JPH02165302A - Control device - Google Patents

Abstract

PURPOSE:To obtain an effective control result when it is necessary to change a controlled variable with the lapse of time by changing the degree of reversion to respective stages based upon respective control rules with the lapse of time by means of fuzzy inference in accordance with a prescribed pattern and controlling the controlled object by a synthetically found output value. CONSTITUTION:Rules for finding out respective stages including respective outputs are determined by a control rule part 2 in accordance with respective stages including respective inputs and the degree of reversion to respective stages of respective outputs found out based upon respective rules is multiplied by a prescribed ratio to be changed with the lapse of time correspondingly to respective applied rules through a contribution ratio part 3 and an individual operation part 4. A synthetic and final output value relating to respective outputs is found out based upon the operated result by a synthetic operation part 5 and the controlled object 9 is controlled by respective output values. Thus, the effective control result of the controlled object whose controlled variable is to be changed with the lapse of time can be obtained.

Description

[Detailed description of the invention] [Industrial application field]

This invention uses fuzzy inference to change the degree of belonging to each stage based on each control rule according to a predetermined pattern over time, and controls the output value based on the comprehensively determined output value. The present invention relates to a control device suitable for processes such as water purification plant chemical injection equipment.

[Conventional technology]

In general, conventional equipment using fuzzy inference
It is possible to perform good control according to the "vague" but ultimately accurate senses that humans have.

[Problem to be solved by the invention]

In conventional devices, predetermined control rules are constant over time. Therefore, it is inappropriate when the controlled amount needs to change over time, such as in equipment related to chemical reactions or heat treatment.
There is a problem that good control results cannot be obtained. The object of this invention is to provide a control device using fuzzy inference that solves the problems of the conventional technology and that can obtain good control results for a controlled object that requires changing the control amount over time. Our goal is to provide the following.

[Means to solve the problem]

In order to solve this problem, a control device according to the present invention is a device that controls a controlled object by one or more outputs corresponding to one or more inputs from the controlled object, each of the inputs,
When each output is divided into one or more stages depending on the range of each value, a relationship defining means for defining the relationship between each of the inputs and outputs and the degree of belonging of each to the each stage; control rule means for determining rules for determining the stages to which each of the outputs belongs in accordance with the stages to which each of the inputs belongs; a ratio calculation means for multiplying the degree of belonging by a predetermined ratio that corresponds to each of the applied rules and changes over time; and an output for calculating a comprehensive output value for each of the outputs based on the calculation result from the ratio calculation means. and a calculation means, and the control object is controlled by each of the comprehensive output values from the output calculation means.

[For use]

The relationship defining means defines the relationship between each individual and the degree of belonging of each individual to each stage for each input and output. The control rule means determines a rule for determining each stage to which each output belongs in accordance with each stage to which each input belongs. The ratio calculation means multiplies the degree of belonging of each output to each stage determined based on each rule by a predetermined ratio that changes over time in accordance with each applied rule. Note that this predetermined ratio physically has a meaning such as the degree of importance of how control is applied or the degree of contribution. The output calculation means calculates a comprehensive and most comprehensive output value for each output based on the calculation result from the ratio calculation means, and the controlled object is controlled based on each output value.

