JP2781886B2 - Control device and control method - Google Patents

Description

The present invention relates to a control device such as a controller constituting a closed loop system.

(B) Conventional technology This type of conventional controller mainly has a configuration as shown in FIG. 10 (A) or (B). The controller shown in FIG. 2A has an FF element (feedforward element) and an FB element (feedback element), and outputs a control amount in response to an external signal consisting of a target value. Further, the controller shown in FIG. 3B is made up of an FB element including an observer, and performs gain control by estimating the state quantity of the control target with the observer. Each controller determines the operation amount at each time from information on the output of the control target and information on the external signal every moment.

(C) Problems to be Solved by the Invention However, when feedback control is performed only from the momentary information, the dead time (the time when it is determined that no output appears on the control target despite the addition of the operation amount) is large. Or when the time constant is large, it is often difficult to control. That is, there is a problem that the overshoot increases and it takes time to reach the target value.

Therefore, an object of the present invention is to provide a control device and a control method which can solve the above problem by determining the correction amount of the operation amount by determining the type of the time pattern of the operation amount. is there.

(D) Means for Solving the Problems The invention described in claim 1 is a basic operation in which a plurality of time patterns of operation amounts within a predetermined time range given to a control target are previously stored as basic operation amount time patterns. An amount time pattern holding unit, a current operation amount time pattern holding unit that holds a time pattern of the operation amount given to the control object currently being controlled as a current operation amount time pattern, and a basic operation amount time pattern holding unit. The stored basic manipulated variable time pattern and the current manipulated variable time pattern are compared with each other, and a determination unit that determines the type of the current manipulated variable time pattern is used as a control target using the determination result of the determination unit. An operation amount correction unit that corrects the operation amount to be given and outputs the operation amount to a control target;
It is characterized by comprising.

According to a second aspect of the present invention, in the control device according to the first aspect, a basic control amount in which a plurality of time patterns of the control amount within a predetermined time range outputted from the control target are previously stored as the basic control amount time pattern. A time pattern holding unit, a current control amount time pattern holding unit that holds a time pattern of a control amount output from a currently controlled control object as a current control amount time pattern, and a basic control amount time pattern holding unit. A second control unit that compares the basic control amount time pattern and the current control amount time pattern, and determines the type of the current control amount time pattern. The operation amount given to the control target is corrected using the determination results of the unit and the second determination unit, and the control target is output.

The invention described in claim 3 is characterized in that the operation amount correction unit includes a fuzzy inference unit.

(E) Function According to the first aspect of the present invention, a time pattern of an operation amount in a predetermined time range output to a control target that is currently being controlled is held, and the held pattern is used as a basic pattern. To determine the type of the time pattern of the current manipulated variable. Then, the operation amount output from the main controller or the like is modified according to the result of the determination. In this case, the operation amount is corrected, and the control parameter is not changed. Therefore, there is no need to determine the dynamic characteristics of the control target, and there is no change in the control parameters, so that the control system does not become unstable. As described above, since the determination of the operation amount is performed based on the time pattern type of the operation amount, the overshoot is reduced and the quick response is improved.

According to the invention described in claim 2, a time pattern of a control amount in a range of a predetermined time outputted from a control object is held, and the held pattern is compared with a basic pattern to compare the time pattern of the current control amount. Is determined. Then, the operation amount output to the control target is modified according to both the determination result of the time pattern type determination result of the operation amount and the determination result of the time pattern type determination result of the current control amount. Thereby, controllability can be further improved.

According to the third aspect of the invention, the operation amount is corrected using fuzzy inference. Therefore, controllability is further improved, for example, responsiveness is further improved.

(F) Embodiment FIG. 1 is a configuration diagram of an embodiment of the present invention.

