ETC2019 17th European Turbulence Conference Submission

10:45 Wall Bounded Turbulence 1

10:45 15 mins #110	Particle image velocimetry measurement on the turbulent boundary layer over convergent-divergent riblets Fang Xu, Shan Zhong, Shanying Zhang Abstract: The surface pattern of convergent-divergent riblets (C-D riblets), as a nature-inspired surface pattern with spanwise heterogeneity, is composed of left-yawed and right-yawed riblet strips in a spanwise alternating manner. As a spanwise heterogeneous surface pattern, C-D riblets can lead to a profound modification to the entire boundary layer. With an upwelling over the converging region, a downwelling over the diverging region and a surface flow in the spanwise direction from the diverging region to the converging region, a time-averaged boundary layer-scale secondary flow can be observed in the cross-stream plane \cite{Nugroho2013, Kevin2017, Xu2018, Xu2018b}. An in-depth understanding of the control mechanism of C-D riblets in the turbulent boundary layer is of great significance both in theory and in practice. To date, our understandings on the effect of C-D riblets on turbulence are quite limited. There is still a lack of adequate researches on how the spanwise vortices are influenced by C-D riblets, and a parameter study on the riblet height in the turbulent boundary layer at a fixed freestream velocity is not yet available. In the present study, we intend to provide a deeper understanding on these issues by investigating vortical structures and the secondary flow in a turbulent boundary layer at $Re_{\theta} = 723$ over C-D riblets. Mono- and stereoscopic particle image velocimetry measurements were conducted in the longitudinal plane and the cross-stream plane respectively. The parameter study on the riblet height is carried out by considering three sets of C-D riblets with different heights at the same freestream velocity, i.e. the same baseline boundary layer. The mechanism of the riblet height effect is revealed by changing the physical length. By contrast, in the work of \cite{Nugroho2013}, the riblet height in wall units was changed by varying the freestream velocity. Therefore, these two works investigate the riblet height effect from two complementary aspects. Compared with the previous researches (also available in the present study) which measured the flow field in a cross-stream plane \cite{Nugroho2013, Kevin2017, Xu2018}, the measurement in the longitudinal plane is directly related to the spatial arrangement of instantaneous vortical structures. Due to the creation of new coherent structures \cite{Kevin2019}, the population and the spatial arrangement of vortical structures exhibit pronounced modifications over the converging region. It is overall demonstrated that a variation of the riblet height is mainly reflected in the near-wall area over the diverging region and in the area away from the wall over the converging region. The present results help to understand how a spanwise heterogeneous surface pattern affects the turbulent boundary layer, especially the coherent structures.
11:00 15 mins #222	Diagnostic plot scaling accounting for adverse pressure gradient history effect Artur Dróżdż, Paweł Niegodajew, Witold Elsner, Ricardo Vinuesa, Ramis Örlü, Philipp Schlatter Abstract: The paper presents the adaptation of diagnostic plot scaling (Alfredsson and Örlü 2010) of turbulence intensity profiles to adverse pressure gradient flows. In the present study a corrected formula with shape factor has been proposed which extends diagnostic plot scaling onto strong pressure gradient flows with Reynolds number of practical interest. Formula corrects the pressure gradient history effect using the coefficient of streamwise gradient of Cp.
