JPH02230402A - Adaptive controller - Google Patents

Abstract

PURPOSE:To obtain an adaptive controller having high universal applicability at a low cost by selecting plural selectable adaptive controllers via a controller deciding part based on the characteristics having a controlled system and switching the manipulated variable outputted from the selected controller via a switching part. CONSTITUTION:Each of plural adaptive controllers 22 outputs a proper controlled variable with only the specific characteristics of a controlled system 1. A controller deciding part 21a fetches the manipulated variable to be given to the controlled system and the outputted controlled variable and selects an optimum controller 22. At the same time, the optimum controller 22 is connected to the controlled system via a switching part 23. As a result, a controller 22 having the optimum manipulated variable is selected among plural controllers 22. Thus a satisfactory manipulated variable is secured and the stable control characteristics can be obtained even though the controlled system has a big difference of characteristics.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an adaptive controller that achieves good control over a plurality of control objects having different characteristics or control objects whose characteristics change.

[Conventional technology]

FIG. 4 shows, for example, Tamura, “Model Normative Adaptive Control and Self-Tuning Regulator,” Measurement and Control, Vol. 2
3, No. 5, P. 428~P. 432 (1988), in which l is the controlled object, 2 is the adaptive controller, 11 is the target value of the controlled object l, 12 is the output from the adaptive controller 2, Operation amount input to controlled object 1, 13
is the control amount output from the controlled object 1, and the adaptive controller 2 is composed of a parameter adjustment mechanism 52 and a controller 51.

Next, the operation will be explained. First, adaptive controller 2
The controller 51 takes in the target value 11 and the control amount 13 as control input information, and calculates the manipulated variable 12 which is the input signal to the controlled object 1. The parameter adjustment mechanism 52 takes in the input of the controlled object 1 and the manipulated variable 12 and the controlled amount 13, which are output signals, and estimates the characteristics of the controlled object 1. Then, parameters for calculating the operation amount in the controller 51 suitable for the estimated characteristics are determined and changed. In addition, in the adaptive controller 2 consisting of the parameter adjustment mechanism 52 and the controller 51, it is necessary to assume the order of the controlled object 1, the size of dead time, etc. in advance.

Next, an example of an adaptive controller using fuzzy rules will be explained. FIG. 5 shows, for example, Mori, "PID (proportional, integral, differential) general-purpose one-lube controller based on fuzzy theory J Instrumentation, Vol. 31 No. 5P. 21~P
．． 25 is a block diagram showing a conventional adaptive controller shown in FIG. The quantity 13 is a controlled quantity output from the controlled object 1, and the adaptive controller 2 is composed of a PID control section 53 and a fuzzy inference section 54.

Next, the operation will be explained. First, the PID control unit 53 performs a PID calculation on the deviation value between the externally specified target value 11 and the manipulated variable 12, and calculates the manipulated variable 12 to the controlled object 1. In the fuzzy inference unit 54, the target value 1
1, the control amount 13, and the operation amount 12 as input information, respectively, and the PID control unit 5 is adapted to suit the characteristics of the controlled object 1.
Find the PID parameters for 3 by fuzzy reasoning,
The PID parameters used by the PID control unit 53 are sequentially changed and the manipulated variable 12 to the controlled object 1 is output.

[Problems to be Solved by the Invention] Since the conventional adaptive controller 1/roller is configured as described above, it is difficult to solve the problem of the order of computation necessary to adaptively control the controlled object 1 and the size of control dead time. There has been a problem in that a good control amount 13 cannot be output if it differs from a pre-assumed controlled object model.

In addition, in the case of an adaptive controller that uses conventional fuzzy rules, if the characteristics of the controlled object 1 deviate significantly from the initial assumption, it becomes difficult to determine the fuzzy rules, and it is not possible to output a good manipulated variable. There was a problem that the range was limited.

This invention was made to solve the above problems, and even when the characteristics of the controlled objects are significantly different,
The objective is to obtain an adaptive controller that can output a good manipulated variable and achieve stable control characteristics.

