PSI - Issue 24

Pierluigi Fanelli et al. / Procedia Structural Integrity 24 (2019) 939–948 Author name / Structural Integrity Procedia 00 (2019) 000–000

945 7

For the air flow, instead, a 1 . 2 kg / m 3 density ρ By referring to the dimensional wind velocity ( Υ ) as:

5 m 2 / s kinematic viscosity ν

f and a 1 . 510 −

f have been assumed.

2 L 3

ρ f U

(19)

Υ =

R

with the aim of comparing the numerical and the literature results, it has been chosen to work with a non-dimensional wind velocity of 1 . 5, which implies a flow velocity U of 3 m / s. The wind tunnel (figure (2)) in which experimental tests have been taken place has a 1 . 2 m length ( L ), 1 . 2 m heigth ( H ) and a 1 . 2 m width ( W ).

U

H

h

l

L

Fig. 2. FSI case system

In reference to the initial conditions of the simulation, it has been imposed a zero movingWallVeocity value for the fluid patch with the aim of reproducing the analysed case. It has been set a fixedVelocity to the inlet, while the zeroGradient velocity condition has been chosen for the outlet. Concerning pressure initial conditions, it has been applied a null value to the outlet and zeroGradient condition for all the other patches. Since the problem under investigation is bidimensional, the empty condition is applied to the frontAndBack faces, in order to avoid the solution computation along the z direction. In order to evaluate simulation results, displacement values of the lamina free edge (last Control Volume node) have been extrapolated, with the aim of comparing vibrations trends and values. The first 75 ms of the experimental test have been simulated, because of the high computational burden of the process. In this time range, the lamina free edge, in according to the experimental reference values, oscillates around an equilibrium position by following a sinusoidal trend. Vibration peaks are in the order of 10 − 9 m , which are comparable with the reference values (figure (3)).