MODE-SHAPES CORRELATION AND CUSUM TEST FOR ON-LINE FLUTTER MONITORING

Investigations of the flutter monitoring problem, stated as statistical hypotheses testing problems regarding specified parameters (eigenfrequencies, damping coefficients, flutter margin), have resulted in on-line algorithms built on a residual associated with subspace identification and the Cusum test. Other investigations exploiting eigenvectors orthogonality have shown the relevance of such a criterion for coupled-mode flutter prediction. In this paper, the subspace-based Cusum test is extended to the handling of mode-shapes correlation, and the relevance of this new algorithm for flutter onset detection is investigated. Numerical results obtained on data simulated with a cantilever wing model are reported.