into the subsequent addition means 5. This addition means 5 adds and synthesizes the operation signal MV' and the disturbance signal R, and applies the obtained addition and synthesis signal to the controlled object 2, so that the control target value 5V0=control amount PV'
Control so that

The target value filter 1 includes a lead/lag element 11 that gives an appropriate lead or lag to the target value Sv, and a first-order lag element 12 whose time constant is an integral time TI that gives a first-order lag to the target value SV. , an incomplete differentiator 13 for imperfectly differentiating the output of the first-order lag element 12, and the lead/lag element 1. The subtraction means 14 subtracts the output of the incomplete differentiator 13 from the output of the subtraction means 14 to obtain the control target value Svo, and then introduces the control target value Svo into the deviation calculation means 3.

Therefore, according to the PI control device configured as above, P
If the transfer function between V and MV is Cpv(s) and the transfer function between Sv and MV is C5v(s), then the following equations are obtained. In the above equation, α is a two-degree-of-freedom coefficient for proportional gain, and β is a two-degree-of-freedom coefficient for integral time.

Therefore, as is clear from the above equation, by determining K p ST + so that the disturbance suppression characteristics are optimal, and then determining α and β so that the target value tracking characteristics are optimal, the two free degree can be achieved.

(Problem to be solved by the invention) However, the above 2 DOF PI or P
Although the ID control device has various features, the following problems have been pointed out.

■ The optimum value of the two-degree-of-freedom coefficient β of the integration time cannot be determined.

The value of the two-degree-of-freedom coefficient α of the proportional gain is determined by CHR (Chjen) in the optimal adjustment method of PID parameters.

However, the value of the two-degree-of-freedom coefficient β of the integration time cannot be specified.

What happens is that (1/T1 -5)(β/
(1+T, ・S)l, (1/T 1S
) is the transfer function, and the human output characteristic is expressed as (E1) in Figure 6. If the 11 value is gradually increased from this (E1) state, (E2), (E3)... On the other hand, when β/(1+T, .multidot.S) is used as the transfer function, the characteristic is as shown in (b) of FIG. 8.

In other words, one side is a straight line and the other side is a curved line. As a result, in the case of a transfer function consisting of (1/T1 ・5)-(β/ (1+T+-5)), the input/output characteristic changes in slope as shown in (c), and the integration time has two degrees of freedom. The value of β cannot be determined. Therefore, in general, it is necessary to perform a simulation to specify each characteristic of the controlled object, which is extremely time-consuming and complicated.

■ The relationship between the value of the two-degree-of-freedom coefficient β of the integration time and the characteristics of the controlled object is not clear.

Even if the two-degree-of-freedom coefficient β of the integration time is found through simulation, the relationship with the characteristics of the controlled object is not clear. The reason for this is that, as is clear from equations (2) and (3), the time constant and dead time of the controlled object must be expressed only by the two-degree-of-freedom coefficient β of the integral time.

Therefore, in order to smoothly transition from a 1-degree-of-freedom PID to a 2-degree-of-freedom PID, it is necessary to completely eliminate the above-mentioned conventional defects, and from both a practical and academic point of view, it is necessary to change the integral time's 2-degree-of-freedom coefficient. β must be easily specified and its relationship to the characteristics of the controlled object must be clarified.

The present invention has been made in view of the above circumstances, and employs a controlled object model for the target value filter, completely specifies the two degrees of freedom of the integral time, and combines the two degrees of freedom of the integral time and the controlled object. It is an object of the present invention to provide a two-degree-of-freedom control device that clearly relates the relationship between the two degrees of freedom.

[Structure of the Invention] (Means for Solving the Problems) In order to solve the above problems, the two-degree-of-freedom control device according to the present invention introduces a target value and uses the control target value obtained through the target value filter means and the control Among the PID (P: proportional, I: integral, D= differential) adjustment calculations, at least the PI adjustment means executes adjustment calculations based on the deviation from the control amount from the control object, and the obtained operation output is applied to the control object. In a two-degree-of-freedom control device that controls the control amount so that it becomes a target value by applying
A configuration consisting of a combination of a first-order lag element having the integration time of the adjustment means as a time constant, and controlled object model means having a transfer function that includes a unit transfer function of the controlled object model among the transfer functions of the controlled object. It is.

(Function) Therefore, by taking the above-mentioned measures, the present invention achieves a transfer including a first-order lag element whose time constant is the integration time of the target value filter means or the PI adjustment means and a unit transfer function of the controlled object. By providing a control target model means for the function, it is possible to identify the optimal value of the two-degree-of-freedom coefficient for the integration time.
In addition, since the control target value is applied to the deviation calculation means while increasing the control target value at a time corresponding to one cycle of the response of the controlled object in response to a change in the target value, the integration time has two degrees of freedom and the relationship with the controlled object. can be clearly related.

