Efficient fuzzy logic fault accommodation algorithm

• US 6,098,011 A
• Filed: 05/18/1998
• Issued: 08/01/2000
• Est. Priority Date: 05/18/1998
• Status: Expired due to Term

• Current Assignee: Alliedsignal Inc.
• Original Assignee: Alliedsignal Inc.
• Inventors: Scott, Michael W.
• Primary Examiner(s): Cuchlinski, Jr., William A.
• Assistant Examiner(s): Beaulieu, Yonel
• Application Number:
  US09/081,037
• Time in Patent Office: 806 Days
• Field of Search: 701/100, 701/106, 701/40, 701/36, 701/54, 701/57, 60/239, 60/624, 60/39.75, 318/564, 318/565, 706/900, 706/904, 706/905, 706/906, 706/907
• US Class Current: 701/100
• CPC Class Codes: G01D 3/08 with provision for safeguarding the apparatus, e.g. against abnormal operation, against breakdown
• CPC Class Codes: G05B 9/03 with multiple-channel loop, i.e. redundant control systems
• CPC Class Codes: Y10S 706/90 Fuzzy logic

Abstract

A simplified fault logic accomodation control which utilizes fuzzy logic to assist in calculating a preselected weighted average of the values of two redundant sensors that are sensing the same parameter. A synthesized, third, predicted value of the parameter is used in the fuzzy logic to arbitrate between the two sensed values.

Claims

1. In a control system having redundant sensors providing at least two sensed values of a single parameter being measured, and having means for generating a synthesized value of said parameter;

a method of determining a single output value of said parameter for use in the control system, comprising the steps of;

determining a first error by comparing said sensed values to one another;

determining a second error by comparing said sensed values to said synthesized value;

determining a graded membership of said first and second errors from a fuzzy logic look-up table; and

calculating a preselected weighted average of said sensed values based on said graded membership, said preselected weighted average being the single output value of said parameter for use in the control system.

2. A method as set forth in claim 1, wherein said step of determining the first error includes determining the difference between said sensed values.

3. A method as set forth in claim 1, wherein said step of determining the second error includes:

selecting one of the sensed values according to a predetermined criteria; and

determining the difference between said synthesized value and said selected one of the sensed values.

4. A method as set forth in claim 3, wherein said step of determining the first error includes determining the difference between said sensed values.

5. A method as set forth in claim 4, wherein said selecting step comprises selecting the higher of said sensed values.

6. A method as set forth in claim 4, wherein said selecting step comprises selecting the lower of said sensed values.

7. A method as set forth in claim 4, wherein said step of determining a graded membership comprises classifying said first error as being small, medium or large, and classifying said second error as being small, medium or large.

8. A method as set forth in claim 7, wherein said graded membership is determined by said fuzzy logic look-up table according to the following fuzzy logic rules:

1) if the first error is small and the second error is small, use the average of said sensed values;

2) if the first error is small and the second error is medium, use the average of said sensed values;

3) if the first error is small and the second error is large, use the average of said sensed values;

4) if the first error is medium and the second error is small, use the higher of said sensed values;

5) if the first error is medium and the second error is medium, use the average of said sensed values;

6) if the first error is medium and the second error is large, use the lower of said sensed values;

7) if the first error is large and the second error is small, use the higher of said sensed values;

8) if the first error is large and the second error is medium, use the average of said sensed values;

9) if the first error is large and the second error is large, use the lower of said sensed values.

9. A method as set forth in claim 8, wherein said graded membership is a numerical output having a value between zero and one, and wherein said step of calculating a preselected weighted average comprises:

multiplying the higher of said sensed values by said numerical output to produce a first weighted output;

multiplying the lower of said sensed values by one minus said numerical output to produce a second weighted output; and

summing said first and second weighted outputs to produce said preselected weighted average.

15. A method as set forth in claim 1, wherein said control system controls operation of machinery, and said single parameter is associated with operation of the machinery.

16. A method as set forth in claim 15, wherein said machinery is a gas turbine engine, said single parameter being an operating pressure at a preselected station within the engine.

10. In a control system having redundant sensors providing at least two sensed values of a single parameter being measured, and having means for generating a synthesized value of said parameter;

a method of determining a single output value of said parameter for use in the control system, comprising the steps of;

determining a first error based on the difference between said sensed values;

selecting one of the sensed values according to a predetermined criteria;

determining a second error based on the difference between said synthesized value and said selected one of the sensed values;

determining a graded membership of said first and second errors from a fuzzy logic look-up table; and

calculating a preselected weighted average of said sensed values based on said graded membership, said preselected weighted average being the single output value of said parameter for use in the control system.

11. A method as set forth in claim 10, wherein said graded membership is a numerical output having a value between zero and one, as determined by said fuzzy logic look-up table.

12. A method as set forth in claim 11, wherein said step of calculating a preselected weighted average comprises:

multiplying said selected one of the sensed values by said numerical output to produce a first weighted output;

multiplying the other of said sensed values by one minus said numerical output to produce a second weighted output; and

summing said first and second weighted outputs to produce said preselected weighted average.

13. A method as set forth in claim 12, wherein selecting step comprises selecting the higher of said sensed values.

14. A method as set forth in claim 12, wherein said selecting step comprises selecting the lower of said sensed values.

17. A gas turbomachine including:

redundant sensors producing at least two sensed values of a single parameter associated with turbomachine operation;

means for generating a synthesized value of said parameter;

a control system for the turbomachine, said control system receiving said two sensed values and said synthesized value; and

means for producing a single output value of said parameter for further use in the control system, comprising;

a first comparator for comparing said sensed values to one another to generate a first error signal;

a second comparator for comparing said sensed values to said synthesized value to generate a second error signal;

a fuzzy logic look-up table for determining a graded membership of said first and second error signals; and

means for calculating said single output value of said parameter as a preselected weighted average of said sensed values based on said graded membership.