Optimized tuner selection for engine performance estimation

• US 8,386,121 B1
• Filed: 06/02/2010
• Issued: 02/26/2013
• Est. Priority Date: 09/30/2009
• Status: Active Grant

• Current Assignee: Nasa Headquarters
• Original Assignee: United States Administrator of National Aeronautics and Space Administration
• Inventors: Simon, Donald L., Garg, Sanjay
• Primary Examiner(s): EL CHANTI, HUSSEIN A
• Application Number:
  US12/791,907
• Time in Patent Office: 1,000 Days
• Field of Search: 701/3, 701/54, 701/29.1, 701/30.7
• US Class Current: 701/30.7
• CPC Class Codes: G01M 15/02 Details or accessories of testing apparatus

Abstract

A methodology for minimizing the error in on-line Kalman filter-based aircraft engine performance estimation applications is presented. This technique specifically addresses the underdetermined estimation problem, where there are more unknown parameters than available sensor measurements. A systematic approach is applied to produce a model tuning parameter vector of appropriate dimension to enable estimation by a Kalman filter, while minimizing the estimation error in the parameters of interest. Tuning parameter selection is performed using a multi-variable iterative search routine which seeks to minimize the theoretical mean-squared estimation error. Theoretical Kalman filter estimation error bias and variance values are derived at steady-state operating conditions, and the tuner selection routine is applied to minimize these values. The new methodology yields an improvement in on-line engine performance estimation accuracy.

Claims

1. A method of optimizing the estimated performance and estimated condition of an engine comprising:

estimating a plurality of engine parameters based on a plurality of sensor inputs, wherein said engine parameters include a set of health parameters and a set of performance parameters, and wherein the number of health parameters in said set is greater than the number of sensors;

determining a relationship between said set of health parameters and a tuning vector, wherein said tuning vector is a linear combination of said entire set of health parameters;

determining at least one state equation in terms of said tuning vector;

solving said at least one state equation to determine an estimate for said set of health parameters;

estimating said performance parameters based on said health parameter estimates;

calculating the error in said estimation of said health parameters and said performance parameters;

optimizing said relationship between said tuning vector and said set of health parameters based on said error calculation;

determining an optimized tuning vector based on said optimized relationship;

determining at least one optimized state equation in terms of said optimized tuning vector;

solving said at least one optimized state equation to determine an optimized estimate of said engine parameters.

2. The method of claim 1, wherein calculating the error includes calculating the mean estimation error.

3. The method of claim 1, wherein said performance parameters are a function of said health parameters.

4. The method of claim 1, wherein said relationship between said tuning vector and said set of health parameters is optimized to minimize the error in estimating a given subset of said health parameters.

5. The method of claim 1, wherein said relationship between said tuning vector and said set of health parameters is optimized to minimize the error in estimating a given subset of said performance parameters.

6. The method of claim 1, wherein said relationship between said tuning vector and said set of health parameters is optimized to minimize the error in estimating a combination of a given subset of said health parameters and a given subset of said performance parameters.

7. The method of claim 1, wherein the dimension of said tuning vector is less than or equal to the number of said sensors.

8. The method of claim 1, wherein said relationship between said health parameters and said tuning vector is defined by a matrix.

9. The method of claim 8, wherein a first dimensions of said matrix is equal to the dimension of said tuning vector.

10. The method of claim 9, wherein a second dimension of said matrix is less than or equal to the number of said health parameters.

11. The method of claim 1 further comprising the step of modifying control of said engine based on said optimized estimate of said engine parameters.

12. The method of claim 1 wherein solving said state equation includes constructing a Kalman filter.

13. An optimal tuning system comprising:

a plurality of sensors capable of sensing engine parameters;

a logic in operable communication with said plurality of sensors and configured to receive input from said plurality of sensors, said logic configured to determine optimized parameters related to the estimated performance and condition of said engine;

wherein said logic estimates a plurality of engine parameters based on said sensors, said engine parameters including a set of health parameters and a set of performance parameters, and wherein the number of health parameters in said set is greater than the number of sensors;

wherein said logic determines a relationship between said set of health parameters and a tuning vector that is a linear combination of said entire set of health parameters;

wherein said logic determines at least one state equation in terms of said tuning vector and solves said at least one state equation to determine an estimate for said set of health parameters;

wherein said logic estimates said performance parameters based on said health parameter estimates and calculates the error in said estimation of said health parameters and said performance parameters;

wherein said logic optimizes said relationship between said tuning vector and said set of health parameters based on said error calculation;

wherein said logic determines an optimized tuning vector based on said optimized relationship;

wherein said logic determines at least one optimized state equation in terms of said optimized tuning vector

wherein said logic solves said at least one optimized state equation to determine an optimized estimate of said engine parameters.

14. The optimal tuning system of claim 13, wherein said logic comprises multiple logics.

15. The optimal tuning system of claim 14, wherein said optimized tuning vector is defined within one of said multiple logics.

16. The optimal tuning system of claim 13, wherein said logic calculates the mean estimation error in said estimation of said health parameters and said performance parameters.

17. The optimal tuning system of claim 13, wherein said performance parameters are a function of said health parameters.

18. The optimal tuning system of claim 13, wherein said logic optimizes said relationship between said tuning vector and said set of health parameters to minimize the error in estimating a given subset of said health parameters.

19. The optimal tuning system of claim 13, wherein said logic optimizes said relationship between said tuning vector and said set of health parameters to minimize the error in estimating a combination of a given subset of said health parameters and a given subset of said performance parameters.

20. The optimal tuning system of claim 13, wherein said logic solves said state equation by constructing a Kalman filter.

21. The optimal tuning system of claim 13, wherein said logic is capable of modifying control of said engine based on said optimized estimate of said engine parameters.

22. The optimal tuning system of claim 13, wherein the dimension of said tuning vector is less than or equal to the number of said sensors.