JPS62241003A - Auto-tuning controller - Google Patents

Abstract

PURPOSE:To attain an automatic control of various controlled systems without performing the identification of the controlled systems, by storing previously the human experience or perception as an inference rule and the inferring optimum parameters from the feature value of waveforms like the controlled variable, etc. based on said inference rule. CONSTITUTION:A feature value extractor 10 receives the deviation e(k) or a target signal r(k) and a controlled variable y(k) and outputs a feature value Si showing the characteristics of a controls system. A velocity type fuzzy inference unit 11 receives the value Si and infers the controlled variable against the present value in order to optimize the control parameters, i.e., the gain KC, the integration action time TI and the derivative action time TD according to an inference rule Rj obtained from the human experience or perception. Then the unit 11 outputs the controlled variable DELTAKC, DELTATI and DELTATD respectively and an integrator 12 receives these values for integration to deliver actual parameter values. These parameters are given to a PID controller 4 and used again to calculate a manipulated variable u(k) from the value e(k). Thus this controller can perform an automatic control at an optional time point while carrying out the control owing to the velocity type inference unit 11.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention is directed to an auto-tuning controller used for example in process control, which is equipped with a function to automatically adjust control parameters according to the characteristics of a controlled object. The present invention relates to controllers, and in particular to automatic adjustment of their control parameters.

[Conventional technology]

Conventionally, as an auto-tuning controller, one shown in FIG. 5, for example, has been proposed.
, B. Collivio, P., M. and I. Mumpkins “Industrial Applications of a Self-Tuning Feedback Control Algorithm”,
ISA TRANSAC', TANS vol, 20.1
981. Pages 3 to 10 (A, B, Corripio
, P, M, Tompkins+ “[ndu
trial Application of a S
elf-Tuning Feedback Control
l Algorithm”, ISA Transa
tions, vol.

20, No, 2. 1981. pp, 3-
10). In the figure, 1 is a target signal generator, 2 is an auto-tuning controller, 3 is a controlled object, 4 is a PID controller, 5 is a mathematical model calculator, 6 is an identifier, and 7 is an adjustment calculator.

Conventional auto-tuning configured as above
The operation of the controller will be explained.

The auto-tuning controller 2 inputs the target value signal r(kl) given from the target value signal generator l and the control amount 7(k) which is the output of the controlled object 3, and becomes the input of the controlled object 3. The manipulated variable u (kl is output. Here, the value in parentheses represents the discretized time for each sampling interval T.

At this time, the following operations are performed inside the auto-tuning controller 2.

First, the deviation e (k) is calculated from the target value signal r(k) and the control amount yTkl.

e (kl −r (k) −y (k)
...(1) The PID controller 4 calculates the above deviation e.
(k) is input, the manipulated variable u(k) is calculated and output according to the set control parameters. In the PID controller 4, the control parameters are the gain Kc, the integration time 'r+, and the m minute time T, and according to these parameters, the manipulated variable u(k) is calculated as follows.

l゛The above-mentioned manipulated variable u(k) becomes an input to the controlled object 3 and also becomes an input to the mathematical model calculator 5 and the identifier 6.

The mathematical model calculator 5 performs the above operation 'Wku (k)
When input, the output v (kl is calculated by a mathematical model such as the following equation.

v(kl 5ml, v(k-1) +az V
(k-2)+b+u(k-m-1)+bt u(
km-m-2) = (3) Here, m is an integer greater than or equal to 0, and means the dead time of the controlled object 3.

The identifier 6 includes the controlled object 3 and the mathematical model calculator 5.
The (3rd
) Find the coefficients a, + aZ + bI + bt of the equation.

For this purpose, the identifier 6 uses the manipulated variable u(kl) and the control m1y(kl, the output V(t) of the mathematical model).
Enter.

In order to explain the operation of the identifier 6, the following vectors X(k), z(k), and φ(kl) are defined here.

x” (k-1) = (y(k-1),
y(k-2), u(k-m-1).

u (ks-2) ) ...(4)
z' (k-1) = (v(k-1), v(k
-2), u(k-+++-1).

u (k-+1-2) ) ...(5
)φ伽) ” (at * ! + 1)
l + bt] ...(6) Here,
The subscript T on the right shoulder of a vector means the transposition of the vector.

