Adaptive Statistical Approach to Flutter Detection

One important issue to be handled online during flight testing is flutter monitoring, here addressed as a detection problem. From subspace detection algorithms proposed for vibration-based monitoring, several online flutter monitoring algorithms have been designed by the authors. They are based on a recursive version of the subspace-based residual and on an hypothesis test for detecting changes in a specific instability indicator with respect to a fixed-reference modal parameter (identified on a safe structure). However, the flutter onset time provided by those algorithms turns out to be too conservative. In this paper, a moving-reference approach is proposed to overcome that issue. Two adaptive flutter-monitoring algorithms are proposed that update the reference modal state during the online test. The usefulness of the proposed approach is discussed based on experimental results obtained on simulation data provided by two academic and industry-relevant simulated aircraft models.