US 5,295,061 A · Control parameter tuning unit and a method of tuning parameters for a control unit

Abstract

A fuzzy control unit includes a fuzzy inference unit forming a manipulated variable for controlling a plant such that a control response from the plant is made equal to a previously set target value, based on fuzzy control knowledge. An automatic tuning unit corrects the fuzzy control rules in accordance with the control response from the plant. The automatic tuning unit includes a response waveform memory storing a control response value from the plant, a rule grade memory for storing grade values of the fuzzy control rules for the control response from the plant, an ideal waveform memory for storing ideal control response values to be attained by the plant under control by the fuzzy control unit, and a control evaluating circuit evaluating the control operation of the fuzzy control unit based on the ideal response value and the control response value. The automatic tuning unit also includes a circuit for correcting a membership function of a consequent portion of the fuzzy control rules in accordance with the amount of correction and the rule grade.

Claims48

Dependent claims

A tuning unit for a fuzzy control unit controlling a plant to be controlled based on fuzzy control knowledge including membership functions and fuzzy control rules such that a control response value applied to said plant is made equal to a predetermined target value, said tuning unit being for correcting said fuzzy control knowledge and comprising:

ideal response waveform storing means for storing a sequence of ideal response values which are desired outputs of said plant to be controlled;
response waveform storing means for storing a sequence of control response values from said plant to be controlled;
rule grade storing means for storing a grade value of the fuzzy control rules included in said fuzzy control knowledge with respect to each said control response value from said plant to be controlled;
control evaluating means for calculating from an ideal response valued stored in said ideal response waveform storing means and from a control response value stored in said response waveform storing means, a response deviation indicating a difference between said ideal response value and a corresponding control response value, and an amount of change of a response deviation indicating magnitude of change of the response deviation in a predetermined period;
means for storing tuning fuzzy knowledge including membership functions and fuzzy rules for correcting the membership function in a consequent portion of said fuzzy control rules in said fuzzy control knowledge, according to respective magnitudes of said response deviation and the amount of change in said response deviation;
and
correcting means responsive to the response deviation information and the amount of change of the response deviation calculated by said control evaluating means and to rule grade information from said rule grade storing means, for correcting the membership functions of said consequent portion of said fuzzy control rules of effecting a predetermined calculation in accordance with said tuning fuzzy knowledge.

4 more elements

A tuning unit for a fuzzy control unit controlling a plant such that a control response value from said plant is made equal to a predetermined target value, based on fuzzy control knowledge including membership functions and a fuzzy control rules, said a tuning unit being for correcting said fuzzy control rules and comprising:

control target value setting means for setting a control target value to be realized by a control response from said plant to be controlled;
control evaluating means for calculating characteristic variables to evaluate control operation of said plant to be controlled from the control response from the plant to be controlled, and for generating a control target deviation indicating a difference between thus calculated characteristic variable and the control target value set by said control target value setting means;
and
fuzzy knowledge correcting means responsive to the control target deviation from said control evaluating means for correcting said fuzzy control rule in said fuzzy control knowledge in accordance with a simplex method.

1 more element

Apparatus for constructing fuzzy control knowledge for a fuzzy control unit controlling a plant to be controlled such that a control response value from said plant under control is made equal to a predetermined target value by using fuzzy control knowledge base including membership functions and fuzzy control rules, comprising:

parameter storage means for storing at least one control parameter of a proportional parameter, an integral parameter and a differential parameter for controlling said plant to be controlled in accordance with PID operation including proportional operation, integration operation and differential operation;
division information storage means for storing information of a range of an input variable in an antecedent portion of said fuzzy control rules and of the number of divisions of the range;
input variable dividing means for fuzzy-dividing said input variable range of the fuzzy control rules, which is to constitute said fuzzy knowledge, into plural regions in accordance with the information of said the number of divisions and of said range stored in said division information storage means and for generating a membership function of said input variable corresponding to each resultant division;
fuzzy rule generating means for determining a consequent portion of said fuzzy control rule in accordance with a selected representative value of each said fuzzy-divided region generated by said input variable dividing means and to the value of the parameter stored in said parameter storing means for generating the fuzzy control rule;
and
fuzzy control knowledge storage means for storing the membership functions generated by said input variable dividing means and the fuzzy rules generated by said fuzzy rule generating means as said fuzzy control knowledge.