【Example】

Embodiments of a control device according to the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of this embodiment. In FIG. 1, 1 is a membership function section, 2 is a control rule section, 3 is a contribution rate section, 4 is an individual calculation section, and 5 is a general calculation section, and the embodiment is configured as described above. 9 is a controlled object. In addition, the relationship defining means in the invention is provided in the membership function section 1, and the control rule means is provided in the control rule section 2.
Similarly, the ratio calculation means includes the contribution rate section 3 and the individual calculation section 4,
Similarly, the output calculation means correspond to the general calculation section 5, respectively. The membership function unit 1 stores membership functions related to each input and output, as will be described in detail later. That is, each input and output is divided into one or more stages depending on its value range, and each input. For each output, each individual and the degree of belonging of each individual to each stage (
The relationship between membership and membership value is defined. The control rule section 2 stores rules for determining each stage to which each output belongs according to each stage to which each input belongs. This rule is basically an empirical rule, and it is expressed in an IF/THEN format as to which stage each input should take and which stage each output should take. Moreover, this rule is modified based on operational experience and improved to be more suitable for practice. The contribution rate unit 3 changes the degree of attribution of each output to each stage determined based on each rule in the control rule unit 2 according to a predetermined ratio pattern over time corresponding to each applied rule. Define this temporal ratio pattern when you want to do so. This predetermined ratio has the meaning of the degree of importance of how control is applied, or the degree of contribution,
Hereinafter, this will be referred to as the contribution rate. The individual calculation section 4 includes a membership function section 1. Control rule section 2. Based on each input from the contribution rate section 3 and each input from the controlled object 9, the degree of belonging of each output to each stage is determined over time. The comprehensive calculation section 5 calculates a comprehensive output value related to each output based on the calculation results from the individual calculation section 4, and uses this as the final control value. Now, this embodiment is applied to the control of a cement kiln. The inputs are the inlet temperature Ti and the outlet temperature To, and the outputs are the fuel flow rate F and the air flow rate A. Now, each membership function will be explained with reference to FIG. The same figure (a), (b), (C),
(d) shows membership functions for Ti, To, F, and A, respectively. The horizontal axis represents each input and output value, and the vertical axis represents the corresponding membership function value. Each input and each output is divided into one or more stages depending on its value range. For example, in the case of Ti, from the lowest value to low (S), slightly low (SM), normal (M), There are 5 levels: slightly high (M L), high (L), To, F,
In the case of A, each level is divided into 7 levels: very low; greatly reduced (SS); low; decreased (S); slightly low; slightly decreased (SM); normal: maintained (M); slightly high; Increase (M
L), high; increased (L), very high; greatly increased (LL,
). Note that among the two expressions indicating the degree of each stage separated by ;, the former corresponds to To, and the latter corresponds to F and A, respectively. Further, each membership function is represented by a part of a triangular chevron or trapezoid. Note that the units of Ti and To are °
The units of C1F and A are rrr/h and km, respectively.
' / h. The control rules will be explained with reference to FIG. In FIG. 3, control rule (R) numbers 1, 2, 3, . ..., IF section in each column in the upper horizontal direction,
THEN part, take the contribution rate (K) respectively. Also, if
The part and THEN part are T i , T o and F, respectively.
It is classified into A. The reading of this control rule is as follows. For example, in rule 1, Ti and To are M and S, respectively.
If M, F and A represent ML and L, respectively. Third
As will be described later with reference to FIG. 4, the value of the contribution rate shown in the figure indicates the value at time TI after the start of control, and naturally changes with the passage of time. The contribution rate will be explained with reference to FIG. 4. Fourth
In the figure, when time T is plotted on the horizontal axis and contribution rate K is plotted on the vertical axis, control rules 1, 2, 3, . ... Separately, temporal changes in the contribution rate are shown. For example, at time T1, rule R1 = 1, rule R2 = 0.9, rule R3
Debani = 0.5, in rule 4 = 0.6, and in rule R5 = 0.7. Next, regarding the operation of this example, Ti = 277'C
, To =263'C, an example will be explained in which the values of F and A are determined. In Figure 2(a), Ti =277°
The membership function value (degree of belonging to each stage) for C is 0.7 for stage SM and 0.3 for stage M. Similarly, in FIG. 2(b), T
The membership function values for o = 263°C are 0.1 for stage SS and 0.1 for stage S.
Becomes 9. That is, Ti:SM...Membership function value 20.7M ・
・・ Same as above: 0.3To:SS-Same as above
:0. IS... Same as above
:0.9 Applying this condition to the control rule in Figure 3, we get
The following results are obtained. F: ML (Rule 1)... Membership function value = OL
(Rule 2)... Same as above: o, 7SS (
Rule 3)... Same as above 20L (Rule 4)
... Same as above: o, IML (Rule 5)...
・ Same as above: 0.3 ALL (Rule 1)・
... Same as above 20L (Rule 2)... Same as above: o, 7S (Rule 3)... Same as above
:OM (Rule 4)... Same as above.
:o, IML (Rule 5)... Same as above =
0.3 For example, the meaning of F';L (Rule 2) is that a step is obtained as a result of applying Rule 2 to F, and each membership function value in F and A is determined by each rule. The smaller membership function value of Ti and To corresponding to is selected. Also, as a result of applying Rule 1.3, since the membership function value of either Ti or To is 0, the membership function value of F or A is also O.
become. The above results are shown in FIG. In other words, Fig. 5 (
The membership function for the fuel flow rate F is shown in a), and the membership function for the air flow rate A is shown in steps in (b). Note that since the membership function value at time TI is here, the contribution rate K at time T1 shown in FIG. 3 is the membership function for F and A obtained by applying the control rule in FIG. Multiplyed by value. For example, regarding the rule R2) in FIG. 5(a), the contribution rate in rule 2 in FIG. 3 is set to 0.9 by =0.9.
7 xo,9 =0.63 is obtained, which defines the height of the membership function triangle. Furthermore, this:
The two ends of the base of the square are the same as those of the step shown in FIG. 2(C). Similarly, in FIG. 5(a), L
(Rule R4) and ML (Rule R5), triangles with heights of 0.06 and 0.21 are obtained, respectively. Similarly, for A, in FIG. 5(b), L (rule R
2), M (rule R4), and ML (rule R5), triangles each having a height of 0°63.0.06.0.21 are obtained. As mentioned above, for each of F and A,
The calculation for individually determining the three graded triangles is performed by the individual calculation unit 5 shown in FIG. Next, the overall output value for each of the fuel flow rate F and the air flow rate A is obtained by calculating the abscissa of the center of gravity of each of the three triangular mountain series shown in Figure 5 (a) and (b). obtained by As a result, the fuel flow rate F is 116rd/h, and the air flow rate A is 1.4 km'/h. Now, the calculation for determining the horizontal coordinate of the center of gravity position of the mountain range is performed by the general calculation section 6 in FIG. 1. Now, the fuel flow rate F obtained as above. Each air flow rate A becomes the final control signal value for the controlled object 9 at time T1. Thereafter, similar calculations are performed at each time, and control signal values that change over time for the fuel flow rate F and the air flow rate A are output, and control is performed based on these.