11:15 15 mins #407	Turbulence dynamics in separated flows: the generalised Kolmogorov equation for inhomogeneous anisotropic conditions Jean-Paul Mollicone, Francesco Battista, Paolo Gualtieri, Carlo Massimo Casciola Abstract: The generalised Kolmogorov equation (GKE) is used to describe the scale-by-scale turbulence dynamics in the shear layer and in the separation bubble generated by a bulge at one of the walls in a turbulent channel flow. The second order structure function, which is the basis of such equation, is used as a proxy to define a scale-energy content, that is an interpretation of the energy associated to a given scale. Production and dissipation regions and the flux interchange between them, in both physical and separation space, are identified. Results show how the GKE, a five-dimensional equation in our anisotropic and strongly inhomogeneous flow, can describe the turbulent flow behaviour and related energy mechanisms. The bulge (also referred to as bump) is introduced at one of the walls in a periodic turbulent channel flow in order to induce massive flow separation and strongly localised turbulent kinetic energy production. The database is taken from a direct numerical simulation recently performed by Mollicone et al. (2017, 2018) using the NEK5000 code, which is based on the spectral element method. Our focus is on the intense shear layer and the recirculation bubble that forms just after the bump. The five-dimensional fields involved in the GKE are represented by selecting appropriate two-dimensional planes for the analysis of results. Such complex statistical observables are linked to a visual inspection of instantaneous turbulent structures detected by means of the Q-criterion. Part of these turbulent structures are trapped in the recirculation where they undergo a pseudo-cyclic process of disruption and reformation. The rest are convected downstream, grow and tend to larger streamwise scales in an inverse cascade. The classical picture of homogeneous isotropic turbulence in which energy is fed at large scales and transferred to dissipate at small scales does not simply apply to this flow where the energy dynamics strongly depend on position, orientation and length scale.
11:30 15 mins #597	EXPERIMENTS IN NON-EQUILIBRIUM TURBULENT BOUNDARY LAYERS WITH FAVORABLE PRESSURE GRADIENTS Ralph Volino Abstract: Turbulent boundary layers on smooth walls with zero pressure gradient have been well studied. A large body of information is available in the literature concerning the statistics and structure of the turbulence in these canonical flows. Notable work has also been done in boundary layers with non-zero pressure gradients. The studies of Aubertine and Eaton [1], Castillo and George [2], and Harun et al. [3] are just a few examples. Much remains to be done, however, to understand how non-equilibrium flows develop and differ from the canonical case. Experiments were done in a recirculating water tunnel. The test section was 2 m long × 0.2m wide and 0.1 m tall at the inlet. A smooth flat plate on the lower surface was the test wall. The upper surface consisted of four flat plates that were positioned independently to set the pressure gradient. The first section was 0.6 m long and set for a zero pressure gradient (ZPG), followed by a 0.5 m long section set for a favorable pressure gradient (FPG). This was followed by a ZPG recovery region and an adverse pressure gradient (APG). The focus of the present paper is the FPG region. Three different configurations of the upper wall were used with varying inlet velocities, for eight experimental cases. The flat upper surfaces resulted in a constant acceleration parameter, K=(/Ue2)(dUe/dx), ranging from 0.125×106 to 2×106. The paper will present the results obtained from velocity profiles acquired with a two-component LDV system at 6 streamwise stations along the length of the FPG region. Figure 1a shows profiles of the Reynolds shear stress normalized on the local freestream velocity, in outer coordinates for the strongest acceleration case. The profile is clearly suppressed by the FPG, proceeding toward an eventual equilibrium as expected for a sink flow. To quantify the change in the profile, Fig. 1b shows the suppression for all cases at a representative location, y/99=0.4. The difference between the measured value and the same value in a canonical ZPG case at the same momentum thickness Reynolds number, Re, is shown as a function of streamwise location normalized on the sink flow region length, L=/(U∞,oK) (o denotes the start of the FPG). Multiplying the ordinate by the local Re1/2 helps to collapse the data. At large x the values approach the equilibrium sink flow results of Spalart [4]. Other quantities, such as the mean velocity and Reynolds normal stresses, show similar behavior and will be shown and further discussed in the final paper. Space allowing, the behavior in the ZPG and APG regions may also be shown. References [1] C. D. Aubertine and J. K. Eaton. Turbulence development in a non-equilibrium turbulent boundary layer with mild advserse pressure gradient. Journal of Fluid Mechanics 532: 345–364, 2005. [2] L. Castillo and W. K. George. Similarity analysis for turbulent boundary layer with pressure gradient: outer flow. AIAA Journal 39: 41–47, 2001. [3] Z. Harun, J. P. Monty, R. Mathis, and I. Marusic. Pressure gradient effects on the large-scale structure of turbulent boundary layers. Journal of Fluid Mechanics 715: 477–498, 2013. [4] P. R. Spalart. Numerical study of sink-flow boundary layers. Journal of Fluid Mechanics 172: 307–328, 1986.
