Modeling and detection of limit cycle oscillations in thin-wing aircraft using adaptable linear models

A method for modeling the flutter response of a thin winged aircraft is presented in this dissertation. Flutter is a behavior of a dynamical system subjected to aeroelastic forces. Adaptive systems can model the flutter response, but require complicated Auto Regressive, Moving Average (ARMA) architectures to do so.; To simplify the task, this work proposes a methodology to separate the AR and MA components into a hybrid physical model consisting of a single-input multi-output (SIMO) adaptive AR model feeding into an adaptive multi-input single-output (MISO) adaptive MA model providing the necessary signal changes due to forces encountered by the wing during flight. Precise AR pole placement is accomplished by adaptive oscillators set at the modal frequencies identified by a free vibration analysis of the wing, thus representing a physical model of the structure.; Zero placement is then accomplished by feeding the oscillator outputs into a bank of adaptive linear finite impulse response (FIR) filters that are adapted to the flight test sensor data. Together these modules form a hybrid model for the wing flutter under flight conditions.; In order to cope with the nonstationary nature of the task, the output of the models is segmented using statistical methods based the Generalized Likelihood Ratio Test (GLRT) using maximum likelihood estimates of the error distributions. The overall sequence of models provides a complete synthesis of the nonlinear response recorded by the accelerometer as flight conditions change. The sequence of network parameters is analyzed for frequency, phase and gain of each mode.; The paradigm is used with two portions of flight test data. It performs very well against data taken at increasing Mach numbers in level flight. Network parameters are shown to correlate very well with the Mach numbers. The paradigm shows promising results when used against more complicated flight scenario with changes in normal accelerations.