[Means for Solving the Problems] An adaptive controller according to the present invention includes a controller determination unit that takes in an operation amount and a control amount, and selects one suitable for a controlled object from among a plurality of adaptive controllers; When the overall characteristics of are classified into multiple characteristics, the selected adaptive controller is determined to be the control target based on the output from the multiple adaptive controllers provided corresponding to each characteristic and the controller determination unit. The switching unit is equipped with a switching unit connected to the switch.

[Operation] The adaptive controller in the present invention takes in the target value and the control amount, and includes a plurality of adaptive controllers, such as (A) a parameter adjustment mechanism and a controller, (B) a fuzzy control section, or (
C) A controller determination unit that takes in the input/output information of the controlled object, ie, the manipulated variable and the controlled variable, from the fuzzy inference unit, for example, (
A) model classification unit, (B) fuzzy control determination unit, or (
C) Good control is achieved by the fuzzy inference decision unit selecting the optimal one (however, the above (A) to (C)
are assumed to correspond to each other).

[Embodiments of the invention] A first embodiment of the present invention will be described below with reference to the drawings.

In FIG. 1, the same parts as in FIG. 4 are designated by the same reference numerals. In FIG. A model classification unit 21a that classifies the model into one of the models, a parameter adjustment mechanism and controller 22 created corresponding to each of the models, and a parameter adjustment mechanism and controller 22 that are adopted based on the output from the model classification unit 21a. and a switching section 23 for switching. In addition, multiple parameter adjustment mechanisms and controller 2
2, the adaptive controller is also connected to the model classifier 2.1. A controller determination unit is realized in a.

Next, the operation of this invention will be explained. First, when classifying the models of the controlled object 1 according to the amount of dead time, consider a plurality of models having different dead times. here,
The parameter adjustment mechanism and controller 22 are made to have characteristics suitable for the controlled object model, and each parameter adjustment mechanism and controller 22 are operated in the same manner as the conventional adaptive controller 2. In the model classification section 21a, the switching section 23
Based on the manipulated variable 12 output from the controlled variable 12 and the controlled variable 13 outputted from the controlled object 1, the size of the dead time is fixed, and it is determined which model the current controlled object 1 corresponds to. Then, the switching section 23 adds the output from the controller for the model determined by the model classification section 21a to the controlled object l as the manipulated variable 12. By doing so, it is possible to perform good control even when the preset dead time of the controlled object 1 is significantly different from the actual value.

In addition, in the above embodiment, the case where the model of the controlled object 1 is classified by the size of the dead time is described, but when the model is classified by the ratio of the dead time and the response time, or by the order of the arithmetic expression of the controlled object 1. We have described an example in which one output is adopted as the manipulated variable 12, such as when classifying them into a mixed model, but if the model classification unit 21a recognizes that it is an intermediate model, for example, the weighted average is used. The manipulated variable 12 may be determined by combining the outputs of a plurality of parameter adjustment mechanisms and the controller 22 using a fuzzy synthesis method or the like.

Next, a second embodiment of the invention will be described with reference to the drawings.

In FIG. 2, reference numeral 2b denotes an adaptive controller according to the present invention, which is composed of a fuzzy control determining section 21b, a switching section 23, and a plurality of fuzzy control sections 26. In this example, the fuzzy control section 26 implements an adaptive controller, and the fuzzy control determination section 2lb implements a controller determination section.

Next, the operation will be explained. First, when performing fuzzy control using fuzzy rules, if there are large variations in the characteristics of the controlled objects 1, it is very difficult to control all the controlled objects 1 well with one fuzzy control unit 26.

Therefore, in this embodiment, a plurality of fuzzy control units 26 are provided based on fuzzy rules suitable for the controlled object 1 having different characteristics, and the fuzzy control determination unit 2lb determines the manipulated variable 12 and the controlled variable 13 which are the input and output of the controlled object 1. A fuzzy control section 26 suitable for the current controlled object 1 is determined by taking in the fuzzy control section 26, and a switching section 23 switches the manipulated variable 12 to the controlled object 1 based on the result of the fuzzy control determining section 2lb. By using the adaptive controller 2b based on fuzzy control with such a system configuration, it is possible to realize good control over a wider range of control objects 1.

Further, a third embodiment of the present invention will be described with reference to the drawings.

