Multioutput Autoregressive Methods for Aeroelastic System Identification and Flutter Prediction

Tomer Ben Asher
Work towards MSc degree under the supervision of Professor Daniella Raveh (Technion)
Department of Aerospace Engineering
Technion – Israel Institute of Technology

Flutter is an aeroelastic instability phenomenon that might lead to structural failure and catastrophic results. As such, every new aircraft design is analyzed and flight tested to assure that its operational envelope is flutter-free. In flutter tests, structural response data recorded at pre-flutter, safe airspeed are used to assess the aeroelastic dynamic properties and predict the flutter onset. Accurate estimates of the aeroelastic characteristics and flutter onset are vital for safe flutter tests and for accurate determination of a vehicle’s operational envelope.

The literature offers different approaches to experimental flutter prediction. These methods typically rely on the identification of the variation of the aeroelastic system’s dynamic parameters (namely, the aeroelastic frequencies and damping values) with airspeed, and an extrapolation  of a stability parameter to predict the flutter onset. Several forms of stability parameter were derived that (mainly) consider the frequency approaching of two modes in two-mode flutter mechanisms. Such stability parameters decay monotonically with airspeed, thus providing a good indication of the onset conditions.

The current study proposes improvements to the experimental system identification of aeroelastic configurations, by concurrently accounting for multiple data available in tests, such as accelerations and strain data from optical fibers. We demonstrate how creating a Vector Autoregressive (VAR) and Vector Autoregressive Moving Average (VARMA) models from multiple data yields a smooth variation of the system’s frequencies and damping with airspeed, leading to accurate flutter onset prediction. The study examines in detail the applicative aspects of aeroelastic system identification, such as model order selection, mode tracking with airspeed, and the identification of the flutter modes. The method is demonstrated with two experimental wind-tunnel test cases. The first is a simple rectangular wing. The second is a half fuselage model of the Active Aeroelastic Aircraft Testbed (A3TB) flying wing configuration.

The talk will be given in Hebrew

Mon, 25-07-2022, 13:30-14:30 (Gathering at 13:15)

Classroom 165, ground floor, Library, Aerospace Eng. &

Light refreshments will be served before the lecture at the classroom

Multioutput Autoregressive Methods for Aeroelastic System Identification and Flutter Prediction