Table of contents

Editor: Javier Jiménez

PREFACE

This third Summer School on Turbulence at the School of Aeronautics of the Universidad Politécnica de Madrid follows two previous ones, in 2013 and 2015. Those previous workshops were funded by the European Research Council Multiflow grant, whose goal was to improve our understanding of the multiscale dynamics of turbulence in fluids. After the initial grant ran its course, funding for the Summer School was transferred to the ERC grant Coturb, directed towards an understanding of turbulence from a structural point of view. The third Madrid Summer School took place from May 29 to June 30, 2017. As for the first two, its goal was to provide a meeting place for theoreticians, experimentalists and simulators, in which to develop and test new ideas on turbulence physics and structure. Around forty, mostly young, participants from twenty one international groups met for five weeks of collaborative work, primarily using the computational data archived in the receiving institution but, in many cases, also contributing their own. Although the format included some invited formal seminars and periodic plenary meetings, most of the work took place in small groups whose composition occasionally changed during the workshop. These proceedings reflect the results of the work of these groups, which, in many cases, later continued in the form of new collaborations.

List of organiser are available in this PDF.

All papers published in this volume of Journal of Physics: Conference Series have been peer reviewed through processes administered by the proceedings Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.

Mean Reynolds stress profiles and instantaneous Reynolds stress structures are investigated in a self-similar adverse pressure gradient turbulent boundary layer (APG-TBL) at the verge of separation using data from direct numerical simulations. The use of a self-similar APG-TBL provides a flow domain in which the flow gradually approaches a constant non-dimensional pressure gradient, resulting in a flow in which the relative contribution of each term in the governing equations is independent of streamwise position over a domain larger than two boundary layer thickness. This allows the flow structures to undergo a development that is less dependent on the upstream flow history when compared to more rapidly decelerated boundary layers. This APG-TBL maintains an almost constant shape factor of H = 2.3 to 2.35 over a momentum thickness based Reynolds number range of Re_δ2 = 8420 to 12400. In the APG-TBL the production of turbulent kinetic energy is still mostly due to the correlation of streamwise and wall-normal fluctuations, 〈uv〉, however the contribution form the other components of the Reynolds stress tensor are no longer negligible. Statistical properties associated with the scale and location of sweeps and ejections in this APG-TBL are compared with those of a zero pressure gradient turbulent boundary layer developing from the same inlet profile, resulting in momentum thickness based range of Re_δ2 = 3400 to 3770. In the APG-TBL the peak in both the mean Reynolds stress and the production of turbulent kinetic energy move from the near wall region out to a point consistent with the displacement thickness height. This is associated with a narrower distribution of the Reynolds stress and a 1.6 times higher relative number of wall-detached negative uv structures. These structures occupy 5 times less of the boundary layer volume and show a similar reduction in their streamwise extent with respect to the boundary layer thickness. A significantly lower percentage of wall-attached structures is observed in the present case when compared with a similar investigation of a rapidly decelerating APG-TBL, suggesting that these wall-attached features could be the remanent from the lower pressure gradient domain upstream.

A study of the back flow events and critical points in the flow through a toroidal pipe at friction Reynolds number Re_τ ≈ 650 is performed and compared with the results in a turbulent channel flow at Re_τ ≈ 934. The statistics and topological properties of the back flow events are analysed and discussed. Conditionally-averaged flow fields in the vicinity of the back flow event are obtained, and the results for the torus show a similar streamwise wall-shear stress topology which varies considerably for the spanwise wall-shear stress when compared to the channel flow. The comparison between the toroidal pipe and channel flows also shows fewer back flow events and critical points in the torus. This cannot be solely attributed to differences in Reynolds number, but is a clear effect of the secondary flow present in the toroidal pipe. A possible mechanism is the effect of the secondary flow present in the torus, which convects momentum from the inner to the outer bend through the core of the pipe, and back from the outer to the inner bend through the pipe walls. In the region around the critical points, the skin-friction streamlines and vorticity lines exhibit similar flow characteristics with a node and saddle pair for both flows. These results indicate that back flow events and critical points are genuine features of wall-bounded turbulence, and are not artifacts of specific boundary or inflow conditions in simulations and/or measurement uncertainties in experiments.

