Characterization of streak development for boundary layer transition caused by isolated and distributed roughness

Abstract

In this work, we characterize the early stages of transition in a flat-plate boundary layer caused by two types of surface roughness: distributed roughness (a strip of 24-grit emery) and isolated roughness (a cylindrical element). Towards this, we carry out extensive particle image velocimetry (PIV) measurements. The distributed roughness consists of a range of grit heights and introduces a broad range of disturbances downstream of it. In the pre-transitional region, both isolated and distributed roughness cause steady streaks, which for the latter case show a lack of spanwise symmetry. This is in contrast to the freestream turbulence (FST), which introduces unsteady streaks in the pre-transitional boundary layer. We propose tilting of spanwise vortices downstream of the distributed roughness to be the mechanism for generation of streamwise vortices, which are likely precursors to the onset of transition. For both the roughness configurations, the steady streaks develop instability resulting in localized streak breakdowns. The wall-normal PIV visualizations show streak instability features qualitatively similar to those for the FST-induced transition. For the distributed roughness, both the outer instability of lifted-up streaks and the inner instability due to streak interaction are present, whereas for the isolated roughness, the inner instability is dominant while the outer instability is much weaker. Conditional sampling of fluctuating velocity shows an asymmetry in the positive and negative fluctuations after the onset of transition, in a manner similar to that observed for the FST-induced transition. These observations suggest that the wall-normal distribution of unsteady velocity fluctuations in the early stages of transition is qualitatively similar irrespective of the source of disturbance (roughness/FST) causing transition. We expect that this commonality of features among different types of transition will be helpful in modelling flow over aerodynamic surfaces such as turbine blades and aircraft wings.

Delaying transition induced by a strip of distributed roughness using additional fine grit roughness

“…At x = 200 mm, the closest streamwise location to the roughness where measurements are taken, there is a peak in the spectrum at 158 Hz, corresponding to a Strouhal number (fk/U 𝑘 ) of 0.037. A similar value of Strouhal number (fk/U 𝑘 ) of 0.032 was reported in hotwire measurements conducted downstream of the same grit roughness at U ∞ =7.0 m/s in our results presented in Joseph et al [19]. We attributed the peak in the spectrum to vortices shed from the forward-backward step caused by the roughness strip consistent with previous studies such as Pinson and Wang [1], who conducted measurements downstream of a roughness strip placed at the leading edge, and Mushyam et al [30] who conducted numerical simulations downstream of a backward-facing step have reported Strouhal numbers of the same order.…”

“…Here, we observe the presence of streaks, which results in spanwise variation of streamwise velocity (𝜕𝑈 /𝜕𝑧). Note that the streaks are steady and this is demonstrated in [19].…”

“…The R24 roughness consists of roughness grits (grit size 24, diameter ≈ 0.7mm) pasted on a backing cloth with the total thickness of the roughness strip(measured using Vernier scale) ≈ 1.5 mm. A detailed characterization of the R24 roughness, conducted using a laser scanner is provided in Joseph et al [19]. Both R24 and R180 roughness spans the entire width of the flat plate and are pasted adjacent to each other without any gaps between them.…”

“…Note here that streamwise vortices were found to be 'optimal' disturbances in a flat plate boundary layer that results in maximum 'transient growth' ([34], [35]) and in a recent study on isolated roughness induced transition, [36] demonstrated that streamwise vortices play a key role in streak amplification via the lift-up effect ( [37]) and subsequent breakdown. Based on this, we expect that the mechanism of distributed roughness-induced transition involves streak instability (A more detailed discussion on the transition mechanism in distributed roughness is presented in [19]). Figure 9 shows the various expected steps in distributed roughness-induced transition, based on the discussion above.…”

“…Joseph and Diwan [18] investigated the growth of disturbances in a pre-transitional boundary layer downstream of a sandpaper roughness strip and contrasted them with the pre-transitional region of FST-induced transition. Finally, Joseph et al [19] used particle image velocimetry to visualize the mechanism of boundary layer transition caused by a sandpaper strip.…”