A water flow loop facility is designed, built and commissioned to investigate the effects of flow unsteadiness on the mean and turbulent characteristics of smooth and rough wall channel flows. Measurements of flow and turbulence are made by means of non-intrusive measurement techniques such as Particle Image Velocimetry (PIV) and Laser Doppler Velocimetry (LDV). The main objective of these investigations is to advance our understanding on the behaviour of turbulence under transient conditions.
The unsteady flows considered, consist of an excursion of flow rate from an initial turbulent state to another over smooth or rough surfaces. A systematic study of turbulence under various initial and final conditions reveals novel insights into the turbulence dynamics of unsteady flows. It is shown that the unsteady flows behave strikingly similar to the so-called boundary layer bypass transition due to free-stream-turbulence. Consistent with the direct numerical simulations (DNS) of He and Seddighi (J. Fluid Mech., 715: 60-102), the process begins with the elongation of streaks much similar to the Klebanoff modes in the buffeted laminar boundary layer in a bypass transition. During the second stage, the formation and propagation of the isolated turbulent spots eventually lead to a complete breakdown of the organised streaky structures resulting in a new turbulent flow corresponding to the final Reynolds number. The present investigation covers a range of initial and final Reynolds numbers over smooth and rough surfaces to elucidate the underlying mechanisms involved in transient flows.
The mean velocity profiles obtained from various unsteady cases are shown to correlate with each other during the pre-transition regime, coinciding with the Stokes solution for unsteady laminar boundary layer flows. The fluctuating velocity profiles are also correlated with each other during this period. The response of the perturbing wall-normal fluctuating velocity is shown to mark the onset of transition, providing a good measure for the duration of the pre-transition phase.
It is shown that an equivalent critical Reynolds number can be defined to express the duration of the pre-transition regime in unsteady flows. The critical Reynolds number is shown to have a power-law relationship with the initial free-stream turbulence intensity levels.
The unsteady rough flows investigated herein encompass a hydrodynamically smooth initial flow that is increased to either a transitionally or fully rough final state. The measurements of the fluctuating streamwise velocities in the wall region reveal a similar transition-like behaviour to the smooth-wall flows. It is shown that the particular roughness pattern investigated herein causes a significant decrease in the duration of the pre-transition regime, promoting an early transition. For relatively high intensity levels, an inner-scaled non-dimensional time correlates the period of the pre-transition regime.
In addition to the experimental investigations, numerical simulations are performed to assess the applicability and robustness of various turbulence models under unsteady conditions. For this purpose, performance of a number of low-Reynolds number turbulence models is evaluated against DNS data. All models are applied to an unsteady flow comprising a ramp-type excursion of flow rate inside a channel. The flow rate is increased linearly with time from an initial Reynolds number of Re_0=9,308 (based on hydraulic diameter and bulk velocity) to a final Reynolds number of Re_1=29,650. The acceleration rate is varied to cover low, intermediate and high accelerations. It is shown that among the models investigated, the k−e models of Launder and Sharma (Lett. Heat Mass Transfer, 1(2), 131-137) and Chang et al. (J. Fluids Eng., Trans. ASME, 117(3): 417-423) and gamma-Re of Langtry and Menter (AIAA Journal, 47(12): 2894-2906) capture well the key flow features of these unsteady turbulent flows. These three models yield predictions of wall shear stress that agree well with the corresponding DNS data, though for the case of high acceleration, the model of Chang et al. exhibits instabilities.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)