The identifier 6 executes the following algorithm.

G(k)=(1+z'(k)P(k)x(k))
−' z” (k)P(k) ...(?)φ(k
l1) =φ(k) + (y(kl1)−φOc)
x (k) ) G (k) = (81P (kl1
) = P(k) −P(k) x(k)G(k)
(9) With this algorithm, the vector φ(g)), that is, the coefficients ai + ax l b, 1b of the mathematical model equation (3), is sequentially obtained.

The above vector φ(k) obtained in the above manner is
The output from the identifier 6 is sent to the mathematical model calculator 5 for use in modifying the mathematical model, and is also sent to the adjustment calculator 7 to obtain control parameters, that is, gain KCI, integration time TI.

Differential time T. It is used to ask for. In order to obtain these control parameters, the adjustment calculator 7 executes the following calculations.

Kc −s (a + + 2 am) q/b+
-alHere, Q appearing in equation 00 and equation (2) is Q-1-e -17m
...Defined by α force. where B is an adjustment parameter and is the desired time constant in the closed group.

The gain ICc obtained as above is 4! lJ minute time T+, m minute time T, are sent to the PII) controller 4, and the deviation e(g) station 1 and the like are again determined by the operation f! u (Used to calculate kl.

[Problem that the invention seeks to solve]

Since the conventional auto-tuning controller is configured as described above, it was necessary to identify the object to be controlled. However, this identification has the following problems.

(11 calculations are complicated.

(2) The amount of calculation is large.

(3) It takes time to converge.

(4) Nonlinearity of the controlled object cannot be handled.

(5) Since the type of the mathematical model is limited to, for example, equation (3), it is inappropriate for identifying other types of controlled objects.

(6) In order to obtain the three control parameters Kp, Tl, and To, the ai l aRr bI +
the law of nature! There is no use in identifying the four coefficients of .

The above-mentioned identification problem has become a problem of its own in conventional auto-tuning controllers.

This invention was made to solve the above-mentioned problems. Instead of identifying the controlled object, human experience and intuition are memorized as inference rules, and these are used to calculate the control amount. Deviation from test signal.

Another purpose is to obtain an auto-tuning controller that can infer optimal control parameters for a controlled object from control variables, test signals, and the like.

[Means for solving problems]

The auto-tuning controller according to the present invention, instead of having a mathematical model calculator, an identifier, and an adjustment calculator, includes an inference rule storage device that stores inference rules for optimizing control parameters, and an inference rule storage device that stores inference rules for optimizing control parameters. According to the present invention, an adjustment section that adjusts control parameters by fuzzy inference is provided.

[Effect]

In the auto-tuning controller of the present invention, human experience and intuition are memorized as inference rules, and control parameters are determined phenomenologically from waveform features such as deviations, target value signals, and control amounts. It attempts to infer how much adjustment should be made from the current value in order to optimize it, and then performs automatic adjustment.

〔Example〕

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

FIG. 1 is a block diagram showing an auto-tuning controller according to an embodiment of the present invention. 1st I
l! In I, 1 is a target value signal generator, which generates a target value signal r(k). Reference numeral 2 denotes an auto-tuning controller, which inputs the target value signal rOt) and the control mfy(g)) which is the output of the controlled object 3, and outputs the manipulated variable u(k). As mentioned above, this control amount y(
k) is fed back to the auto-tuning controller 2.

Next, the internal configuration of the above auto-tuning controller will be explained. 4 is a controller, and in this embodiment a PID controller is used. This PID controller 4 manually calculates the deviation e(k) between the target value signal r(kl and the control amount y(k)) and calculates the set control parameters, that is, gain KC+integration time 71.
The operation t u (k) is generated according to the differential time T ゎ.

8 is an inference rule storage device, in which inference rules Rj for optimizing the control parameters of the controller 4 are stored;
j=1.2. ..., m are stored. 9 is an adjustment unit, which performs fuzzy inference according to the above inference rules,
The control parameters of the controller 4 are adjusted.

