Impact of System Nonlinearities on Output-Based Whirl Flutter Prediction

This paper investigates the impact of system nonlinearities on output-based whirl flutter prediction. The investigation considers four widely-used damping identification and extrapolation methods-logarithmic decrement, exponential fit, matrix-pencil estimation, and moving-block analysis-alongside bifurcation forecasting, a recently proposed method. The examined methods are compared by predicting the whirl flutter speed of a propeller-nacelle system with structural nonlinearities using computational output data. The comparisons consider various hardening and softening behaviors, excitation amplitudes, and pre-flutter forward speeds, including values as low as 30% below the flutter point. Results indicate that all methods accurately predict the whirl flutter speed for weak nonlinearities and small-amplitude perturbations. While the first four methods markedly lose accuracy with stronger nonlinear effects, bifurcation forecasting maintains consistent accuracy across all examined conditions, even when significant nonlinearities are present.