Stochastic approach to flutter data processing and application to real F-18 A/B flight test data

Sami Abou-Kebeh LLano, Senior Flight Test Engineer, CLAEX, Spain
Roberto Gil Pita, Senior Lecturer, Alcala de Henares Univ., Spain
Manuel Rosa Zurera, Universitary Professor, Alcala de Henares Univ., Spain

Abstract

Flutter data processing has been historically considered as a very hard problem to solve 1. Most of the techniques used for processing follow a deterministic approach and may show different results between each other. That discrepancy was blamed, among others, on non-linearities, damping showing a behavior different to the model employed, and several other causes.

In this paper, a different approach to the traditional techniques will be tested and validated. Instead of following a deterministic model, a statistical data processing method will be used. The frequency and damping of the natural modes of an aeroelastic structure will be considered to be a random variable, assuming the population of frequency to be a part of a Cauchy distribution instead of a Gaussian distribution.

This paper will present a new aeroelastic data processing technique, show a mathematical relationship between the traditional model for flutter (viscous damping, as a sinusoidal damped by an exponential) and a Cauchy probability distribution, present a comparison between different filtering techniques and afterwards, validate those methods, when applied to real data, from flutter tests on an F-18 A/B of the Spanish A.F.