16:15
Boundary Free Turbulence 2
16:15
15 mins
#58
|
Universality of enstrophy dynamics within the turbulent/non-turbulent interface layer
Marco Zecchetto, Carlos B. da Silva
Abstract: The universality of the small scale motions within the turbulent/non-turbulent interface (TNTI) layer, that exists at the edges of free shear flows, and boundary layers, and that separates the turbulent from the irrotational (or non-turbulent - NT) flow regions, is demonstrated by direct numerical simulations (DNS) of jets (JET), wakes (WAKE) and shear free turbulence (SFT).
Direct numerical simulations of jets (JET), wakes (WAKE) and shear free turbulence (SFT) are used to demonstrate the universality of the small scale motions inside the so called turbulent/non-turbulent interface (TNTI) layer, that exists at the edges of free shear flows, and also in boundary layers, which separates the turbulent from the irrotational (or non-turbulent - NT) flow regions.
The DNSs use a very accurate pseudo-spectral methods solver and attain Reynolds numbers of the order of Reynolds number based on Talyor micro-scale of 400, while keeping with extremely fine grids to allow the study of the TNTI layer.
Recently, it has been established that the TNTI layers contains two sublayers within itself: a viscous superlayer (VSL) where viscous effects dominate, and a turbulent sublayer (TSL) where inertial effects dominate, and that the thickness of these two sublayers (as well as the total TNTI layer) scales with the Kolmogorov micro-scale, for sufficiently high Reynolds numbers.
In the present work we go one point further and show that the small scale quantities governing the dynamics of the TNTI layer such as the enstrophy and its governing terms, display statistical universality, when normalised with the local mean Kolmogorov length, time and velocity scales.
This is illustrated in, among other things, the conditional mean profiles computed as function of the distance from the irrotational boundary (IB) which is an outer surface delimit ting the TNTI layer (in the NT side of the flow). The conditional mean profiles are carried out in a local coordinate system positioned at the IB (at the edge of the the TNTI layer), and the averaging is computed as function of the distance from the IB. The distance to the IB, is normalised by the Kolmogorov micro-scale from the turbulent core region of the flow. The enstrophy is zero in the NT region and roughly constant in the T core region.
Bridging the two regions we observe a sharp jump with a thickness close to 10 Kolmogorov micro-scales,
followed by a smooth rise associated with large scale inhomogeneities from the particular flow configuration. When the conditional mean is normalised by the (mean) local Kolmogorov micro-scale we see that the conditional mean profiles collapse for the three flow configurations studied here. Similar observations are made for other quantities and statistics.
In the present work we show how a new normalisation, based on the Kolmogorov time and length scales allows the emergence of universal statistics for the small scale quantities of the flow at particular and meaningful locations inside the VSL and TSL regions of the TNTI layer, thus attesting that the small scale motions within the TNTI layer are universal and independent from the particular flow configuration analysed.
|
16:30
15 mins
#268
|
Entrainment and self-similarity in negatively buoyant turbulent jets
Liam Milton-McGurk, Nicholas Williamson, Krishna Talluru, Steven Armfield, Michael Kirkpatrick
Abstract: Turbulent negatively buoyant jets occur when a jet's buoyancy opposes its initial momentum. For a vertically aligned negatively buoyant jet with initial momentum in the upward direction, the fluid will rise until the downward buoyancy forces overcome its momentum and its vertical velocity is reduced to zero. Here the fluid reverses direction and flows annularly back down towards to the source, forming a fountain. This study focuses on the initial `negatively buoyant jet' stage, prior to the return flow forming and a fountain developing. The flow is investigated experimentally using a mixture of freshwater and Ethanol entering a salt-water ambient in a 1m^3 tank, and measured using two-dimensional Particle Image Velocimetry (PIV) and Laser Induced Fluorescence (LIF). Negatively buoyant jets at a range of source Froude and Reynolds numbers (20< Fr<60 and 6,000< Re<,000) are measured at different streamwise distances from the inlet (10< z/D<100), and are compared to neutral jets in this Re range. This study will report on the self-similarity of negatively buoyant jets and the effect of negative buoyancy on entrainment. The present results show that entrainment is reduced in negatively buoyant jets when compared to neutral jets.
