Design and Optimization of Auxiliary Control in Active Fault Detection for Aircraft Systems

  • Dawei Sun
  • , Koffi M.D. Motchon
  • , Daochun Li*
  • , Jinwu Xiang
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Active fault detection excites a dynamic system by an auxiliary control input such that undetectable faults are revealed. For aircraft systems, it has been pointed out by some relevant works that certain actuator faults cannot be effectively detected in time without actively exciting the aircraft dynamics. However, existing works using auxiliary control for aircraft fault diagnosis mostly consider rather simple signals, which cannot guarantee the fault detection without degrading the performance of nominal flight operation. Since optimal auxiliary control input design for aircraft systems has not been extensively studied yet, we are motivated to formulate the design problem and propose an optimization scheme in this paper. Specifically, the proposed auxiliary control design problem addresses the active fault detection to multiple potential faults simultaneously under unknown-but-constrained inputs and nonlinear flight dynamics, and minimization of the intervention to nominal flight performance is considered. The auxiliary control design problem can be solved by the proposed bi-level optimization scheme without conservative relaxations, and an effective strategy for finding initial guesses is proposed. Numerical results and tests are presented to demonstrate the effectiveness of the proposed formulation and the optimization scheme for auxiliary control design.

Original languageEnglish
JournalIEEE Transactions on Aerospace and Electronic Systems
DOIs
StateAccepted/In press - 2026

Keywords

  • Active fault detection
  • aircraft systems
  • auxiliary control design
  • bi-level optimization

Fingerprint

Dive into the research topics of 'Design and Optimization of Auxiliary Control in Active Fault Detection for Aircraft Systems'. Together they form a unique fingerprint.

Cite this