In the figure, reference numeral 1 denotes a main controller, which has a configuration as shown in FIGS. 10 (A) and 10 (B), for example. Exogenous signals 1 and 2 consisting of target values are input information, and the controller output is corrected by the manipulated variable corrector 2 and output to the controlled object 3 as a manipulated variable. The operation amount is guided to the operation amount holding unit 4, where the operation amount is held for a certain time. The operation amount can be given for a certain period of time or by the user using an interface such as a keyboard. The operation amount held by the operation amount holding unit 4 is guided to the history determination unit 5, where the history of the operation amount is determined. The history is determined by, for example, fuzzy measure calculation.

Here, the history of the operation amount means the type of the time pattern of the operation amount. This type is determined by fuzzy measure calculation.

FIG. 2 is a block diagram of the history determination unit 5. The history determination unit includes a fuzzy measure calculation unit 50 and a basic pattern storage unit 51. The fuzzy measure calculation unit 50 outputs the degree of coincidence between each of the plurality of basic patterns stored in the basic pattern storage unit 51 and the pattern captured as the holding operation amount in the range of (0 to 1). The basic pattern refers to a time pattern (basic operation amount time pattern) of an operation amount in a predetermined time range given to the control target. The fuzzy measure calculation unit 50 uses the membership function as a distribution function indicating the possibility, and calculates the possibility measure of the captured pattern (the pattern of the holding operation amount) from the membership function set corresponding to each basic pattern. Calculate and output. That is, the degree of conformity of the held operation amount pattern to each basic pattern is calculated.

FIG. 3 is a diagram for explaining the operation of the fuzzy measure calculation unit 50. Fig (A) is t ₁ from the monitoring time width T (t ₀
In this case, a basic pattern in which the manipulated variable during the period up to) is continuously large. FIG. 7B shows a basic pattern in which the operation amount oscillates in the monitoring time width T. FIG. 3C shows a basic pattern in which the operation amount continuously increases in the monitoring time width T. FIG. 4D shows a fuzzy measure calculation unit 50.
1 shows an example of the holding operation amount taken in by the operation.
In this example, the result of the fuzzy measure calculation is the degree of conformity to pattern 1 0.01 the degree of suitability to pattern 2 0.8 the degree of conformity to pattern 3 0.0, and these degrees of conformity are the results of the fuzzy measure calculation in FIG. Output to the manipulated variable corrector 2.

In the above example, the degree of conformity to the basic pattern is obtained by the fuzzy measure operation unit 50 using the fuzzy measure. However, the degree of conformity to the basic pattern can be obtained by another method. As described above, the type of the time pattern of the operation amount is determined by calculating the degree of conformity to each basic pattern.

 The following method shows the procedure of this other method.

step1: The average operation amount (u) within the monitoring time width T is obtained.

step2: Integrate the absolute value of the deviation from the average manipulated variable within the monitoring time width T (u _e ) step3: Divide the monitoring time width T into n, take the average for each section, and then perform linear approximation (least square method ) To approximate y = ax + b (see FIG. 4).

step4: Determining the degree of suitability for large, medium, and small manipulated variables. This determination is made from the membership function shown in FIG.

step5: The degree of conformity for each of the oscillating and smooth operation amounts is determined. This determination is made from the membership function shown in FIG.

step6: Determine the degree of adaptation to whether the manipulated variable is increasing or decreasing. Similarly a membership function (not shown)
This determination is made from.

In the above steps 4 to 6, the characteristic amount of the input held operation amount is extracted and output to the operation amount corrector.

1 has a well-known fuzzy inference unit including a fuzzy calculation unit and a defuzzification unit, and further includes a timer, a phase delay element, and a limiter for executing a definite value output from the fuzzy inference unit. It has.

Here, a general fuzzy inference unit will be described as follows.

The fuzzy inference unit is composed of a fuzzy operation unit and a defuzzification unit that performs a deterministic value operation, as is well known. A membership value (affiliation value) is calculated, and an inference value calculated based on the result is output to the defuzzification unit. The fuzzy rules are expressed in the form if (x ₁ = A and x ₂ = B ...) then (y = Z), where (x ₁ = A and x ₂ = B ...) is the antecedent, (Y = Z) is called the consequent.

FIG. 9 is a view for explaining one known method of outputting an inference result in accordance with the above fuzzy rules.

