An envelope method for detection of turbulence intermittency in a transitional boundary layer

Chew, Y. T.; Shah, D. A.; Wan, Jin

doi:10.1016/s0169-5983(98)00012-4

Cited by 5 publications

(1 citation statement)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A2 (h)). This can present difficulties in removing false zeros and ones, which again is a well-recognized fact in the literature [13,43,[72][73][74].…”

Section: Discussionmentioning

confidence: 96%

Instability mechanisms and intermittency distribution in adverse pressure gradient attached and separated boundary layers

2021

View full text Add to dashboard Cite

Direct Numerical Simulation of a boundary layer subjected to an adverse pressure gradient has been carried out to investigate the instability mechanisms and transition-zone characteristics, as the boundary layer changes its character from an attached to a separated flow. The freestream turbulence intensity is fixed at a moderately low value of 0.3% and the inlet Reynolds number (Re) is reduced to obtain one attached and two separated flow (small and large separation) cases. A detailed characterisation of the pre-transitional boundary layer has been carried out that reveals a "mixed-mode" instability, involving contribution from instability waves associated with inflectional instability and streamwise streaks generated by the lift-up effect. A unified picture is presented of the changes in the relative significance of these two modes, as Re is varied, leading up to the transition onset marked by a breakdown of spanwise vortical rollers for all the three cases. Next, we carry out a time-frequency analysis of the transitional velocity signals and show that as Re decreases, the character of the time traces evolves continuously from a "spotty" behaviour (i.e., exhibiting distinct turbulent spots) for the attached case to a "non-spotty" behaviour (i.e., involving a more "uniform" distribution of turbulent fluctuations in time) for the large separation case, encompassing the entire spectrum of transition scenarios. The intermittency factor (𝛾𝛾) has been calculated within the transition zone and its variation is seen to compare well with the "universal" 𝛾𝛾-distribution of Dhawan and Narasimha (J. Fluid Mech., Vol.3, pp.418, 1958) for all the three cases. One of the basic premises of the universal 𝛾𝛾-distribution is the concentrated breakdown hypothesis (requiring generation of spots randomly in time and span) and we investigate whether the transition 2 scenario in the large separation case conforms to this hypothesis. We find that although the time variation of velocity for large separation is non-spotty (or more "uniform"), the spanwise variation of velocity is "spotty" (or "random") in character showing a clear clustering of highwavenumber fluctuations separated by quasi-laminar regions. This suggests that the concentrated breakdown hypothesis is "partially" satisfied, which might explain why the 𝛾𝛾distribution compares reasonably well with the universal distribution for this case. On the other hand, for the attached-flow case, the velocity signal is "spotty" in both time and span, conforming more closely to the premise of this hypothesis. Finally, we present a physical cartoon for the transition scenarios for the attached and separated cases, using the ideas of vortex-wall interaction and the instability of spanwise vortical structures.

show abstract

“…A2 (h)). This can present difficulties in removing false zeros and ones, which again is a well-recognized fact in the literature [13,43,[72][73][74].…”

Section: Discussionmentioning

confidence: 96%