Next, the internal configuration of the adjustment section 9 will be explained. Reference numeral 10 denotes a feature extractor which inputs the deviation e (kl or the target value signal r(k), the operation 11y(k), etc.), and extracts the feature amount Si ;s −L2+ . . . representing the characteristics of the control system.・、
Output n. Reference numeral 11 denotes a speed type fuzzy inference machine, which inputs the above time @ MSi and calculates the above control parameters, that is, gain KC + integration time T1 . In order to optimize the differential time T, we infer how much to adjust from the current value, and calculate the amount of adjustment,
ΔKc and ΔTI, respectively.

Outputs ΔT. 12 is an integrator, and the above ΔKC+
ΔTI and ΔT are input, integrated, and output as actual parameter values. The control parameters output from the integrator 12 are given to the controller 4,
Again, it is used to calculate the operation flu (k) from the deviation e(kl).

FIG. 2 is a flowchart showing the operation of the auto-tuning controller.

Auto-tuning using speed-type fuzzy inference machine
The controller can perform automatic adjustment at any time while performing control. Here, an example is shown in which automatic adjustment is always performed while performing control.

First, in step 13, the value of K is set to 0.

Next, in step 14, control parameters are initialized.

The value is set to a value that is sufficiently safe in terms of stability rather than continuity or precision.

Unless a decision is made in step 15 to stop the control system, the auto-tuning controller of this embodiment repeats the operations in steps 16 to 22.

In step 17, the PID controller 4 calculates equation (2) using the current control parameters and controls the controlled object 3.

Accordingly, in step 18, the deviation e (k
-N), manipulated variable u (K-N), controlled variable y (K-N
) etc. and create the corresponding new data e(kl.

Record u(kl, )l(k).

Step 19 is the data e(·) as described above.

It is determined whether U(.) and y(.) have been collected for N samples. The process returns to step 15 until the data is accumulated.

In step 20, the feature extractor 10 extracts data e(K-N+1), . . . collected for N samples.

a(k), r (K-N+ 1), ”-, r
From (k),, Y (K-N-1),..., y(k), etc., special f! Calculate the amount S directly; i-1, 2, . . . 1 mouth. The above-mentioned feature amount is, for example, as follows.

S! = (average change rate of error) In step 21, the speed type fuzzy inference device 11 inputs the feature amount S, and infers the inference rules Rj; J −L2+..., l stored in the inference rule storage device 8. According to
Outputs KC to the gain adjustment amount + integral time adjustment amount ΔT11 and T0 to the differential time adjustment amount.

In step 22, the integrator 12 calculates the adjustment amount ΔKc,
Input ΔTl and ΔT0, integrate them, and obtain the actual control parameters as Kc, 'r'+, and T, respectively.

It is output as and given to the PID controller 4.

Thereafter, by repeating the operations in step 15 and above again, the auto-tuning controller controls the controlled object 3 while automatically adjusting the control parameters.

Now, the inference rules RJ stored in the inference rule storage 8 and the operation of the speed-type fuzzy inference device 11 will be explained below.

First, the above-mentioned inference rule R is a rule based on the empirical rules and intuition that humans use when adjusting control parameters, and is, for example, as follows.

R+:jIf S is small and SZ is also small, Kc
Keep it that way. ” R2: “If Sl is small and SZ is large, Kc
Make it smaller. ” In this way, the inference rule R4 has the form “If ..., then ...”. Of these, the part that says ``if...'' is called the antecedent proposition, and the part that says 'do...' is called the consequent proposition.

If the form of this consequent proposition specifies the amount of change in value, such as ``make ~ smaller'' or ``make ~ a little larger,'' as in the above rules R5゜R8, then this This theory is especially called speed-type fuzzy inference. On the other hand, when the form of the consequent proposition specifies the value itself, such as "take it" or "set it to...", the fuzzy inference is particularly called positional fuzzy inference.

The present invention includes a speed-type fuzzy inference device 11 that executes speed-type fuzzy inference. Its operation will be explained. In fuzzy reasoning, first, the degree to which the current state satisfies the conditions of the antecedent proposition is evaluated using a membership function and expressed as a number between 0 and 1.

FIG. 3 shows an example of evaluation using membership functions. Here, the proposition that "the average error Sl is large" is given. Assume that the average error actually calculated by the feature extractor 10 is 3,wS,''.In this case, as shown in FIG. 3, the degree to which the antecedent proposition holds true is 0.75. be evaluated.