Fig (A), (B) 2 single variable antecedent is (x _1, x ₂₎
(C) shows the membership function corresponding to the consequent part. Here, two membership functions in the antecedent part are shown, but as the types of variables in the antecedent part increase, the membership functions increase accordingly. In each figure, the horizontal axis represents the value of the variable, and the vertical axis represents the membership value (degree of affiliation).

Now, the value of the variable x ₁ of the first item of antecedent is assumed to be x ₁ ', the appertaining at that time is 0.5 (see FIG. (A)). If the value of the variable x ₂ of the second item in the antecedent part is x ₂ ′, the affiliation at that time is 0.3 (see FIG. 13B). In such a case, the fuzzy calculation unit takes the smallest value among the respective degrees of affiliation. That is, in the above example, the degree of belonging 0.3 is selected. Next, the member shop function corresponding to Z is truncated at the above-mentioned affiliation degree 0.3, and the center of gravity y 'of the lower trapezoid S is obtained. Then, y 'is output as an inference result.

The above inference is performed for one rule, but a plurality of rules are generally set. In this case, the inference result shown in FIG. 9 (C) is output for each rule. In this case, the trapezoids output for each rule are ORed, and the center of gravity y ″ of the ORed portion (the hatched area in FIG. 9D) is obtained.
Is output as the final value of the inference.

In the above inference method, the rule of the logical product operation of the degree of belonging to the antecedent part (the operation of selecting the smaller degree of belonging) and the rule of the logical sum of the trapezoid part to the consequent part are called mini-max rules.

In the present embodiment, the operation amount of the output of the main controller 1 is corrected by executing the above-described inference method in the operation amount corrector 2.

The fuzzy rules set in the manipulated variable modifier 2 are, for example, as follows.

FIG. 6A shows a membership function of the timer set value, and FIG. 6B shows a membership function of the correction amount. FIG. 3C shows a phase lag membership function. These membership functions are membership functions corresponding to the consequent part of the above rule. Although not shown, a membership function corresponding to the antecedent part is of course set. A fuzzy inference finalized value is output using these membership functions, and a correction amount is added to the operation amount output from the main controller 1 by passing through a timer, a phase delay element, and a limiter.

FIG. 7 shows another embodiment of the present invention. In this example, the history is also determined for the control amount which is the output of the control target in the same manner as the operation amount. That is, the control amount is held in the control amount holding unit 6 for a certain period of time, and the history is determined by the history determination unit 7 based on the held control amount. Then, the result is output to the manipulated variable corrector 2. When this method is used, the fuzzy rules in the manipulated variable modifier 2 are set as follows, for example, using the basic pattern shown in FIG.

Basic patterns S1 shown in FIGS. 8 (A), (B) and (C)
In S3, the basic pattern S1 indicates a pattern with a small control amount, and the basic pattern S2 indicates a pattern with an increased control amount. The basic pattern S3 indicates a pattern in which the control amount is oscillating.

When the history is determined for the control amount as described above, the overshoot is further reduced, and the control performance is further improved, for example, the responsiveness is improved.

Although the fuzzy inference is performed in the manipulated variable corrector 2, it is not always necessary. However, if the operation amount is corrected by fuzzy inference, more accurate correction becomes possible and the responsiveness is improved.

(G) Effects of the Invention According to the inventions set forth in claims 1 and 4, determination is made by monitoring the time pattern of the operation amount in a predetermined time range and comparing the time pattern type of the operation amount during that period with the basic pattern. This makes it possible to correct the operation amount output from the main controller or the like, so that good control characteristics can be obtained particularly in a system with a large dead time of control or a system with a large time constant. This has the advantage that the overshoot is small and the responsiveness can be very high. Further, since the operation amount is corrected without changing the control parameters, the control system does not become unstable even if the operation amount changes due to non-linearity.