The effect of different definitions of the momentum flux on the properties of the coherent structures of the logarithmic region of wall-bounded turbulence is investigated by comparing the structures of intense tangential Reynolds stress with those of the alternative flux proposed in [Jimenez (2016) J. Fluid Mech. 809:585]. Despite the fairly different statistical properties of the two flux definitions, it is found that their intense structures show many similarities, such as the dominance of 'wall-attached' objects, and geometric self-similarity. However, the new structures are wider, although not taller, than the classical ones, and include both high- and low-momentum regions within the same object. It is concluded that they represent the same phenomenon as the classical group of a sweep, an ejection, and a roller, which should thus be considered as the fundamental coherent structure of the momentum flux. The present results suggest that the properties of these momentum structures are robust with respect to the definition of the fluxes.

The present work studies the fluctuating characteristics of the streamwise wall-shear stress in a DNS of a turbulent boundary layer at Re_τ=1500 from a structural view. The two-dimensional field of the fluctuating friction velocity u'_τ(x,z) is decomposed into the large- and small-scale components via a recently proposed scale separation algorithm, Quasi-bivariate Variational Mode Decomposition (QB-VMD). Both components are found to be dominated by streak-like structures, which can be regarded as the wall signature of the inner-layer streaks and the outer-layer LSMs, respectively. Extreme positive/negative wall-shear stress fluctuation events are detected in the large-scale component. The former's occurrence frequency is nearly one order of magnitude higher than the latter; therefore, they contribute a significant portion of the long tail of the wall-shear stress distribution. Both two-point correlations and conditional averages show that these extreme positive wall-shear stress events are embedded in the large-scale positive u'_τ streaks. They seem to be formed by near-wall 'splatting' process, which are related to strong finger-like sweeping (Q4) events originated from the outer-layer positive LSMs.

Drag reduction induced by polymer additives in wall-bounded turbulence has been studied for decades. A small dosage of polymer additives can drastically reduce the energy dissipation in turbulent flows and alter the flow structures at the same time. As the polymer-induced fluid elasticity increases, drag reduction goes through several stages of transition with drastically different flow statistics. While much attention in the area of polymer-turbulence interactions has been focused on the onset and the asymptotic stage of maximum drag reduction, the transition between the two intermediate stages – low-extent drag reduction (LDR) and high-extent drag reduction (HDR) – likely reflects a qualitative change in the underlying vortex dynamics according to our recent study [1]. In particular, we proposed that polymers start to suppress the lift-up and bursting of vortices at HDR, leading to the localization of turbulent structures. To test our hypothesis, a statistically robust conditional sampling algorithm, based on Jenong and Hussain [2]'s work, was adopted in this study. The comparison of conditional eddies between the Newtonian and the highly elastic turbulence shows that (i) the lifting "strength" of vortices is suppressed by polymers as reflected by the decreasing lifting angle of the conditional eddy and (ii) the curvature of vortices is also eliminated as the orientation of the head of the conditional eddy changes. In summary, the results of conditional sampling support our hypothesis of polymer-turbulence interactions during the LDR-HDR transition.

The estimation problem for a fully-developed turbulent channel flow at Re_τ = 2000 is considered. Specifically, a Kalman filter is designed using a Navier–Stokes-based linear model. The estimator uses time-resolved velocity measurements at a single wall-normal location (provided by DNS) to estimate the time-resolved velocity field at other wall-normal locations. The estimator is able to reproduce the largest scales with reasonable accuracy for a range of wavenumber pairs, measurement locations and estimation locations. Importantly, the linear model is also able to predict with reasonable accuracy the performance that will be achieved by the estimator when applied to the DNS. A more practical estimation scheme using the shear stress at the wall as measurement is also considered. The estimator is still able to estimate the largest scales with reasonable accuracy, although the estimator's performance is reduced.

During the last decades, a number of reduced order models based on coherent structures have been proposed to describe wall-bounded turbulence. Many of these models emphasize the importance of coherent wall-normal velocity eddies (ν-eddies), which drive the generation of the very long streamwise velocity structures observed in the logarithmic and outer region. In order to use these models to improve our ability to control wall-bounded turbulence in realistic applications, these ν-eddies need to be identified from the wall in a non-intrusive way. In this paper, the possibility of using the pressure signal at the wall to identify these ν-eddies is explored, analyzing the cross-correlation between the wall-normal velocity component and the pressure fluctuations at the wall in a DNS of a turbulent channel flow at Re_τ = 939. The results show that the cross-correlation has a region of negative correlation upstream, and a region of positive correlation backwards. In the spanwise direction the correlation decays monotonously, except very close to the wall where a change of sign of the correlation coefficient is observed. Moreover, filtering the pressure fluctuations at the wall in space results in an increase of the region where the cross-correlation is strong, both for the positively and the negatively correlated regions. The use of a time filter for the pressure fluctuations at the wall yields different results, displacing the regions of strong correlation without changing much their sizes. The results suggest that space-filtering the pressure at the wall is a feasible way to identify ν-eddies of different sizes, which could be used to trigger turbulent control strategies.

