Parameter identification of wake-oscillator from wind tunnel data

This paper proposes a procedure for the parameter identification of Tamura's wake-oscillator model. A multiple timescale analysis of the dimensionless model shows that the response is governed by two dimensionless groups D0 and D1, highlighting the importance of the forcing terms in the two governing equations, the total (aerodynamic and structural) damping and the coefficient ɛ of the fluid Van der Pol oscillator. In particular, this approach provides a simple closed form expression for the steady state amplitude of the structural displacement, which is usually measured in wind tunnel experiments. The proposed method of identification consists in fitting the parameters of the model by adjusting the closed-form expression of the VIV curve on experimental points. It is developed into two variants: a least-square fitting and a fitting based on some simple geometrical indicators (height, width, asymmetry). The model is sufficiently versatile to estimate the maximum amplitude and lock-in range. Applications of VIV in air for different geometries and Scruton numbers show that the two variants give equivalent results thanks to the robustness of the method. The paper is first intended for experimenters looking for a simple robust procedure to identify the parameters of the wake-oscillator, which can then be used in a prediction phase. The derivation of the slow phase version of Tamura's model might also be appealing to better understand the main features of this model.