3 more elements

A PID control unit for controlling a plant to be controlled such that a control response value from said plant to be controlled is made equal to a predetermined target value in accordance with at least one control parameter of a proportional parameter, an integral parameter and a differential parameter indicating proportional operation, integral operation and differential operation, respectively, comprising:

parameter storage means for storing said control parameter;
arithmetic operation controlling means responsive to the control response value from said plant to be controlled and to said predetermined target value for carrying out control arithmetic operation based on the control parameter stored in said parameter storage means and for controlling said plant in accordance with the result of said control arithmetic operation;
target value setting means for setting a control target value which is to be attained by said plant under control of said arithmetic operation controlling means;
control evaluating means for calculating a characteristic variables for evaluating operation of control of said plant to be controlled from the control response value from said plant to be controlled, and for outputting a control target deviation indicating a difference between the calculated characteristic variable and the control target value set by said target value setting means;
a tuning fuzzy knowledge base storage means having control target for storing deviation as tuning input information for tuning said control parameters in response to an amount of said control target deviation;
correcting means for carrying out fuzzy inference in accordance with the fuzzy knowledge stored in said fuzzy knowledge for tuning storage means, using said control target deviation as input variables, and for correcting said control parameters in accordance with the result of the fuzzy inference.

3 more elements

For a fuzzy control unit controlling a plant such that a control response value from said plant is made equal to a predetermined target value based on a fuzzy control knowledge including membership functions and fuzzy control rules, a method of tuning for correcting said fuzzy control knowledge, comprising the steps of:

storing a sequence of control response values from said plant to be controlled;
storing grade information of fuzzy control rules included in said fuzzy control knowledge corresponding to each said control response value from said plant to be controlled;
calculating, from a stored ideal response value and the stored control response value, a response deviation indicating a difference between said ideal response value and the corresponding control response value, and an amount of change of the response deviation indicating a magnitude of change in the response deviation in a predetermined period;
and
carrying out a fuzzy inference calculation in accordance with a previously prepared fuzzy knowledge for tuning, in accordance with respective magnitudes of the calculated response deviation and the amount of change of the response deviation and with the stored rule grade information for correcting a membership function in a consequent portion of said fuzzy control rules;
wherein
said step of correcting includes the step of using tuning fuzzy knowledge for correcting a membership function of a consequent portion of said fuzzy control rules, using said response deviation and said amount of change of the response deviation as antecedent portion input variables.

4 more elements

A tuning method for a fuzzy control unit for controlling a plant to be controlled such that a control response value from said plant is made equal to a predetermined target value based on fuzzy control knowledge including membership functions and fuzzy control rules, said tuning method being for correcting said fuzzy control knowledge and comprising the steps of:

setting a control target value to be realized by the control response from said plant to be controlled;
calculating a characteristic variable for evaluating control operation of said plant to be controlled from the control response from said plant to be controlled and generating a control target deviation indicating a difference between thus calculated characteristic variable value and said set control target value;
and
correcting said fuzzy control rule in said fuzzy control knowledge in accordance with a simplex method in response to said control target deviation.

1 more element

A method of constructing fuzzy control knowledge for a fuzzy control unit controlling a plant to be controlled such that a control response value from said plant to be controlled is made equal to a previously set target value by using fuzzy control knowledge including membership functions and fuzzy control rules, comprising the steps of:

storing at least one control parameter out of a proportional parameter, an integral parameter and a differential parameter for controlling said plant to be controlled in accordance with PID operation including proportional operating, integral operation and differential operation;
storing information of a value range of an input variable in an antecedent portion of said fuzzy control rule and of the number of divisions of the value range;
fuzzy-dividing said input variable of said fuzzy control rules which is to constitute said fuzzy knowledge into plural regions in accordance with said stored information of said number of divisions and said range, and generating a membership function of said input variable corresponding to each of resulting divisions;
determining a consequent portion of said fuzzy control rule in accordance with a selected representative value of each of thus generated fuzzy-divided regions and to the value of the stored parameter for generating the fuzzy control rule;
and
storing thus generated membership functions and thus generated fuzzy rules as said fuzzy control knowledge.

3 more elements

A method of PID control for controlling a plant to be controlled such that a control response value from said plant to be controlled is made equal to a previously set target value in accordance with at least one control parameter of a proportional parameter, an integral parameter and a differential parameter indicating proportional operation, integral operation and differential operation, respectively, comprising the steps of:

storing said control parameter;
performing a control arithmetic operation in response to a control response from said plant to be controlled and to the previously set target value based on said stored control parameter and controlling said plant to be controlled in accordance with the result of the control arithmetic operation;
setting a control target value to be attained by said plant under control based on said control arithmetic operation;
calculating a characteristic variable for evaluating state of control of said plant to be controlled, from the control response from said plant to be controlled, and outputting a control target deviation indicating a difference between thus calculated characteristic variable and said set control target value;
and
carrying out a fuzzy inference using said control target deviation as an input variable for correcting the stored control parameter in accordance with the result of the fuzzy inference, said step of correcting the stored control parameter including the step of using fuzzy knowledge for tuning said control parameter according to said control target deviation.

3 more elements

Dependencies are inferred from the claim text.

Control parameter tuning unit and a method of tuning parameters for a control unit

Life of this patent

Control parameter tuning unit and a method of tuning parameters for a control unit

Life of this patent

Abstract

Claims48