FIG. 4 shows the inference mechanism performed by the positional fuzzy inference unit 11. Here, a case is shown in which St and St are selected as special quantities and R1 and R2 are used as inference rules. For simplicity, only gain adjustment will be described here, but the same applies to integral time and differential time.

First, as described above, the fuzzy inference rules R1, R
Evaluate the degree to which the antecedent proposition of z holds true. In the example of FIG. 4, the rule R, ``If Sl is small,'' is satisfied by 0.5, and ``If Sz is also small,'' is satisfied by 0.2. Take the lowest of these, rule R1
It is assumed that only 0.2 of the antecedent propositions are true as a whole.

The consequent proposition is also expressed by a membership function. This membership function is weighted according to the degree to which the antecedent proposition holds true. Rule R+ is weighted 0.2 times.

Finally, the membership functions of the weighted consequent propositions of each rule are superimposed. Then, the center of gravity is calculated, and the center of gravity is used as the optimum gain adjustment amount ΔKC.

Similarly, the optimal integration time adjustment amount ΔTl lThe optimal differentiation time adjustment amount Δ]゛. is also inferred.

In the manner described above, the speed type f-shi inference unit 11 infers the optimal amount of control parameter adjustment. These values are passed through the integrator 12 and given to the controller 4 as actual control parameter values.

In addition, in the above embodiment, a PID controller is used as the controller, and its gain and integration time are determined.

Although the auto-tuning controller that automatically adjusts the differential time has been described, the present invention can also be applied to other types of auto-tuning controllers. , it is also effective to apply it to an auto-tuning controller that automatically adjusts the width of the dead zone, and an auto-tuning that uses an optimal controller based on modern control theory as the controller and automatically adjusts the parameters of its evaluation function.・It is also effective when applied to a controller.

Furthermore, in the above embodiment, the automatic adjustment of the control parameters is as follows:
This automatic adjustment has always been performed while performing control, but the present invention is also effective in performing the following automatic adjustment. The automatic adjustment may be started only when a change occurs.Also, the control operation and control characteristics are usually only monitored, and the automatic adjustment is activated only when a disturbance in the control characteristics is detected. You can do it like this.

〔Effect of the invention〕

As described above, according to the present invention, human experience and intuition are memorized as inference rules, and control parameters are fuzzy inferred from waveform features such as deviations, target value signals, and control amounts in accordance with the inference rules. It's complicated,
There is no need to perform identification that would limit the type of control target, and automatic adjustment based on human experience and intuition can be performed through simple membership function calculations, resulting in a light calculation load.

Automatic adjustment in a short time 1 This has the effect of enabling automatic adjustment of a wide variety of control objects.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an auto-tuning controller according to an embodiment of the present invention, FIG. 2 is a flowchart showing its operation, FIG. 3 is a diagram showing an example of evaluation using a membership function, and FIG. The figure shows the mechanism of fuzzy inference, and Figure 5 shows the conventional auto-tuning system.
It is a figure showing an example of a controller. In the figure, 1 is a target value signal generator, 2 is an auto-tuning controller, 3 is a controlled object, and 4 is a controller; here, as an example, a PID controller, 8 is an inference rule storage, and 9 is a feature amount. extractor, 11
is a speed-type fuzzy inference device, and 13 is an integrator. Note that the same reference numerals in the figures indicate the same or equivalent parts.

Claims (3)

[Claims] (1) A controller that controls a controlled object, an inference rule storage device that stores inference rules for inferring optimal control parameters for the controller, and adjusts the control parameters of the controller by fuzzy inference according to the inference rules. An auto tuning controller characterized by comprising an adjustment section. (2) The adjustment section inputs the target value signal given from the target value signal generator, the controlled variable that is the output of the controlled object, or the deviation between the target value signal and the controlled variable, and adjusts the state of the control system. and a fuzzy inference device that inputs the feature amount and adjusts the control parameter according to the inference rule. Auto tuning controller as described. (3) The fuzzy inference device is a speed-type fuzzy inference device that infers the adjustment amount of the control parameter, and further includes an integrator that integrates the output of the speed-type fuzzy inference device and inputs the output to the controller. An auto-tuning controller according to claim 2, characterized in that: