TY - GEN
T1 - Nonlinear system identification of a pitching wing in a surging flow
AU - De Troyer, Tim
AU - Csurcsia, Péter Zoltán
AU - Greenblatt, David
PY - 2020/1/1
Y1 - 2020/1/1
N2 - Many engineering applications involve lifting surfaces (wings) that oscillate about in pitch, heave, and/or stream-wise (referred to as surge). For small pitch amplitudes, under the stall angle of attack, the lift force will respond linearly to such oscillations. However, at sufficiently large amplitudes, unsteady flow separation and dynamic stall set in, and the lift force becomes a nonlinear function. Data-driven models are typically linear, and therefore, they often fail to capture the nonlinear response related to unsteady separation. In this work, we use wind tunnel experiments of a wing under harmonic pitching and surging such that it oscillates in and out of the nonlinear regime. We then use nonlinear system identification techniques to 1) identify the best linear approximation (BLA) between output (lift) and input (angle of attack, wind speed), 2) compare the BLA estimates with the classical power spectra estimates, 3) study the (in)effectiveness of (non)linear state space models in the linear and nonlinear flow regimes, 4) evaluate the possibility of interpolating between different measurements.
AB - Many engineering applications involve lifting surfaces (wings) that oscillate about in pitch, heave, and/or stream-wise (referred to as surge). For small pitch amplitudes, under the stall angle of attack, the lift force will respond linearly to such oscillations. However, at sufficiently large amplitudes, unsteady flow separation and dynamic stall set in, and the lift force becomes a nonlinear function. Data-driven models are typically linear, and therefore, they often fail to capture the nonlinear response related to unsteady separation. In this work, we use wind tunnel experiments of a wing under harmonic pitching and surging such that it oscillates in and out of the nonlinear regime. We then use nonlinear system identification techniques to 1) identify the best linear approximation (BLA) between output (lift) and input (angle of attack, wind speed), 2) compare the BLA estimates with the classical power spectra estimates, 3) study the (in)effectiveness of (non)linear state space models in the linear and nonlinear flow regimes, 4) evaluate the possibility of interpolating between different measurements.
UR - http://www.scopus.com/inward/record.url?scp=85105796498&partnerID=8YFLogxK
M3 - Conference paper
T3 - Proceedings of ISMA 2020 - International Conference on Noise and Vibration Engineering and USD 2020 - International Conference on Uncertainty in Structural Dynamics
SP - 939
EP - 954
BT - Proceedings of the International Conference on Noise and Vibration Engineering (ISMA)
A2 - Desmet, W.
A2 - Pluymers, B.
A2 - Moens, D.
A2 - Vandemaele, S.
ER -