Long rigid fibres in a turbulent channel flow: comparison between experiments and simulations
The dynamics of long rigid fibres transported by turbulent channel flow are investigated experimentally and numerically. Experiments use polystyrene fibres of three lengths, $\ell/h=0.25$, $0.5$ and $1$, with moderate inertia, $St^+\approx20$. Their settling velocity is comparable to the friction velocity, causing accumulation near the bottom wall. Measurements are compared systematically with simulations based on a rigid slender-body model. Statistics conditioned on the distance from the wall are used to characterise the effects of fibre length and confinement on translation, orientation and tumbling. Away from the wall, the experimental fibres lag the fluid, with no clear dependence of the velocity deficit on length. Near the wall, the shortest fibres move faster than the local mean flow, whereas longer fibres remain slower, indicating length-dependent sampling of near-wall turbulence. Confinement also strongly constrains orientation and rotation. Fibres close to the wall predominantly undergo "kayaking" motion, tumbling in planes approximately parallel to it. Orientation statistics collapse when wall distance is normalised by fibre length, identifying $y^+/\ell^+$ as the relevant geometrical variable. Where fibres can acquire a significant wall-normal orientation, "pole-vaulting" events produce a local tumbling-rate maximum at $y^+\approx\ell^+/2$. Its magnitude decreases approximately as $(\ell^+)^{-2}$, consistently with a dimensional estimate based on the near-wall velocity variation sampled along the fibre. Experiments and simulations agree well for orientation and tumbling but differ more for translational velocity. The discrepancies highlight the effect of settling, finite fibre thickness and finite slip Reynolds number that are not fully represented by the model.