Rule set for root cause diagnostics

• US 20060095394A1
• Filed: 12/20/2005
• Published: 05/04/2006
• Est. Priority Date: 03/28/1996
• Status: Active Grant

• Current Assignee: Fisher-Rosemount Systems Incorporated (Emerson Electric Company)
• Original Assignee: Fisher-Rosemount Systems Incorporated (Emerson Electric Company)
• Inventors: Miller, John P., Eryurek, Evren
• Granted Patent: US 7,623,932 B2
• Time in Patent Office: Days
• US Class Current: 706/45
• CPC Class Codes: G05B 13/0275 using fuzzy logic only
• CPC Class Codes: G05B 21/02 electric
• CPC Class Codes: G05B 23/0213 Modular or universal configuration of the monitoring system, e.g. monitoring system having modules that may be combined to build monitoring program; monitoring system that can be applied to legacy systems; adaptable monitoring system; using different communication protocols
• CPC Class Codes: G05B 23/0278 Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA
• CPC Class Codes: G06N 5/025 Extracting rules from data

Abstract

A reduced rule set for identifying a root cause of an aberration in an industrial process is generated from a complete rule base. The rule base includes plurality of rules for the industrial process, and each rule comprises a condition of at least one process signal of the industrial process and a fault which corresponds to the condition of at least one process signal. Available process signals are identified. Rules are selectively removed from the rule base to produce the reduced rule set.

Claims

1. A method for generating a reduced rule set for identifying a root cause of an aberration in an industrial process, comprising:

(a) obtaining a rule base comprising a plurality of rules for the industrial process, each rule comprising a condition of at least one process signal of the industrial process and a fault which corresponds to a condition of at least one process signal;

(b) identifying which process signals of the industrial process are available;

(c) selectively removing a rule from the plurality of rules of the rule base to produce the reduced rule set, wherein the step of selectively removing a rule is based upon the step of identifying which process signals of the industrial are available.

2. The method of claim 1 including identifying rules within the reduced rule set having identical process signal conditions.

3. The method of claim 2 including combining rules in the reduced rule set which have identical process signal conditions.

4. The method of claim 3 including creating a new fault for a combined rule.

5. The method of claim 1 wherein the process signals comprise process variables.

6. The method of claim 1 including automating steps (a) through (c).

7. The method of claim 6 and including performing (a) through (c) in a computer.

8. The method of claim 3 including repeating the step of combining until only unique rules are left in the reduced rule set.

9. The method of claim 1 wherein the condition of at least one of the process signals comprises the group of conditions consisting of less than a constant, greater than a constant, trending up, trending down, no change and blank.

10. A computer system configured to implement the method of claim 1.

11. A storage medium storing computer program instructions to execute the steps of:

a plurality of process models, each model related to a physical implementation of an industrial process;

a model selection input configured to receive a selected model, the selected model uniquely identifying one of the process models;

a process signal input configured to receive a plurality of process signals related to the process; and

a root cause output indicative of a source of the aberration in the process, the root cause output a function of the selected model and the process signals.

12. The method of claim 11 including identifying rules within the reduced rule set having identical process signal conditions.

13. The method of claim 12 including combining rules in the reduced rule set which have identical process signal conditions.

14. The method of claim 13 including creating a new fault for a combined rule.

15. The method of claim 11 wherein the process signals comprise process variables.

16. The method of claim 11 including automating steps (a) through (c).

17. The method of claim 16 and including performing (a) through (c) in a computer.

18. The method of claim 13 including repeating the step of combining until only unique rules are left in the reduced rule set.

19. The method of claim 11 wherein the condition of at least one of the process signals comprises the group of conditions consisting of less than a constant, greater than a constant, trending up, trending down, no change and blank.

20. An apparatus for determining a reduced rule set for identifying a root cause of an aberration in an industrial process, comprising;

(a) obtaining a rule base comprising a plurality of rules for the industrial process, each rule comprising a condition of at least one process signal of the industrial process and a fault which corresponds to a condition of at least one process signal;

(b) identifying which process signals of the industrial process are available;

(c) selectively removing a rule from the plurality of rules of the rule base to produce the reduced rule set, wherein the step of selectively removing a rule is based upon the step of identifying which process signals of the industrial are available.

21. The method of claim 20 including identifying rules within the reduced rule set having identical process signal conditions.

22. The method of claim 21 including combining rules in the reduced rule set which have identical process signal conditions.

23. The method of claim 22 including creating a new fault for a combined rule.

24. The method of claim 20 wherein the process signals comprise process variables.

25. The method of claim 20 including automating steps (a) through (c).

26. The method of claim 25 and including performing (a) through (c) in a computer.

27. The method of claim 22 including repeating the step of combining until only unique rules are left in the reduced rule set.

28. The method of claim 20 wherein the condition of at least one of the process signals comprises the group of conditions consisting of less than a constant, greater than a constant, trending up, trending down, no change and blank.

29. A computer system configured to implement the method of claim 20.