11:45 15 mins #456	THE EFFECT OF MOMENTUM EXCHANGE BY COHERENT STRUCTURES ON THE FRICTION FACTOR AND MEAN VELOCITY PROFILE AT EXTREME REYNOLDS NUMBERS Hamidreza Anbarlooei, Fabio Ramos, Daniel Cruz Abstract: In turbulent flows inside pipes and channels, the dissipation of mean axial kinetic energy can be described as an interplay between diffusive and inertial effects. As the Reynolds number (a non-dimensional ratio of inertial to viscous forces) increases above a critical level, the flow develops a population of coherent structures exchanging momentum between the fast inertial centerline and the slow viscous near-wall regions. Despite the rich literature on coherent structures in turbulent flows, predicting the decrease in flow pressure due to wall friction based on the dynamics of coherent structures remains an elusive task. We report on a novel power-law formula for evaluating the friction factor of turbulent incompressible Newtonian fluid flows, f ∼ Re −2/13 . The formula is based on a new phenomenology for coherent structures dominating the momentum exchange in mesoregions that are set by the geometric mean of the viscous and outer length scales that have recently been shown to provide the missing link between the flow’s viscous and inertial regions. Comparison with the Princeton superpipe data [4] and Furuichi et al. [2] shows excellent agreement (figure 1). The new formula also implies a new asymptotic mean velocity profile of the form u + ∼ Ay +1/12 for the leading order, which is in total agreement ( both qualitatively and quantitatively) with the power-law profile of Barenblatt [1]. Our theory states that for extreme Reynolds numbers and small roughness sizes, the deviation from Blasius region to fully rough regime does not start because of the roughness effects. On the other hand, it gives us a mean to study the transition from the newly found regime to the fully rough one. This work shows that by considering the new region, some of the underlying problems of the Gioia and Chakraborty’s theory [3] can be solved and also very good agreement with the experimental results could be obtained. In the present work the influence of wall roughness on the start of the transition and also friction factor is studied. Figure 2 shows a comparison of the present result (whole theory from Blasius region to the newly found one and finally fully rough regime) with the Haaland empirical equation. As can be seen, the agreement is very good.
12:00 15 mins #184	Experimental investigation of spatially developing turbulent boundary layers over longitudinal grooves Wenfeng Li, Michael Klaas, Wolfgang Schröder Abstract: This investigation focuses on spatially developing turbulent boundary layer flows over a surface structured with grooves aligned in the streamwise direction. The longitudinal groove, i.e., riblet, has been recognized as one of the most promising passive flow control techniques that reduce friction drag in turbulent wall-bounded flows [1]. Using uτ for the friction velocity, v the kinematic viscosity, and s the lateral distance between riblet tips to define the non-dimensional spacing s+ = suτ /v, riblets reduce the friction drag by up to 10% if s+ is less than 40 wall units. Riblets are considered as a ‘negative roughness’ that decreases the friction drag. For larger non-dimensional riblet spacings, riblets can be regarded as roughness elements that increase the friction drag. The drag reducing and drag increasing effects are related to the scale of turbulent structures in the unperturbed turbulent wall region [2]. To improve the understanding of the turbulent boundary layers above drag decreasing and drag increasing riblet surfaces, an experimental investigation is conducted for a semi-circular riblet surface at several Reynolds numbers. A 2D-2C PIV setup as shown in figure 1(a) is used to simultaneously measure the flow field with a large field of view (FOV) which extends over 15 boundary layer thicknesses in the streamwise direction. To ensure a sufficient spatial resolution, 6 PCO edge cameras with a resolution of 5.5 megapixels are used to capture the particle images. The large FOV decomposed into 6 smaller FOVs possesses an area of 540 × 65 mm and a wall-normal spatial resolution of 0.44 mm based on the PIV interrogation window size. The Reynolds number Reθ based on the momentum thickness θ at the leading edge of the riblet surface is either 1410 or 3820 corresponding to freestream velocities U∞ = 8 and 20 m/s. This results in riblet spacings of s+ =16.8 in the drag reducing regime and 49.5 in the drag increasing regime. For each flow configuration, 10000 snapshots of the flow field are captured to reach converged turbulent statistics. The preliminary result in figure 1(c) shows an instantaneous flow field of the turbulent boundary layer above a riblet structured surface at Reθ = 1410. The experimental data are to be analyzed to determine, e.g., the statistics of uniform momentum zones, the large-scale modulation on the small-scale structures, and the flow recovery effect from the ribbed surface to a smooth surface.