(Example) Hereinafter, an example of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing the configuration of a 2DOF PID control device. In this figure, the same parts as in FIG. 7 are given the same reference numerals, and detailed explanation thereof will be omitted. This control device is provided with a target value filter means 10, which performs arithmetic processing to convert the target value into a proportional gain and two degrees of freedom regarding the integral time T1 to obtain a control target value Svo. It is introduced into the deviation calculation means 3 together with the control amount P■ from 2, and the control deviation E=SVO-PV is determined here and supplied to the PI adjustment means 4. This PI adjustment means 4 executes a PI adjustment calculation based on the control deviation E, and introduces the obtained PI adjustment calculation output, that is, the operation signal MV, into the addition means 5, which adds and synthesizes the operation signal MV and the disturbance. . And addition means 5
Apply the addition composite signal from 5v-sv to the controlled object 2.
This is a configuration in which control is performed so that o = pv.

The target value filter means 10 multiplies the target value SV by a two-degree-of-freedom coefficient α of the proportional gain in the coefficient means 11, and then inputs it into the addition means 12, and the subtraction means 13 multiplies the target value SV.
After subtracting the output of the coefficient means 11 from V, the output is passed through the controlled object model means 14 with a gain of 1 and the first-order delay element 15 whose time constant is the integration time T1, and then the addition means 1
2, and is added and combined with the output of the coefficient means 11 to obtain the control target value Svo, which is introduced into the deviation calculation means 3.

Next, the reason why the target value filter means 10 having the above configuration is adopted will be explained. Now, FIG. 2 is a diagram showing the configuration of a target value filter type two-degree-of-freedom PID control device,
Here, the transfer function G (s) of the controlled object 2 is: G (s) = KM - gu (s)
−(4). KM is the gain of the controlled object model, and gM(s) is the unit transfer function of the controlled object model.

Therefore, in the control device shown in FIG.
Let us examine what properties the target value filter means 10 must have in order for PM introduced into the filter to have the same properties as S V o .

Now, if the transfer function between the disturbance D and PV is G op (S), then GDP (S) = PV/D = G (s) = KM - gM (S)
...(5) Therefore, the transfer function G 5s(
S) is a transfer function including G55(s) -8Vo /SV-H(s)-gM It is necessary that the relationship is as shown in (6). That is, from equation (5), S V =
The transfer function G55(S) between S Vo, that is, the transfer function H(s) of the target value filter means 10 is the disturbance D-PV
In order to have the same properties as the transfer function G Dr(S) between, the transfer function must include gM(S). The transfer function including this gu(s) can be determined in advance from the step response of the controlled object.

By the way, the lead/lag element 11 in the target value filter means 1 in the conventional 2DOFPID control device shown in FIG.
If we perform equivalent transformation on , we will obtain , which is expressed as a functional block as shown in Fig. 3.

Therefore, by inserting the filter of equation (7) above into the input side of the deviation calculation means 3, only the P motion becomes two degrees of freedom, but when a unit step is inserted into the input side of the filter, as shown in FIG. b) Output response characteristics can be obtained. However, this output response characteristic does not take into account the characteristics of the controlled object 2, that is, the dead time and delay of the controlled object 2, and PV and Svo, which are input to the deviation calculation means 3, do not have the same characteristics.

On the other hand, the unit transfer function gM(
If s) is introduced, the following is obtained. Similarly, if the unit step is input to the filter of equation (8), the time T corresponding to exactly one round of the controlled object 2 is obtained, as shown in (b) of Fig. 4. In other words, the control target value SVo is set by the deviation calculation means 3 so as to be started at the time when the control result appears with dead time and delay of the controlled object 2.
can be entered.

Therefore, the target value filter means 10 in FIG. 1 is constructed using the transfer function of equation (8) obtained under the above-mentioned preconditions. Therefore, from the above-mentioned background, in the control device shown in Fig. 1, the transfer function c from PV to MV is
PM(S) becomes, and the transfer function between S→MV is 0
5M(s) becomes, which means that P and I have two degrees of freedom.

In other words, equation (10) can be made to have a relationship that depends only on the two-degree-of-freedom coefficient α of the proportional gain by removing the two-degree-of-freedom coefficient β of the integral time, and moreover, the two-degree-of-freedom coefficient α of the proportional gain can be specified by the CHR method as described above, so that the integration of the integral term in equation (10) into two degrees of freedom can also be easily specified.