This work is aimed at a first characterization of coherent structures in turbulent square duct flows. Coherent structures are defined as connected components in the domain identified as places where a quantity of interest (such as Reynolds stress or vorticity) is larger than a prescribed non-uniform threshold. Firstly, we qualitatively discuss how a percolation analysis can be used to assess the effectiveness of the threshold function, and how it can be affected by statistical uncertainty. Secondly, various physical quantities that are expected to play an important role in the dynamics of the secondary flow of Prandtl's second kind are studied. Furthermore, a characterization of intense Reynolds-stress events in square duct flow, together with a comparison of their shape for analogous events in channel flow at the same Reynolds number, is presented.

We carry out direct numerical simulations (DNS) of square duct flow spanning the friction Reynolds number range ${Re}_{\tau }^{* }=150-1055$, to study the nature and the role of secondary motions. We preliminarily find that secondary motions are not the mere result of the time averaging procedure, but rather they are present in the instantaneous flow realizations, corresponding to large eddies persistent in both space and time. Numerical experiments have also been carried out whereby the secondary motions are suppressed, hence allowing to quantifying their effect on the mean flow field. At sufficiently high Reynolds number, secondary motions are found to increase the friction coefficient by about 3%, hence proportionally to their relative strength with respect to the bulk flow. Simulations without secondary motions are found to yield larger deviations on the mean velocity profiles from the standard law-of-the-wall, revealing that secondary motions act as a self-regulating mechanism of turbulence whereby the effect of the corners is mitigated.

A Dynamic Mode Decomposition (DMD) of Direct Numerical Simulations (DNS) of fully developed channel flows is undertaken in order to study the main differences in flow features between a plane-channel flow and a passively "controlled" flow wherein the mean friction was reduced relative to the baseline by modifying the geometry in order to generate a streamwise-periodic spanwise pressure gradient, as is the case for an oblique wavy wall. The present analysis reports POD and DMD modes for the plane channel, jointly with the application of a sparsity-promoting method, as well as a reconstruction of the Reynolds shear stress with the dynamic modes. Additionally, a dynamic link between the streamwise velocity fluctuations and the friction on the wall is sought by means of a composite approach both in the plane and wavy cases. One of the DMD modes associated with the wavy-wall friction exhibits a meandering motion which was hardly identifiable on the instantaneous friction fluctuations.

We study the effect on near-wall turbulence of tangential slip and wall-normal transpiration, typically produced by textured surfaces and other surface manipulations. For this, we conduct direct numerical simulations (DNSs) with different virtual origins for the different velocity components. The different origins result in a relative wall-normal displacement of the near-wall, quasi-streamwise vortices with respect to the mean flow, which in turn produces a change in drag. The objective of this work is to extend the existing understanding on how these virtual origins affect the flow. In the literature, the virtual origins for the tangential velocities are typically characterised by slip boundary conditions, while the wall-normal velocity is assumed to be zero at the boundary plane. Here we explore different techniques to define and implement the three virtual origins, with special emphasis on the wall-normal one. We investigate impedance conditions relating the wall-normal velocity to the pressure, and linear relations between the velocity components and their wall-normal gradients, as is typically done to impose slip conditions. These models are first tested to represent a smooth wall below the boundary plane, with all virtual origins equal, and later for different tangential and wall-normal origins. Our results confirm that the change in drag is determined by the offset between the origins perceived by mean flow and the quasi-streamwise vortices or, more generally, the near-wall turbulent cycle. The origin for the latter, however, is not set by the spanwise virtual origin alone, as previously proposed, but by a combination of the spanwise and wall-normal origins, and mainly determined by the shallowest of the two. These observations allow us to extend the existing expression to predict the change in drag, accounting for the wall-normal effect when the transpiration is not negligible.

Turbulent flows over sparse and dense canopies exerting a similar drag force on the flow are investigated using Direct Numerical Simulations. The dense canopies are modelled using a homogeneous drag force, while for the sparse canopy, the geometry of the canopy elements is represented. It is found that on using the friction velocity based on the local shear at each height, the streamwise velocity fluctuations and the Reynolds stress within the sparse canopy are similar to those from a comparable smooth-wall case. In addition, when scaled with the local friction velocity, the intensity of the off-wall peak in the streamwise vorticity for sparse canopies also recovers a value similar to a smooth-wall. This indicates that the sparse canopy does not significantly disturb the near-wall turbulence cycle, but causes its rescaling to an intensity consistent with a lower friction velocity within the canopy. In comparison, the dense canopy is found to have a higher damping effect on the turbulent fluctuations. For the case of the sparse canopy, a peak in the spectral energy density of the wall-normal velocity, and Reynolds stress is observed, which may indicate the formation of Kelvin-Helmholtz-like instabilities. It is also found that a sparse canopy is better modelled by a homogeneous drag applied on the mean flow alone, and not the turbulent fluctuations.