12:15 15 mins #579	THE QUEST FOR HIGH REYNOLDS NUMBER TURBULENCE: RESULTS FROM AND FUTURE PERSPECTIVES OF CICLOPE gabriele bellani, Henrik Alfredsson, Jens Fransson, Hassan Nagib, Ramis Örlü, Alessandro Talamelli Abstract: The Long Pipe facility at the Center for International Cooperation in Long Pipe Experiments is designed to provide a canonical high-Reynolds number flow where turbulence quantities can be measured with the highest possible accuracy and resolution using various standard measurement techniques, such as hot-wire anemometry and laser Doppler velocimetry. The CICLoPE facility is located in a long underground tunnel with length of 130 m which allows a pipe length (L) of 105 m. Since for a given Reynolds the viscous length scale is directly proportional to the pipe diameter (D), the diameter was chosen as large as possible but with the limitation that L=D is still sufficiently long to establish a fully developed turbulent pipe flow over at least the ten final diameters. This resulted in a pipe diameter of 0.9 m which at the highest Reynolds numbers available (Re = 60000), corresponds to a centreline velocity of more than 60 m/s. At this Re the viscous length scale in CICLoPE is 7.5 m which, and is e.g. seven times larger than in the Princeton Superpipe. The Long Pipe in CICLoPE has the potential to address “many, if not all” of the most important outstanding issues in wall-bounded turbulence [1], at least for pipe flows. Some of the issues were presented in a situation paper that was the outcome of a series of turbulence meetings in the beginning of this century [2]. Since its inauguration in late 2014, the Long Pipe facility has hosted several international research activities, mostly within the framework of the European High-Performance Infrastructures in Turbulence (EuHIT), which have been reported in separate papers [3, 4, 5, 6, 7]. An important aspect of the Long Pipe is that the Reynolds number range is such that full scale separation is achieved, i.e. that scales associated with large – geometry dependent – scales are sufficiently separated from the small-scale motion. This is extremely important because most scaling arguments and asymptotic theories rely on this hypothesis. However, the increasing availability of high-Reynolds number data has undermined this view. Reducing measurement uncertainty, both on velocity fluctuations and on friction velocity, is a key factor to determine whether scaling anomalies are significant or a measurement artifact. The early experiments performed in CICLOPE have indeed highlighted that scaling mean and fluctuating velocty might be in contrast with the classic arguments [8], providing further supporting evidence that large scale motions strongly interact with the near-wall small scales. In this regard, recent work [9] has shown the possibility of an inverse inter-scale transport of the Reynolds shear stress, i.e. it can evolve towards larger scales and also be transported from the inner to the outer region (the central region of the pipe), maybe resulting in the creation of large-scale structures in the outer flow. The Long Pipe is an ideal facility to further investigate these issues. Results from the experimental campaigns of the first years of activity of CICLoPE will be presented, and the new perspectives opened will be discussed.
12:30 15 mins #16	NUMERICAL INVESTIGATION OF FLOW CONTROL BY EMBEDDED VORTICES IN A DIFFUSER Yang Zhang, Ugo Piomelli, Gang Chen, Jiakuan Xu Abstract: The diffuser is widely used to increase the pressure of fluid flow in industrial applications, such as the draft tubes of the hydroelectric turbine, for example. The adverse pressure gradient created by the area expansion, however, can result in flow separation, which has negative effects on the pressure recovery and increases the losses. To control the separation, various devices can be used, such as vortex generators (VGs). Our objective is to study how vortex generators can affect the flow, and find a configuration (in terms of placement of the vortex generators and strength of the vortices they induce) that can maximize the pressure recovery. Several parameters for the installing of the streamwise vortices are studied in the present paper. An assignment of streamwise vortices on both top and bottom surfaces at the throat of the diffuser is proposed and a very good pressure recovery is achieved by this configuration.