Moreover, the unit transfer function gM(s) of the controlled object model is −
Boat fishing involves wasted time + first-order delay, that is, wasted time that is ten higher-order advances/lags.

Next, FIG. 5 is a block diagram showing another embodiment of the apparatus of the present invention. This control device has the same components as the target value filter means shown in FIG. In this example, a controlled object model means 14' is provided between the target value SV input terminal and the subtraction means 13.

Therefore, although it is expressed as a transfer function between 5V and 8Vo at this time, in reality it may have only a first-order lag, or more complicatedly, it may have dead time with ten higher-order leads/lags. TM is the time constant of the controlled object model, and LM is the dead time of the controlled object model. Therefore, the unit transfer function g of the controlled model
M(s) includes only first-order delay, dead time +1
By setting the next delayed one and... (12), the transfer function between Sv-+Mv is obtained. Therefore, as is clear from equation (13),
The proportional gain KP is determined by the two-degree-of-freedom coefficient α of the proportional gain.
On the other hand, the integral time T1 is made into two degrees of freedom by the unit transfer function gM(s) of the controlled object model, which is compatible with the basics of feedback control. Furthermore, by providing within the target value filter means 10 a first-order lag element 15 having an integration time as a time constant and a controlled object model means 14' having a transfer function including a unit transfer function gM(s) of the controlled object model. , the transfer function of the target value filter means 10 can be matched to the properties of the controlled object 2.

Furthermore, FIG. 6 is a configuration diagram showing another embodiment of the present invention. The target value filter means 10 of this control device serially connects coefficient means 11.16 to the input terminal of the target value SV, and multiplies the target value SV by the two-degree-of-freedom coefficient α of the proportional gain in the coefficient means 11. Further, the coefficient means 16 multiplies the integration time by a two-degree-of-freedom coefficient β, and the obtained multiplication output is subtracted from the target value Sv by the subtraction means 17, and is supplied to the controlled object model means 14. Further, in the subtraction means 18, the multiplication output from the coefficient means 16 is calculated from the coefficient means 11. The multiplication output of .

That is, in this device, the temporal components in the target value filter means 11 are the same as those in FIG. and these coefficient means 11.1
By changing the settings of α and β in 6, one degree of freedom PI
D control - The embodiment shown in FIG. 1 - The embodiment shown in FIG. 5 can be used as a universal combination system that can be continuously varied.

Incidentally, in this example, by setting the transfer function between SV - "SVo, it becomes the transfer function between Sv → MV. In this equation (15), β - 1 is expressed as (10)
When β=0, it becomes equal to equation (13), and becomes the universal combination method described above.

Note that the present invention is not limited to the above-mentioned embodiments, and can be implemented with various modifications without changing the gist thereof.

[Brief explanation of the drawing]

1 to 4 are shown to explain an embodiment of a two-degree-of-freedom control device according to the present invention, and FIG. 1 is a configuration diagram of one embodiment, and FIG. 2 is a target value filter means. Schematic configuration diagram of a general two-degree-of-freedom PI control device with added
The figure is an equivalent conversion diagram of the lead/lag element which is one component of the target value filter means in a conventional device. Figure 4 shows the target value when the target value filter means does not have the unit transfer function of the controlled object model and when it does have it. Response characteristic diagram for value changes, 5th
6 and 6 are respectively configuration diagrams showing other embodiments of the device of the present invention, and FIG.
... Deviation calculation means, 4 ... PI adjustment means, 5 ... 1 stage of addition, 10 ... Target value filter means, 11.16.
... Coefficient means, 12... Addition means, 13... Subtraction means, 14.14'... Controlled object model means, 15...
・First-order delay means, 17.18... Subtraction means, 19...
- Addition means. This is a diagram explaining that the applicant's agent, patent attorney Takehiko Suzue, cannot be identified.

Claims (1)

[Claims] PID (P: proportional, I: integral, D: differential) adjustment based on the deviation between the control target value obtained through the target value filter means and the controlled amount from the controlled object by introducing the target value. In a two-degree-of-freedom control device that performs a PI adjustment calculation among the calculations by at least a PI adjustment means and applies the obtained operation output to the controlled object so that the control amount becomes a target value, the target value is The filter means includes: a first-order lag element whose time constant is the integration time of the PI adjustment means; and a controlled object model means having a transfer function that includes a unit transfer function of the controlled object model among the controlled object transfer functions. A two-degree-of-freedom control device characterized by being configured by combining the following.