Problem state monitoring

• US 5,193,143 A
• Filed: 11/07/1989
• Issued: 03/09/1993
• Est. Priority Date: 01/12/1988
• Status: Expired due to Term

• Current Assignee: Honeywell Incorporated (Honeywell International Inc.)
• Original Assignee: Honeywell Incorporated (Honeywell International Inc.)
• Inventors: Kaemmerer, William F., Allard, James R.
• Primary Examiner(s): MacDonald, Allen R.
• Application Number:
  US07/433,401
• Time in Patent Office: 1,218 Days
• Field of Search: 364/200, 364/900, 364/513, 395/51-55, 395/911, 395/912, 395/914, 395/915, 395/903, 395/904, 395/906, 395/907
• US Class Current: 706/45
• CPC Class Codes: G05B 13/028 using expert systems only
• CPC Class Codes: Y10S 706/911 Nonmedical diagnostics
• CPC Class Codes: Y10S 706/912 Manufacturing or machine, e.g. agricultural machinery, machine tool

Abstract

The invention relates to a problem-state monitoring system which includes an automated reasoning technique for providing moment-by-moment advice concerning the operation of a process. The reasoning technique involves the representation of states a problem can attain during the problem solving process which provides greater flexibility for the monitoring and advisory functions.

Claims

1. A method for use with an expert system for diagnosing and providing remedies for problems incurred in the operation of a manufacturing process, wherein said manufacturing process has a plurality of data points from which the occurrence of said problems may be determined, and wherein said expert system has an inference engine means for inferencing which interfaces with and monitors said data points,

said inference engine means being operable relative to said manufacturing process to facilitate (1) the determination of an unconfirmed root problem associated with one of said data points, designated a root node, and (2) the generation of a hierarchical problem space relative to said root node listing a set of descendant nodes having unconfirmed operating problems associated therewith potentially causing said root problem;

and wherein said inference engine means has available an investigation procedure during the course of the use of such procedure a determination can be made for each of said unconfirmed operating problems whether it is an actual problem and whether said inference engine means has a remedy for said actual problem,

said method for use with such an expert system comprising the steps of;

A. monitoring said manufacturing process and reading said data points with the use of said inference engine means;

B. determining from said data point readings in association with said inference engine means the existence of any root problem and generating in association with said inference engine means a hierarchical problem space listing of a set of descendant unconfirmed operating problems potentially causing said root problem;

C. determining for each of said unconfirmed operating problems in association with said investigation procedure (1) whether said unconfirmed operating problem is in fact a cause of said root problem and (2) whether said inference engine has a remedy for said cause of said root problem;

D. providing a transition network having a format of transition states for said root problem during the course of said investigation procedure from the time said root problem is detected until it is either resolved or it is determined to be irresolvable; and

E. returning to step A.

2. A method for use with an expert system for diagnosing and providing remedies for problems incurred in the operation of a manufacturing process, wherein said manufacturing process has a plurality of data points from which the occurrence of said problems may be determined, and wherein said expert system has an inference engine means for inferencing which interfaces with and monitors said data points,

said inference engine means being operable relative to said manufacturing process to facilitate (1) the determination of an unconfirmed root problem associated with one of said data points, designated a root node, and (2) the generation of a hierarchical problem space relative to said root node listing a set of descendant nodes having unconfirmed operating problems associated therewith potentially causing said root problem;

and wherein said inference engine means has available an investigation procedure during the course of the use of such procedure a determination can be made for each of said unconfirmed operating problems whether it is an actual problem and whether said inference engine means has a remedy for said actual problem,

said method for use with such an expert system comprising the steps of;

A. monitoring said manufacturing process and reading said data points with the use of said inference engine means;

B. determining from said data point readings in association with said inference engine means the existence of any root problem and generating in association with said inference engine means a hierarchical problem space comprising a hierarchically organized set of descendant unconfirmed operating problems potentially causing said root problem;

said hierarchical problem space having a top nearest said root and a bottom furthest from said root problem,