Despite the large amount of information provided by direct numerical simulations of turbulent flows, their underlying dynamics remain elusive even in the most simple and canonical configurations. Most common approaches to investigate the turbulence phenomena do not provide a clear causal inference between events, which is essential to determine the dynamics of self-sustaining processes. In the present work, we examine the causal interactions between streaks, rolls and mean shear in the logarithmic layer of a minimal turbulent channel flow. Causality between structures is assessed in a non-intrusive manner by transfer entropy, i.e., how much the uncertainty of one structure is reduced by knowing the past states of the others. We choose to represent streaks by the first Fourier modes of the streamwise velocity, while rolls are defined by the wall-normal and spanwise velocity modes. The results show that the process is mainly unidirectional rather than cyclic, and that the log-layer motions are sustained by extracting energy from the mean shear which controls the dynamics and time-scales. The well-known lift-up effect is also identified, but shown to be of secondary importance in the causal network between shear, streaks and rolls.

Turbulence in wall-bounded shear flow results from a synergistic interaction between linear non-normality and nonlinearity in which non-normal growth of a subset of perturbations configured to transfer energy from the externally forced component of the turbulent state to the perturbation component maintains the perturbation energy, while the subset of energy-transferring perturbations is replenished by nonlinearity. Although it is accepted that both linear non-normality mediated energy transfer from the forced component of the mean flow and nonlinear interactions among perturbations are required to maintain the turbulent state, the detailed physical mechanism by which these processes interact in maintaining turbulence has not been determined. In this work a statistical state dynamics based analysis is performed on turbulent Couette flow at R = 600 and a comparison to DNS is used to demonstrate that the perturbation component in Couette flow turbulence is replenished by a non-normality mediated parametric growth process in which the fluctuating streamwise mean flow has been adjusted to marginal Lyapunov stability. It is further shown that the alternative mechanism in which the subspace of non-normally growing perturbations is maintained directly by perturbation-perturbation nonlinearity does not contribute to maintaining the turbulent state. This work identifies parametric interaction between the fluctuating streamwise mean flow and the streamwise varying perturbations to be the mechanism of the nonlinear interaction maintaining the perturbation component of the turbulent state, and identifies the associated Lyapunov vectors with positive energetics as the structures of the perturbation subspace supporting the turbulence.

A mathematical and computational framework is developed for the detection and identification of coherent structures in turbulent wall-bounded shear flows. In a first step, this data-based technique will use an embedding methodology to formulate the fluid motion as a phase-space trajectory, from which state-transition probabilities can be computed. Within this formalism, a second step then applies repeated clustering and graph-community techniques to determine a hierarchy of coherent structures ranked by their persistencies. This latter information will be used to detect highly transitory states that act as precursors to violent and intermittent events in turbulent fluid motion (e.g., bursts). Used as an analysis tool, this technique allows the objective identification of intermittent (but important) events in turbulent fluid motion; however, it also lays the foundation for advanced control strategies for their manipulation. The techniques are applied to low-dimensional model equations for turbulent transport, such as the self-sustaining process (SSP), for varying levels of complexity.

We present results from large-eddy simulations (LES) of turbulent pipe flow in a computational domain of 42 radii in length. Wide ranges of shear the Reynolds number and Smagorinsky model parameter are covered, 180 ≤ Re_τ ≤ 1500 and 0.05 ≤ C_s ≤ 1.2, respectively. The aim is to asses the effect of C_s on the resolved flow field and turbulence statistics as well as to test whether very large scale motions (VLSM) in pipe flow can be isolated from the near-wall cycle by enhancing the dissipative character of the static Smagorinsky model with elevated C_s values. We found that the optimal C_s to achieve best agreement with reference data varies with Re_τ and further depends on the wall normal location and the quantity of interest. Furthermore, for increasing Re_τ, the optimal C_s for pipe flow LES seems to approach the theoretically optimal value for LES of isotropic turbulence. In agreement with previous studies, we found that for increasing C_s small-scale streaks in simple flow field visualisations are gradually quenched and replaced by much larger smooth streaks. Our analysis of low-order turbulence statistics suggests, that these structures originate from an effective reduction of the Reynolds number and thus represent modified low-Reynolds number near-wall streaks rather than VLSM. We argue that overdamped LES with the static Smagorinsky model cannot be used to unambiguously determine the origin and the dynamics of VLSM in pipe flow. The approach might be salvaged by e.g. using more sophisticated LES models accounting for energy flux towards large scales or explicit anisotropic filter kernels.

