Fault Tolerant Digital Flight Control with Analytical Redundancy
This report has yet to be scanned by Contrails staff
Report Number: AFFDL-TR-77-25
Author(s): Cunningham, Tom, Carlson, Don, Hendrick, Russ, Shaner, Don, Hartmann, Gary, Stein, Gunter
Corporate Author(s): Honeywell Systems & Research Center
Laboratory: Air Force Flight Dynamics Laboratory
Date of Publication: 1977-05
Pages: 396
Contract: F33615-76-C-3031
DoD Project: 1987
DoD Task: 198701
Identifier: ADA045671
Abstract:
Analytical redundancy offers high potential for solving the sensor redundancy problem. The current state-of-the-art in analytical redundancy contains a number of candidate filters which are developed here from a fundamental relationship of variables through hybrid simulation. Three analytical redundancy concepts of varying complexity were developed for the A-7D aircraft. Two monitor techniques were used: a multiple trip level exceedance criteria currently in use with voting systems, and a sequential likelihood ratio hypothesis test (SLRT) on the mean value of the error signal. Results indicate that fault detection, whether designed by classical methods or Kalman filtering, perform similarly and that gust estimation improves fault detection performance. Also, failure isolation for a single set of unlike sensors is difficult to achieve with a reasonable computation load. The output recommendation is to proceed to flight test with a set of nonlinear diagnostic filters to be used with comparison monitors for dual sensor failure isolation, creating a fail-operative dual sensor system. Fail-safe operation will also be achieved by continuing to monitor the filters and to detect, but not isolate, a second failure.
Provenance: University of Colorado Colorado Springs, Kraemer Family Library
Author(s): Cunningham, Tom, Carlson, Don, Hendrick, Russ, Shaner, Don, Hartmann, Gary, Stein, Gunter
Corporate Author(s): Honeywell Systems & Research Center
Laboratory: Air Force Flight Dynamics Laboratory
Date of Publication: 1977-05
Pages: 396
Contract: F33615-76-C-3031
DoD Project: 1987
DoD Task: 198701
Identifier: ADA045671
Abstract:
Analytical redundancy offers high potential for solving the sensor redundancy problem. The current state-of-the-art in analytical redundancy contains a number of candidate filters which are developed here from a fundamental relationship of variables through hybrid simulation. Three analytical redundancy concepts of varying complexity were developed for the A-7D aircraft. Two monitor techniques were used: a multiple trip level exceedance criteria currently in use with voting systems, and a sequential likelihood ratio hypothesis test (SLRT) on the mean value of the error signal. Results indicate that fault detection, whether designed by classical methods or Kalman filtering, perform similarly and that gust estimation improves fault detection performance. Also, failure isolation for a single set of unlike sensors is difficult to achieve with a reasonable computation load. The output recommendation is to proceed to flight test with a set of nonlinear diagnostic filters to be used with comparison monitors for dual sensor failure isolation, creating a fail-operative dual sensor system. Fail-safe operation will also be achieved by continuing to monitor the filters and to detect, but not isolate, a second failure.
Provenance: University of Colorado Colorado Springs, Kraemer Family Library