C. providing a transition network having a set of transition states for each said descendant unconfirmed problem during the course of said investigation procedure from the time said descendant unconfirmed problem is detected until it is either resolved or it is determined to be irresolvable; and

D. determining from the bottom of said hierarchically organized set upwards in parallel toward said root for each of said unconfirmed problems in association with said investigation procedure;

whether said unconfirmed problem is in fact a cause of said root problem and, if true;

whether said inference engine has a remedy for said cause of said root problem;

E. providing a transition network having a format of transition states for said root problem during the course of said investigation procedure from the time said root problem is detected until it is either resolved or it is determined to be irresolvable; and

F. returning to step A.

3. A method for use with an expert system for diagnosing and providing remedies for problems in a controlled system, said controlled system having a plurality of data points from which said problems may be determined, said expert system having an inference engine means for inferencing which interfaces with and monitors said data points, said inference engine operating relative to said controlled system to facilitate both the determination of an unconfirmed root problem associated with one of said data points, designation of a root node, and the generation of a hierarchical problem space relative to said root node, said space having a set comprising a listing of descending nodes having unconfirmed problems associated therewith which have the potential to be a cause of said root problem;

said inference engine means having investigation procedure available during the course of the use of such procedure a determination can be made for each of said unconfirmed problems whether said unconfirmed problem is an actual problem, and whether said inference engine has a remedy for said actual problem, said method comprising the steps of;

A. monitoring said manufacturing process and reading the data points with the use of said inference engine means;

B. determining from said data point readings in association with said inference engine means the existence of any root problem and generating in association with said inference engine means a hierarchical problem space comprising a set of descendant unconfirmed problems potentially causing said root problem;

C. determining for each of said unconfirmed problems in association with said investigation procedure;

whether said unconfirmed problem is in fact a cause of said root problem and, if true;

whether said inference engine has a remedy for said cause of said root problem;

D. providing a transition network having a set of transition states for said root problem during the course of said investigation procedure from the time said root problem is detected until it is either resolved or it is determined to be irresolvable; and

E. returning to step A.

4. A method for use with an expert system for diagnosing and providing remedies for problems in a controlled system, said controlled system having a plurality of data points from which said problems may be determined, said expert system having an inference engine means for inferencing which interfaces with and monitors said data points, said inference engine operating relative to said controlled system to facilitate both the determination of an unconfirmed root problem associated with one of said data points, designation of a root node, and the generation of a hierarchical problem space relative to said root node, said space having a set comprising a listing of descending nodes having unconfirmed problems associated therewith which have the potential to be a cause of said root problem;

said inference engine means having investigation procedure available during the course of the use of such procedure a determination can be made for each of said unconfirmed problems whether said unconfirmed problem is an actual problem, and whether said inference engine has a remedy for said actual problem, said method comprising the steps of;

A. monitoring said manufacturing process and reading the data points with the use of said inference engine means;

B. determining from said data point readings in association with said inference engine means any root problem and generating in association with said inference engine means a hierarchical problem space comprising a hierarchically organized set of descendant unconfirmed problems potentially causing said root problem;

said hierarchical problem space having a top nearest said root and a bottom furthest from said root problem,

C. providing a transition network having a set of transition states for each said descendant unconfirmed problem during the course of said investigation procedure from the time said descendant unconfirmed problem is detected until it is either resolved or it is determined to be irresolvable; and

D. determining from the bottom of said hierarchically organized set upwards in parallel toward said root for each of said unconfirmed problems in association with said investigation procedure;

whether said unconfirmed problem is in fact a cause of said root problem and, if true;

whether said inference engine has a remedy for said cause of said root problem;

E. providing a transition network having a format of transition states for said root problem during the course of said investigation procedure from the time said root problem is detected until it is either resolved or it is determined to be irresolvable; and

F. returning to step A.