Large eddy simulations (LES) of Taylor-Couette (TC) flow, the flow between two co-axial and independently rotating cylinders are performed in an attempt to explore the large-scale axially-pinned structures seen in experiments and simulations. Both static and dynamic LES models are used. The Reynolds number is kept fixed at Re = 3.4 · 10⁴, and the radius ratio η = r_i/r_o is set to η = 0.909, limiting the effects of curvature and resulting in frictional Reynolds numbers of around Re_τ ≈ 500. Four rotation ratios from Rot = −0.0909 to Rot = 0.3 are simulated. First, the LES of TC is benchmarked for different rotation ratios. Both the Smagorinsky model with a constant of c_s = 0.1 and the dynamic model are found to produce reasonable results for no mean rotation and cyclonic rotation, but deviations increase for increasing rotation. This is attributed to the increasing anisotropic character of the fluctuations. Second, "over-damped" LES, i.e. LES with a large Smagorinsky constant is performed and is shown to reproduce some features of the large-scale structures, even when the near-wall region is not adequately modeled. This shows the potential for using over-damped LES for fast explorations of the parameter space where large-scale structures are found.

We explore the possibility of computing simple invariant solutions, like travelling waves or periodic orbits, in Large Eddy Simulation (LES) on a periodic domain with constant external forcing. The absence of material boundaries and the simple forcing mechanism make this system a comparatively simple target for the study of turbulent dynamics through invariant solutions. We show, that in spite of the application of eddy viscosity the computations are still rather challenging and must be performed on GPU cards rather than conventional coupled CPUs. We investigate the onset of turbulence in this system by means of bifurcation analysis, and present a long-period, large-amplitude unstable periodic orbit that is filtered from a turbulent time series. Although this orbit is computed on a coarse grid, with only a small separation between the integral scale and the LES filter length, the periodic dynamics seem to capture a regeneration process of the large-scale vortices.

Non-modal instability analysis of the shear layer near the nozzle of a supersonic under-expanded impinging jet is studied. The shear layer instability is considered to be one of the main components of the feedback loop in supersonic jets. The feedback loop is observed in instantaneous visualisations of the density field where it is noted that acoustic waves scattered by the nozzle lip internalise as shear layer instabilities.

A modal analysis describes the asymptotic limit of the instability disturbances and fails to capture short-time responses. Therefore, a non-modal analysis which allows the quantitative description of the short-time amplification or decay of a disturbance is performed by means of a local far-field pressure pulse. An impulse response analysis is performed which allows a wide range of frequencies to be excited. The temporal and spatial growths of the disturbances in the shear layer near the nozzle are studied by decomposing the response using dynamic mode decomposition and Hilbert transform analysis.

The short-time response shows that disturbances with non-dimensionalised temporal frequencies in the range of 1 to 4 have positive growth rates in the shear layer. The Hilbert transform analysis shows that high non-dimensionalised temporal frequencies (>4) are dampened immediately, whereas low non-dimensionalised temporal frequencies (<1) are neutral. Both dynamic mode decomposition and Hilbert transform analysis show that spatial frequencies between 1 and 3 have positive spatial growth rates. Finally, the envelope of the streamwise velocity disturbances reveals the presence of a convective instability.

The turbulence statistics and dynamics in the spanwise-minimal plane Couette flow with pressure gradients, so-called, Couette-Poiseuille (C-P) flow, are investigated using direct numerical simulation. The large-scale motion is limited in the spanwise box dimension as in the minimal-span channel turbulence of Flores & Jiménez (Phys. Fluids, vol. 22, 2010, 071704). The effect of the top wall, where normal pressure-driven Poiseuille flow is realised, is distinguished from the events on the bottom wall, where the pressure gradient results in mild or almost-zero wall-shear stress. A proper scaling of turbulence statistics in minimal-span C-P flows is presented. Also the 'shear-less' wall-bounded turbulence, where the Corrsin shear parameter is very weak compared to normal wall-bounded turbulence, represents local separation, which is also observed as spanwise streaks of reversed flow in full-size plane C-P turbulence. The local separation is a multi-scale event, which grows up to the order of the channel height even in the minimal-span geometry.