|
16:45
15 mins
#318
|
STATISTICS OF THE IRROTATIONAL FLOW REGION NEAR THE TURBULENT/NON-TURBULENT INTERFACE LAYER
Ricardo P. Xavier, Carlos B. da Silva, Miguel A. C. Teixeira
Abstract: Free shear flows such as wakes, jets and mixing layers (as well as boundary layers) are characterised by the coexistence of turbulent (T) and irrotational (or non-turbulent - NT) flow regions, separated by a sharp turbulent/non-turbulent interface (TNTI) layer[1]. Several small scale features of these layers have recently been shown to be universal[2]. However, it is well known that large-scale features of the TNTI and its governing features are flow dependent[3], such as entrainment
and spreading rates (e.g. planar jets grow faster than circular jets). In the present work we used new direct numerical simulations (DNS) of wakes, jets, and shear free turbulence (SFT - turbulence without mean shear) with Reynolds numbers of the order of up to Re λ ≈ 400, to assess the characteristics of the large scale motions of different flows near the TNTI
layer. It turns out that the effects of the different flow types in the dynamics of the TNTI, are better assessed by observing the flow statistics in the NT region, whereas in the T region, the intense small scale turbulence fluctuations tend to blur these differences. Indeed, important potential velocity fluctuations exist in the NT region, induced by the large scale vortices from the T region. Moreover, predictions from rapid distortion theory (RDT) allow to highlight these differences in the NT region.
Figure 1a shows a side view of the enstrophy in one of the SFT simulations, evidencing a sharp separation between T and NT flow regions. Figure 1b, displays conditional mean profiles of strain components S ij computed as function of the distance from the irrotational boundary (IB), which is an outer surface delimiting the TNTI layer (in the NT side of the flow). The conditional mean profiles are carried out in a local coordinate system positionedat the IB, and the averaging is computed as function of the distance from the IB. In the T region the strain exhibit = and <2*S12^2> = 3/2 relations, as in isotropic turbulence. However, these are not observed in the NT region due to inhomogeneities, i.e mean shear arising from the large scales, which is flow dependent. Figure 1c shows the predicted profiles for the same flow
obtained from RDT, showing a good qualitative comparison in the NT region.
The presentation will discuss how different flow features observed in the NT region can be linked to the different spreading and entrainment rates observed in different flow configurations.
|
17:00
15 mins
#347
|
Velocity and scalar structure near the turbulent/non-turbulent interface compared to internal turbulence
Gerrit Elsinga, Carlos da Silva
Abstract: The average patterns of the velocity and scalar fields near turbulent/non-turbulent interfaces (TNTI) are assessed in the strain eigenframe and results are compared to internal turbulence. These flow patterns were obtained from direct numerical simulations (DNS) of shear free turbulence (Reλ = 86) and help to clarify many aspects of the flow dynamics, including a passive scalar (Sc = 0.7), near a TNTI layer. The averaged flow field near the TNTI layer exhibits a saddle-node flow topology associated with a vortex in one half of the interface, while the other half of the interface consists of a shear layer (figure 1-left). The flow pattern near the TNTI is thus very different from the shear-layer structure consisting of two aligned vortical motions bounded by two large-scale regions of uniform flow, that typically characterizes the average strain field in the fully developed turbulent regions (figure 1-right). This finding contrasts earlier suggestions that the TNTI may be similar to (internal) shear layers [1].
The particular flow pattern observed near the TNTI layer has important consequences for the dynamics of a passive scalar field. The saddle-node topology explains the preferential alignment between the scalar gradient and the most compressive strain (figure 2), which is why regions of particularly high scalar gradient (magnitude) are typically found at TNTIs separating fluid with different levels of scalar concentration. Within the fully developed internal turbulent region, the scalar gradient exhibits an angle with the most compressive straining direction with a peak probability at around 20 degrees (figure 2) [2], thus the two vectors are not preferentially aligned, as has been considered for many years, which has important implications for the amplification rate of scalar gradients in turbulent flows.
|
17:15
15 mins
#542
|
PARTICLE ENTRAINMENT THROUGH A TURBULENT/NON-TURBULENT INTERFACE.