According to the invention described in claims 2 and 5, in addition to the time pattern of the operation amount in the predetermined time range, the time pattern of the control amount in the predetermined time range is monitored, and these time pattern types are respectively defined as the basic pattern and the basic pattern. Since the correction of the operation amount is performed by making a determination by comparing, the responsiveness is further increased, and the correction is more accurate.

According to the third and sixth aspects of the present invention, the correction of the operation amount is performed by fuzzy inference, so that the responsiveness is further improved and the nonlinear control can be properly dealt with.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a controller according to an embodiment of the present invention, FIG. 2 is a configuration diagram of a history determination unit, and FIGS.
(D) is a diagram showing an example of a basic pattern and a holding operation amount pattern. FIGS. 4, 5 (A) and (B) are diagrams for explaining a method of obtaining a degree of conformity to the basic pattern, and FIG. 6 (A). ) ~
(C) is a diagram showing a membership function set in the manipulated variable corrector, FIG. 7 is a configuration diagram of another embodiment of the present invention, and FIGS. 8 (A) to (C) are history of control variables. FIG. 9 (A) is a diagram showing a basic pattern used to determine
10 (D) are diagrams for explaining a known fuzzy inference method, and FIGS. 10 (A) and 10 (B) are configuration diagrams of a conventional controller. 2 ... operation amount corrector, 3 ... control target, 4 ... operation amount holding unit, 5 ... history determination unit.

────────────────────────────────────────────────── ─── Continued on the front page (58) Field surveyed (Int. Cl. ⁶ , DB name) G05B 13/00-13/02 G05B 11/36 505 G05B 11/36 507 G06F 9/44 554 JICST file (JOIS )

Claims (6)

(57) [Claims] 1. A basic operation amount time pattern holding unit that previously holds a plurality of time patterns of operation amounts within a predetermined time range given to a control object as basic operation amount time patterns, and a control object that is actually controlled. A current operation amount time pattern holding unit that holds a time pattern of the operation amount given to the current operation amount time pattern; a basic operation amount time pattern stored in the basic operation amount time pattern holding unit; A determination unit that compares the current operation amount time pattern with the amount of time pattern, and corrects the operation amount to be applied to the control object using the determination result of the determination unit and outputs the operation amount to the control object. A control device comprising: a quantity correction unit. 2. The control unit according to claim 1, further comprising: a basic control amount time pattern holding unit for holding a plurality of time patterns of control amounts in a predetermined time range output from a control object as basic control amount time patterns in advance. A current control amount time pattern holding unit that holds a time pattern of the control amount output from the control object currently being controlled as a current control amount time pattern; and a basic control amount stored in the basic control amount time pattern holding unit. A second determination unit that compares the control amount time pattern with the current control amount time pattern to determine the type of the current control amount time pattern, wherein the operation amount correction unit includes the control unit and the second control unit; A controller that corrects an operation amount given to a control target using the determination result of the determination unit and outputs the correction to the control target. 3. The control device according to claim 1, wherein the operation amount correction unit includes a fuzzy inference unit. 4. A method according to claim 1, wherein a plurality of time patterns of operation amounts within a predetermined time range given to the control object are previously stored as basic operation amount time patterns, and a plurality of time patterns of the operation amount given to the control object currently being controlled are stored. Holding a time pattern as a current manipulated variable time pattern; comparing the plurality of basic manipulated variable time patterns with the current manipulated variable time pattern to determine the type of the current manipulated variable time pattern; A control method, comprising modifying an operation amount given to a control target using a determination result and outputting the correction to the control target. 5. The control method according to claim 1, wherein a plurality of time patterns of the control amount output from the control object within a predetermined time range are previously held as basic control amount time patterns, and the actual control is performed. Holding the time pattern of the control amount output from the control target as the current control amount time pattern, comparing the plurality of basic control amount time patterns with the current control amount time pattern, Determining the type, modifying the operation amount to be given to the control target using the determination result of the type of the current manipulated variable time pattern and the determination result of the type of the current control amount time pattern, and outputting the correction to the control target. How to control. 6. The control method according to claim 4, wherein the operation amount correction is performed by fuzzy inference.