Work

An online identification method for time-varying modal parameters of morphing aircraft

Year: 2026
Type: conference-paper
Abstract: During the morphing of novel-configuration aircraft, the structural stiffness and mass distribution evolve continuously, so the aeroelastic modes are strongly time-varying and non-stationary within a very short observation window, and conventional modal-identification methods that assume a linear time-invariant system have intrinsic limitations. This paper proposes a three-stage online method comprising adaptive pre-processing, recursive identification, and intelligent cleaning (MAC-aware adaptive recursive subspace tracker, MAC-ARST). In pre-processing, an adaptive notch filter and zero-phase band-pass filtering suppress mechanism harmonics. In identification, a mode-aware variable forgetting factor (VFF) driven by the modal assurance criterion (MAC) distance and the relative frequency increment is built on the variable forgetting factor-generalized YAST (VFF-GYAST) recursive subspace tracker, together with a lagged-reference subspace-residual detector with an adaptive-baseline cumulative sum (CUSUM) test, covariance hot-restart, and an adaptive Kalman tracker, giving stable tracking under slow and abrupt changes. In post-processing, spurious modes are removed automatically, and confidence intervals and flutter margins are output online. On a three-degree-of-freedom time-varying benchmark with a stiffness jump, a 2.13 Hz jump is detected with a 0.70 s delay and no detector false alarm in this single simulation realization. The re-convergence time is shortened by about 80% relative to sliding-window subspace identification, and the 0.90 ms median recursive step cost meets real-time requirements.
Author: Yipin Su
Language: English
Open Access status: diamond
OpenAlex ID: W7213307447
FWCI:
Cites: 15
Cited by:
Related to: 9