Tai Tai Wada, Christos Vassilicos
Abstract: Turbulent entrainment is an important process closely linked to the dynamics of the Turbulent/Non-Turbulent Interface (TNTI) [1]. In this work we study entrainment in terms of fluid element trajectories at the vicinity of the TNTI which may or may not be crossing it. We run Direct Numerical Simulations (DNS) of turbulent planar jets at inlet Reynolds number Re = 4000 and integrate fluid element as well as particle trajectories in the DNS velocity fields with and without gravity. Our results reveal the existence of both detrainment and entrainment events across the TNTI. There are also many fluid elements which stay in the vicinity of the TNTI for significantly long times. Depending on Stokes and Froude numbers, this can also be the case for inertial particles. This barrier effect of the TNTI causes its own clustering and preferential concentration effects.
|
17:30
15 mins
#558
|
The statistical topology of a turbulent--non-turbulent interface
John Craske, Jean-Paul Mollicone, Vishnu Satheesh Kumar Nair, Maarten van Reeuwijk
Abstract: The process of entrainment across a turbulent--non-turbulent interface (TNTI) underpins the behaviour of a wide range of flows in nature and industry such as jets, plumes, gravity currents, wakes and boundary layers. Over a period of approximately 70 years, work in this field ranges from integral parameterisations [6,7] to investigations of the TNTI at the microscale [2]. From the perspective of the latter, the geometrical properties of the TNTI, such as area and fractal scaling [5], play a key role in understanding how entrainment is affected by stratification [3]. More broadly, the TNTI's geometry provides an opportunity for unification and identification of the ways that entrainment may or may not depend on context.
In comparison with the geometrical insights mentioned above, the topological structure of the TNTI is unexplored. Yet, in describing how a space is connected, topology captures more general aspects of geometry than those associated with length, area and volume. To investigate the topology of a TNTI we compute
spatially-averaged Betti numbers associated with thresholds on enstrophy from direct simulations of a temporal jet at Re=1000 (see figure 1) using CHomP [1,4]. The first three Betti numbers count the number of connected components, circular holes and cavities of a space, respectively, and can be combined to describe a space's structure as an Euler characteristic. We investigate how the scaling of the Euler characteristic depends on time and the threshold used on enstrophy.
|
17:45
15 mins
#577
|
SCALE-BY-SCALE ANALYSIS OF A TURBULENT TEMPORAL JET
Elisabetta De Angelis, Andrea Cimarelli, Jean-Paul Mollicone, Maarten van Reeuwijk, Thorsten Stoesser
Abstract: The most relevant feature of turbulent free flows is given by the fact that the phenomenon of turbulence is driven not solely by processes embedded in the turbulent core itself but also by interaction phenomena with the surrounding laminar flow region. Both aspects are strongly scale and position dependent. In particular, the turbulent core is characterized by large scale production and small scale dissipation. These processes are strongly position dependent so that in general do not balance each other locally leaving space for spatially evolving turbulent transport phenomena compensating the resulting inhomogeneity. On the other hand, also the turbulent/non-turbulent interactions are characterized phenomena occurring at large engulfing and small scale nibbling again strongly dependent on the position with respect to the putative turbulent/non-turbulent interface. The dual nature, in the space of scales and positions of these processes, challenges for a rational approach able to give a clear understanding of the flow physics. To overcome this issue, in the present work we propose the use of a theoretical framework based on exact equations for the second order moments of two-point velocity and temperature increments, the so-called generalized Kolmogorov and Yaglom equations [1, 2, 3]. The flow case considered is a temporal jet with temperature field [4] which represents the simplest paradigm for the study of free flow with simple shear, see figure 1. Starting from basic single-point statistics for the characterization of the main flow properties, we will present the analysis of the Kolmogorov and Yaglom equations focusing mainly on the production, cascade and dissipation processes of the turbulent core region and on the nature (viscous or inertial) of the transport processes in the interfacial layer.
|
|