In FIG. 3, reference numeral 2c denotes an adaptive controller according to the present invention, which is composed of a plurality of fuzzy inference sections 24, a fuzzy inference determination section 21c, a switching section 23, and a PID control section 25. In this example, the fuzzy inference section 24 and the PID control section 25 implement an adaptive controller, and the fuzzy inference determination section 21C implements a controller determination section.

In this embodiment, in consideration of the case where the characteristics of the controlled object 1 vary greatly over time, or the case where the characteristics of the controlled object 1 for which this adaptive controller 2c is used are large, the control device 1 is designed to be suitable for the controlled object 1 in any case. PID
Allow control. multiple fuzzy inference units 24
Each has different fuzzy rules suitable for controlled objects with different characteristics, and performs fuzzy inference to determine the variation of PID based on the target value 11, the manipulated variable 12, and the controlled variable 13.

The fuzzy inference determining unit 21c uses the human output to the controlled object 1. Controlled object 1 is obtained by taking in the manipulated variable l2 and controlled variable 13.
The characteristics of the fuzzy inference unit 24 are estimated and the result obtained by which fuzzy inference unit 24 is used as the PID parameter is determined. The switching unit 23 sends the adopted PID as a parameter to the PID control unit 25 based on the result from the fuzzy inference determining unit 21c. The PID control unit 25 performs normal PID control using the sent PID parameters.

Above, in the second and third embodiments, the case where the output of one fuzzy inference unit 24 or the fuzzy control unit 26 is adopted as the operation of the switching unit 23 has been described, but the fuzzy inference determination unit 21c or the fuzzy control determination Part 2lb
The outputs of a plurality of fuzzy inference units 24 or fuzzy control units 26 may be combined using a fuzzy synthesis method or the like. Further, as the fuzzy inference described in the above embodiment, a recursive fuzzy inference method using fuzzy rules may be adopted. Furthermore, as the fuzzy inference determining unit 21c or the fuzzy control determining unit 2lb in the second and third embodiments, the first
A model classification unit having the model concept described in the embodiment may be used.

[Effects of the Invention] As described above, according to the present invention, the adaptive controller is controlled by a plurality of selectable adaptive controls B. = Since the configuration includes a controller determining section that selects an adaptive controller based on the characteristics of the controlled object, and a switching section that switches the output of the adaptive control word, the characteristics of the controlled object can vary greatly over time. It can be easily handled in cases where the characteristics of the controlled object to be used vary greatly, and
This has the effect of being able to provide a highly versatile adaptive controller/roller at a low cost.

[Brief explanation of the drawing]

Fig. 2 is a block diagram showing an adaptive controller according to a first embodiment of the present invention, Fig. 2 is a block diagram showing an adaptive controller according to a second embodiment of the invention, and Fig. 3 is a block diagram showing an adaptive controller according to a second embodiment of the present invention.
The figure is a block diagram showing an adaptive controller according to a third embodiment of the present invention, and FIGS. 4 and 5 are block diagrams showing conventional adaptive controllers. In the figure, 1 is the controlled object, 2a, 2b, and 2c are the adaptive controllers, 11 is the target value, 12 is the manipulated variable, 13 is the controlled variable, 21a is the model classification section (controller judgment section), and 2lb is the fuzzy control decision section. (Controller judgment unit). 21c a fuzzy inference determining unit (controller determining unit), 22 a parameter adjustment mechanism and controller (adaptive controller), 23 a switching unit, 24
is a fuzzy inference unit (adaptive controller), 25 is a PID control unit (adaptive controller), and 26 is a fuzzy control unit (adaptive controller)
. In addition, in the figures, the same reference numerals indicate the same or equivalent parts.

Claims (1)

[Claims] In an adaptive controller that takes in a target value and a controlled variable output from a controlled object and controls the manipulated variable for the controlled object by changing a calculation formula or calculation parameter according to the characteristics, the overall characteristics of the controlled object are controlled. A plurality of adaptive controllers are classified into a plurality of characteristics and designed for the classified characteristics, and input/output information from the control object is taken in and applied to the control object from among the plurality of adaptive controllers. An adaptive controller comprising: a controller determining section that selects a suitable controller; and a switching section that switches an adaptive controller based on information from the controller determining section.