【Effect of the invention】

This inventive device uses fuzzy inference to change the degree of belonging to each stage based on each control rule according to a predetermined pattern over time, and controls the controlled object with the comprehensively determined output value. In general, control is performed according to the "vague" but ultimately accurate sense that humans have, and the controlled amount changes over time, especially when the controlled object is equipment related to chemical reactions or heat treatment. It is suitable for situations where it is necessary.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment according to the present invention, FIG. 2 shows membership functions related to input and output,
Figure (a) relates to the inlet temperature, Figure (b) relates to the outlet temperature, Figure (C) relates to the fuel flow rate, Figure (d) relates to the air flow rate. The figure shows the correspondence diagram regarding the control rules, Figure 4 shows the temporal change of the contribution rate, Figure 5 shows the membership function on the applied control rule side at a certain time of output, and (a) of the same figure shows the relationship with respect to the fuel flow rate. (Illustrator) relates to air flow rate. Code explanation Ti: Inlet temperature, To: Outlet temperature, F: Fuel flow rate,
A: Air flow rate, K: Contribution rate, R: Control rule, 1: Membership function part, 2: Control rule part, 3: Contribution rate part, 4:
Individual calculation unit, 5: General calculation unit, 9: Controlled object. Figure 3 Figure 4 (Time) Taku (%) (K too)

Claims (1)

[Claims] 1) In a device that controls the controlled object by one or more outputs corresponding to one or more inputs from the controlled object, when each of the inputs and outputs is divided into one or more stages depending on the range of each value, each of the above relationship defining means for defining the relationship between each value for each input and each output and the degree of belonging of each value to each of the stages; control rule means for determining rules for determining each of the steps; the degree of belonging of each of the outputs to each step, determined based on each of the rules, changes over time in response to each of the applied rules; Ratio calculation means for multiplying by a predetermined ratio; Output calculation means for calculating a comprehensive output value related to each of the outputs based on the calculation result from the ratio calculation means; and each of the total output values from the output calculation means. A control device characterized in that the control target is controlled by: