Method and system for process control with complex inference mechanism using qualitative and quantitative reasoning

• US 5,251,285 A
• Filed: 06/07/1991
• Issued: 10/05/1993
• Est. Priority Date: 03/25/1988
• Status: Expired due to Term

• Current Assignee: Hitachi Engineering Co.,Ltd (Hitachi, Ltd.), Hitachi, Ltd.
• Original Assignee: Hitachi Engineering Co.,Ltd (Hitachi, Ltd.), Hitachi, Ltd.
• Inventors: Inoue, Haruki, Satoh, Yoshiyuki, Funabashi, Motohisa, Yahiro, Masakazu
• Primary Examiner(s): MacDonald, Allen R.
• Application Number:
  US07/712,104
• Time in Patent Office: 851 Days
• Field of Search: 395/10, 395/11, 395/61, 395/900, 395/906, 395/908, 395/910, 395/775, 364/160
• US Class Current: 706/10
• CPC Class Codes: G05B 13/0295 using fuzzy logic and expert systems
• CPC Class Codes: G06N 5/048 Fuzzy inferencing
• CPC Class Codes: G06N 7/04 Physical realisation
• CPC Class Codes: Y10S 706/90 Fuzzy logic
• CPC Class Codes: Y10S 706/906 Process plant
• CPC Class Codes: Y10S 706/908 Electronic or computer, internal or network, circuit
• CPC Class Codes: Y10S 706/91 Elevator

Abstract

A prediction control of a process containing a non-linear behavior is effected by predicting a transition of process quantities (control quantities) of an object to be controlled a predetermined time period after the current time to provide predicted values and by determining manipulation quantities of at least a control effector in accordance with a difference between the predicted values and predetermined target values of the controlled object. A quantitative operation is carried out arithmetically identifying a process as a linear behavior, and a fuzzy inference or qualitative operation, including a fuzzy rule based on empirical knowledge, is carried out for simulating a process, wherein for input process information and current manipulation quantities to be maintained, both operations parallelly determine predicted values of control quantities. A process behavior determination is arithmetically carried out to determine whether the process linearly behaves (i.e. the normal state or the state dominant to statistic distribution) or not (i.e. the extreme transition state such as a transient response, or an indeterminate state) from the specified process information on the basis of a process behavior determination rule. As a result, a predicted value acquired by the quantitative operation is selected in response to a determination of a linear behavior, and a predicted value acquired by the fuzzy inference operation is selected in response to a determination of a non-linear behavior.

Claims

1. A method of performing process control in accordance with a target value of a controlled object with a transition of the process being predicted, comprising the steps of:

inputting process information;

determining whether a process behavior at a current control timing is linear based on an evaluation rule for evaluating linearity of the process behavior;

in response to a determination indicating a linear behavior in said determining step, predicting a control quantity of the controlled object a predetermined period of time after the current control timing, using means for performing a quantitative arithmetic operation which simulates the process in a linear relation;

in response to a determination indicating no linear behavior in said determining step, predicting the control quantity of the controlled object a predetermined period of time after the current control timing, using means for performing a qualitative arithmetic operation which simulates the process in a fuzzy rule relation;

determining a manipulation quantity of a control effector from a difference between a predicted value obtained in a selected one of said predicting steps and the target value; and

controlling the process using the determined manipulation quantity.

2. The method according to claim 1, wherein said step of inputting process information includes inputting at least one of time-series data including past values and current values in said quantitative arithmetic operation means and differential data between the current value and the most recent value in said qualitative arithmetic operation means.

3. The method according to claim 1, further comprising the steps of operating said quantitative arithmetic operation means and said qualitative arithmetic operation means in parallel to obtain the values of said predicted control quantity.

4. The method according to claim 1, wherein said evaluation rule is composed of a set of production rules which evaluate linearity or non-linearity of the process behavior based on empirical knowledge, and further including representing a conclusion portion determined from an inference based on input information as an assumption portion by a certainty on which the significance of the quantitative arithmetic operation and the qualitative arithmetic operation depends, and performing a conditional matching operation for adjusting conflicting conclusions between the rules to determine whether the process behavior is linear or non-linear.

5. A method of performing a process control in accordance with a plurality of controlled objects with a transition of a process being predicted, comprising the steps of:

using input process information and an assumed manipulation quantity for arithmetically operating on predicted values of control quantities for the plurality of controlled objects a predetermined period of time after a current control timing by using means for performing a quantitative arithmetic operation which simulates the process in a linear relation and means for performing a qualitative arithmetic operation which simulates the process in a fuzzy rule relation, respectively;

determining whether a process behavior at the current control timing is linear based on an evaluation rule which evaluates the linearity of the process behavior;

selecting one of the predicted values obtained by the quantitative arithmetic operation means and obtained by said qualitative arithmetic operation means based on a result of said determining step;

determining the grade of satisfaction in which said plurality of controlled objects are as a whole satisfied with respective predicted values selected by said selecting step; and

if said determined satisfaction grade meets a predetermined reference, controlling the process using the assumed manipulation quantities.

6. The method according to claim 5, wherein said satisfaction grade determining step includes determining said satisfaction grade in accordance with weights predetermined for respective controlled objects and sum-of-product operations of said predicted values.

7. The method according to claim 5, wherein, when a manipulation quantity is outputted which is a combination of manipulation quantities of a plurality of control effectors exhibiting the highest satisfaction grade, arithmetically operating on said predicted values based on a combination of assumed manipulation quantities.

8. A process control system, including a computer system, for controlling a manipulation quantity of a control effector in accordance with a target quantity of a controlled object with a transition of a process being predicted, comprising:

quantitative arithmetic operation means for simulating a process in a linear relation to predict a control quantity of the controlled object a predetermined period of time after inputting of process information;

qualitative arithmetic operation means for simulating a process in a fuzzy rule relation to predict a control quantity of the controlled object a predetermined period of time after inputting of process information;

process behavior inference means, having a set of production rules which evaluate linearity of non-linearity of a process behavior based on empirical knowledge, for determining whether the process behavior is linear or non-linear from the process information at a current timing;

manipulation quantity output means, responsive to a determination by said process behavior inference means, for selecting one of said quantitative arithmetic operation means and said qualitative arithmetic operation means and for outputting a determined manipulation quantity, based on a difference between said predicted value and said target quantity; and

control means for controlling the process on the basis of said determined manipulation quantity.

9. The process control system according to claim 8, wherein said qualitative arithmetic operation means includes means providing a plurality of fuzzy rules presenting a relation between process information and a controlled object in the form of a production rule, a membership function evaluating the process information and the controlled object, respectively, in an assumption portion and a conclusion portion in said rules, for making an inference to said conclusion portion based on a minimum value in a result of evaluation of said assumption portion and predicting a control quantity of said controlled object based on a maximum value in a result of evaluation of the conclusion portion of each rule.

10. A process control system, including a computer system, for controlling manipulation quantities of a plurality of control effectors for a process in accordance with a plurality of controlled objects while a transition of the process is being predicted, comprising:

manipulation quantity candidate generation means for generating candidates on combination of predetermined manipulation quantities which are changeable from a current combination of manipulation quantities;

prediction means including quantitative arithmetic operation means for simulating the process in a linear relation and qualitative arithmetic operation means for simulating the process in a fuzzy rile relation, and for parallelly arithmetically operating on a control quantity of each of the controlled objects a predetermined period of time after a current time based on input process information and candidates on the combination of manipulation quantities;

process behavior inference means, including a set of production rules based on empirical knowledge evaluating the degree of linearity or the degree of non-linearity of the process behavior, for determining whether the process behavior is linear based on predetermined process information at a relevant timing;

manipulation quantity output means for selecting predicted values of one of said quantitative arithmetic operation means and said qualitative arithmetic operation means in accordance with a determination result of said process behavior inference means, and outputting said candidates on the combination of manipulation quantities to a plurality of corresponding control effectors when the degree of satisfaction obtained by the selected predicted values to thereby satisfy the plurality of controlled objects collectively meets a predetermined reference; and

means for controlling the process on the basis of said manipulation quantities.